full project

.gitignore

@@ -0,0 +1,46 @@
+# Miscellaneous
+*.class
+*.log
+*.pyc
+*.swp
+.DS_Store
+.atom/
+.buildlog/
+.history
+.svn/
+migrate_working_dir/
+
+# IntelliJ related
+*.iml
+*.ipr
+*.iws
+.idea/
+
+# The .vscode folder contains launch configuration and tasks you configure in
+# VS Code which you may wish to be included in version control, so this line
+# is commented out by default.
+#.vscode/
+
+# Flutter/Dart/Pub related
+# Libraries should not include pubspec.lock, per https://dart.dev/guides/libraries/private-files#pubspeclock.
+/pubspec.lock
+**/doc/api/
+.dart_tool/
+build/
+#/src/Stockfish
+trash
+.github
+.vscode
+
+# Files from build
+**/*.o
+**/*.s
+src/.depend
+
+# Built binary
+src/stockfish*
+src/-lstdc++.res
+
+# Neural network for the NNUE evaluation
+**/*.nnue
+

.metadata

@@ -0,0 +1,27 @@
+# This file tracks properties of this Flutter project.
+# Used by Flutter tool to assess capabilities and perform upgrades etc.
+#
+# This file should be version controlled and should not be manually edited.
+
+version:
+  revision: "476aa717cd342d11e16439b71f4f4c9209c50712"
+  channel: "beta"
+
+project_type: plugin
+
+# Tracks metadata for the flutter migrate command
+migration:
+  platforms:
+    - platform: root
+      create_revision: 476aa717cd342d11e16439b71f4f4c9209c50712
+      base_revision: 476aa717cd342d11e16439b71f4f4c9209c50712
+
+  # User provided section
+
+  # List of Local paths (relative to this file) that should be
+  # ignored by the migrate tool.
+  #
+  # Files that are not part of the templates will be ignored by default.
+  unmanaged_files:
+    - 'lib/main.dart'
+    - 'ios/Runner.xcodeproj/project.pbxproj'

CHANGELOG.md

@@ -0,0 +1,5 @@
+## 1.0.0
+
+First Release:
+* Usable
+* Stockfish logger is not usable due to cout/cin stream manipulationen.

LICENSE

@@ -0,0 +1,674 @@
+                    GNU GENERAL PUBLIC LICENSE
+                       Version 3, 29 June 2007
+
+ Copyright (C) 2007 Free Software Foundation, Inc. <https://fsf.org/>
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+                            Preamble
+
+  The GNU General Public License is a free, copyleft license for
+software and other kinds of works.
+
+  The licenses for most software and other practical works are designed
+to take away your freedom to share and change the works.  By contrast,
+the GNU General Public License is intended to guarantee your freedom to
+share and change all versions of a program--to make sure it remains free
+software for all its users.  We, the Free Software Foundation, use the
+GNU General Public License for most of our software; it applies also to
+any other work released this way by its authors.  You can apply it to
+your programs, too.
+
+  When we speak of free software, we are referring to freedom, not
+price.  Our General Public Licenses are designed to make sure that you
+have the freedom to distribute copies of free software (and charge for
+them if you wish), that you receive source code or can get it if you
+want it, that you can change the software or use pieces of it in new
+free programs, and that you know you can do these things.
+
+  To protect your rights, we need to prevent others from denying you
+these rights or asking you to surrender the rights.  Therefore, you have
+certain responsibilities if you distribute copies of the software, or if
+you modify it: responsibilities to respect the freedom of others.
+
+  For example, if you distribute copies of such a program, whether
+gratis or for a fee, you must pass on to the recipients the same
+freedoms that you received.  You must make sure that they, too, receive
+or can get the source code.  And you must show them these terms so they
+know their rights.
+
+  Developers that use the GNU GPL protect your rights with two steps:
+(1) assert copyright on the software, and (2) offer you this License
+giving you legal permission to copy, distribute and/or modify it.
+
+  For the developers' and authors' protection, the GPL clearly explains
+that there is no warranty for this free software.  For both users' and
+authors' sake, the GPL requires that modified versions be marked as
+changed, so that their problems will not be attributed erroneously to
+authors of previous versions.
+
+  Some devices are designed to deny users access to install or run
+modified versions of the software inside them, although the manufacturer
+can do so.  This is fundamentally incompatible with the aim of
+protecting users' freedom to change the software.  The systematic
+pattern of such abuse occurs in the area of products for individuals to
+use, which is precisely where it is most unacceptable.  Therefore, we
+have designed this version of the GPL to prohibit the practice for those
+products.  If such problems arise substantially in other domains, we
+stand ready to extend this provision to those domains in future versions
+of the GPL, as needed to protect the freedom of users.
+
+  Finally, every program is threatened constantly by software patents.
+States should not allow patents to restrict development and use of
+software on general-purpose computers, but in those that do, we wish to
+avoid the special danger that patents applied to a free program could
+make it effectively proprietary.  To prevent this, the GPL assures that
+patents cannot be used to render the program non-free.
+
+  The precise terms and conditions for copying, distribution and
+modification follow.
+
+                       TERMS AND CONDITIONS
+
+  0. Definitions.
+
+  "This License" refers to version 3 of the GNU General Public License.
+
+  "Copyright" also means copyright-like laws that apply to other kinds of
+works, such as semiconductor masks.
+
+  "The Program" refers to any copyrightable work licensed under this
+License.  Each licensee is addressed as "you".  "Licensees" and
+"recipients" may be individuals or organizations.
+
+  To "modify" a work means to copy from or adapt all or part of the work
+in a fashion requiring copyright permission, other than the making of an
+exact copy.  The resulting work is called a "modified version" of the
+earlier work or a work "based on" the earlier work.
+
+  A "covered work" means either the unmodified Program or a work based
+on the Program.
+
+  To "propagate" a work means to do anything with it that, without
+permission, would make you directly or secondarily liable for
+infringement under applicable copyright law, except executing it on a
+computer or modifying a private copy.  Propagation includes copying,
+distribution (with or without modification), making available to the
+public, and in some countries other activities as well.
+
+  To "convey" a work means any kind of propagation that enables other
+parties to make or receive copies.  Mere interaction with a user through
+a computer network, with no transfer of a copy, is not conveying.
+
+  An interactive user interface displays "Appropriate Legal Notices"
+to the extent that it includes a convenient and prominently visible
+feature that (1) displays an appropriate copyright notice, and (2)
+tells the user that there is no warranty for the work (except to the
+extent that warranties are provided), that licensees may convey the
+work under this License, and how to view a copy of this License.  If
+the interface presents a list of user commands or options, such as a
+menu, a prominent item in the list meets this criterion.
+
+  1. Source Code.
+
+  The "source code" for a work means the preferred form of the work
+for making modifications to it.  "Object code" means any non-source
+form of a work.
+
+  A "Standard Interface" means an interface that either is an official
+standard defined by a recognized standards body, or, in the case of
+interfaces specified for a particular programming language, one that
+is widely used among developers working in that language.
+
+  The "System Libraries" of an executable work include anything, other
+than the work as a whole, that (a) is included in the normal form of
+packaging a Major Component, but which is not part of that Major
+Component, and (b) serves only to enable use of the work with that
+Major Component, or to implement a Standard Interface for which an
+implementation is available to the public in source code form.  A
+"Major Component", in this context, means a major essential component
+(kernel, window system, and so on) of the specific operating system
+(if any) on which the executable work runs, or a compiler used to
+produce the work, or an object code interpreter used to run it.
+
+  The "Corresponding Source" for a work in object code form means all
+the source code needed to generate, install, and (for an executable
+work) run the object code and to modify the work, including scripts to
+control those activities.  However, it does not include the work's
+System Libraries, or general-purpose tools or generally available free
+programs which are used unmodified in performing those activities but
+which are not part of the work.  For example, Corresponding Source
+includes interface definition files associated with source files for
+the work, and the source code for shared libraries and dynamically
+linked subprograms that the work is specifically designed to require,
+such as by intimate data communication or control flow between those
+subprograms and other parts of the work.
+
+  The Corresponding Source need not include anything that users
+can regenerate automatically from other parts of the Corresponding
+Source.
+
+  The Corresponding Source for a work in source code form is that
+same work.
+
+  2. Basic Permissions.
+
+  All rights granted under this License are granted for the term of
+copyright on the Program, and are irrevocable provided the stated
+conditions are met.  This License explicitly affirms your unlimited
+permission to run the unmodified Program.  The output from running a
+covered work is covered by this License only if the output, given its
+content, constitutes a covered work.  This License acknowledges your
+rights of fair use or other equivalent, as provided by copyright law.
+
+  You may make, run and propagate covered works that you do not
+convey, without conditions so long as your license otherwise remains
+in force.  You may convey covered works to others for the sole purpose
+of having them make modifications exclusively for you, or provide you
+with facilities for running those works, provided that you comply with
+the terms of this License in conveying all material for which you do
+not control copyright.  Those thus making or running the covered works
+for you must do so exclusively on your behalf, under your direction
+and control, on terms that prohibit them from making any copies of
+your copyrighted material outside their relationship with you.
+
+  Conveying under any other circumstances is permitted solely under
+the conditions stated below.  Sublicensing is not allowed; section 10
+makes it unnecessary.
+
+  3. Protecting Users' Legal Rights From Anti-Circumvention Law.
+
+  No covered work shall be deemed part of an effective technological
+measure under any applicable law fulfilling obligations under article
+11 of the WIPO copyright treaty adopted on 20 December 1996, or
+similar laws prohibiting or restricting circumvention of such
+measures.
+
+  When you convey a covered work, you waive any legal power to forbid
+circumvention of technological measures to the extent such circumvention
+is effected by exercising rights under this License with respect to
+the covered work, and you disclaim any intention to limit operation or
+modification of the work as a means of enforcing, against the work's
+users, your or third parties' legal rights to forbid circumvention of
+technological measures.
+
+  4. Conveying Verbatim Copies.
+
+  You may convey verbatim copies of the Program's source code as you
+receive it, in any medium, provided that you conspicuously and
+appropriately publish on each copy an appropriate copyright notice;
+keep intact all notices stating that this License and any
+non-permissive terms added in accord with section 7 apply to the code;
+keep intact all notices of the absence of any warranty; and give all
+recipients a copy of this License along with the Program.
+
+  You may charge any price or no price for each copy that you convey,
+and you may offer support or warranty protection for a fee.
+
+  5. Conveying Modified Source Versions.
+
+  You may convey a work based on the Program, or the modifications to
+produce it from the Program, in the form of source code under the
+terms of section 4, provided that you also meet all of these conditions:
+
+    a) The work must carry prominent notices stating that you modified
+    it, and giving a relevant date.
+
+    b) The work must carry prominent notices stating that it is
+    released under this License and any conditions added under section
+    7.  This requirement modifies the requirement in section 4 to
+    "keep intact all notices".
+
+    c) You must license the entire work, as a whole, under this
+    License to anyone who comes into possession of a copy.  This
+    License will therefore apply, along with any applicable section 7
+    additional terms, to the whole of the work, and all its parts,
+    regardless of how they are packaged.  This License gives no
+    permission to license the work in any other way, but it does not
+    invalidate such permission if you have separately received it.
+
+    d) If the work has interactive user interfaces, each must display
+    Appropriate Legal Notices; however, if the Program has interactive
+    interfaces that do not display Appropriate Legal Notices, your
+    work need not make them do so.
+
+  A compilation of a covered work with other separate and independent
+works, which are not by their nature extensions of the covered work,
+and which are not combined with it such as to form a larger program,
+in or on a volume of a storage or distribution medium, is called an
+"aggregate" if the compilation and its resulting copyright are not
+used to limit the access or legal rights of the compilation's users
+beyond what the individual works permit.  Inclusion of a covered work
+in an aggregate does not cause this License to apply to the other
+parts of the aggregate.
+
+  6. Conveying Non-Source Forms.
+
+  You may convey a covered work in object code form under the terms
+of sections 4 and 5, provided that you also convey the
+machine-readable Corresponding Source under the terms of this License,
+in one of these ways:
+
+    a) Convey the object code in, or embodied in, a physical product
+    (including a physical distribution medium), accompanied by the
+    Corresponding Source fixed on a durable physical medium
+    customarily used for software interchange.
+
+    b) Convey the object code in, or embodied in, a physical product
+    (including a physical distribution medium), accompanied by a
+    written offer, valid for at least three years and valid for as
+    long as you offer spare parts or customer support for that product
+    model, to give anyone who possesses the object code either (1) a
+    copy of the Corresponding Source for all the software in the
+    product that is covered by this License, on a durable physical
+    medium customarily used for software interchange, for a price no
+    more than your reasonable cost of physically performing this
+    conveying of source, or (2) access to copy the
+    Corresponding Source from a network server at no charge.
+
+    c) Convey individual copies of the object code with a copy of the
+    written offer to provide the Corresponding Source.  This
+    alternative is allowed only occasionally and noncommercially, and
+    only if you received the object code with such an offer, in accord
+    with subsection 6b.
+
+    d) Convey the object code by offering access from a designated
+    place (gratis or for a charge), and offer equivalent access to the
+    Corresponding Source in the same way through the same place at no
+    further charge.  You need not require recipients to copy the
+    Corresponding Source along with the object code.  If the place to
+    copy the object code is a network server, the Corresponding Source
+    may be on a different server (operated by you or a third party)
+    that supports equivalent copying facilities, provided you maintain
+    clear directions next to the object code saying where to find the
+    Corresponding Source.  Regardless of what server hosts the
+    Corresponding Source, you remain obligated to ensure that it is
+    available for as long as needed to satisfy these requirements.
+
+    e) Convey the object code using peer-to-peer transmission, provided
+    you inform other peers where the object code and Corresponding
+    Source of the work are being offered to the general public at no
+    charge under subsection 6d.
+
+  A separable portion of the object code, whose source code is excluded
+from the Corresponding Source as a System Library, need not be
+included in conveying the object code work.
+
+  A "User Product" is either (1) a "consumer product", which means any
+tangible personal property which is normally used for personal, family,
+or household purposes, or (2) anything designed or sold for incorporation
+into a dwelling.  In determining whether a product is a consumer product,
+doubtful cases shall be resolved in favor of coverage.  For a particular
+product received by a particular user, "normally used" refers to a
+typical or common use of that class of product, regardless of the status
+of the particular user or of the way in which the particular user
+actually uses, or expects or is expected to use, the product.  A product
+is a consumer product regardless of whether the product has substantial
+commercial, industrial or non-consumer uses, unless such uses represent
+the only significant mode of use of the product.
+
+  "Installation Information" for a User Product means any methods,
+procedures, authorization keys, or other information required to install
+and execute modified versions of a covered work in that User Product from
+a modified version of its Corresponding Source.  The information must
+suffice to ensure that the continued functioning of the modified object
+code is in no case prevented or interfered with solely because
+modification has been made.
+
+  If you convey an object code work under this section in, or with, or
+specifically for use in, a User Product, and the conveying occurs as
+part of a transaction in which the right of possession and use of the
+User Product is transferred to the recipient in perpetuity or for a
+fixed term (regardless of how the transaction is characterized), the
+Corresponding Source conveyed under this section must be accompanied
+by the Installation Information.  But this requirement does not apply
+if neither you nor any third party retains the ability to install
+modified object code on the User Product (for example, the work has
+been installed in ROM).
+
+  The requirement to provide Installation Information does not include a
+requirement to continue to provide support service, warranty, or updates
+for a work that has been modified or installed by the recipient, or for
+the User Product in which it has been modified or installed.  Access to a
+network may be denied when the modification itself materially and
+adversely affects the operation of the network or violates the rules and
+protocols for communication across the network.
+
+  Corresponding Source conveyed, and Installation Information provided,
+in accord with this section must be in a format that is publicly
+documented (and with an implementation available to the public in
+source code form), and must require no special password or key for
+unpacking, reading or copying.
+
+  7. Additional Terms.
+
+  "Additional permissions" are terms that supplement the terms of this
+License by making exceptions from one or more of its conditions.
+Additional permissions that are applicable to the entire Program shall
+be treated as though they were included in this License, to the extent
+that they are valid under applicable law.  If additional permissions
+apply only to part of the Program, that part may be used separately
+under those permissions, but the entire Program remains governed by
+this License without regard to the additional permissions.
+
+  When you convey a copy of a covered work, you may at your option
+remove any additional permissions from that copy, or from any part of
+it.  (Additional permissions may be written to require their own
+removal in certain cases when you modify the work.)  You may place
+additional permissions on material, added by you to a covered work,
+for which you have or can give appropriate copyright permission.
+
+  Notwithstanding any other provision of this License, for material you
+add to a covered work, you may (if authorized by the copyright holders of
+that material) supplement the terms of this License with terms:
+
+    a) Disclaiming warranty or limiting liability differently from the
+    terms of sections 15 and 16 of this License; or
+
+    b) Requiring preservation of specified reasonable legal notices or
+    author attributions in that material or in the Appropriate Legal
+    Notices displayed by works containing it; or
+
+    c) Prohibiting misrepresentation of the origin of that material, or
+    requiring that modified versions of such material be marked in
+    reasonable ways as different from the original version; or
+
+    d) Limiting the use for publicity purposes of names of licensors or
+    authors of the material; or
+
+    e) Declining to grant rights under trademark law for use of some
+    trade names, trademarks, or service marks; or
+
+    f) Requiring indemnification of licensors and authors of that
+    material by anyone who conveys the material (or modified versions of
+    it) with contractual assumptions of liability to the recipient, for
+    any liability that these contractual assumptions directly impose on
+    those licensors and authors.
+
+  All other non-permissive additional terms are considered "further
+restrictions" within the meaning of section 10.  If the Program as you
+received it, or any part of it, contains a notice stating that it is
+governed by this License along with a term that is a further
+restriction, you may remove that term.  If a license document contains
+a further restriction but permits relicensing or conveying under this
+License, you may add to a covered work material governed by the terms
+of that license document, provided that the further restriction does
+not survive such relicensing or conveying.
+
+  If you add terms to a covered work in accord with this section, you
+must place, in the relevant source files, a statement of the
+additional terms that apply to those files, or a notice indicating
+where to find the applicable terms.
+
+  Additional terms, permissive or non-permissive, may be stated in the
+form of a separately written license, or stated as exceptions;
+the above requirements apply either way.
+
+  8. Termination.
+
+  You may not propagate or modify a covered work except as expressly
+provided under this License.  Any attempt otherwise to propagate or
+modify it is void, and will automatically terminate your rights under
+this License (including any patent licenses granted under the third
+paragraph of section 11).
+
+  However, if you cease all violation of this License, then your
+license from a particular copyright holder is reinstated (a)
+provisionally, unless and until the copyright holder explicitly and
+finally terminates your license, and (b) permanently, if the copyright
+holder fails to notify you of the violation by some reasonable means
+prior to 60 days after the cessation.
+
+  Moreover, your license from a particular copyright holder is
+reinstated permanently if the copyright holder notifies you of the
+violation by some reasonable means, this is the first time you have
+received notice of violation of this License (for any work) from that
+copyright holder, and you cure the violation prior to 30 days after
+your receipt of the notice.
+
+  Termination of your rights under this section does not terminate the
+licenses of parties who have received copies or rights from you under
+this License.  If your rights have been terminated and not permanently
+reinstated, you do not qualify to receive new licenses for the same
+material under section 10.
+
+  9. Acceptance Not Required for Having Copies.
+
+  You are not required to accept this License in order to receive or
+run a copy of the Program.  Ancillary propagation of a covered work
+occurring solely as a consequence of using peer-to-peer transmission
+to receive a copy likewise does not require acceptance.  However,
+nothing other than this License grants you permission to propagate or
+modify any covered work.  These actions infringe copyright if you do
+not accept this License.  Therefore, by modifying or propagating a
+covered work, you indicate your acceptance of this License to do so.
+
+  10. Automatic Licensing of Downstream Recipients.
+
+  Each time you convey a covered work, the recipient automatically
+receives a license from the original licensors, to run, modify and
+propagate that work, subject to this License.  You are not responsible
+for enforcing compliance by third parties with this License.
+
+  An "entity transaction" is a transaction transferring control of an
+organization, or substantially all assets of one, or subdividing an
+organization, or merging organizations.  If propagation of a covered
+work results from an entity transaction, each party to that
+transaction who receives a copy of the work also receives whatever
+licenses to the work the party's predecessor in interest had or could
+give under the previous paragraph, plus a right to possession of the
+Corresponding Source of the work from the predecessor in interest, if
+the predecessor has it or can get it with reasonable efforts.
+
+  You may not impose any further restrictions on the exercise of the
+rights granted or affirmed under this License.  For example, you may
+not impose a license fee, royalty, or other charge for exercise of
+rights granted under this License, and you may not initiate litigation
+(including a cross-claim or counterclaim in a lawsuit) alleging that
+any patent claim is infringed by making, using, selling, offering for
+sale, or importing the Program or any portion of it.
+
+  11. Patents.
+
+  A "contributor" is a copyright holder who authorizes use under this
+License of the Program or a work on which the Program is based.  The
+work thus licensed is called the contributor's "contributor version".
+
+  A contributor's "essential patent claims" are all patent claims
+owned or controlled by the contributor, whether already acquired or
+hereafter acquired, that would be infringed by some manner, permitted
+by this License, of making, using, or selling its contributor version,
+but do not include claims that would be infringed only as a
+consequence of further modification of the contributor version.  For
+purposes of this definition, "control" includes the right to grant
+patent sublicenses in a manner consistent with the requirements of
+this License.
+
+  Each contributor grants you a non-exclusive, worldwide, royalty-free
+patent license under the contributor's essential patent claims, to
+make, use, sell, offer for sale, import and otherwise run, modify and
+propagate the contents of its contributor version.
+
+  In the following three paragraphs, a "patent license" is any express
+agreement or commitment, however denominated, not to enforce a patent
+(such as an express permission to practice a patent or covenant not to
+sue for patent infringement).  To "grant" such a patent license to a
+party means to make such an agreement or commitment not to enforce a
+patent against the party.
+
+  If you convey a covered work, knowingly relying on a patent license,
+and the Corresponding Source of the work is not available for anyone
+to copy, free of charge and under the terms of this License, through a
+publicly available network server or other readily accessible means,
+then you must either (1) cause the Corresponding Source to be so
+available, or (2) arrange to deprive yourself of the benefit of the
+patent license for this particular work, or (3) arrange, in a manner
+consistent with the requirements of this License, to extend the patent
+license to downstream recipients.  "Knowingly relying" means you have
+actual knowledge that, but for the patent license, your conveying the
+covered work in a country, or your recipient's use of the covered work
+in a country, would infringe one or more identifiable patents in that
+country that you have reason to believe are valid.
+
+  If, pursuant to or in connection with a single transaction or
+arrangement, you convey, or propagate by procuring conveyance of, a
+covered work, and grant a patent license to some of the parties
+receiving the covered work authorizing them to use, propagate, modify
+or convey a specific copy of the covered work, then the patent license
+you grant is automatically extended to all recipients of the covered
+work and works based on it.
+
+  A patent license is "discriminatory" if it does not include within
+the scope of its coverage, prohibits the exercise of, or is
+conditioned on the non-exercise of one or more of the rights that are
+specifically granted under this License.  You may not convey a covered
+work if you are a party to an arrangement with a third party that is
+in the business of distributing software, under which you make payment
+to the third party based on the extent of your activity of conveying
+the work, and under which the third party grants, to any of the
+parties who would receive the covered work from you, a discriminatory
+patent license (a) in connection with copies of the covered work
+conveyed by you (or copies made from those copies), or (b) primarily
+for and in connection with specific products or compilations that
+contain the covered work, unless you entered into that arrangement,
+or that patent license was granted, prior to 28 March 2007.
+
+  Nothing in this License shall be construed as excluding or limiting
+any implied license or other defenses to infringement that may
+otherwise be available to you under applicable patent law.
+
+  12. No Surrender of Others' Freedom.
+
+  If conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License.  If you cannot convey a
+covered work so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you may
+not convey it at all.  For example, if you agree to terms that obligate you
+to collect a royalty for further conveying from those to whom you convey
+the Program, the only way you could satisfy both those terms and this
+License would be to refrain entirely from conveying the Program.
+
+  13. Use with the GNU Affero General Public License.
+
+  Notwithstanding any other provision of this License, you have
+permission to link or combine any covered work with a work licensed
+under version 3 of the GNU Affero General Public License into a single
+combined work, and to convey the resulting work.  The terms of this
+License will continue to apply to the part which is the covered work,
+but the special requirements of the GNU Affero General Public License,
+section 13, concerning interaction through a network will apply to the
+combination as such.
+
+  14. Revised Versions of this License.
+
+  The Free Software Foundation may publish revised and/or new versions of
+the GNU General Public License from time to time.  Such new versions will
+be similar in spirit to the present version, but may differ in detail to
+address new problems or concerns.
+
+  Each version is given a distinguishing version number.  If the
+Program specifies that a certain numbered version of the GNU General
+Public License "or any later version" applies to it, you have the
+option of following the terms and conditions either of that numbered
+version or of any later version published by the Free Software
+Foundation.  If the Program does not specify a version number of the
+GNU General Public License, you may choose any version ever published
+by the Free Software Foundation.
+
+  If the Program specifies that a proxy can decide which future
+versions of the GNU General Public License can be used, that proxy's
+public statement of acceptance of a version permanently authorizes you
+to choose that version for the Program.
+
+  Later license versions may give you additional or different
+permissions.  However, no additional obligations are imposed on any
+author or copyright holder as a result of your choosing to follow a
+later version.
+
+  15. Disclaimer of Warranty.
+
+  THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY
+APPLICABLE LAW.  EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT
+HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY
+OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO,
+THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+PURPOSE.  THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM
+IS WITH YOU.  SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF
+ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
+
+  16. Limitation of Liability.
+
+  IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS
+THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY
+GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE
+USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF
+DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD
+PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS),
+EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF
+SUCH DAMAGES.
+
+  17. Interpretation of Sections 15 and 16.
+
+  If the disclaimer of warranty and limitation of liability provided
+above cannot be given local legal effect according to their terms,
+reviewing courts shall apply local law that most closely approximates
+an absolute waiver of all civil liability in connection with the
+Program, unless a warranty or assumption of liability accompanies a
+copy of the Program in return for a fee.
+
+                     END OF TERMS AND CONDITIONS
+
+            How to Apply These Terms to Your New Programs
+
+  If you develop a new program, and you want it to be of the greatest
+possible use to the public, the best way to achieve this is to make it
+free software which everyone can redistribute and change under these terms.
+
+  To do so, attach the following notices to the program.  It is safest
+to attach them to the start of each source file to most effectively
+state the exclusion of warranty; and each file should have at least
+the "copyright" line and a pointer to where the full notice is found.
+
+    <one line to give the program's name and a brief idea of what it does.>
+    Copyright (C) <year>  <name of author>
+
+    This program is free software: you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation, either version 3 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+Also add information on how to contact you by electronic and paper mail.
+
+  If the program does terminal interaction, make it output a short
+notice like this when it starts in an interactive mode:
+
+    <program>  Copyright (C) <year>  <name of author>
+    This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+    This is free software, and you are welcome to redistribute it
+    under certain conditions; type `show c' for details.
+
+The hypothetical commands `show w' and `show c' should show the appropriate
+parts of the General Public License.  Of course, your program's commands
+might be different; for a GUI interface, you would use an "about box".
+
+  You should also get your employer (if you work as a programmer) or school,
+if any, to sign a "copyright disclaimer" for the program, if necessary.
+For more information on this, and how to apply and follow the GNU GPL, see
+<https://www.gnu.org/licenses/>.
+
+  The GNU General Public License does not permit incorporating your program
+into proprietary programs.  If your program is a subroutine library, you
+may consider it more useful to permit linking proprietary applications with
+the library.  If this is what you want to do, use the GNU Lesser General
+Public License instead of this License.  But first, please read
+<https://www.gnu.org/licenses/why-not-lgpl.html>.

README.md

@@ -0,0 +1,42 @@
+# flutter_stockfish_plugin
+
+A Flutter plugin for Stockfish Chess engine.
+
+The current Version is based on Stockfish 16.
+
+## Usage
+
+```dart
+final stockfish = new Stockfish()
+
+// Listen on stdout of Stockfish engine
+final stockfishSubscription = stockfish.stdout.listen((line) {
+    print("received: $line");
+});
+
+// Sending UCI command to get Stockfish ready
+stockfish.stdin = 'isready'
+
+stockfish.stdin = 'position startpos moves e2e4' // set up start position
+stockfish.stdin = 'go depth 20' // search bestmove with a max septh of 20
+
+// Don't remember to dispose Stockfish when you're done.
+// Make shure to dispose Stockfish when closing the app. May use WindowListener.
+stockfishSubscription.cancel();
+stockfish.dispose();
+```
+
+## Goal of this fork of stockfish_chess_engine
+
+* Avoid limitation. This Version does not redirect stdout and stdin of the app for communication with stockfish.
+* stdin and stdout where replaced with a fakestream element.
+* Stockfish internel Logging might not work (could be fixed).
+
+## Credits
+* Based on and using source code from [stockfish_chess_engine](https://github.com/loloof64/StockfishChessEngineFlutter)
+* Using source code from [Stockfish](https://stockfishchess.org).
+* Using source code from [Flutter Stockfish](https://github.com/ArjanAswal/Stockfish).
+
+Directory src/Stockfish contains the latest current release.
+The code is modified to use a different communication interface.
+Original License of [Stockfish](https://stockfishchess.org) can be found in their [Github](https://github.com/official-stockfish/Stockfish) Repository.

analysis_options.yaml

@@ -0,0 +1,4 @@
+include: package:flutter_lints/flutter.yaml
+
+# Additional information about this file can be found at
+# https://dart.dev/guides/language/analysis-options

android/.gitignore

@@ -0,0 +1,9 @@
+*.iml
+.gradle
+/local.properties
+/.idea/workspace.xml
+/.idea/libraries
+.DS_Store
+/build
+/captures
+.cxx

android/build.gradle

@@ -0,0 +1,59 @@
+// The Android Gradle Plugin builds the native code with the Android NDK.
+
+group 'de.jusax.stockfish'
+version '1.0'
+
+buildscript {
+    repositories {
+        google()
+        mavenCentral()
+    }
+
+    dependencies {
+        // The Android Gradle Plugin knows how to build native code with the NDK.
+        classpath 'com.android.tools.build:gradle:7.1.2'
+    }
+}
+
+rootProject.allprojects {
+    repositories {
+        google()
+        mavenCentral()
+    }
+}
+
+apply plugin: 'com.android.library'
+
+android {
+    // Bumping the plugin compileSdkVersion requires all clients of this plugin
+    // to bump the version in their app.
+    compileSdkVersion 31
+
+    // Bumping the plugin ndkVersion requires all clients of this plugin to bump
+    // the version in their app and to download a newer version of the NDK.
+    ndkVersion "21.1.6352462"
+
+    // Invoke the shared CMake build with the Android Gradle Plugin.
+    externalNativeBuild {
+        cmake {
+            path "../src/CMakeLists.txt"
+
+            // The default CMake version for the Android Gradle Plugin is 3.10.2.
+            // https://developer.android.com/studio/projects/install-ndk#vanilla_cmake
+            //
+            // The Flutter tooling requires that developers have CMake 3.10 or later
+            // installed. You should not increase this version, as doing so will cause
+            // the plugin to fail to compile for some customers of the plugin.
+            // version "3.10.2"
+        }
+    }
+
+    compileOptions {
+        sourceCompatibility JavaVersion.VERSION_1_8
+        targetCompatibility JavaVersion.VERSION_1_8
+    }
+
+    defaultConfig {
+        minSdkVersion 16
+    }
+}

android/settings.gradle

@@ -0,0 +1 @@
+rootProject.name = 'stockfish'

android/src/main/AndroidManifest.xml

@@ -0,0 +1,3 @@
+<manifest xmlns:android="http://schemas.android.com/apk/res/android"
+  package="de.jusax.stockfish">
+</manifest>

ffigen.yaml

@@ -0,0 +1,21 @@
+# Run with `flutter pub run ffigen --config ffigen_linux.yaml`.
+name: StockfishChessEngineBindings
+description: |
+  Bindings for `src/stockfish.h`.
+
+  Regenerate bindings with `dart run ffigen --config ffigen.yaml`.
+output: 'lib/stockfish_bindings_generated.dart'
+headers:
+  entry-points:
+    - 'src/stockfish.h'
+  include-directives:
+    - 'src/stockfish.h'
+preamble: |
+  // ignore_for_file: always_specify_types
+  // ignore_for_file: camel_case_types
+  // ignore_for_file: non_constant_identifier_names
+comments:
+  style: any
+  length: full
+llvm-path:
+  - /usr/lib/llvm-14

lib/stockfish.dart

@@ -165,7 +165,6 @@ void _isolateMain(SendPort mainPort) {
   mainPort.send(exitCode);
 
   developer.log('nativeMain returns $exitCode', name: 'Stockfish');
-  print("stoped");
 }
 
 void _isolateStdout(SendPort stdoutPort) {

pubspec.yaml

@@ -0,0 +1,76 @@
+name: flutter_stockfish_plugin
+description: "A Stockfish Plugin for Flutter."
+version: 1.0.0
+homepage:
+
+environment:
+  sdk: '>=3.2.0-210.3.beta <4.0.0'
+  flutter: '>=3.3.0'
+
+dependencies:
+  ffi: ^2.1.0
+  flutter:
+    sdk: flutter
+  plugin_platform_interface: ^2.0.2
+
+dev_dependencies:
+  flutter_test:
+    sdk: flutter
+  flutter_lints: ^2.0.0
+  ffigen: ^9.0.1
+
+# For information on the generic Dart part of this file, see the
+# following page: https://dart.dev/tools/pub/pubspec
+
+# The following section is specific to Flutter packages.
+flutter:
+  # This section identifies this Flutter project as a plugin project.
+  # The 'pluginClass' specifies the class (in Java, Kotlin, Swift, Objective-C, etc.)
+  # which should be registered in the plugin registry. This is required for
+  # using method channels.
+  # The Android 'package' specifies package in which the registered class is.
+  # This is required for using method channels on Android.
+  # The 'ffiPlugin' specifies that native code should be built and bundled.
+  # This is required for using `dart:ffi`.
+  # All these are used by the tooling to maintain consistency when
+  # adding or updating assets for this project.
+  plugin:
+    platforms:
+      android:
+        ffiPlugin: true
+      linux:
+        ffiPlugin: true
+      windows:
+        ffiPlugin: true
+    # -------------------
+
+  # To add assets to your plugin package, add an assets section, like this:
+  # assets:
+  #   - images/a_dot_burr.jpeg
+  #   - images/a_dot_ham.jpeg
+  #
+  # For details regarding assets in packages, see
+  # https://flutter.dev/assets-and-images/#from-packages
+  #
+  # An image asset can refer to one or more resolution-specific "variants", see
+  # https://flutter.dev/assets-and-images/#resolution-aware
+
+  # To add custom fonts to your plugin package, add a fonts section here,
+  # in this "flutter" section. Each entry in this list should have a
+  # "family" key with the font family name, and a "fonts" key with a
+  # list giving the asset and other descriptors for the font. For
+  # example:
+  # fonts:
+  #   - family: Schyler
+  #     fonts:
+  #       - asset: fonts/Schyler-Regular.ttf
+  #       - asset: fonts/Schyler-Italic.ttf
+  #         style: italic
+  #   - family: Trajan Pro
+  #     fonts:
+  #       - asset: fonts/TrajanPro.ttf
+  #       - asset: fonts/TrajanPro_Bold.ttf
+  #         weight: 700
+  #
+  # For details regarding fonts in packages, see
+  # https://flutter.dev/custom-fonts/#from-packages

src/CMakeLists.txt

@@ -6,7 +6,10 @@ cmake_minimum_required(VERSION 3.18)
 project(stockfish VERSION 0.0.1 LANGUAGES CXX)
 file(GLOB_RECURSE cppPaths "Stockfish/src/*.cpp")
 set(CMAKE_CXX_STANDARD 17)
-set(NNUE_NAME nn-0000000000a0.nnue)
+set(CMAKE_CXX_FLAGS "-O3 -flto -march=native")
+#old: -Wshadow -Wmissing-declarations -fno-exceptions -std=c++17 -fPIC -Wall -Wextra -march=native
+#set(CMAKE_CXX_FLAGS_RELEASE "-O3 -flto -march=native")
+set(NNUE_NAME nn-5af11540bbfe.nnue)
 
 add_library(stockfish SHARED
   "stockfish.cpp"

src/Stockfish/.gitignore

@@ -0,0 +1,12 @@
+# Files from build
+**/*.o
+**/*.s
+src/.depend
+
+# Built binary
+src/stockfish*
+src/-lstdc++.res
+
+# Neural network for the NNUE evaluation
+**/*.nnue
+

src/Stockfish/AUTHORS

@@ -0,0 +1,226 @@
+# Founders of the Stockfish project and Fishtest infrastructure
+Tord Romstad (romstad)
+Marco Costalba (mcostalba)
+Joona Kiiski (zamar)
+Gary Linscott (glinscott)
+
+# Authors and inventors of NNUE, training, and NNUE port
+Yu Nasu (ynasu87)
+Motohiro Isozaki (yaneurao)
+Hisayori Noda (nodchip)
+
+# All other authors of Stockfish code (in alphabetical order)
+Aditya (absimaldata)
+Adrian Petrescu (apetresc)
+Ajith Chandy Jose (ajithcj)
+Alain Savard (Rocky640)
+Alayan Feh (Alayan-stk-2)
+Alexander Kure
+Alexander Pagel (Lolligerhans)
+Alfredo Menezes (lonfom169)
+Ali AlZhrani (Cooffe)
+Andreas Matthies (Matthies)
+Andrei Vetrov (proukornew)
+Andrew Grant (AndyGrant)
+Andrey Neporada (nepal)
+Andy Duplain
+Antoine Champion (antoinechampion)
+Aram Tumanian (atumanian)
+Arjun Temurnikar
+Artem Solopiy (EntityFX)
+Auguste Pop
+Balint Pfliegel
+Ben Chaney (Chaneybenjamini)
+Ben Koshy (BKSpurgeon)
+Bill Henry (VoyagerOne)
+Bojun Guo (noobpwnftw, Nooby)
+borg323
+Boštjan Mejak (PedanticHacker)
+braich
+Brian Sheppard (SapphireBrand, briansheppard-toast)
+Bruno de Melo Costa (BM123499)
+Bruno Pellanda (pellanda)
+Bryan Cross (crossbr)
+candirufish
+Chess13234
+Chris Cain (ceebo)
+clefrks
+Dale Weiler (graphitemaster)
+Daniel Axtens (daxtens)
+Daniel Dugovic (ddugovic)
+Dan Schmidt (dfannius)
+Dariusz Orzechowski (dorzechowski)
+David (dav1312)
+David Zar
+Daylen Yang (daylen)
+Deshawn Mohan-Smith (GoldenRare)
+Dieter Dobbelaere (ddobbelaere)
+DiscanX
+Dominik Schlösser (domschl)
+double-beep
+Douglas Matos Gomes (dsmsgms)
+Dubslow
+Eduardo Cáceres (eduherminio)
+Eelco de Groot (KingDefender)
+Elvin Liu (solarlight2)
+erbsenzaehler
+Ernesto Gatti
+Fabian Beuke (madnight)
+Fabian Fichter (ianfab)
+Fanael Linithien (Fanael)
+fanon
+Fauzi Akram Dabat (FauziAkram)
+Felix Wittmann
+gamander
+Gary Heckman (gheckman)
+George Sobala (gsobala)
+gguliash
+Giacomo Lorenzetti (G-Lorenz)
+Gian-Carlo Pascutto (gcp)
+Goh CJ (cj5716)
+Gontran Lemaire (gonlem)
+Goodkov Vasiliy Aleksandrovich (goodkov)
+Gregor Cramer
+GuardianRM
+Guy Vreuls (gvreuls)
+Günther Demetz (pb00067, pb00068)
+Henri Wiechers
+Hiraoka Takuya (HiraokaTakuya)
+homoSapiensSapiens
+Hongzhi Cheng
+Ivan Ivec (IIvec)
+Jacques B. (Timshel)
+Jake Senne (w1wwwwww)
+Jan Ondruš (hxim)
+Jared Kish (Kurtbusch, kurt22i)
+Jarrod Torriero (DU-jdto)
+Jean-Francois Romang (jromang)
+Jean Gauthier (OuaisBla)
+Jekaa
+Jerry Donald Watson (jerrydonaldwatson)
+jjoshua2
+Jonathan Buladas Dumale (SFisGOD)
+Jonathan Calovski (Mysseno)
+Jonathan McDermid (jonathanmcdermid)
+Joost VandeVondele (vondele)
+Joseph Ellis (jhellis3)
+Joseph R. Prostko
+Jörg Oster (joergoster)
+Julian Willemer (NightlyKing)
+jundery
+Justin Blanchard (UncombedCoconut)
+Kelly Wilson
+Ken Takusagawa
+Kian E (KJE-98)
+kinderchocolate
+Kiran Panditrao (Krgp)
+Kojirion
+Krystian Kuzniarek (kuzkry)
+Leonardo Ljubičić (ICCF World Champion)
+Leonid Pechenik (lp--)
+Liam Keegan (lkeegan)
+Linmiao Xu (linrock)
+Linus Arver (listx)
+loco-loco
+Lub van den Berg (ElbertoOne)
+Luca Brivio (lucabrivio)
+Lucas Braesch (lucasart)
+Lyudmil Antonov (lantonov)
+Maciej Żenczykowski (zenczykowski)
+Malcolm Campbell (xoto10)
+Mark Tenzer (31m059)
+marotear
+Matt Ginsberg (mattginsberg)
+Matthew Lai (matthewlai)
+Matthew Sullivan (Matt14916)
+Max A. (Disservin)
+Maxim Masiutin (maximmasiutin)
+Maxim Molchanov (Maxim)
+Michael An (man)
+Michael Byrne (MichaelB7)
+Michael Chaly (Vizvezdenec)
+Michael Stembera (mstembera)
+Michael Whiteley (protonspring)
+Michel Van den Bergh (vdbergh)
+Miguel Lahoz (miguel-l)
+Mikael Bäckman (mbootsector)
+Mike Babigian (Farseer)
+Mira
+Miroslav Fontán (Hexik)
+Moez Jellouli (MJZ1977)
+Mohammed Li (tthsqe12)
+Muzhen J (XInTheDark)
+Nathan Rugg (nmrugg)
+Nguyen Pham (nguyenpham)
+Nicklas Persson (NicklasPersson)
+Nick Pelling (nickpelling)
+Niklas Fiekas (niklasf)
+Nikolay Kostov (NikolayIT)
+Norman Schmidt (FireFather)
+notruck
+Ofek Shochat (OfekShochat, ghostway)
+Ondrej Mosnáček (WOnder93)
+Ondřej Mišina (AndrovT)
+Oskar Werkelin Ahlin
+Pablo Vazquez
+Panthee
+Pascal Romaret
+Pasquale Pigazzini (ppigazzini)
+Patrick Jansen (mibere)
+Peter Schneider (pschneider1968)
+Peter Zsifkovits (CoffeeOne)
+PikaCat
+Praveen Kumar Tummala (praveentml)
+Prokop Randáček (ProkopRandacek)
+Rahul Dsilva (silversolver1)
+Ralph Stößer (Ralph Stoesser)
+Raminder Singh
+renouve
+Reuven Peleg (R-Peleg)
+Richard Lloyd (Richard-Lloyd)
+Rodrigo Exterckötter Tjäder
+Rodrigo Roim (roim)
+Ronald de Man (syzygy1, syzygy)
+Ron Britvich (Britvich)
+rqs
+Rui Coelho (ruicoelhopedro)
+Ryan Schmitt
+Ryan Takker
+Sami Kiminki (skiminki)
+Sebastian Buchwald (UniQP)
+Sergei Antonov (saproj)
+Sergei Ivanov (svivanov72)
+Sergio Vieri (sergiovieri)
+sf-x
+Shahin M. Shahin (peregrine)
+Shane Booth (shane31)
+Shawn Varghese (xXH4CKST3RXx)
+Siad Daboul (Topologist)
+Stefan Geschwentner (locutus2)
+Stefano Cardanobile (Stefano80)
+Stefano Di Martino (StefanoD)
+Steinar Gunderson (sesse)
+Stéphane Nicolet (snicolet)
+Syine Mineta (MinetaS)
+Thanar2
+thaspel
+theo77186
+Tomasz Sobczyk (Sopel97)
+Tom Truscott
+Tom Vijlbrief (tomtor)
+Torsten Franz (torfranz, tfranzer)
+Torsten Hellwig (Torom)
+Tracey Emery (basepr1me)
+tttak
+Unai Corzo (unaiic)
+Uri Blass (uriblass)
+Vince Negri (cuddlestmonkey)
+Viren
+windfishballad
+xefoci7612
+zz4032
+
+# Additionally, we acknowledge the authors and maintainers of fishtest,
+# an amazing and essential framework for Stockfish development!
+#
+# https://github.com/glinscott/fishtest/blob/master/AUTHORS

src/Stockfish/CITATION.cff

@@ -0,0 +1,23 @@
+# This CITATION.cff file was generated with cffinit.
+# Visit https://bit.ly/cffinit to generate yours today!
+
+cff-version: 1.2.0
+title: Stockfish
+message: >-
+  Please cite this software using the metadata from this
+  file.
+type: software
+authors:
+  - name: The Stockfish developers (see AUTHORS file)
+repository-code: 'https://github.com/official-stockfish/Stockfish'
+url: 'https://stockfishchess.org/'
+repository-artifact: 'https://stockfishchess.org/download/'
+abstract: Stockfish is a free and strong UCI chess engine.
+keywords:
+  - chess
+  - artificial intelligence (AI)
+  - tree search
+  - alpha-beta search
+  - neural networks (NN)
+  - efficiently updatable neural networks (NNUE)
+license: GPL-3.0

src/Stockfish/Copying.txt

@@ -0,0 +1,674 @@
+                    GNU GENERAL PUBLIC LICENSE
+                       Version 3, 29 June 2007
+
+ Copyright (C) 2007 Free Software Foundation, Inc. <http://fsf.org/>
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+                            Preamble
+
+  The GNU General Public License is a free, copyleft license for
+software and other kinds of works.
+
+  The licenses for most software and other practical works are designed
+to take away your freedom to share and change the works.  By contrast,
+the GNU General Public License is intended to guarantee your freedom to
+share and change all versions of a program--to make sure it remains free
+software for all its users.  We, the Free Software Foundation, use the
+GNU General Public License for most of our software; it applies also to
+any other work released this way by its authors.  You can apply it to
+your programs, too.
+
+  When we speak of free software, we are referring to freedom, not
+price.  Our General Public Licenses are designed to make sure that you
+have the freedom to distribute copies of free software (and charge for
+them if you wish), that you receive source code or can get it if you
+want it, that you can change the software or use pieces of it in new
+free programs, and that you know you can do these things.
+
+  To protect your rights, we need to prevent others from denying you
+these rights or asking you to surrender the rights.  Therefore, you have
+certain responsibilities if you distribute copies of the software, or if
+you modify it: responsibilities to respect the freedom of others.
+
+  For example, if you distribute copies of such a program, whether
+gratis or for a fee, you must pass on to the recipients the same
+freedoms that you received.  You must make sure that they, too, receive
+or can get the source code.  And you must show them these terms so they
+know their rights.
+
+  Developers that use the GNU GPL protect your rights with two steps:
+(1) assert copyright on the software, and (2) offer you this License
+giving you legal permission to copy, distribute and/or modify it.
+
+  For the developers' and authors' protection, the GPL clearly explains
+that there is no warranty for this free software.  For both users' and
+authors' sake, the GPL requires that modified versions be marked as
+changed, so that their problems will not be attributed erroneously to
+authors of previous versions.
+
+  Some devices are designed to deny users access to install or run
+modified versions of the software inside them, although the manufacturer
+can do so.  This is fundamentally incompatible with the aim of
+protecting users' freedom to change the software.  The systematic
+pattern of such abuse occurs in the area of products for individuals to
+use, which is precisely where it is most unacceptable.  Therefore, we
+have designed this version of the GPL to prohibit the practice for those
+products.  If such problems arise substantially in other domains, we
+stand ready to extend this provision to those domains in future versions
+of the GPL, as needed to protect the freedom of users.
+
+  Finally, every program is threatened constantly by software patents.
+States should not allow patents to restrict development and use of
+software on general-purpose computers, but in those that do, we wish to
+avoid the special danger that patents applied to a free program could
+make it effectively proprietary.  To prevent this, the GPL assures that
+patents cannot be used to render the program non-free.
+
+  The precise terms and conditions for copying, distribution and
+modification follow.
+
+                       TERMS AND CONDITIONS
+
+  0. Definitions.
+
+  "This License" refers to version 3 of the GNU General Public License.
+
+  "Copyright" also means copyright-like laws that apply to other kinds of
+works, such as semiconductor masks.
+
+  "The Program" refers to any copyrightable work licensed under this
+License.  Each licensee is addressed as "you".  "Licensees" and
+"recipients" may be individuals or organizations.
+
+  To "modify" a work means to copy from or adapt all or part of the work
+in a fashion requiring copyright permission, other than the making of an
+exact copy.  The resulting work is called a "modified version" of the
+earlier work or a work "based on" the earlier work.
+
+  A "covered work" means either the unmodified Program or a work based
+on the Program.
+
+  To "propagate" a work means to do anything with it that, without
+permission, would make you directly or secondarily liable for
+infringement under applicable copyright law, except executing it on a
+computer or modifying a private copy.  Propagation includes copying,
+distribution (with or without modification), making available to the
+public, and in some countries other activities as well.
+
+  To "convey" a work means any kind of propagation that enables other
+parties to make or receive copies.  Mere interaction with a user through
+a computer network, with no transfer of a copy, is not conveying.
+
+  An interactive user interface displays "Appropriate Legal Notices"
+to the extent that it includes a convenient and prominently visible
+feature that (1) displays an appropriate copyright notice, and (2)
+tells the user that there is no warranty for the work (except to the
+extent that warranties are provided), that licensees may convey the
+work under this License, and how to view a copy of this License.  If
+the interface presents a list of user commands or options, such as a
+menu, a prominent item in the list meets this criterion.
+
+  1. Source Code.
+
+  The "source code" for a work means the preferred form of the work
+for making modifications to it.  "Object code" means any non-source
+form of a work.
+
+  A "Standard Interface" means an interface that either is an official
+standard defined by a recognized standards body, or, in the case of
+interfaces specified for a particular programming language, one that
+is widely used among developers working in that language.
+
+  The "System Libraries" of an executable work include anything, other
+than the work as a whole, that (a) is included in the normal form of
+packaging a Major Component, but which is not part of that Major
+Component, and (b) serves only to enable use of the work with that
+Major Component, or to implement a Standard Interface for which an
+implementation is available to the public in source code form.  A
+"Major Component", in this context, means a major essential component
+(kernel, window system, and so on) of the specific operating system
+(if any) on which the executable work runs, or a compiler used to
+produce the work, or an object code interpreter used to run it.
+
+  The "Corresponding Source" for a work in object code form means all
+the source code needed to generate, install, and (for an executable
+work) run the object code and to modify the work, including scripts to
+control those activities.  However, it does not include the work's
+System Libraries, or general-purpose tools or generally available free
+programs which are used unmodified in performing those activities but
+which are not part of the work.  For example, Corresponding Source
+includes interface definition files associated with source files for
+the work, and the source code for shared libraries and dynamically
+linked subprograms that the work is specifically designed to require,
+such as by intimate data communication or control flow between those
+subprograms and other parts of the work.
+
+  The Corresponding Source need not include anything that users
+can regenerate automatically from other parts of the Corresponding
+Source.
+
+  The Corresponding Source for a work in source code form is that
+same work.
+
+  2. Basic Permissions.
+
+  All rights granted under this License are granted for the term of
+copyright on the Program, and are irrevocable provided the stated
+conditions are met.  This License explicitly affirms your unlimited
+permission to run the unmodified Program.  The output from running a
+covered work is covered by this License only if the output, given its
+content, constitutes a covered work.  This License acknowledges your
+rights of fair use or other equivalent, as provided by copyright law.
+
+  You may make, run and propagate covered works that you do not
+convey, without conditions so long as your license otherwise remains
+in force.  You may convey covered works to others for the sole purpose
+of having them make modifications exclusively for you, or provide you
+with facilities for running those works, provided that you comply with
+the terms of this License in conveying all material for which you do
+not control copyright.  Those thus making or running the covered works
+for you must do so exclusively on your behalf, under your direction
+and control, on terms that prohibit them from making any copies of
+your copyrighted material outside their relationship with you.
+
+  Conveying under any other circumstances is permitted solely under
+the conditions stated below.  Sublicensing is not allowed; section 10
+makes it unnecessary.
+
+  3. Protecting Users' Legal Rights From Anti-Circumvention Law.
+
+  No covered work shall be deemed part of an effective technological
+measure under any applicable law fulfilling obligations under article
+11 of the WIPO copyright treaty adopted on 20 December 1996, or
+similar laws prohibiting or restricting circumvention of such
+measures.
+
+  When you convey a covered work, you waive any legal power to forbid
+circumvention of technological measures to the extent such circumvention
+is effected by exercising rights under this License with respect to
+the covered work, and you disclaim any intention to limit operation or
+modification of the work as a means of enforcing, against the work's
+users, your or third parties' legal rights to forbid circumvention of
+technological measures.
+
+  4. Conveying Verbatim Copies.
+
+  You may convey verbatim copies of the Program's source code as you
+receive it, in any medium, provided that you conspicuously and
+appropriately publish on each copy an appropriate copyright notice;
+keep intact all notices stating that this License and any
+non-permissive terms added in accord with section 7 apply to the code;
+keep intact all notices of the absence of any warranty; and give all
+recipients a copy of this License along with the Program.
+
+  You may charge any price or no price for each copy that you convey,
+and you may offer support or warranty protection for a fee.
+
+  5. Conveying Modified Source Versions.
+
+  You may convey a work based on the Program, or the modifications to
+produce it from the Program, in the form of source code under the
+terms of section 4, provided that you also meet all of these conditions:
+
+    a) The work must carry prominent notices stating that you modified
+    it, and giving a relevant date.
+
+    b) The work must carry prominent notices stating that it is
+    released under this License and any conditions added under section
+    7.  This requirement modifies the requirement in section 4 to
+    "keep intact all notices".
+
+    c) You must license the entire work, as a whole, under this
+    License to anyone who comes into possession of a copy.  This
+    License will therefore apply, along with any applicable section 7
+    additional terms, to the whole of the work, and all its parts,
+    regardless of how they are packaged.  This License gives no
+    permission to license the work in any other way, but it does not
+    invalidate such permission if you have separately received it.
+
+    d) If the work has interactive user interfaces, each must display
+    Appropriate Legal Notices; however, if the Program has interactive
+    interfaces that do not display Appropriate Legal Notices, your
+    work need not make them do so.
+
+  A compilation of a covered work with other separate and independent
+works, which are not by their nature extensions of the covered work,
+and which are not combined with it such as to form a larger program,
+in or on a volume of a storage or distribution medium, is called an
+"aggregate" if the compilation and its resulting copyright are not
+used to limit the access or legal rights of the compilation's users
+beyond what the individual works permit.  Inclusion of a covered work
+in an aggregate does not cause this License to apply to the other
+parts of the aggregate.
+
+  6. Conveying Non-Source Forms.
+
+  You may convey a covered work in object code form under the terms
+of sections 4 and 5, provided that you also convey the
+machine-readable Corresponding Source under the terms of this License,
+in one of these ways:
+
+    a) Convey the object code in, or embodied in, a physical product
+    (including a physical distribution medium), accompanied by the
+    Corresponding Source fixed on a durable physical medium
+    customarily used for software interchange.
+
+    b) Convey the object code in, or embodied in, a physical product
+    (including a physical distribution medium), accompanied by a
+    written offer, valid for at least three years and valid for as
+    long as you offer spare parts or customer support for that product
+    model, to give anyone who possesses the object code either (1) a
+    copy of the Corresponding Source for all the software in the
+    product that is covered by this License, on a durable physical
+    medium customarily used for software interchange, for a price no
+    more than your reasonable cost of physically performing this
+    conveying of source, or (2) access to copy the
+    Corresponding Source from a network server at no charge.
+
+    c) Convey individual copies of the object code with a copy of the
+    written offer to provide the Corresponding Source.  This
+    alternative is allowed only occasionally and noncommercially, and
+    only if you received the object code with such an offer, in accord
+    with subsection 6b.
+
+    d) Convey the object code by offering access from a designated
+    place (gratis or for a charge), and offer equivalent access to the
+    Corresponding Source in the same way through the same place at no
+    further charge.  You need not require recipients to copy the
+    Corresponding Source along with the object code.  If the place to
+    copy the object code is a network server, the Corresponding Source
+    may be on a different server (operated by you or a third party)
+    that supports equivalent copying facilities, provided you maintain
+    clear directions next to the object code saying where to find the
+    Corresponding Source.  Regardless of what server hosts the
+    Corresponding Source, you remain obligated to ensure that it is
+    available for as long as needed to satisfy these requirements.
+
+    e) Convey the object code using peer-to-peer transmission, provided
+    you inform other peers where the object code and Corresponding
+    Source of the work are being offered to the general public at no
+    charge under subsection 6d.
+
+  A separable portion of the object code, whose source code is excluded
+from the Corresponding Source as a System Library, need not be
+included in conveying the object code work.
+
+  A "User Product" is either (1) a "consumer product", which means any
+tangible personal property which is normally used for personal, family,
+or household purposes, or (2) anything designed or sold for incorporation
+into a dwelling.  In determining whether a product is a consumer product,
+doubtful cases shall be resolved in favor of coverage.  For a particular
+product received by a particular user, "normally used" refers to a
+typical or common use of that class of product, regardless of the status
+of the particular user or of the way in which the particular user
+actually uses, or expects or is expected to use, the product.  A product
+is a consumer product regardless of whether the product has substantial
+commercial, industrial or non-consumer uses, unless such uses represent
+the only significant mode of use of the product.
+
+  "Installation Information" for a User Product means any methods,
+procedures, authorization keys, or other information required to install
+and execute modified versions of a covered work in that User Product from
+a modified version of its Corresponding Source.  The information must
+suffice to ensure that the continued functioning of the modified object
+code is in no case prevented or interfered with solely because
+modification has been made.
+
+  If you convey an object code work under this section in, or with, or
+specifically for use in, a User Product, and the conveying occurs as
+part of a transaction in which the right of possession and use of the
+User Product is transferred to the recipient in perpetuity or for a
+fixed term (regardless of how the transaction is characterized), the
+Corresponding Source conveyed under this section must be accompanied
+by the Installation Information.  But this requirement does not apply
+if neither you nor any third party retains the ability to install
+modified object code on the User Product (for example, the work has
+been installed in ROM).
+
+  The requirement to provide Installation Information does not include a
+requirement to continue to provide support service, warranty, or updates
+for a work that has been modified or installed by the recipient, or for
+the User Product in which it has been modified or installed.  Access to a
+network may be denied when the modification itself materially and
+adversely affects the operation of the network or violates the rules and
+protocols for communication across the network.
+
+  Corresponding Source conveyed, and Installation Information provided,
+in accord with this section must be in a format that is publicly
+documented (and with an implementation available to the public in
+source code form), and must require no special password or key for
+unpacking, reading or copying.
+
+  7. Additional Terms.
+
+  "Additional permissions" are terms that supplement the terms of this
+License by making exceptions from one or more of its conditions.
+Additional permissions that are applicable to the entire Program shall
+be treated as though they were included in this License, to the extent
+that they are valid under applicable law.  If additional permissions
+apply only to part of the Program, that part may be used separately
+under those permissions, but the entire Program remains governed by
+this License without regard to the additional permissions.
+
+  When you convey a copy of a covered work, you may at your option
+remove any additional permissions from that copy, or from any part of
+it.  (Additional permissions may be written to require their own
+removal in certain cases when you modify the work.)  You may place
+additional permissions on material, added by you to a covered work,
+for which you have or can give appropriate copyright permission.
+
+  Notwithstanding any other provision of this License, for material you
+add to a covered work, you may (if authorized by the copyright holders of
+that material) supplement the terms of this License with terms:
+
+    a) Disclaiming warranty or limiting liability differently from the
+    terms of sections 15 and 16 of this License; or
+
+    b) Requiring preservation of specified reasonable legal notices or
+    author attributions in that material or in the Appropriate Legal
+    Notices displayed by works containing it; or
+
+    c) Prohibiting misrepresentation of the origin of that material, or
+    requiring that modified versions of such material be marked in
+    reasonable ways as different from the original version; or
+
+    d) Limiting the use for publicity purposes of names of licensors or
+    authors of the material; or
+
+    e) Declining to grant rights under trademark law for use of some
+    trade names, trademarks, or service marks; or
+
+    f) Requiring indemnification of licensors and authors of that
+    material by anyone who conveys the material (or modified versions of
+    it) with contractual assumptions of liability to the recipient, for
+    any liability that these contractual assumptions directly impose on
+    those licensors and authors.
+
+  All other non-permissive additional terms are considered "further
+restrictions" within the meaning of section 10.  If the Program as you
+received it, or any part of it, contains a notice stating that it is
+governed by this License along with a term that is a further
+restriction, you may remove that term.  If a license document contains
+a further restriction but permits relicensing or conveying under this
+License, you may add to a covered work material governed by the terms
+of that license document, provided that the further restriction does
+not survive such relicensing or conveying.
+
+  If you add terms to a covered work in accord with this section, you
+must place, in the relevant source files, a statement of the
+additional terms that apply to those files, or a notice indicating
+where to find the applicable terms.
+
+  Additional terms, permissive or non-permissive, may be stated in the
+form of a separately written license, or stated as exceptions;
+the above requirements apply either way.
+
+  8. Termination.
+
+  You may not propagate or modify a covered work except as expressly
+provided under this License.  Any attempt otherwise to propagate or
+modify it is void, and will automatically terminate your rights under
+this License (including any patent licenses granted under the third
+paragraph of section 11).
+
+  However, if you cease all violation of this License, then your
+license from a particular copyright holder is reinstated (a)
+provisionally, unless and until the copyright holder explicitly and
+finally terminates your license, and (b) permanently, if the copyright
+holder fails to notify you of the violation by some reasonable means
+prior to 60 days after the cessation.
+
+  Moreover, your license from a particular copyright holder is
+reinstated permanently if the copyright holder notifies you of the
+violation by some reasonable means, this is the first time you have
+received notice of violation of this License (for any work) from that
+copyright holder, and you cure the violation prior to 30 days after
+your receipt of the notice.
+
+  Termination of your rights under this section does not terminate the
+licenses of parties who have received copies or rights from you under
+this License.  If your rights have been terminated and not permanently
+reinstated, you do not qualify to receive new licenses for the same
+material under section 10.
+
+  9. Acceptance Not Required for Having Copies.
+
+  You are not required to accept this License in order to receive or
+run a copy of the Program.  Ancillary propagation of a covered work
+occurring solely as a consequence of using peer-to-peer transmission
+to receive a copy likewise does not require acceptance.  However,
+nothing other than this License grants you permission to propagate or
+modify any covered work.  These actions infringe copyright if you do
+not accept this License.  Therefore, by modifying or propagating a
+covered work, you indicate your acceptance of this License to do so.
+
+  10. Automatic Licensing of Downstream Recipients.
+
+  Each time you convey a covered work, the recipient automatically
+receives a license from the original licensors, to run, modify and
+propagate that work, subject to this License.  You are not responsible
+for enforcing compliance by third parties with this License.
+
+  An "entity transaction" is a transaction transferring control of an
+organization, or substantially all assets of one, or subdividing an
+organization, or merging organizations.  If propagation of a covered
+work results from an entity transaction, each party to that
+transaction who receives a copy of the work also receives whatever
+licenses to the work the party's predecessor in interest had or could
+give under the previous paragraph, plus a right to possession of the
+Corresponding Source of the work from the predecessor in interest, if
+the predecessor has it or can get it with reasonable efforts.
+
+  You may not impose any further restrictions on the exercise of the
+rights granted or affirmed under this License.  For example, you may
+not impose a license fee, royalty, or other charge for exercise of
+rights granted under this License, and you may not initiate litigation
+(including a cross-claim or counterclaim in a lawsuit) alleging that
+any patent claim is infringed by making, using, selling, offering for
+sale, or importing the Program or any portion of it.
+
+  11. Patents.
+
+  A "contributor" is a copyright holder who authorizes use under this
+License of the Program or a work on which the Program is based.  The
+work thus licensed is called the contributor's "contributor version".
+
+  A contributor's "essential patent claims" are all patent claims
+owned or controlled by the contributor, whether already acquired or
+hereafter acquired, that would be infringed by some manner, permitted
+by this License, of making, using, or selling its contributor version,
+but do not include claims that would be infringed only as a
+consequence of further modification of the contributor version.  For
+purposes of this definition, "control" includes the right to grant
+patent sublicenses in a manner consistent with the requirements of
+this License.
+
+  Each contributor grants you a non-exclusive, worldwide, royalty-free
+patent license under the contributor's essential patent claims, to
+make, use, sell, offer for sale, import and otherwise run, modify and
+propagate the contents of its contributor version.
+
+  In the following three paragraphs, a "patent license" is any express
+agreement or commitment, however denominated, not to enforce a patent
+(such as an express permission to practice a patent or covenant not to
+sue for patent infringement).  To "grant" such a patent license to a
+party means to make such an agreement or commitment not to enforce a
+patent against the party.
+
+  If you convey a covered work, knowingly relying on a patent license,
+and the Corresponding Source of the work is not available for anyone
+to copy, free of charge and under the terms of this License, through a
+publicly available network server or other readily accessible means,
+then you must either (1) cause the Corresponding Source to be so
+available, or (2) arrange to deprive yourself of the benefit of the
+patent license for this particular work, or (3) arrange, in a manner
+consistent with the requirements of this License, to extend the patent
+license to downstream recipients.  "Knowingly relying" means you have
+actual knowledge that, but for the patent license, your conveying the
+covered work in a country, or your recipient's use of the covered work
+in a country, would infringe one or more identifiable patents in that
+country that you have reason to believe are valid.
+
+  If, pursuant to or in connection with a single transaction or
+arrangement, you convey, or propagate by procuring conveyance of, a
+covered work, and grant a patent license to some of the parties
+receiving the covered work authorizing them to use, propagate, modify
+or convey a specific copy of the covered work, then the patent license
+you grant is automatically extended to all recipients of the covered
+work and works based on it.
+
+  A patent license is "discriminatory" if it does not include within
+the scope of its coverage, prohibits the exercise of, or is
+conditioned on the non-exercise of one or more of the rights that are
+specifically granted under this License.  You may not convey a covered
+work if you are a party to an arrangement with a third party that is
+in the business of distributing software, under which you make payment
+to the third party based on the extent of your activity of conveying
+the work, and under which the third party grants, to any of the
+parties who would receive the covered work from you, a discriminatory
+patent license (a) in connection with copies of the covered work
+conveyed by you (or copies made from those copies), or (b) primarily
+for and in connection with specific products or compilations that
+contain the covered work, unless you entered into that arrangement,
+or that patent license was granted, prior to 28 March 2007.
+
+  Nothing in this License shall be construed as excluding or limiting
+any implied license or other defenses to infringement that may
+otherwise be available to you under applicable patent law.
+
+  12. No Surrender of Others' Freedom.
+
+  If conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License.  If you cannot convey a
+covered work so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you may
+not convey it at all.  For example, if you agree to terms that obligate you
+to collect a royalty for further conveying from those to whom you convey
+the Program, the only way you could satisfy both those terms and this
+License would be to refrain entirely from conveying the Program.
+
+  13. Use with the GNU Affero General Public License.
+
+  Notwithstanding any other provision of this License, you have
+permission to link or combine any covered work with a work licensed
+under version 3 of the GNU Affero General Public License into a single
+combined work, and to convey the resulting work.  The terms of this
+License will continue to apply to the part which is the covered work,
+but the special requirements of the GNU Affero General Public License,
+section 13, concerning interaction through a network will apply to the
+combination as such.
+
+  14. Revised Versions of this License.
+
+  The Free Software Foundation may publish revised and/or new versions of
+the GNU General Public License from time to time.  Such new versions will
+be similar in spirit to the present version, but may differ in detail to
+address new problems or concerns.
+
+  Each version is given a distinguishing version number.  If the
+Program specifies that a certain numbered version of the GNU General
+Public License "or any later version" applies to it, you have the
+option of following the terms and conditions either of that numbered
+version or of any later version published by the Free Software
+Foundation.  If the Program does not specify a version number of the
+GNU General Public License, you may choose any version ever published
+by the Free Software Foundation.
+
+  If the Program specifies that a proxy can decide which future
+versions of the GNU General Public License can be used, that proxy's
+public statement of acceptance of a version permanently authorizes you
+to choose that version for the Program.
+
+  Later license versions may give you additional or different
+permissions.  However, no additional obligations are imposed on any
+author or copyright holder as a result of your choosing to follow a
+later version.
+
+  15. Disclaimer of Warranty.
+
+  THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY
+APPLICABLE LAW.  EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT
+HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY
+OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO,
+THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+PURPOSE.  THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM
+IS WITH YOU.  SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF
+ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
+
+  16. Limitation of Liability.
+
+  IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS
+THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY
+GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE
+USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF
+DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD
+PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS),
+EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF
+SUCH DAMAGES.
+
+  17. Interpretation of Sections 15 and 16.
+
+  If the disclaimer of warranty and limitation of liability provided
+above cannot be given local legal effect according to their terms,
+reviewing courts shall apply local law that most closely approximates
+an absolute waiver of all civil liability in connection with the
+Program, unless a warranty or assumption of liability accompanies a
+copy of the Program in return for a fee.
+
+                     END OF TERMS AND CONDITIONS
+
+            How to Apply These Terms to Your New Programs
+
+  If you develop a new program, and you want it to be of the greatest
+possible use to the public, the best way to achieve this is to make it
+free software which everyone can redistribute and change under these terms.
+
+  To do so, attach the following notices to the program.  It is safest
+to attach them to the start of each source file to most effectively
+state the exclusion of warranty; and each file should have at least
+the "copyright" line and a pointer to where the full notice is found.
+
+    <one line to give the program's name and a brief idea of what it does.>
+    Copyright (C) <year>  <name of author>
+
+    This program is free software: you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation, either version 3 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+Also add information on how to contact you by electronic and paper mail.
+
+  If the program does terminal interaction, make it output a short
+notice like this when it starts in an interactive mode:
+
+    <program>  Copyright (C) <year>  <name of author>
+    This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+    This is free software, and you are welcome to redistribute it
+    under certain conditions; type `show c' for details.
+
+The hypothetical commands `show w' and `show c' should show the appropriate
+parts of the General Public License.  Of course, your program's commands
+might be different; for a GUI interface, you would use an "about box".
+
+  You should also get your employer (if you work as a programmer) or school,
+if any, to sign a "copyright disclaimer" for the program, if necessary.
+For more information on this, and how to apply and follow the GNU GPL, see
+<http://www.gnu.org/licenses/>.
+
+  The GNU General Public License does not permit incorporating your program
+into proprietary programs.  If your program is a subroutine library, you
+may consider it more useful to permit linking proprietary applications with
+the library.  If this is what you want to do, use the GNU Lesser General
+Public License instead of this License.  But first, please read
+<http://www.gnu.org/philosophy/why-not-lgpl.html>.

src/Stockfish/README.md

@@ -0,0 +1,164 @@
+<div align="center">
+
+  [![Stockfish][stockfish128-logo]][website-link]
+
+  <h3>Stockfish</h3>
+
+  A free and strong UCI chess engine.
+  <br>
+  <strong>[Explore Stockfish docs »][wiki-link]</strong>
+  <br>
+  <br>
+  [Report bug][issue-link]
+  ·
+  [Open a discussion][discussions-link]
+  ·
+  [Discord][discord-link]
+  ·
+  [Blog][website-blog-link]
+
+  [![Build][build-badge]][build-link]
+  [![License][license-badge]][license-link]
+  <br>
+  [![Release][release-badge]][release-link]
+  [![Commits][commits-badge]][commits-link]
+  <br>
+  [![Website][website-badge]][website-link]
+  [![Fishtest][fishtest-badge]][fishtest-link]
+  [![Discord][discord-badge]][discord-link]
+
+</div>
+
+## Overview
+
+[Stockfish][website-link] is a **free and strong UCI chess engine** derived from
+Glaurung 2.1 that analyzes chess positions and computes the optimal moves.
+
+Stockfish **does not include a graphical user interface** (GUI) that is required
+to display a chessboard and to make it easy to input moves. These GUIs are
+developed independently from Stockfish and are available online. **Read the
+documentation for your GUI** of choice for information about how to use
+Stockfish with it.
+
+See also the Stockfish [documentation][wiki-usage-link] for further usage help.
+
+## Files
+
+This distribution of Stockfish consists of the following files:
+
+  * [README.md][readme-link], the file you are currently reading.
+
+  * [Copying.txt][license-link], a text file containing the GNU General Public
+    License version 3.
+
+  * [AUTHORS][authors-link], a text file with the list of authors for the project.
+
+  * [src][src-link], a subdirectory containing the full source code, including a
+    Makefile that can be used to compile Stockfish on Unix-like systems.
+
+  * a file with the .nnue extension, storing the neural network for the NNUE
+    evaluation. Binary distributions will have this file embedded.
+
+## The UCI protocol
+
+The [Universal Chess Interface][uci-link] (UCI) is a standard text-based protocol
+used to communicate with a chess engine and is the recommended way to do so for
+typical graphical user interfaces (GUI) or chess tools. Stockfish implements the
+majority of its options.
+
+Developers can see the default values for the UCI options available in Stockfish
+by typing `./stockfish uci` in a terminal, but most users should typically use a
+chess GUI to interact with Stockfish.
+
+For more information on UCI or debug commands, see our [documentation][wiki-commands-link].
+
+## Compiling Stockfish
+
+Stockfish has support for 32 or 64-bit CPUs, certain hardware instructions,
+big-endian machines such as Power PC, and other platforms.
+
+On Unix-like systems, it should be easy to compile Stockfish directly from the
+source code with the included Makefile in the folder `src`. In general, it is
+recommended to run `make help` to see a list of make targets with corresponding
+descriptions.
+
+```
+cd src
+make -j build ARCH=x86-64-modern
+```
+
+Detailed compilation instructions for all platforms can be found in our
+[documentation][wiki-compile-link].
+
+## Contributing
+
+### Donating hardware
+
+Improving Stockfish requires a massive amount of testing. You can donate your
+hardware resources by installing the [Fishtest Worker][worker-link] and viewing
+the current tests on [Fishtest][fishtest-link].
+
+### Improving the code
+
+In the [chessprogramming wiki][programming-link], many techniques used in
+Stockfish are explained with a lot of background information.
+The [section on Stockfish][programmingsf-link] describes many features
+and techniques used by Stockfish. However, it is generic rather than
+focused on Stockfish's precise implementation.
+
+The engine testing is done on [Fishtest][fishtest-link].
+If you want to help improve Stockfish, please read this [guideline][guideline-link]
+first, where the basics of Stockfish development are explained.
+
+Discussions about Stockfish take place these days mainly in the Stockfish
+[Discord server][discord-link]. This is also the best place to ask questions
+about the codebase and how to improve it.
+
+## Terms of use
+
+Stockfish is free and distributed under the
+[**GNU General Public License version 3**][license-link] (GPL v3). Essentially,
+this means you are free to do almost exactly what you want with the program,
+including distributing it among your friends, making it available for download
+from your website, selling it (either by itself or as part of some bigger
+software package), or using it as the starting point for a software project of
+your own.
+
+The only real limitation is that whenever you distribute Stockfish in some way,
+you MUST always include the license and the full source code (or a pointer to
+where the source code can be found) to generate the exact binary you are
+distributing. If you make any changes to the source code, these changes must
+also be made available under GPL v3.
+
+
+[authors-link]:       https://github.com/official-stockfish/Stockfish/blob/master/AUTHORS
+[build-link]:         https://github.com/official-stockfish/Stockfish/actions/workflows/stockfish.yml
+[commits-link]:       https://github.com/official-stockfish/Stockfish/commits/master
+[discord-link]:       https://discord.gg/GWDRS3kU6R
+[issue-link]:         https://github.com/official-stockfish/Stockfish/issues/new?assignees=&labels=&template=BUG-REPORT.yml
+[discussions-link]:   https://github.com/official-stockfish/Stockfish/discussions/new
+[fishtest-link]:      https://tests.stockfishchess.org/tests
+[guideline-link]:     https://github.com/glinscott/fishtest/wiki/Creating-my-first-test
+[license-link]:       https://github.com/official-stockfish/Stockfish/blob/master/Copying.txt
+[programming-link]:   https://www.chessprogramming.org/Main_Page
+[programmingsf-link]: https://www.chessprogramming.org/Stockfish
+[readme-link]:        https://github.com/official-stockfish/Stockfish/blob/master/README.md
+[release-link]:       https://github.com/official-stockfish/Stockfish/releases/latest
+[src-link]:           https://github.com/official-stockfish/Stockfish/tree/master/src
+[stockfish128-logo]:  https://stockfishchess.org/images/logo/icon_128x128.png
+[uci-link]:           https://backscattering.de/chess/uci/
+[website-link]:       https://stockfishchess.org
+[website-blog-link]:  https://stockfishchess.org/blog/
+[wiki-link]:          https://github.com/official-stockfish/Stockfish/wiki
+[wiki-usage-link]:    https://github.com/official-stockfish/Stockfish/wiki/Download-and-usage
+[wiki-compile-link]:  https://github.com/official-stockfish/Stockfish/wiki/Compiling-from-source
+[wiki-commands-link]: https://github.com/official-stockfish/Stockfish/wiki/Commands
+[worker-link]:        https://github.com/glinscott/fishtest/wiki/Running-the-worker
+
+[build-badge]:        https://img.shields.io/github/actions/workflow/status/official-stockfish/Stockfish/stockfish.yml?branch=master&style=for-the-badge&label=stockfish&logo=github
+[commits-badge]:      https://img.shields.io/github/commits-since/official-stockfish/Stockfish/latest?style=for-the-badge
+[discord-badge]:      https://img.shields.io/discord/435943710472011776?style=for-the-badge&label=discord&logo=Discord
+[fishtest-badge]:     https://img.shields.io/website?style=for-the-badge&down_color=red&down_message=Offline&label=Fishtest&up_color=success&up_message=Online&url=https%3A%2F%2Ftests.stockfishchess.org%2Ftests%2Ffinished
+[license-badge]:      https://img.shields.io/github/license/official-stockfish/Stockfish?style=for-the-badge&label=license&color=success
+[release-badge]:      https://img.shields.io/github/v/release/official-stockfish/Stockfish?style=for-the-badge&label=official%20release
+[website-badge]:      https://img.shields.io/website?style=for-the-badge&down_color=red&down_message=Offline&label=website&up_color=success&up_message=Online&url=https%3A%2F%2Fstockfishchess.org

src/Stockfish/Top CPU Contributors.txt

@@ -0,0 +1,267 @@
+Contributors to Fishtest with >10,000 CPU hours, as of 2023-06-20.
+Thank you!
+
+Username                                CPU Hours     Games played
+------------------------------------------------------------------
+noobpwnftw                               37457426       2850540907
+technologov                              14135647        742892808
+linrock                                   4423514        303254809
+mlang                                     3026000        200065824
+dew                                       1689162        100033738
+okrout                                    1578136        148855886
+pemo                                      1508508         48814305
+grandphish2                               1461406         91540343
+TueRens                                   1194790         70400852
+JojoM                                      947612         61773190
+tvijlbrief                                 796125         51897690
+sebastronomy                               742434         38218524
+mibere                                     703840         46867607
+gvreuls                                    651026         42988582
+oz                                         543438         39314736
+cw                                         517858         34869755
+fastgm                                     503862         30260818
+leszek                                     467278         33514883
+CSU_Dynasty                                464940         31177118
+ctoks                                      434416         28506889
+crunchy                                    427035         27344275
+maximmasiutin                              424795         26577722
+bcross                                     415722         29060963
+olafm                                      395922         32268020
+rpngn                                      348378         24560289
+velislav                                   342567         22138992
+Fisherman                                  327231         21829379
+mgrabiak                                   300612         20608380
+Dantist                                    296386         18031762
+nordlandia                                 246201         16189678
+robal                                      241300         15656382
+marrco                                     234581         17714473
+ncfish1                                    227517         15233777
+glinscott                                  208125         13277240
+drabel                                     204167         13930674
+mhoram                                     202894         12601997
+bking_US                                   198894         11876016
+Thanar                                     179852         12365359
+vdv                                        175544          9904472
+spams                                      157128         10319326
+sqrt2                                      147963          9724586
+DesolatedDodo                              146350          9536172
+Calis007                                   143165          9478764
+vdbergh                                    138650          9064413
+CoffeeOne                                  137100          5024116
+armo9494                                   136191          9460264
+malala                                     136182          8002293
+xoto                                       133759          9159372
+davar                                      129023          8376525
+DMBK                                       122960          8980062
+dsmith                                     122059          7570238
+amicic                                     119661          7938029
+Data                                       113305          8220352
+BrunoBanani                                112960          7436849
+CypressChess                               108331          7759788
+skiminki                                   107583          7218170
+jcAEie                                     105675          8238962
+MaZePallas                                 102823          6633619
+sterni1971                                 100532          5880772
+sunu                                       100167          7040199
+zeryl                                       99331          6221261
+thirdlife                                   99124          2242380
+ElbertoOne                                  99028          7023771
+cuistot                                     98853          6069816
+bigpen0r                                    94809          6529203
+brabos                                      92118          6186135
+Wolfgang                                    91939          6105872
+psk                                         89957          5984901
+sschnee                                     88235          5268000
+racerschmacer                               85805          6122790
+Fifis                                       85722          5709729
+Dubslow                                     84986          6042456
+Vizvezdenec                                 83761          5344740
+0x3C33                                      82614          5271253
+BRAVONE                                     81239          5054681
+nssy                                        76497          5259388
+jromang                                     76106          5236025
+teddybaer                                   75125          5407666
+tolkki963                                   74762          5149662
+megaman7de                                  74351          4940352
+Wencey                                      74181          4711488
+Pking_cda                                   73776          5293873
+yurikvelo                                   73150          5004382
+markkulix                                   72607          5304642
+Bobo1239                                    70579          4794999
+solarlight                                  70517          5028306
+dv8silencer                                 70287          3883992
+manap                                       66273          4121774
+tinker                                      64333          4268790
+qurashee                                    61208          3429862
+Mineta                                      59357          4418202
+Spprtr                                      58723          3911011
+AGI                                         58147          4325994
+robnjr                                      57262          4053117
+Freja                                       56938          3733019
+MaxKlaxxMiner                               56879          3423958
+MarcusTullius                               56746          3762951
+ttruscott                                   56010          3680085
+rkl                                         55132          4164467
+renouve                                     53811          3501516
+javran                                      53785          4627608
+finfish                                     51360          3370515
+eva42                                       51272          3599691
+eastorwest                                  51117          3454811
+rap                                         49985          3219146
+pb00067                                     49733          3298934
+OuaisBla                                    48626          3445134
+ronaldjerum                                 47654          3240695
+biffhero                                    46564          3111352
+VoyagerOne                                  45476          3452465
+jmdana                                      44893          3065205
+maposora                                    44597          4039578
+oryx                                        44570          3454238
+speedycpu                                   43842          3003273
+jbwiebe                                     43305          2805433
+GPUex                                       42378          3133332
+Antihistamine                               41788          2761312
+mhunt                                       41735          2691355
+homyur                                      39893          2850481
+gri                                         39871          2515779
+Garf                                        37741          2999686
+SC                                          37299          2731694
+csnodgrass                                  36207          2688994
+strelock                                    34716          2074055
+szupaw                                      34102          2880346
+EthanOConnor                                33370          2090311
+slakovv                                     32915          2021889
+Gelma                                       31771          1551204
+gopeto                                      31671          2060990
+kdave                                       31157          2198362
+manapbk                                     30987          1810399
+Prcuvu                                      30377          2170122
+anst                                        30301          2190091
+jkiiski                                     30136          1904470
+spcc                                        29925          1901692
+hyperbolic.tom                              29840          2017394
+chuckstablers                               29659          2093438
+Pyafue                                      29650          1902349
+belzedar94                                  28846          1811530
+chriswk                                     26902          1868317
+xwziegtm                                    26897          2124586
+achambord                                   26582          1767323
+Patrick_G                                   26276          1801617
+yorkman                                     26193          1992080
+Ulysses                                     25288          1689730
+SFTUser                                     25182          1675689
+nabildanial                                 24942          1519409
+Sharaf_DG                                   24765          1786697
+Maxim                                       24705          1502062
+rodneyc                                     24376          1416402
+agg177                                      23890          1395014
+Goatminola                                  23763          1956036
+Ente                                        23639          1671638
+Jopo12321                                   23467          1483172
+JanErik                                     23408          1703875
+Isidor                                      23388          1680691
+Norabor                                     23371          1603244
+cisco2015                                   22920          1763301
+jsys14                                      22824          1591906
+Zirie                                       22542          1472937
+team-oh                                     22272          1636708
+Roady                                       22220          1465606
+MazeOfGalious                               21978          1629593
+sg4032                                      21947          1643353
+ianh2105                                    21725          1632562
+xor12                                       21628          1680365
+dex                                         21612          1467203
+nesoneg                                     21494          1463031
+user213718                                  21454          1404128
+sphinx                                      21211          1384728
+AndreasKrug                                 21097          1634811
+jjoshua2                                    21001          1423089
+Zake9298                                    20938          1565848
+horst.prack                                 20878          1465656
+0xB00B1ES                                   20590          1208666
+j3corre                                     20405           941444
+Adrian.Schmidt123                           20316          1281436
+wei                                         19973          1745989
+notchris                                    19958          1800128
+Serpensin                                   19840          1697528
+Gaster319                                   19712          1677310
+fishtester                                  19617          1257388
+rstoesser                                   19569          1293588
+eudhan                                      19274          1283717
+votoanthuan                                 19108          1609992
+vulcan                                      18871          1729392
+Karpovbot                                   18766          1053178
+qoo_charly_cai                              18543          1284937
+jundery                                     18445          1115855
+ville                                       17883          1384026
+chris                                       17698          1487385
+purplefishies                               17595          1092533
+dju                                         17414           981289
+iisiraider                                  17275          1049015
+DragonLord                                  17014          1162790
+redstone59                                  16842          1461780
+Alb11747                                    16787          1213926
+IgorLeMasson                                16064          1147232
+Karby                                       15982           979610
+scuzzi                                      15757           968735
+ako027ako                                   15671          1173203
+Nikolay.IT                                  15154          1068349
+Andrew Grant                                15114           895539
+Naven94                                     15054           834762
+OssumOpossum                                14857          1007129
+ZacHFX                                      14783          1021842
+enedene                                     14476           905279
+bpfliegel                                   14233           882523
+mpx86                                       14019           759568
+jpulman                                     13982           870599
+Skiff84                                     13826           721996
+crocogoat                                   13803          1117422
+Nesa92                                      13786          1114691
+joster                                      13710           946160
+mbeier                                      13650          1044928
+Hjax                                        13535           915487
+Nullvalue                                   13468          1140498
+Dark_wizzie                                 13422          1007152
+Rudolphous                                  13244           883140
+pirt                                        13100          1009897
+Machariel                                   13010           863104
+infinigon                                   12991           943216
+mabichito                                   12903           749391
+thijsk                                      12886           722107
+AdrianSA                                    12860           804972
+Flopzee                                     12698           894821
+korposzczur                                 12606           838168
+fatmurphy                                   12547           853210
+SapphireBrand                               12416           969604
+Oakwen                                      12399           844109
+deflectooor                                 12386           579392
+modolief                                    12386           896470
+Farseer                                     12249           694108
+Jackfish                                    12213           805008
+pgontarz                                    12151           848794
+dbernier                                    12103           860824
+getraideBFF                                 12072          1024966
+stocky                                      11954           699440
+mschmidt                                    11941           803401
+MooTheCow                                   11870           773598
+FormazChar                                  11766           885707
+whelanh                                     11557           245188
+3cho                                        11494          1031076
+infinity                                    11470           727027
+aga                                         11412           695127
+torbjo                                      11395           729145
+Thomas A. Anderson                          11372           732094
+savage84                                    11358           670860
+d64                                         11263           789184
+ali-al-zhrani                               11245           779246
+snicolet                                    11106           869170
+dapper                                      11032           771402
+ols                                         10947           624903
+Karmatron                                   10828           677458
+basepi                                      10637           744851
+Cubox                                       10621           826448
+michaelrpg                                  10509           739239
+OIVAS7572                                   10420           995586
+jojo2357                                    10419           929708
+WoodMan777                                  10380           873720
+Garruk                                      10365           706465
+dzjp                                        10343           732529

src/Stockfish/src/benchmark.cpp

@@ -0,0 +1,177 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include "benchmark.h"
+
+#include <fstream>
+#include <iostream>
+#include <istream>
+#include <vector>
+
+#include "position.h"
+
+using namespace std;
+
+namespace {
+
+const vector<string> Defaults = {
+  "setoption name UCI_Chess960 value false",
+  "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1",
+  "r3k2r/p1ppqpb1/bn2pnp1/3PN3/1p2P3/2N2Q1p/PPPBBPPP/R3K2R w KQkq - 0 10",
+  "8/2p5/3p4/KP5r/1R3p1k/8/4P1P1/8 w - - 0 11",
+  "4rrk1/pp1n3p/3q2pQ/2p1pb2/2PP4/2P3N1/P2B2PP/4RRK1 b - - 7 19",
+  "rq3rk1/ppp2ppp/1bnpb3/3N2B1/3NP3/7P/PPPQ1PP1/2KR3R w - - 7 14 moves d4e6",
+  "r1bq1r1k/1pp1n1pp/1p1p4/4p2Q/4Pp2/1BNP4/PPP2PPP/3R1RK1 w - - 2 14 moves g2g4",
+  "r3r1k1/2p2ppp/p1p1bn2/8/1q2P3/2NPQN2/PPP3PP/R4RK1 b - - 2 15",
+  "r1bbk1nr/pp3p1p/2n5/1N4p1/2Np1B2/8/PPP2PPP/2KR1B1R w kq - 0 13",
+  "r1bq1rk1/ppp1nppp/4n3/3p3Q/3P4/1BP1B3/PP1N2PP/R4RK1 w - - 1 16",
+  "4r1k1/r1q2ppp/ppp2n2/4P3/5Rb1/1N1BQ3/PPP3PP/R5K1 w - - 1 17",
+  "2rqkb1r/ppp2p2/2npb1p1/1N1Nn2p/2P1PP2/8/PP2B1PP/R1BQK2R b KQ - 0 11",
+  "r1bq1r1k/b1p1npp1/p2p3p/1p6/3PP3/1B2NN2/PP3PPP/R2Q1RK1 w - - 1 16",
+  "3r1rk1/p5pp/bpp1pp2/8/q1PP1P2/b3P3/P2NQRPP/1R2B1K1 b - - 6 22",
+  "r1q2rk1/2p1bppp/2Pp4/p6b/Q1PNp3/4B3/PP1R1PPP/2K4R w - - 2 18",
+  "4k2r/1pb2ppp/1p2p3/1R1p4/3P4/2r1PN2/P4PPP/1R4K1 b - - 3 22",
+  "3q2k1/pb3p1p/4pbp1/2r5/PpN2N2/1P2P2P/5PP1/Q2R2K1 b - - 4 26",
+  "6k1/6p1/6Pp/ppp5/3pn2P/1P3K2/1PP2P2/3N4 b - - 0 1",
+  "3b4/5kp1/1p1p1p1p/pP1PpP1P/P1P1P3/3KN3/8/8 w - - 0 1",
+  "2K5/p7/7P/5pR1/8/5k2/r7/8 w - - 0 1 moves g5g6 f3e3 g6g5 e3f3",
+  "8/6pk/1p6/8/PP3p1p/5P2/4KP1q/3Q4 w - - 0 1",
+  "7k/3p2pp/4q3/8/4Q3/5Kp1/P6b/8 w - - 0 1",
+  "8/2p5/8/2kPKp1p/2p4P/2P5/3P4/8 w - - 0 1",
+  "8/1p3pp1/7p/5P1P/2k3P1/8/2K2P2/8 w - - 0 1",
+  "8/pp2r1k1/2p1p3/3pP2p/1P1P1P1P/P5KR/8/8 w - - 0 1",
+  "8/3p4/p1bk3p/Pp6/1Kp1PpPp/2P2P1P/2P5/5B2 b - - 0 1",
+  "5k2/7R/4P2p/5K2/p1r2P1p/8/8/8 b - - 0 1",
+  "6k1/6p1/P6p/r1N5/5p2/7P/1b3PP1/4R1K1 w - - 0 1",
+  "1r3k2/4q3/2Pp3b/3Bp3/2Q2p2/1p1P2P1/1P2KP2/3N4 w - - 0 1",
+  "6k1/4pp1p/3p2p1/P1pPb3/R7/1r2P1PP/3B1P2/6K1 w - - 0 1",
+  "8/3p3B/5p2/5P2/p7/PP5b/k7/6K1 w - - 0 1",
+  "5rk1/q6p/2p3bR/1pPp1rP1/1P1Pp3/P3B1Q1/1K3P2/R7 w - - 93 90",
+  "4rrk1/1p1nq3/p7/2p1P1pp/3P2bp/3Q1Bn1/PPPB4/1K2R1NR w - - 40 21",
+  "r3k2r/3nnpbp/q2pp1p1/p7/Pp1PPPP1/4BNN1/1P5P/R2Q1RK1 w kq - 0 16",
+  "3Qb1k1/1r2ppb1/pN1n2q1/Pp1Pp1Pr/4P2p/4BP2/4B1R1/1R5K b - - 11 40",
+  "4k3/3q1r2/1N2r1b1/3ppN2/2nPP3/1B1R2n1/2R1Q3/3K4 w - - 5 1",
+
+  // 5-man positions
+  "8/8/8/8/5kp1/P7/8/1K1N4 w - - 0 1",     // Kc2 - mate
+  "8/8/8/5N2/8/p7/8/2NK3k w - - 0 1",      // Na2 - mate
+  "8/3k4/8/8/8/4B3/4KB2/2B5 w - - 0 1",    // draw
+
+  // 6-man positions
+  "8/8/1P6/5pr1/8/4R3/7k/2K5 w - - 0 1",   // Re5 - mate
+  "8/2p4P/8/kr6/6R1/8/8/1K6 w - - 0 1",    // Ka2 - mate
+  "8/8/3P3k/8/1p6/8/1P6/1K3n2 b - - 0 1",  // Nd2 - draw
+
+  // 7-man positions
+  "8/R7/2q5/8/6k1/8/1P5p/K6R w - - 0 124", // Draw
+
+  // Mate and stalemate positions
+  "6k1/3b3r/1p1p4/p1n2p2/1PPNpP1q/P3Q1p1/1R1RB1P1/5K2 b - - 0 1",
+  "r2r1n2/pp2bk2/2p1p2p/3q4/3PN1QP/2P3R1/P4PP1/5RK1 w - - 0 1",
+  "8/8/8/8/8/6k1/6p1/6K1 w - -",
+  "7k/7P/6K1/8/3B4/8/8/8 b - -",
+
+  // Chess 960
+  "setoption name UCI_Chess960 value true",
+  "bbqnnrkr/pppppppp/8/8/8/8/PPPPPPPP/BBQNNRKR w HFhf - 0 1 moves g2g3 d7d5 d2d4 c8h3 c1g5 e8d6 g5e7 f7f6",
+  "nqbnrkrb/pppppppp/8/8/8/8/PPPPPPPP/NQBNRKRB w KQkq - 0 1",
+  "setoption name UCI_Chess960 value false"
+};
+
+} // namespace
+
+namespace Stockfish {
+
+/// setup_bench() builds a list of UCI commands to be run by bench. There
+/// are five parameters: TT size in MB, number of search threads that
+/// should be used, the limit value spent for each position, a file name
+/// where to look for positions in FEN format, the type of the limit:
+/// depth, perft, nodes and movetime (in millisecs), and evaluation type
+/// mixed (default), classical, NNUE.
+///
+/// bench -> search default positions up to depth 13
+/// bench 64 1 15 -> search default positions up to depth 15 (TT = 64MB)
+/// bench 64 4 5000 current movetime -> search current position with 4 threads for 5 sec
+/// bench 64 1 100000 default nodes -> search default positions for 100K nodes each
+/// bench 16 1 5 default perft -> run a perft 5 on default positions
+
+vector<string> setup_bench(const Position& current, istream& is) {
+
+  vector<string> fens, list;
+  string go, token;
+
+  // Assign default values to missing arguments
+  string ttSize    = (is >> token) ? token : "16";
+  string threads   = (is >> token) ? token : "1";
+  string limit     = (is >> token) ? token : "13";
+  string fenFile   = (is >> token) ? token : "default";
+  string limitType = (is >> token) ? token : "depth";
+  string evalType  = (is >> token) ? token : "mixed";
+
+  go = limitType == "eval" ? "eval" : "go " + limitType + " " + limit;
+
+  if (fenFile == "default")
+      fens = Defaults;
+
+  else if (fenFile == "current")
+      fens.push_back(current.fen());
+
+  else
+  {
+      string fen;
+      ifstream file(fenFile);
+
+      if (!file.is_open())
+      {
+          cerr << "Unable to open file " << fenFile << endl;
+          exit(EXIT_FAILURE);
+      }
+
+      while (getline(file, fen))
+          if (!fen.empty())
+              fens.push_back(fen);
+
+      file.close();
+  }
+
+  list.emplace_back("setoption name Threads value " + threads);
+  list.emplace_back("setoption name Hash value " + ttSize);
+  list.emplace_back("ucinewgame");
+
+  size_t posCounter = 0;
+
+  for (const string& fen : fens)
+      if (fen.find("setoption") != string::npos)
+          list.emplace_back(fen);
+      else
+      {
+          if (evalType == "classical" || (evalType == "mixed" && posCounter % 2 == 0))
+              list.emplace_back("setoption name Use NNUE value false");
+          else if (evalType == "NNUE" || (evalType == "mixed" && posCounter % 2 != 0))
+              list.emplace_back("setoption name Use NNUE value true");
+          list.emplace_back("position fen " + fen);
+          list.emplace_back(go);
+          ++posCounter;
+      }
+
+  list.emplace_back("setoption name Use NNUE value true");
+
+  return list;
+}
+
+} // namespace Stockfish

src/Stockfish/src/benchmark.h

@@ -0,0 +1,34 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef BENCHMARK_H_INCLUDED
+#define BENCHMARK_H_INCLUDED
+
+#include <iosfwd>
+#include <string>
+#include <vector>
+
+namespace Stockfish {
+
+class Position;
+
+std::vector<std::string> setup_bench(const Position&, std::istream&);
+
+} // namespace Stockfish
+
+#endif // #ifndef BENCHMARK_H_INCLUDED

src/Stockfish/src/bitbase.cpp

@@ -0,0 +1,172 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include <cassert>
+#include <vector>
+#include <bitset>
+
+#include "bitboard.h"
+#include "types.h"
+
+namespace Stockfish {
+
+namespace {
+
+  // There are 24 possible pawn squares: files A to D and ranks from 2 to 7.
+  // Positions with the pawn on files E to H will be mirrored before probing.
+  constexpr unsigned MAX_INDEX = 2*24*64*64; // stm * psq * wksq * bksq = 196608
+
+  std::bitset<MAX_INDEX> KPKBitbase;
+
+  // A KPK bitbase index is an integer in [0, IndexMax] range
+  //
+  // Information is mapped in a way that minimizes the number of iterations:
+  //
+  // bit  0- 5: white king square (from SQ_A1 to SQ_H8)
+  // bit  6-11: black king square (from SQ_A1 to SQ_H8)
+  // bit    12: side to move (WHITE or BLACK)
+  // bit 13-14: white pawn file (from FILE_A to FILE_D)
+  // bit 15-17: white pawn RANK_7 - rank (from RANK_7 - RANK_7 to RANK_7 - RANK_2)
+  unsigned index(Color stm, Square bksq, Square wksq, Square psq) {
+    return int(wksq) | (bksq << 6) | (stm << 12) | (file_of(psq) << 13) | ((RANK_7 - rank_of(psq)) << 15);
+  }
+
+  enum Result {
+    INVALID = 0,
+    UNKNOWN = 1,
+    DRAW    = 2,
+    WIN     = 4
+  };
+
+  Result& operator|=(Result& r, Result v) { return r = Result(r | v); }
+
+  struct KPKPosition {
+    KPKPosition() = default;
+    explicit KPKPosition(unsigned idx);
+    operator Result() const { return result; }
+    Result classify(const std::vector<KPKPosition>& db);
+
+    Color stm;
+    Square ksq[COLOR_NB], psq;
+    Result result;
+  };
+
+} // namespace
+
+bool Bitbases::probe(Square wksq, Square wpsq, Square bksq, Color stm) {
+
+  assert(file_of(wpsq) <= FILE_D);
+
+  return KPKBitbase[index(stm, bksq, wksq, wpsq)];
+}
+
+
+void Bitbases::init() {
+
+  std::vector<KPKPosition> db(MAX_INDEX);
+  unsigned idx, repeat = 1;
+
+  // Initialize db with known win / draw positions
+  for (idx = 0; idx < MAX_INDEX; ++idx)
+      db[idx] = KPKPosition(idx);
+
+  // Iterate through the positions until none of the unknown positions can be
+  // changed to either wins or draws (15 cycles needed).
+  while (repeat)
+      for (repeat = idx = 0; idx < MAX_INDEX; ++idx)
+          repeat |= (db[idx] == UNKNOWN && db[idx].classify(db) != UNKNOWN);
+
+  // Fill the bitbase with the decisive results
+  for (idx = 0; idx < MAX_INDEX; ++idx)
+      if (db[idx] == WIN)
+          KPKBitbase.set(idx);
+}
+
+namespace {
+
+  KPKPosition::KPKPosition(unsigned idx) {
+
+    ksq[WHITE] = Square((idx >>  0) & 0x3F);
+    ksq[BLACK] = Square((idx >>  6) & 0x3F);
+    stm        = Color ((idx >> 12) & 0x01);
+    psq        = make_square(File((idx >> 13) & 0x3), Rank(RANK_7 - ((idx >> 15) & 0x7)));
+
+    // Invalid if two pieces are on the same square or if a king can be captured
+    if (   distance(ksq[WHITE], ksq[BLACK]) <= 1
+        || ksq[WHITE] == psq
+        || ksq[BLACK] == psq
+        || (stm == WHITE && (pawn_attacks_bb(WHITE, psq) & ksq[BLACK])))
+        result = INVALID;
+
+    // Win if the pawn can be promoted without getting captured
+    else if (   stm == WHITE
+             && rank_of(psq) == RANK_7
+             && ksq[WHITE] != psq + NORTH
+             && (    distance(ksq[BLACK], psq + NORTH) > 1
+                 || (distance(ksq[WHITE], psq + NORTH) == 1)))
+        result = WIN;
+
+    // Draw if it is stalemate or the black king can capture the pawn
+    else if (   stm == BLACK
+             && (  !(attacks_bb<KING>(ksq[BLACK]) & ~(attacks_bb<KING>(ksq[WHITE]) | pawn_attacks_bb(WHITE, psq)))
+                 || (attacks_bb<KING>(ksq[BLACK]) & ~attacks_bb<KING>(ksq[WHITE]) & psq)))
+        result = DRAW;
+
+    // Position will be classified later
+    else
+        result = UNKNOWN;
+  }
+
+  Result KPKPosition::classify(const std::vector<KPKPosition>& db) {
+
+    // White to move: If one move leads to a position classified as WIN, the result
+    // of the current position is WIN. If all moves lead to positions classified
+    // as DRAW, the current position is classified as DRAW, otherwise the current
+    // position is classified as UNKNOWN.
+    //
+    // Black to move: If one move leads to a position classified as DRAW, the result
+    // of the current position is DRAW. If all moves lead to positions classified
+    // as WIN, the position is classified as WIN, otherwise the current position is
+    // classified as UNKNOWN.
+    const Result Good = (stm == WHITE ? WIN   : DRAW);
+    const Result Bad  = (stm == WHITE ? DRAW  : WIN);
+
+    Result r = INVALID;
+    Bitboard b = attacks_bb<KING>(ksq[stm]);
+
+    while (b)
+        r |= stm == WHITE ? db[index(BLACK, ksq[BLACK], pop_lsb(b), psq)]
+                          : db[index(WHITE, pop_lsb(b), ksq[WHITE], psq)];
+
+    if (stm == WHITE)
+    {
+        if (rank_of(psq) < RANK_7)      // Single push
+            r |= db[index(BLACK, ksq[BLACK], ksq[WHITE], psq + NORTH)];
+
+        if (   rank_of(psq) == RANK_2   // Double push
+            && psq + NORTH != ksq[WHITE]
+            && psq + NORTH != ksq[BLACK])
+            r |= db[index(BLACK, ksq[BLACK], ksq[WHITE], psq + NORTH + NORTH)];
+    }
+
+    return result = r & Good  ? Good  : r & UNKNOWN ? UNKNOWN : Bad;
+  }
+
+} // namespace
+
+} // namespace Stockfish

src/Stockfish/src/bitboard.cpp

@@ -0,0 +1,218 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include <algorithm>
+#include <bitset>
+
+#include "bitboard.h"
+#include "misc.h"
+
+namespace Stockfish {
+
+uint8_t PopCnt16[1 << 16];
+uint8_t SquareDistance[SQUARE_NB][SQUARE_NB];
+
+Bitboard LineBB[SQUARE_NB][SQUARE_NB];
+Bitboard BetweenBB[SQUARE_NB][SQUARE_NB];
+Bitboard PseudoAttacks[PIECE_TYPE_NB][SQUARE_NB];
+Bitboard PawnAttacks[COLOR_NB][SQUARE_NB];
+
+Magic RookMagics[SQUARE_NB];
+Magic BishopMagics[SQUARE_NB];
+
+namespace {
+
+  Bitboard RookTable[0x19000];  // To store rook attacks
+  Bitboard BishopTable[0x1480]; // To store bishop attacks
+
+  void init_magics(PieceType pt, Bitboard table[], Magic magics[]);
+
+}
+
+/// safe_destination() returns the bitboard of target square for the given step
+/// from the given square. If the step is off the board, returns empty bitboard.
+
+inline Bitboard safe_destination(Square s, int step) {
+    Square to = Square(s + step);
+    return is_ok(to) && distance(s, to) <= 2 ? square_bb(to) : Bitboard(0);
+}
+
+
+/// Bitboards::pretty() returns an ASCII representation of a bitboard suitable
+/// to be printed to standard output. Useful for debugging.
+
+std::string Bitboards::pretty(Bitboard b) {
+
+  std::string s = "+---+---+---+---+---+---+---+---+\n";
+
+  for (Rank r = RANK_8; r >= RANK_1; --r)
+  {
+      for (File f = FILE_A; f <= FILE_H; ++f)
+          s += b & make_square(f, r) ? "| X " : "|   ";
+
+      s += "| " + std::to_string(1 + r) + "\n+---+---+---+---+---+---+---+---+\n";
+  }
+  s += "  a   b   c   d   e   f   g   h\n";
+
+  return s;
+}
+
+
+/// Bitboards::init() initializes various bitboard tables. It is called at
+/// startup and relies on global objects to be already zero-initialized.
+
+void Bitboards::init() {
+
+  for (unsigned i = 0; i < (1 << 16); ++i)
+      PopCnt16[i] = uint8_t(std::bitset<16>(i).count());
+
+  for (Square s1 = SQ_A1; s1 <= SQ_H8; ++s1)
+      for (Square s2 = SQ_A1; s2 <= SQ_H8; ++s2)
+          SquareDistance[s1][s2] = std::max(distance<File>(s1, s2), distance<Rank>(s1, s2));
+
+  init_magics(ROOK, RookTable, RookMagics);
+  init_magics(BISHOP, BishopTable, BishopMagics);
+
+  for (Square s1 = SQ_A1; s1 <= SQ_H8; ++s1)
+  {
+      PawnAttacks[WHITE][s1] = pawn_attacks_bb<WHITE>(square_bb(s1));
+      PawnAttacks[BLACK][s1] = pawn_attacks_bb<BLACK>(square_bb(s1));
+
+      for (int step : {-9, -8, -7, -1, 1, 7, 8, 9} )
+         PseudoAttacks[KING][s1] |= safe_destination(s1, step);
+
+      for (int step : {-17, -15, -10, -6, 6, 10, 15, 17} )
+         PseudoAttacks[KNIGHT][s1] |= safe_destination(s1, step);
+
+      PseudoAttacks[QUEEN][s1]  = PseudoAttacks[BISHOP][s1] = attacks_bb<BISHOP>(s1, 0);
+      PseudoAttacks[QUEEN][s1] |= PseudoAttacks[  ROOK][s1] = attacks_bb<  ROOK>(s1, 0);
+
+      for (PieceType pt : { BISHOP, ROOK })
+          for (Square s2 = SQ_A1; s2 <= SQ_H8; ++s2)
+          {
+              if (PseudoAttacks[pt][s1] & s2)
+              {
+                  LineBB[s1][s2]    = (attacks_bb(pt, s1, 0) & attacks_bb(pt, s2, 0)) | s1 | s2;
+                  BetweenBB[s1][s2] = (attacks_bb(pt, s1, square_bb(s2)) & attacks_bb(pt, s2, square_bb(s1)));
+              }
+              BetweenBB[s1][s2] |= s2;
+          }
+  }
+}
+
+namespace {
+
+  Bitboard sliding_attack(PieceType pt, Square sq, Bitboard occupied) {
+
+    Bitboard attacks = 0;
+    Direction   RookDirections[4] = {NORTH, SOUTH, EAST, WEST};
+    Direction BishopDirections[4] = {NORTH_EAST, SOUTH_EAST, SOUTH_WEST, NORTH_WEST};
+
+    for (Direction d : (pt == ROOK ? RookDirections : BishopDirections))
+    {
+        Square s = sq;
+        while (safe_destination(s, d) && !(occupied & s))
+            attacks |= (s += d);
+    }
+
+    return attacks;
+  }
+
+
+  // init_magics() computes all rook and bishop attacks at startup. Magic
+  // bitboards are used to look up attacks of sliding pieces. As a reference see
+  // www.chessprogramming.org/Magic_Bitboards. In particular, here we use the so
+  // called "fancy" approach.
+
+  void init_magics(PieceType pt, Bitboard table[], Magic magics[]) {
+
+    // Optimal PRNG seeds to pick the correct magics in the shortest time
+    int seeds[][RANK_NB] = { { 8977, 44560, 54343, 38998,  5731, 95205, 104912, 17020 },
+                             {  728, 10316, 55013, 32803, 12281, 15100,  16645,   255 } };
+
+    Bitboard occupancy[4096], reference[4096], edges, b;
+    int epoch[4096] = {}, cnt = 0, size = 0;
+
+    for (Square s = SQ_A1; s <= SQ_H8; ++s)
+    {
+        // Board edges are not considered in the relevant occupancies
+        edges = ((Rank1BB | Rank8BB) & ~rank_bb(s)) | ((FileABB | FileHBB) & ~file_bb(s));
+
+        // Given a square 's', the mask is the bitboard of sliding attacks from
+        // 's' computed on an empty board. The index must be big enough to contain
+        // all the attacks for each possible subset of the mask and so is 2 power
+        // the number of 1s of the mask. Hence we deduce the size of the shift to
+        // apply to the 64 or 32 bits word to get the index.
+        Magic& m = magics[s];
+        m.mask  = sliding_attack(pt, s, 0) & ~edges;
+        m.shift = (Is64Bit ? 64 : 32) - popcount(m.mask);
+
+        // Set the offset for the attacks table of the square. We have individual
+        // table sizes for each square with "Fancy Magic Bitboards".
+        m.attacks = s == SQ_A1 ? table : magics[s - 1].attacks + size;
+
+        // Use Carry-Rippler trick to enumerate all subsets of masks[s] and
+        // store the corresponding sliding attack bitboard in reference[].
+        b = size = 0;
+        do {
+            occupancy[size] = b;
+            reference[size] = sliding_attack(pt, s, b);
+
+            if (HasPext)
+                m.attacks[pext(b, m.mask)] = reference[size];
+
+            size++;
+            b = (b - m.mask) & m.mask;
+        } while (b);
+
+        if (HasPext)
+            continue;
+
+        PRNG rng(seeds[Is64Bit][rank_of(s)]);
+
+        // Find a magic for square 's' picking up an (almost) random number
+        // until we find the one that passes the verification test.
+        for (int i = 0; i < size; )
+        {
+            for (m.magic = 0; popcount((m.magic * m.mask) >> 56) < 6; )
+                m.magic = rng.sparse_rand<Bitboard>();
+
+            // A good magic must map every possible occupancy to an index that
+            // looks up the correct sliding attack in the attacks[s] database.
+            // Note that we build up the database for square 's' as a side
+            // effect of verifying the magic. Keep track of the attempt count
+            // and save it in epoch[], little speed-up trick to avoid resetting
+            // m.attacks[] after every failed attempt.
+            for (++cnt, i = 0; i < size; ++i)
+            {
+                unsigned idx = m.index(occupancy[i]);
+
+                if (epoch[idx] < cnt)
+                {
+                    epoch[idx] = cnt;
+                    m.attacks[idx] = reference[i];
+                }
+                else if (m.attacks[idx] != reference[i])
+                    break;
+            }
+        }
+    }
+  }
+}
+
+} // namespace Stockfish

src/Stockfish/src/bitboard.h

@@ -0,0 +1,450 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef BITBOARD_H_INCLUDED
+#define BITBOARD_H_INCLUDED
+
+#include <string>
+
+#include "types.h"
+
+namespace Stockfish {
+
+namespace Bitbases {
+
+void init();
+bool probe(Square wksq, Square wpsq, Square bksq, Color us);
+
+} // namespace Stockfish::Bitbases
+
+namespace Bitboards {
+
+void init();
+std::string pretty(Bitboard b);
+
+} // namespace Stockfish::Bitboards
+
+constexpr Bitboard AllSquares = ~Bitboard(0);
+constexpr Bitboard DarkSquares = 0xAA55AA55AA55AA55ULL;
+
+constexpr Bitboard FileABB = 0x0101010101010101ULL;
+constexpr Bitboard FileBBB = FileABB << 1;
+constexpr Bitboard FileCBB = FileABB << 2;
+constexpr Bitboard FileDBB = FileABB << 3;
+constexpr Bitboard FileEBB = FileABB << 4;
+constexpr Bitboard FileFBB = FileABB << 5;
+constexpr Bitboard FileGBB = FileABB << 6;
+constexpr Bitboard FileHBB = FileABB << 7;
+
+constexpr Bitboard Rank1BB = 0xFF;
+constexpr Bitboard Rank2BB = Rank1BB << (8 * 1);
+constexpr Bitboard Rank3BB = Rank1BB << (8 * 2);
+constexpr Bitboard Rank4BB = Rank1BB << (8 * 3);
+constexpr Bitboard Rank5BB = Rank1BB << (8 * 4);
+constexpr Bitboard Rank6BB = Rank1BB << (8 * 5);
+constexpr Bitboard Rank7BB = Rank1BB << (8 * 6);
+constexpr Bitboard Rank8BB = Rank1BB << (8 * 7);
+
+constexpr Bitboard QueenSide   = FileABB | FileBBB | FileCBB | FileDBB;
+constexpr Bitboard CenterFiles = FileCBB | FileDBB | FileEBB | FileFBB;
+constexpr Bitboard KingSide    = FileEBB | FileFBB | FileGBB | FileHBB;
+constexpr Bitboard Center      = (FileDBB | FileEBB) & (Rank4BB | Rank5BB);
+
+constexpr Bitboard KingFlank[FILE_NB] = {
+  QueenSide ^ FileDBB, QueenSide, QueenSide,
+  CenterFiles, CenterFiles,
+  KingSide, KingSide, KingSide ^ FileEBB
+};
+
+extern uint8_t PopCnt16[1 << 16];
+extern uint8_t SquareDistance[SQUARE_NB][SQUARE_NB];
+
+extern Bitboard BetweenBB[SQUARE_NB][SQUARE_NB];
+extern Bitboard LineBB[SQUARE_NB][SQUARE_NB];
+extern Bitboard PseudoAttacks[PIECE_TYPE_NB][SQUARE_NB];
+extern Bitboard PawnAttacks[COLOR_NB][SQUARE_NB];
+
+
+/// Magic holds all magic bitboards relevant data for a single square
+struct Magic {
+  Bitboard  mask;
+  Bitboard  magic;
+  Bitboard* attacks;
+  unsigned  shift;
+
+  // Compute the attack's index using the 'magic bitboards' approach
+  unsigned index(Bitboard occupied) const {
+
+    if (HasPext)
+        return unsigned(pext(occupied, mask));
+
+    if (Is64Bit)
+        return unsigned(((occupied & mask) * magic) >> shift);
+
+    unsigned lo = unsigned(occupied) & unsigned(mask);
+    unsigned hi = unsigned(occupied >> 32) & unsigned(mask >> 32);
+    return (lo * unsigned(magic) ^ hi * unsigned(magic >> 32)) >> shift;
+  }
+};
+
+extern Magic RookMagics[SQUARE_NB];
+extern Magic BishopMagics[SQUARE_NB];
+
+inline Bitboard square_bb(Square s) {
+  assert(is_ok(s));
+  return (1ULL << s);
+}
+
+
+/// Overloads of bitwise operators between a Bitboard and a Square for testing
+/// whether a given bit is set in a bitboard, and for setting and clearing bits.
+
+inline Bitboard  operator&( Bitboard  b, Square s) { return b &  square_bb(s); }
+inline Bitboard  operator|( Bitboard  b, Square s) { return b |  square_bb(s); }
+inline Bitboard  operator^( Bitboard  b, Square s) { return b ^  square_bb(s); }
+inline Bitboard& operator|=(Bitboard& b, Square s) { return b |= square_bb(s); }
+inline Bitboard& operator^=(Bitboard& b, Square s) { return b ^= square_bb(s); }
+
+inline Bitboard  operator&(Square s, Bitboard b) { return b & s; }
+inline Bitboard  operator|(Square s, Bitboard b) { return b | s; }
+inline Bitboard  operator^(Square s, Bitboard b) { return b ^ s; }
+
+inline Bitboard  operator|(Square s1, Square s2) { return square_bb(s1) | s2; }
+
+constexpr bool more_than_one(Bitboard b) {
+  return b & (b - 1);
+}
+
+
+constexpr bool opposite_colors(Square s1, Square s2) {
+  return (s1 + rank_of(s1) + s2 + rank_of(s2)) & 1;
+}
+
+
+/// rank_bb() and file_bb() return a bitboard representing all the squares on
+/// the given file or rank.
+
+constexpr Bitboard rank_bb(Rank r) {
+  return Rank1BB << (8 * r);
+}
+
+constexpr Bitboard rank_bb(Square s) {
+  return rank_bb(rank_of(s));
+}
+
+constexpr Bitboard file_bb(File f) {
+  return FileABB << f;
+}
+
+constexpr Bitboard file_bb(Square s) {
+  return file_bb(file_of(s));
+}
+
+
+/// shift() moves a bitboard one or two steps as specified by the direction D
+
+template<Direction D>
+constexpr Bitboard shift(Bitboard b) {
+  return  D == NORTH      ?  b             << 8 : D == SOUTH      ?  b             >> 8
+        : D == NORTH+NORTH?  b             <<16 : D == SOUTH+SOUTH?  b             >>16
+        : D == EAST       ? (b & ~FileHBB) << 1 : D == WEST       ? (b & ~FileABB) >> 1
+        : D == NORTH_EAST ? (b & ~FileHBB) << 9 : D == NORTH_WEST ? (b & ~FileABB) << 7
+        : D == SOUTH_EAST ? (b & ~FileHBB) >> 7 : D == SOUTH_WEST ? (b & ~FileABB) >> 9
+        : 0;
+}
+
+
+/// pawn_attacks_bb() returns the squares attacked by pawns of the given color
+/// from the squares in the given bitboard.
+
+template<Color C>
+constexpr Bitboard pawn_attacks_bb(Bitboard b) {
+  return C == WHITE ? shift<NORTH_WEST>(b) | shift<NORTH_EAST>(b)
+                    : shift<SOUTH_WEST>(b) | shift<SOUTH_EAST>(b);
+}
+
+inline Bitboard pawn_attacks_bb(Color c, Square s) {
+
+  assert(is_ok(s));
+  return PawnAttacks[c][s];
+}
+
+
+/// pawn_double_attacks_bb() returns the squares doubly attacked by pawns of the
+/// given color from the squares in the given bitboard.
+
+template<Color C>
+constexpr Bitboard pawn_double_attacks_bb(Bitboard b) {
+  return C == WHITE ? shift<NORTH_WEST>(b) & shift<NORTH_EAST>(b)
+                    : shift<SOUTH_WEST>(b) & shift<SOUTH_EAST>(b);
+}
+
+
+/// adjacent_files_bb() returns a bitboard representing all the squares on the
+/// adjacent files of a given square.
+
+constexpr Bitboard adjacent_files_bb(Square s) {
+  return shift<EAST>(file_bb(s)) | shift<WEST>(file_bb(s));
+}
+
+
+/// line_bb() returns a bitboard representing an entire line (from board edge
+/// to board edge) that intersects the two given squares. If the given squares
+/// are not on a same file/rank/diagonal, the function returns 0. For instance,
+/// line_bb(SQ_C4, SQ_F7) will return a bitboard with the A2-G8 diagonal.
+
+inline Bitboard line_bb(Square s1, Square s2) {
+
+  assert(is_ok(s1) && is_ok(s2));
+
+  return LineBB[s1][s2];
+}
+
+
+/// between_bb(s1, s2) returns a bitboard representing the squares in the semi-open
+/// segment between the squares s1 and s2 (excluding s1 but including s2). If the
+/// given squares are not on a same file/rank/diagonal, it returns s2. For instance,
+/// between_bb(SQ_C4, SQ_F7) will return a bitboard with squares D5, E6 and F7, but
+/// between_bb(SQ_E6, SQ_F8) will return a bitboard with the square F8. This trick
+/// allows to generate non-king evasion moves faster: the defending piece must either
+/// interpose itself to cover the check or capture the checking piece.
+
+inline Bitboard between_bb(Square s1, Square s2) {
+
+  assert(is_ok(s1) && is_ok(s2));
+
+  return BetweenBB[s1][s2];
+}
+
+
+/// forward_ranks_bb() returns a bitboard representing the squares on the ranks in
+/// front of the given one, from the point of view of the given color. For instance,
+/// forward_ranks_bb(BLACK, SQ_D3) will return the 16 squares on ranks 1 and 2.
+
+constexpr Bitboard forward_ranks_bb(Color c, Square s) {
+  return c == WHITE ? ~Rank1BB << 8 * relative_rank(WHITE, s)
+                    : ~Rank8BB >> 8 * relative_rank(BLACK, s);
+}
+
+
+/// forward_file_bb() returns a bitboard representing all the squares along the
+/// line in front of the given one, from the point of view of the given color.
+
+constexpr Bitboard forward_file_bb(Color c, Square s) {
+  return forward_ranks_bb(c, s) & file_bb(s);
+}
+
+
+/// pawn_attack_span() returns a bitboard representing all the squares that can
+/// be attacked by a pawn of the given color when it moves along its file, starting
+/// from the given square.
+
+constexpr Bitboard pawn_attack_span(Color c, Square s) {
+  return forward_ranks_bb(c, s) & adjacent_files_bb(s);
+}
+
+
+/// passed_pawn_span() returns a bitboard which can be used to test if a pawn of
+/// the given color and on the given square is a passed pawn.
+
+constexpr Bitboard passed_pawn_span(Color c, Square s) {
+  return pawn_attack_span(c, s) | forward_file_bb(c, s);
+}
+
+
+/// aligned() returns true if the squares s1, s2 and s3 are aligned either on a
+/// straight or on a diagonal line.
+
+inline bool aligned(Square s1, Square s2, Square s3) {
+  return line_bb(s1, s2) & s3;
+}
+
+
+/// distance() functions return the distance between x and y, defined as the
+/// number of steps for a king in x to reach y.
+
+template<typename T1 = Square> inline int distance(Square x, Square y);
+template<> inline int distance<File>(Square x, Square y) { return std::abs(file_of(x) - file_of(y)); }
+template<> inline int distance<Rank>(Square x, Square y) { return std::abs(rank_of(x) - rank_of(y)); }
+template<> inline int distance<Square>(Square x, Square y) { return SquareDistance[x][y]; }
+
+inline int edge_distance(File f) { return std::min(f, File(FILE_H - f)); }
+inline int edge_distance(Rank r) { return std::min(r, Rank(RANK_8 - r)); }
+
+
+/// attacks_bb(Square) returns the pseudo attacks of the give piece type
+/// assuming an empty board.
+
+template<PieceType Pt>
+inline Bitboard attacks_bb(Square s) {
+
+  assert((Pt != PAWN) && (is_ok(s)));
+
+  return PseudoAttacks[Pt][s];
+}
+
+
+/// attacks_bb(Square, Bitboard) returns the attacks by the given piece
+/// assuming the board is occupied according to the passed Bitboard.
+/// Sliding piece attacks do not continue passed an occupied square.
+
+template<PieceType Pt>
+inline Bitboard attacks_bb(Square s, Bitboard occupied) {
+
+  assert((Pt != PAWN) && (is_ok(s)));
+
+  switch (Pt)
+  {
+  case BISHOP: return BishopMagics[s].attacks[BishopMagics[s].index(occupied)];
+  case ROOK  : return   RookMagics[s].attacks[  RookMagics[s].index(occupied)];
+  case QUEEN : return attacks_bb<BISHOP>(s, occupied) | attacks_bb<ROOK>(s, occupied);
+  default    : return PseudoAttacks[Pt][s];
+  }
+}
+
+inline Bitboard attacks_bb(PieceType pt, Square s, Bitboard occupied) {
+
+  assert((pt != PAWN) && (is_ok(s)));
+
+  switch (pt)
+  {
+  case BISHOP: return attacks_bb<BISHOP>(s, occupied);
+  case ROOK  : return attacks_bb<  ROOK>(s, occupied);
+  case QUEEN : return attacks_bb<BISHOP>(s, occupied) | attacks_bb<ROOK>(s, occupied);
+  default    : return PseudoAttacks[pt][s];
+  }
+}
+
+
+/// popcount() counts the number of non-zero bits in a bitboard
+
+inline int popcount(Bitboard b) {
+
+#ifndef USE_POPCNT
+
+  union { Bitboard bb; uint16_t u[4]; } v = { b };
+  return PopCnt16[v.u[0]] + PopCnt16[v.u[1]] + PopCnt16[v.u[2]] + PopCnt16[v.u[3]];
+
+#elif defined(_MSC_VER) || defined(__INTEL_COMPILER)
+
+  return (int)_mm_popcnt_u64(b);
+
+#else // Assumed gcc or compatible compiler
+
+  return __builtin_popcountll(b);
+
+#endif
+}
+
+
+/// lsb() and msb() return the least/most significant bit in a non-zero bitboard
+
+#if defined(__GNUC__)  // GCC, Clang, ICC
+
+inline Square lsb(Bitboard b) {
+  assert(b);
+  return Square(__builtin_ctzll(b));
+}
+
+inline Square msb(Bitboard b) {
+  assert(b);
+  return Square(63 ^ __builtin_clzll(b));
+}
+
+#elif defined(_MSC_VER)  // MSVC
+
+#ifdef _WIN64  // MSVC, WIN64
+
+inline Square lsb(Bitboard b) {
+  assert(b);
+  unsigned long idx;
+  _BitScanForward64(&idx, b);
+  return (Square) idx;
+}
+
+inline Square msb(Bitboard b) {
+  assert(b);
+  unsigned long idx;
+  _BitScanReverse64(&idx, b);
+  return (Square) idx;
+}
+
+#else  // MSVC, WIN32
+
+inline Square lsb(Bitboard b) {
+  assert(b);
+  unsigned long idx;
+
+  if (b & 0xffffffff) {
+      _BitScanForward(&idx, int32_t(b));
+      return Square(idx);
+  } else {
+      _BitScanForward(&idx, int32_t(b >> 32));
+      return Square(idx + 32);
+  }
+}
+
+inline Square msb(Bitboard b) {
+  assert(b);
+  unsigned long idx;
+
+  if (b >> 32) {
+      _BitScanReverse(&idx, int32_t(b >> 32));
+      return Square(idx + 32);
+  } else {
+      _BitScanReverse(&idx, int32_t(b));
+      return Square(idx);
+  }
+}
+
+#endif
+
+#else  // Compiler is neither GCC nor MSVC compatible
+
+#error "Compiler not supported."
+
+#endif
+
+/// least_significant_square_bb() returns the bitboard of the least significant
+/// square of a non-zero bitboard. It is equivalent to square_bb(lsb(bb)).
+
+inline Bitboard least_significant_square_bb(Bitboard b) {
+  assert(b);
+  return b & -b;
+}
+
+/// pop_lsb() finds and clears the least significant bit in a non-zero bitboard
+
+inline Square pop_lsb(Bitboard& b) {
+  assert(b);
+  const Square s = lsb(b);
+  b &= b - 1;
+  return s;
+}
+
+
+/// frontmost_sq() returns the most advanced square for the given color,
+/// requires a non-zero bitboard.
+inline Square frontmost_sq(Color c, Bitboard b) {
+  assert(b);
+  return c == WHITE ? msb(b) : lsb(b);
+}
+
+} // namespace Stockfish
+
+#endif // #ifndef BITBOARD_H_INCLUDED

src/Stockfish/src/endgame.cpp

@@ -0,0 +1,747 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include <cassert>
+
+#include "bitboard.h"
+#include "endgame.h"
+#include "movegen.h"
+
+namespace Stockfish {
+
+namespace {
+
+  // Used to drive the king towards the edge of the board
+  // in KX vs K and KQ vs KR endgames.
+  // Values range from 27 (center squares) to 90 (in the corners)
+  inline int push_to_edge(Square s) {
+      int rd = edge_distance(rank_of(s)), fd = edge_distance(file_of(s));
+      return 90 - (7 * fd * fd / 2 + 7 * rd * rd / 2);
+  }
+
+  // Used to drive the king towards A1H8 corners in KBN vs K endgames.
+  // Values range from 0 on A8H1 diagonal to 7 in A1H8 corners
+  inline int push_to_corner(Square s) {
+      return abs(7 - rank_of(s) - file_of(s));
+  }
+
+  // Drive a piece close to or away from another piece
+  inline int push_close(Square s1, Square s2) { return 140 - 20 * distance(s1, s2); }
+  inline int push_away(Square s1, Square s2) { return 120 - push_close(s1, s2); }
+
+#ifndef NDEBUG
+  bool verify_material(const Position& pos, Color c, Value npm, int pawnsCnt) {
+    return pos.non_pawn_material(c) == npm && pos.count<PAWN>(c) == pawnsCnt;
+  }
+#endif
+
+  // Map the square as if strongSide is white and strongSide's only pawn
+  // is on the left half of the board.
+  Square normalize(const Position& pos, Color strongSide, Square sq) {
+
+    assert(pos.count<PAWN>(strongSide) == 1);
+
+    if (file_of(pos.square<PAWN>(strongSide)) >= FILE_E)
+        sq = flip_file(sq);
+
+    return strongSide == WHITE ? sq : flip_rank(sq);
+  }
+
+} // namespace
+
+
+namespace Endgames {
+
+  std::pair<Map<Value>, Map<ScaleFactor>> maps;
+
+  void init() {
+
+    add<KPK>("KPK");
+    add<KNNK>("KNNK");
+    add<KBNK>("KBNK");
+    add<KRKP>("KRKP");
+    add<KRKB>("KRKB");
+    add<KRKN>("KRKN");
+    add<KQKP>("KQKP");
+    add<KQKR>("KQKR");
+    add<KNNKP>("KNNKP");
+
+    add<KRPKR>("KRPKR");
+    add<KRPKB>("KRPKB");
+    add<KBPKB>("KBPKB");
+    add<KBPKN>("KBPKN");
+    add<KBPPKB>("KBPPKB");
+    add<KRPPKRP>("KRPPKRP");
+  }
+}
+
+
+/// Mate with KX vs K. This function is used to evaluate positions with
+/// king and plenty of material vs a lone king. It simply gives the
+/// attacking side a bonus for driving the defending king towards the edge
+/// of the board, and for keeping the distance between the two kings small.
+template<>
+Value Endgame<KXK>::operator()(const Position& pos) const {
+
+  assert(verify_material(pos, weakSide, VALUE_ZERO, 0));
+  assert(!pos.checkers()); // Eval is never called when in check
+
+  // Stalemate detection with lone king
+  if (pos.side_to_move() == weakSide && !MoveList<LEGAL>(pos).size())
+      return VALUE_DRAW;
+
+  Square strongKing = pos.square<KING>(strongSide);
+  Square weakKing   = pos.square<KING>(weakSide);
+
+  Value result =  pos.non_pawn_material(strongSide)
+                + pos.count<PAWN>(strongSide) * PawnValueEg
+                + push_to_edge(weakKing)
+                + push_close(strongKing, weakKing);
+
+  if (   pos.count<QUEEN>(strongSide)
+      || pos.count<ROOK>(strongSide)
+      ||(pos.count<BISHOP>(strongSide) && pos.count<KNIGHT>(strongSide))
+      || (   (pos.pieces(strongSide, BISHOP) & ~DarkSquares)
+          && (pos.pieces(strongSide, BISHOP) &  DarkSquares)))
+      result = std::min(result + VALUE_KNOWN_WIN, VALUE_TB_WIN_IN_MAX_PLY - 1);
+
+  return strongSide == pos.side_to_move() ? result : -result;
+}
+
+
+/// Mate with KBN vs K. This is similar to KX vs K, but we have to drive the
+/// defending king towards a corner square that our bishop attacks.
+template<>
+Value Endgame<KBNK>::operator()(const Position& pos) const {
+
+  assert(verify_material(pos, strongSide, KnightValueMg + BishopValueMg, 0));
+  assert(verify_material(pos, weakSide, VALUE_ZERO, 0));
+
+  Square strongKing   = pos.square<KING>(strongSide);
+  Square strongBishop = pos.square<BISHOP>(strongSide);
+  Square weakKing     = pos.square<KING>(weakSide);
+
+  // If our bishop does not attack A1/H8, we flip the enemy king square
+  // to drive to opposite corners (A8/H1).
+
+  Value result =  (VALUE_KNOWN_WIN + 3520)
+                + push_close(strongKing, weakKing)
+                + 420 * push_to_corner(opposite_colors(strongBishop, SQ_A1) ? flip_file(weakKing) : weakKing);
+
+  assert(abs(result) < VALUE_TB_WIN_IN_MAX_PLY);
+  return strongSide == pos.side_to_move() ? result : -result;
+}
+
+
+/// KP vs K. This endgame is evaluated with the help of a bitbase
+template<>
+Value Endgame<KPK>::operator()(const Position& pos) const {
+
+  assert(verify_material(pos, strongSide, VALUE_ZERO, 1));
+  assert(verify_material(pos, weakSide, VALUE_ZERO, 0));
+
+  // Assume strongSide is white and the pawn is on files A-D
+  Square strongKing = normalize(pos, strongSide, pos.square<KING>(strongSide));
+  Square strongPawn = normalize(pos, strongSide, pos.square<PAWN>(strongSide));
+  Square weakKing   = normalize(pos, strongSide, pos.square<KING>(weakSide));
+
+  Color us = strongSide == pos.side_to_move() ? WHITE : BLACK;
+
+  if (!Bitbases::probe(strongKing, strongPawn, weakKing, us))
+      return VALUE_DRAW;
+
+  Value result = VALUE_KNOWN_WIN + PawnValueEg + Value(rank_of(strongPawn));
+
+  return strongSide == pos.side_to_move() ? result : -result;
+}
+
+
+/// KR vs KP. This is a somewhat tricky endgame to evaluate precisely without
+/// a bitbase. The function below returns drawish scores when the pawn is
+/// far advanced with support of the king, while the attacking king is far
+/// away.
+template<>
+Value Endgame<KRKP>::operator()(const Position& pos) const {
+
+  assert(verify_material(pos, strongSide, RookValueMg, 0));
+  assert(verify_material(pos, weakSide, VALUE_ZERO, 1));
+
+  Square strongKing = pos.square<KING>(strongSide);
+  Square weakKing   = pos.square<KING>(weakSide);
+  Square strongRook = pos.square<ROOK>(strongSide);
+  Square weakPawn   = pos.square<PAWN>(weakSide);
+  Square queeningSquare = make_square(file_of(weakPawn), relative_rank(weakSide, RANK_8));
+  Value result;
+
+  // If the stronger side's king is in front of the pawn, it's a win
+  if (forward_file_bb(strongSide, strongKing) & weakPawn)
+      result = RookValueEg - distance(strongKing, weakPawn);
+
+  // If the weaker side's king is too far from the pawn and the rook,
+  // it's a win.
+  else if (   distance(weakKing, weakPawn) >= 3 + (pos.side_to_move() == weakSide)
+           && distance(weakKing, strongRook) >= 3)
+      result = RookValueEg - distance(strongKing, weakPawn);
+
+  // If the pawn is far advanced and supported by the defending king,
+  // the position is drawish
+  else if (   relative_rank(strongSide, weakKing) <= RANK_3
+           && distance(weakKing, weakPawn) == 1
+           && relative_rank(strongSide, strongKing) >= RANK_4
+           && distance(strongKing, weakPawn) > 2 + (pos.side_to_move() == strongSide))
+      result = Value(80) - 8 * distance(strongKing, weakPawn);
+
+  else
+      result =  Value(200) - 8 * (  distance(strongKing, weakPawn + pawn_push(weakSide))
+                                  - distance(weakKing, weakPawn + pawn_push(weakSide))
+                                  - distance(weakPawn, queeningSquare));
+
+  return strongSide == pos.side_to_move() ? result : -result;
+}
+
+
+/// KR vs KB. This is very simple, and always returns drawish scores. The
+/// score is slightly bigger when the defending king is close to the edge.
+template<>
+Value Endgame<KRKB>::operator()(const Position& pos) const {
+
+  assert(verify_material(pos, strongSide, RookValueMg, 0));
+  assert(verify_material(pos, weakSide, BishopValueMg, 0));
+
+  Value result = Value(push_to_edge(pos.square<KING>(weakSide)));
+  return strongSide == pos.side_to_move() ? result : -result;
+}
+
+
+/// KR vs KN. The attacking side has slightly better winning chances than
+/// in KR vs KB, particularly if the king and the knight are far apart.
+template<>
+Value Endgame<KRKN>::operator()(const Position& pos) const {
+
+  assert(verify_material(pos, strongSide, RookValueMg, 0));
+  assert(verify_material(pos, weakSide, KnightValueMg, 0));
+
+  Square weakKing   = pos.square<KING>(weakSide);
+  Square weakKnight = pos.square<KNIGHT>(weakSide);
+  Value result = Value(push_to_edge(weakKing) + push_away(weakKing, weakKnight));
+  return strongSide == pos.side_to_move() ? result : -result;
+}
+
+
+/// KQ vs KP. In general, this is a win for the stronger side, but there are a
+/// few important exceptions. A pawn on 7th rank and on the A,C,F or H files
+/// with a king positioned next to it can be a draw, so in that case, we only
+/// use the distance between the kings.
+template<>
+Value Endgame<KQKP>::operator()(const Position& pos) const {
+
+  assert(verify_material(pos, strongSide, QueenValueMg, 0));
+  assert(verify_material(pos, weakSide, VALUE_ZERO, 1));
+
+  Square strongKing = pos.square<KING>(strongSide);
+  Square weakKing   = pos.square<KING>(weakSide);
+  Square weakPawn   = pos.square<PAWN>(weakSide);
+
+  Value result = Value(push_close(strongKing, weakKing));
+
+  if (   relative_rank(weakSide, weakPawn) != RANK_7
+      || distance(weakKing, weakPawn) != 1
+      || ((FileBBB | FileDBB | FileEBB | FileGBB) & weakPawn))
+      result += QueenValueEg - PawnValueEg;
+
+  return strongSide == pos.side_to_move() ? result : -result;
+}
+
+
+/// KQ vs KR. This is almost identical to KX vs K: we give the attacking
+/// king a bonus for having the kings close together, and for forcing the
+/// defending king towards the edge. If we also take care to avoid null move for
+/// the defending side in the search, this is usually sufficient to win KQ vs KR.
+template<>
+Value Endgame<KQKR>::operator()(const Position& pos) const {
+
+  assert(verify_material(pos, strongSide, QueenValueMg, 0));
+  assert(verify_material(pos, weakSide, RookValueMg, 0));
+
+  Square strongKing = pos.square<KING>(strongSide);
+  Square weakKing   = pos.square<KING>(weakSide);
+
+  Value result =  QueenValueEg
+                - RookValueEg
+                + push_to_edge(weakKing)
+                + push_close(strongKing, weakKing);
+
+  return strongSide == pos.side_to_move() ? result : -result;
+}
+
+
+/// KNN vs KP. Very drawish, but there are some mate opportunities if we can
+/// press the weakSide King to a corner before the pawn advances too much.
+template<>
+Value Endgame<KNNKP>::operator()(const Position& pos) const {
+
+  assert(verify_material(pos, strongSide, 2 * KnightValueMg, 0));
+  assert(verify_material(pos, weakSide, VALUE_ZERO, 1));
+
+  Square weakKing = pos.square<KING>(weakSide);
+  Square weakPawn = pos.square<PAWN>(weakSide);
+
+  Value result =      PawnValueEg
+               +  2 * push_to_edge(weakKing)
+               - 10 * relative_rank(weakSide, weakPawn);
+
+  return strongSide == pos.side_to_move() ? result : -result;
+}
+
+
+/// Some cases of trivial draws
+template<> Value Endgame<KNNK>::operator()(const Position&) const { return VALUE_DRAW; }
+
+
+/// KB and one or more pawns vs K. It checks for draws with rook pawns and
+/// a bishop of the wrong color. If such a draw is detected, SCALE_FACTOR_DRAW
+/// is returned. If not, the return value is SCALE_FACTOR_NONE, i.e. no scaling
+/// will be used.
+template<>
+ScaleFactor Endgame<KBPsK>::operator()(const Position& pos) const {
+
+  assert(pos.non_pawn_material(strongSide) == BishopValueMg);
+  assert(pos.count<PAWN>(strongSide) >= 1);
+
+  // No assertions about the material of weakSide, because we want draws to
+  // be detected even when the weaker side has some pawns.
+
+  Bitboard strongPawns = pos.pieces(strongSide, PAWN);
+  Bitboard allPawns = pos.pieces(PAWN);
+
+  Square strongBishop = pos.square<BISHOP>(strongSide);
+  Square weakKing = pos.square<KING>(weakSide);
+  Square strongKing = pos.square<KING>(strongSide);
+
+  // All strongSide pawns are on a single rook file?
+  if (!(strongPawns & ~FileABB) || !(strongPawns & ~FileHBB))
+  {
+      Square queeningSquare = relative_square(strongSide, make_square(file_of(lsb(strongPawns)), RANK_8));
+
+      if (   opposite_colors(queeningSquare, strongBishop)
+          && distance(queeningSquare, weakKing) <= 1)
+          return SCALE_FACTOR_DRAW;
+  }
+
+  // If all the pawns are on the same B or G file, then it's potentially a draw
+  if ((!(allPawns & ~FileBBB) || !(allPawns & ~FileGBB))
+      && pos.non_pawn_material(weakSide) == 0
+      && pos.count<PAWN>(weakSide) >= 1)
+  {
+      // Get the least advanced weakSide pawn
+      Square weakPawn = frontmost_sq(strongSide, pos.pieces(weakSide, PAWN));
+
+      // There's potential for a draw if our pawn is blocked on the 7th rank,
+      // the bishop cannot attack it or they only have one pawn left.
+      if (   relative_rank(strongSide, weakPawn) == RANK_7
+          && (strongPawns & (weakPawn + pawn_push(weakSide)))
+          && (opposite_colors(strongBishop, weakPawn) || !more_than_one(strongPawns)))
+      {
+          int strongKingDist = distance(weakPawn, strongKing);
+          int weakKingDist = distance(weakPawn, weakKing);
+
+          // It's a draw if the weak king is on its back two ranks, within 2
+          // squares of the blocking pawn and the strong king is not
+          // closer. (I think this rule only fails in practically
+          // unreachable positions such as 5k1K/6p1/6P1/8/8/3B4/8/8 w
+          // and positions where qsearch will immediately correct the
+          // problem such as 8/4k1p1/6P1/1K6/3B4/8/8/8 w).
+          if (   relative_rank(strongSide, weakKing) >= RANK_7
+              && weakKingDist <= 2
+              && weakKingDist <= strongKingDist)
+              return SCALE_FACTOR_DRAW;
+      }
+  }
+
+  return SCALE_FACTOR_NONE;
+}
+
+
+/// KQ vs KR and one or more pawns. It tests for fortress draws with a rook on
+/// the third rank defended by a pawn.
+template<>
+ScaleFactor Endgame<KQKRPs>::operator()(const Position& pos) const {
+
+  assert(verify_material(pos, strongSide, QueenValueMg, 0));
+  assert(pos.count<ROOK>(weakSide) == 1);
+  assert(pos.count<PAWN>(weakSide) >= 1);
+
+  Square strongKing = pos.square<KING>(strongSide);
+  Square weakKing   = pos.square<KING>(weakSide);
+  Square weakRook   = pos.square<ROOK>(weakSide);
+
+  if (    relative_rank(weakSide,   weakKing) <= RANK_2
+      &&  relative_rank(weakSide, strongKing) >= RANK_4
+      &&  relative_rank(weakSide,   weakRook) == RANK_3
+      && (  pos.pieces(weakSide, PAWN)
+          & attacks_bb<KING>(weakKing)
+          & pawn_attacks_bb(strongSide, weakRook)))
+          return SCALE_FACTOR_DRAW;
+
+  return SCALE_FACTOR_NONE;
+}
+
+
+/// KRP vs KR. This function knows a handful of the most important classes of
+/// drawn positions, but is far from perfect. It would probably be a good idea
+/// to add more knowledge in the future.
+///
+/// It would also be nice to rewrite the actual code for this function,
+/// which is mostly copied from Glaurung 1.x, and isn't very pretty.
+template<>
+ScaleFactor Endgame<KRPKR>::operator()(const Position& pos) const {
+
+  assert(verify_material(pos, strongSide, RookValueMg, 1));
+  assert(verify_material(pos, weakSide,   RookValueMg, 0));
+
+  // Assume strongSide is white and the pawn is on files A-D
+  Square strongKing = normalize(pos, strongSide, pos.square<KING>(strongSide));
+  Square strongRook = normalize(pos, strongSide, pos.square<ROOK>(strongSide));
+  Square strongPawn = normalize(pos, strongSide, pos.square<PAWN>(strongSide));
+  Square weakKing = normalize(pos, strongSide, pos.square<KING>(weakSide));
+  Square weakRook = normalize(pos, strongSide, pos.square<ROOK>(weakSide));
+
+  File pawnFile = file_of(strongPawn);
+  Rank pawnRank = rank_of(strongPawn);
+  Square queeningSquare = make_square(pawnFile, RANK_8);
+  int tempo = (pos.side_to_move() == strongSide);
+
+  // If the pawn is not too far advanced and the defending king defends the
+  // queening square, use the third-rank defence.
+  if (   pawnRank <= RANK_5
+      && distance(weakKing, queeningSquare) <= 1
+      && strongKing <= SQ_H5
+      && (rank_of(weakRook) == RANK_6 || (pawnRank <= RANK_3 && rank_of(strongRook) != RANK_6)))
+      return SCALE_FACTOR_DRAW;
+
+  // The defending side saves a draw by checking from behind in case the pawn
+  // has advanced to the 6th rank with the king behind.
+  if (   pawnRank == RANK_6
+      && distance(weakKing, queeningSquare) <= 1
+      && rank_of(strongKing) + tempo <= RANK_6
+      && (rank_of(weakRook) == RANK_1 || (!tempo && distance<File>(weakRook, strongPawn) >= 3)))
+      return SCALE_FACTOR_DRAW;
+
+  if (   pawnRank >= RANK_6
+      && weakKing == queeningSquare
+      && rank_of(weakRook) == RANK_1
+      && (!tempo || distance(strongKing, strongPawn) >= 2))
+      return SCALE_FACTOR_DRAW;
+
+  // White pawn on a7 and rook on a8 is a draw if black's king is on g7 or h7
+  // and the black rook is behind the pawn.
+  if (   strongPawn == SQ_A7
+      && strongRook == SQ_A8
+      && (weakKing == SQ_H7 || weakKing == SQ_G7)
+      && file_of(weakRook) == FILE_A
+      && (rank_of(weakRook) <= RANK_3 || file_of(strongKing) >= FILE_D || rank_of(strongKing) <= RANK_5))
+      return SCALE_FACTOR_DRAW;
+
+  // If the defending king blocks the pawn and the attacking king is too far
+  // away, it's a draw.
+  if (   pawnRank <= RANK_5
+      && weakKing == strongPawn + NORTH
+      && distance(strongKing, strongPawn) - tempo >= 2
+      && distance(strongKing, weakRook) - tempo >= 2)
+      return SCALE_FACTOR_DRAW;
+
+  // Pawn on the 7th rank supported by the rook from behind usually wins if the
+  // attacking king is closer to the queening square than the defending king,
+  // and the defending king cannot gain tempi by threatening the attacking rook.
+  if (   pawnRank == RANK_7
+      && pawnFile != FILE_A
+      && file_of(strongRook) == pawnFile
+      && strongRook != queeningSquare
+      && (distance(strongKing, queeningSquare) < distance(weakKing, queeningSquare) - 2 + tempo)
+      && (distance(strongKing, queeningSquare) < distance(weakKing, strongRook) + tempo))
+      return ScaleFactor(SCALE_FACTOR_MAX - 2 * distance(strongKing, queeningSquare));
+
+  // Similar to the above, but with the pawn further back
+  if (   pawnFile != FILE_A
+      && file_of(strongRook) == pawnFile
+      && strongRook < strongPawn
+      && (distance(strongKing, queeningSquare) < distance(weakKing, queeningSquare) - 2 + tempo)
+      && (distance(strongKing, strongPawn + NORTH) < distance(weakKing, strongPawn + NORTH) - 2 + tempo)
+      && (  distance(weakKing, strongRook) + tempo >= 3
+          || (    distance(strongKing, queeningSquare) < distance(weakKing, strongRook) + tempo
+              && (distance(strongKing, strongPawn + NORTH) < distance(weakKing, strongPawn) + tempo))))
+      return ScaleFactor(  SCALE_FACTOR_MAX
+                         - 8 * distance(strongPawn, queeningSquare)
+                         - 2 * distance(strongKing, queeningSquare));
+
+  // If the pawn is not far advanced and the defending king is somewhere in
+  // the pawn's path, it's probably a draw.
+  if (pawnRank <= RANK_4 && weakKing > strongPawn)
+  {
+      if (file_of(weakKing) == file_of(strongPawn))
+          return ScaleFactor(10);
+      if (   distance<File>(weakKing, strongPawn) == 1
+          && distance(strongKing, weakKing) > 2)
+          return ScaleFactor(24 - 2 * distance(strongKing, weakKing));
+  }
+  return SCALE_FACTOR_NONE;
+}
+
+template<>
+ScaleFactor Endgame<KRPKB>::operator()(const Position& pos) const {
+
+  assert(verify_material(pos, strongSide, RookValueMg, 1));
+  assert(verify_material(pos, weakSide, BishopValueMg, 0));
+
+  // Test for a rook pawn
+  if (pos.pieces(PAWN) & (FileABB | FileHBB))
+  {
+      Square weakKing = pos.square<KING>(weakSide);
+      Square weakBishop = pos.square<BISHOP>(weakSide);
+      Square strongKing = pos.square<KING>(strongSide);
+      Square strongPawn = pos.square<PAWN>(strongSide);
+      Rank pawnRank = relative_rank(strongSide, strongPawn);
+      Direction push = pawn_push(strongSide);
+
+      // If the pawn is on the 5th rank and the pawn (currently) is on
+      // the same color square as the bishop then there is a chance of
+      // a fortress. Depending on the king position give a moderate
+      // reduction or a stronger one if the defending king is near the
+      // corner but not trapped there.
+      if (pawnRank == RANK_5 && !opposite_colors(weakBishop, strongPawn))
+      {
+          int d = distance(strongPawn + 3 * push, weakKing);
+
+          if (d <= 2 && !(d == 0 && weakKing == strongKing + 2 * push))
+              return ScaleFactor(24);
+          else
+              return ScaleFactor(48);
+      }
+
+      // When the pawn has moved to the 6th rank we can be fairly sure
+      // it's drawn if the bishop attacks the square in front of the
+      // pawn from a reasonable distance and the defending king is near
+      // the corner
+      if (   pawnRank == RANK_6
+          && distance(strongPawn + 2 * push, weakKing) <= 1
+          && (attacks_bb<BISHOP>(weakBishop) & (strongPawn + push))
+          && distance<File>(weakBishop, strongPawn) >= 2)
+          return ScaleFactor(8);
+  }
+
+  return SCALE_FACTOR_NONE;
+}
+
+/// KRPP vs KRP. There is just a single rule: if the stronger side has no passed
+/// pawns and the defending king is actively placed, the position is drawish.
+template<>
+ScaleFactor Endgame<KRPPKRP>::operator()(const Position& pos) const {
+
+  assert(verify_material(pos, strongSide, RookValueMg, 2));
+  assert(verify_material(pos, weakSide,   RookValueMg, 1));
+
+  Square strongPawn1 = lsb(pos.pieces(strongSide, PAWN));
+  Square strongPawn2 = msb(pos.pieces(strongSide, PAWN));
+  Square weakKing = pos.square<KING>(weakSide);
+
+  // Does the stronger side have a passed pawn?
+  if (pos.pawn_passed(strongSide, strongPawn1) || pos.pawn_passed(strongSide, strongPawn2))
+      return SCALE_FACTOR_NONE;
+
+  Rank pawnRank = std::max(relative_rank(strongSide, strongPawn1), relative_rank(strongSide, strongPawn2));
+
+  if (   distance<File>(weakKing, strongPawn1) <= 1
+      && distance<File>(weakKing, strongPawn2) <= 1
+      && relative_rank(strongSide, weakKing) > pawnRank)
+  {
+      assert(pawnRank > RANK_1 && pawnRank < RANK_7);
+      return ScaleFactor(7 * pawnRank);
+  }
+  return SCALE_FACTOR_NONE;
+}
+
+
+/// K and two or more pawns vs K. There is just a single rule here: if all pawns
+/// are on the same rook file and are blocked by the defending king, it's a draw.
+template<>
+ScaleFactor Endgame<KPsK>::operator()(const Position& pos) const {
+
+  assert(pos.non_pawn_material(strongSide) == VALUE_ZERO);
+  assert(pos.count<PAWN>(strongSide) >= 2);
+  assert(verify_material(pos, weakSide, VALUE_ZERO, 0));
+
+  Square weakKing = pos.square<KING>(weakSide);
+  Bitboard strongPawns = pos.pieces(strongSide, PAWN);
+
+  // If all pawns are ahead of the king on a single rook file, it's a draw.
+  if (   !(strongPawns & ~(FileABB | FileHBB))
+      && !(strongPawns & ~passed_pawn_span(weakSide, weakKing)))
+      return SCALE_FACTOR_DRAW;
+
+  return SCALE_FACTOR_NONE;
+}
+
+
+/// KBP vs KB. There are two rules: if the defending king is somewhere along the
+/// path of the pawn, and the square of the king is not of the same color as the
+/// stronger side's bishop, it's a draw. If the two bishops have opposite color,
+/// it's almost always a draw.
+template<>
+ScaleFactor Endgame<KBPKB>::operator()(const Position& pos) const {
+
+  assert(verify_material(pos, strongSide, BishopValueMg, 1));
+  assert(verify_material(pos, weakSide,   BishopValueMg, 0));
+
+  Square strongPawn = pos.square<PAWN>(strongSide);
+  Square strongBishop = pos.square<BISHOP>(strongSide);
+  Square weakBishop = pos.square<BISHOP>(weakSide);
+  Square weakKing = pos.square<KING>(weakSide);
+
+  // Case 1: Defending king blocks the pawn, and cannot be driven away
+  if (   (forward_file_bb(strongSide, strongPawn) & weakKing)
+      && (   opposite_colors(weakKing, strongBishop)
+          || relative_rank(strongSide, weakKing) <= RANK_6))
+      return SCALE_FACTOR_DRAW;
+
+  // Case 2: Opposite colored bishops
+  if (opposite_colors(strongBishop, weakBishop))
+      return SCALE_FACTOR_DRAW;
+
+  return SCALE_FACTOR_NONE;
+}
+
+
+/// KBPP vs KB. It detects a few basic draws with opposite-colored bishops
+template<>
+ScaleFactor Endgame<KBPPKB>::operator()(const Position& pos) const {
+
+  assert(verify_material(pos, strongSide, BishopValueMg, 2));
+  assert(verify_material(pos, weakSide,   BishopValueMg, 0));
+
+  Square strongBishop = pos.square<BISHOP>(strongSide);
+  Square weakBishop   = pos.square<BISHOP>(weakSide);
+
+  if (!opposite_colors(strongBishop, weakBishop))
+      return SCALE_FACTOR_NONE;
+
+  Square weakKing = pos.square<KING>(weakSide);
+  Square strongPawn1 = lsb(pos.pieces(strongSide, PAWN));
+  Square strongPawn2 = msb(pos.pieces(strongSide, PAWN));
+  Square blockSq1, blockSq2;
+
+  if (relative_rank(strongSide, strongPawn1) > relative_rank(strongSide, strongPawn2))
+  {
+      blockSq1 = strongPawn1 + pawn_push(strongSide);
+      blockSq2 = make_square(file_of(strongPawn2), rank_of(strongPawn1));
+  }
+  else
+  {
+      blockSq1 = strongPawn2 + pawn_push(strongSide);
+      blockSq2 = make_square(file_of(strongPawn1), rank_of(strongPawn2));
+  }
+
+  switch (distance<File>(strongPawn1, strongPawn2))
+  {
+  case 0:
+    // Both pawns are on the same file. It's an easy draw if the defender firmly
+    // controls some square in the frontmost pawn's path.
+    if (   file_of(weakKing) == file_of(blockSq1)
+        && relative_rank(strongSide, weakKing) >= relative_rank(strongSide, blockSq1)
+        && opposite_colors(weakKing, strongBishop))
+        return SCALE_FACTOR_DRAW;
+    else
+        return SCALE_FACTOR_NONE;
+
+  case 1:
+    // Pawns on adjacent files. It's a draw if the defender firmly controls the
+    // square in front of the frontmost pawn's path, and the square diagonally
+    // behind this square on the file of the other pawn.
+    if (   weakKing == blockSq1
+        && opposite_colors(weakKing, strongBishop)
+        && (   weakBishop == blockSq2
+            || (attacks_bb<BISHOP>(blockSq2, pos.pieces()) & pos.pieces(weakSide, BISHOP))
+            || distance<Rank>(strongPawn1, strongPawn2) >= 2))
+        return SCALE_FACTOR_DRAW;
+
+    else if (   weakKing == blockSq2
+             && opposite_colors(weakKing, strongBishop)
+             && (   weakBishop == blockSq1
+                 || (attacks_bb<BISHOP>(blockSq1, pos.pieces()) & pos.pieces(weakSide, BISHOP))))
+        return SCALE_FACTOR_DRAW;
+    else
+        return SCALE_FACTOR_NONE;
+
+  default:
+    // The pawns are not on the same file or adjacent files. No scaling.
+    return SCALE_FACTOR_NONE;
+  }
+}
+
+
+/// KBP vs KN. There is a single rule: if the defending king is somewhere along
+/// the path of the pawn, and the square of the king is not of the same color as
+/// the stronger side's bishop, it's a draw.
+template<>
+ScaleFactor Endgame<KBPKN>::operator()(const Position& pos) const {
+
+  assert(verify_material(pos, strongSide, BishopValueMg, 1));
+  assert(verify_material(pos, weakSide, KnightValueMg, 0));
+
+  Square strongPawn = pos.square<PAWN>(strongSide);
+  Square strongBishop = pos.square<BISHOP>(strongSide);
+  Square weakKing = pos.square<KING>(weakSide);
+
+  if (   file_of(weakKing) == file_of(strongPawn)
+      && relative_rank(strongSide, strongPawn) < relative_rank(strongSide, weakKing)
+      && (   opposite_colors(weakKing, strongBishop)
+          || relative_rank(strongSide, weakKing) <= RANK_6))
+      return SCALE_FACTOR_DRAW;
+
+  return SCALE_FACTOR_NONE;
+}
+
+
+/// KP vs KP. This is done by removing the weakest side's pawn and probing the
+/// KP vs K bitbase: if the weakest side has a draw without the pawn, it probably
+/// has at least a draw with the pawn as well. The exception is when the stronger
+/// side's pawn is far advanced and not on a rook file; in this case it is often
+/// possible to win (e.g. 8/4k3/3p4/3P4/6K1/8/8/8 w - - 0 1).
+template<>
+ScaleFactor Endgame<KPKP>::operator()(const Position& pos) const {
+
+  assert(verify_material(pos, strongSide, VALUE_ZERO, 1));
+  assert(verify_material(pos, weakSide,   VALUE_ZERO, 1));
+
+  // Assume strongSide is white and the pawn is on files A-D
+  Square strongKing = normalize(pos, strongSide, pos.square<KING>(strongSide));
+  Square weakKing   = normalize(pos, strongSide, pos.square<KING>(weakSide));
+  Square strongPawn = normalize(pos, strongSide, pos.square<PAWN>(strongSide));
+
+  Color us = strongSide == pos.side_to_move() ? WHITE : BLACK;
+
+  // If the pawn has advanced to the fifth rank or further, and is not a
+  // rook pawn, it's too dangerous to assume that it's at least a draw.
+  if (rank_of(strongPawn) >= RANK_5 && file_of(strongPawn) != FILE_A)
+      return SCALE_FACTOR_NONE;
+
+  // Probe the KPK bitbase with the weakest side's pawn removed. If it's a draw,
+  // it's probably at least a draw even with the pawn.
+  return Bitbases::probe(strongKing, strongPawn, weakKing, us) ? SCALE_FACTOR_NONE : SCALE_FACTOR_DRAW;
+}
+
+} // namespace Stockfish

src/Stockfish/src/endgame.h

@@ -0,0 +1,126 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef ENDGAME_H_INCLUDED
+#define ENDGAME_H_INCLUDED
+
+#include <memory>
+#include <string>
+#include <type_traits>
+#include <unordered_map>
+#include <utility>
+
+#include "position.h"
+#include "types.h"
+
+namespace Stockfish {
+
+/// EndgameCode lists all supported endgame functions by corresponding codes
+
+enum EndgameCode {
+
+  EVALUATION_FUNCTIONS,
+  KNNK,  // KNN vs K
+  KNNKP, // KNN vs KP
+  KXK,   // Generic "mate lone king" eval
+  KBNK,  // KBN vs K
+  KPK,   // KP vs K
+  KRKP,  // KR vs KP
+  KRKB,  // KR vs KB
+  KRKN,  // KR vs KN
+  KQKP,  // KQ vs KP
+  KQKR,  // KQ vs KR
+
+  SCALING_FUNCTIONS,
+  KBPsK,   // KB and pawns vs K
+  KQKRPs,  // KQ vs KR and pawns
+  KRPKR,   // KRP vs KR
+  KRPKB,   // KRP vs KB
+  KRPPKRP, // KRPP vs KRP
+  KPsK,    // K and pawns vs K
+  KBPKB,   // KBP vs KB
+  KBPPKB,  // KBPP vs KB
+  KBPKN,   // KBP vs KN
+  KPKP     // KP vs KP
+};
+
+
+/// Endgame functions can be of two types depending on whether they return a
+/// Value or a ScaleFactor.
+
+template<EndgameCode E> using
+eg_type = typename std::conditional<(E < SCALING_FUNCTIONS), Value, ScaleFactor>::type;
+
+
+/// Base and derived functors for endgame evaluation and scaling functions
+
+template<typename T>
+struct EndgameBase {
+
+  explicit EndgameBase(Color c) : strongSide(c), weakSide(~c) {}
+  virtual ~EndgameBase() = default;
+  virtual T operator()(const Position&) const = 0;
+
+  const Color strongSide, weakSide;
+};
+
+
+template<EndgameCode E, typename T = eg_type<E>>
+struct Endgame : public EndgameBase<T> {
+
+  explicit Endgame(Color c) : EndgameBase<T>(c) {}
+  T operator()(const Position&) const override;
+};
+
+
+/// The Endgames namespace handles the pointers to endgame evaluation and scaling
+/// base objects in two std::map. We use polymorphism to invoke the actual
+/// endgame function by calling its virtual operator().
+
+namespace Endgames {
+
+  template<typename T> using Ptr = std::unique_ptr<EndgameBase<T>>;
+  template<typename T> using Map = std::unordered_map<Key, Ptr<T>>;
+
+  extern std::pair<Map<Value>, Map<ScaleFactor>> maps;
+
+  void init();
+
+  template<typename T>
+  Map<T>& map() {
+    return std::get<std::is_same<T, ScaleFactor>::value>(maps);
+  }
+
+  template<EndgameCode E, typename T = eg_type<E>>
+  void add(const std::string& code) {
+
+    StateInfo st;
+    map<T>()[Position().set(code, WHITE, &st).material_key()] = Ptr<T>(new Endgame<E>(WHITE));
+    map<T>()[Position().set(code, BLACK, &st).material_key()] = Ptr<T>(new Endgame<E>(BLACK));
+  }
+
+  template<typename T>
+  const EndgameBase<T>* probe(Key key) {
+    auto it = map<T>().find(key);
+    return it != map<T>().end() ? it->second.get() : nullptr;
+  }
+}
+
+} // namespace Stockfish
+
+#endif // #ifndef ENDGAME_H_INCLUDED

src/Stockfish/src/evaluate.h

@@ -0,0 +1,55 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef EVALUATE_H_INCLUDED
+#define EVALUATE_H_INCLUDED
+
+#include <string>
+#include <optional>
+
+#include "types.h"
+
+namespace Stockfish {
+
+class Position;
+
+namespace Eval {
+
+  std::string trace(Position& pos);
+  Value evaluate(const Position& pos);
+
+  extern bool useNNUE;
+  extern std::string currentEvalFileName;
+
+  // The default net name MUST follow the format nn-[SHA256 first 12 digits].nnue
+  // for the build process (profile-build and fishtest) to work. Do not change the
+  // name of the macro, as it is used in the Makefile.
+  #define EvalFileDefaultName   "nn-5af11540bbfe.nnue"
+
+  namespace NNUE {
+
+    void init();
+    void verify();
+
+  } // namespace NNUE
+
+} // namespace Eval
+
+} // namespace Stockfish
+
+#endif // #ifndef EVALUATE_H_INCLUDED

src/Stockfish/src/incbin/UNLICENCE

@@ -0,0 +1,26 @@
+The file "incbin.h" is free and unencumbered software released into
+the public domain by Dale Weiler, see:
+   <https://github.com/graphitemaster/incbin>
+
+Anyone is free to copy, modify, publish, use, compile, sell, or
+distribute this software, either in source code form or as a compiled
+binary, for any purpose, commercial or non-commercial, and by any
+means.
+
+In jurisdictions that recognize copyright laws, the author or authors
+of this software dedicate any and all copyright interest in the
+software to the public domain. We make this dedication for the benefit
+of the public at large and to the detriment of our heirs and
+successors. We intend this dedication to be an overt act of
+relinquishment in perpetuity of all present and future rights to this
+software under copyright law.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR
+OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+OTHER DEALINGS IN THE SOFTWARE.
+
+For more information, please refer to <http://unlicense.org/>

src/Stockfish/src/incbin/incbin.h

@@ -0,0 +1,368 @@
+/**
+ * @file incbin.h
+ * @author Dale Weiler
+ * @brief Utility for including binary files
+ *
+ * Facilities for including binary files into the current translation unit and
+ * making use from them externally in other translation units.
+ */
+#ifndef INCBIN_HDR
+#define INCBIN_HDR
+#include <limits.h>
+#if   defined(__AVX512BW__) || \
+      defined(__AVX512CD__) || \
+      defined(__AVX512DQ__) || \
+      defined(__AVX512ER__) || \
+      defined(__AVX512PF__) || \
+      defined(__AVX512VL__) || \
+      defined(__AVX512F__)
+# define INCBIN_ALIGNMENT_INDEX 6
+#elif defined(__AVX__)      || \
+      defined(__AVX2__)
+# define INCBIN_ALIGNMENT_INDEX 5
+#elif defined(__SSE__)      || \
+      defined(__SSE2__)     || \
+      defined(__SSE3__)     || \
+      defined(__SSSE3__)    || \
+      defined(__SSE4_1__)   || \
+      defined(__SSE4_2__)   || \
+      defined(__neon__)
+# define INCBIN_ALIGNMENT_INDEX 4
+#elif ULONG_MAX != 0xffffffffu
+# define INCBIN_ALIGNMENT_INDEX 3
+# else
+# define INCBIN_ALIGNMENT_INDEX 2
+#endif
+
+/* Lookup table of (1 << n) where `n' is `INCBIN_ALIGNMENT_INDEX' */
+#define INCBIN_ALIGN_SHIFT_0 1
+#define INCBIN_ALIGN_SHIFT_1 2
+#define INCBIN_ALIGN_SHIFT_2 4
+#define INCBIN_ALIGN_SHIFT_3 8
+#define INCBIN_ALIGN_SHIFT_4 16
+#define INCBIN_ALIGN_SHIFT_5 32
+#define INCBIN_ALIGN_SHIFT_6 64
+
+/* Actual alignment value */
+#define INCBIN_ALIGNMENT \
+    INCBIN_CONCATENATE( \
+        INCBIN_CONCATENATE(INCBIN_ALIGN_SHIFT, _), \
+        INCBIN_ALIGNMENT_INDEX)
+
+/* Stringize */
+#define INCBIN_STR(X) \
+    #X
+#define INCBIN_STRINGIZE(X) \
+    INCBIN_STR(X)
+/* Concatenate */
+#define INCBIN_CAT(X, Y) \
+    X ## Y
+#define INCBIN_CONCATENATE(X, Y) \
+    INCBIN_CAT(X, Y)
+/* Deferred macro expansion */
+#define INCBIN_EVAL(X) \
+    X
+#define INCBIN_INVOKE(N, ...) \
+    INCBIN_EVAL(N(__VA_ARGS__))
+
+/* Green Hills uses a different directive for including binary data */
+#if defined(__ghs__)
+#  if (__ghs_asm == 2)
+#    define INCBIN_MACRO ".file"
+/* Or consider the ".myrawdata" entry in the ld file */
+#  else
+#    define INCBIN_MACRO "\tINCBIN"
+#  endif
+#else
+#  define INCBIN_MACRO ".incbin"
+#endif
+
+#ifndef _MSC_VER
+#  define INCBIN_ALIGN \
+    __attribute__((aligned(INCBIN_ALIGNMENT)))
+#else
+#  define INCBIN_ALIGN __declspec(align(INCBIN_ALIGNMENT))
+#endif
+
+#if defined(__arm__) || /* GNU C and RealView */ \
+    defined(__arm) || /* Diab */ \
+    defined(_ARM) /* ImageCraft */
+#  define INCBIN_ARM
+#endif
+
+#ifdef __GNUC__
+/* Utilize .balign where supported */
+#  define INCBIN_ALIGN_HOST ".balign " INCBIN_STRINGIZE(INCBIN_ALIGNMENT) "\n"
+#  define INCBIN_ALIGN_BYTE ".balign 1\n"
+#elif defined(INCBIN_ARM)
+/*
+ * On arm assemblers, the alignment value is calculated as (1 << n) where `n' is
+ * the shift count. This is the value passed to `.align'
+ */
+#  define INCBIN_ALIGN_HOST ".align " INCBIN_STRINGIZE(INCBIN_ALIGNMENT_INDEX) "\n"
+#  define INCBIN_ALIGN_BYTE ".align 0\n"
+#else
+/* We assume other inline assembler's treat `.align' as `.balign' */
+#  define INCBIN_ALIGN_HOST ".align " INCBIN_STRINGIZE(INCBIN_ALIGNMENT) "\n"
+#  define INCBIN_ALIGN_BYTE ".align 1\n"
+#endif
+
+/* INCBIN_CONST is used by incbin.c generated files */
+#if defined(__cplusplus)
+#  define INCBIN_EXTERNAL extern "C"
+#  define INCBIN_CONST    extern const
+#else
+#  define INCBIN_EXTERNAL extern
+#  define INCBIN_CONST    const
+#endif
+
+/**
+ * @brief Optionally override the linker section into which data is emitted.
+ *
+ * @warning If you use this facility, you'll have to deal with platform-specific linker output
+ * section naming on your own
+ *
+ * Overriding the default linker output section, e.g for esp8266/Arduino:
+ * @code
+ * #define INCBIN_OUTPUT_SECTION ".irom.text"
+ * #include "incbin.h"
+ * INCBIN(Foo, "foo.txt");
+ * // Data is emitted into program memory that never gets copied to RAM
+ * @endcode
+ */
+#if !defined(INCBIN_OUTPUT_SECTION)
+#  if defined(__APPLE__)
+#    define INCBIN_OUTPUT_SECTION         ".const_data"
+#  else
+#    define INCBIN_OUTPUT_SECTION         ".rodata"
+#  endif
+#endif
+
+#if defined(__APPLE__)
+/* The directives are different for Apple branded compilers */
+#  define INCBIN_SECTION         INCBIN_OUTPUT_SECTION "\n"
+#  define INCBIN_GLOBAL(NAME)    ".globl " INCBIN_MANGLE INCBIN_STRINGIZE(INCBIN_PREFIX) #NAME "\n"
+#  define INCBIN_INT             ".long "
+#  define INCBIN_MANGLE          "_"
+#  define INCBIN_BYTE            ".byte "
+#  define INCBIN_TYPE(...)
+#else
+#  define INCBIN_SECTION         ".section " INCBIN_OUTPUT_SECTION "\n"
+#  define INCBIN_GLOBAL(NAME)    ".global " INCBIN_STRINGIZE(INCBIN_PREFIX) #NAME "\n"
+#  if defined(__ghs__)
+#    define INCBIN_INT           ".word "
+#  else
+#    define INCBIN_INT           ".int "
+#  endif
+#  if defined(__USER_LABEL_PREFIX__)
+#    define INCBIN_MANGLE        INCBIN_STRINGIZE(__USER_LABEL_PREFIX__)
+#  else
+#    define INCBIN_MANGLE        ""
+#  endif
+#  if defined(INCBIN_ARM)
+/* On arm assemblers, `@' is used as a line comment token */
+#    define INCBIN_TYPE(NAME)    ".type " INCBIN_STRINGIZE(INCBIN_PREFIX) #NAME ", %object\n"
+#  elif defined(__MINGW32__) || defined(__MINGW64__)
+/* Mingw doesn't support this directive either */
+#    define INCBIN_TYPE(NAME)
+#  else
+/* It's safe to use `@' on other architectures */
+#    define INCBIN_TYPE(NAME)    ".type " INCBIN_STRINGIZE(INCBIN_PREFIX) #NAME ", @object\n"
+#  endif
+#  define INCBIN_BYTE            ".byte "
+#endif
+
+/* List of style types used for symbol names */
+#define INCBIN_STYLE_CAMEL 0
+#define INCBIN_STYLE_SNAKE 1
+
+/**
+ * @brief Specify the prefix to use for symbol names.
+ *
+ * By default this is `g', producing symbols of the form:
+ * @code
+ * #include "incbin.h"
+ * INCBIN(Foo, "foo.txt");
+ *
+ * // Now you have the following symbols:
+ * // const unsigned char gFooData[];
+ * // const unsigned char *const gFooEnd;
+ * // const unsigned int gFooSize;
+ * @endcode
+ *
+ * If however you specify a prefix before including: e.g:
+ * @code
+ * #define INCBIN_PREFIX incbin
+ * #include "incbin.h"
+ * INCBIN(Foo, "foo.txt");
+ *
+ * // Now you have the following symbols instead:
+ * // const unsigned char incbinFooData[];
+ * // const unsigned char *const incbinFooEnd;
+ * // const unsigned int incbinFooSize;
+ * @endcode
+ */
+#if !defined(INCBIN_PREFIX)
+#  define INCBIN_PREFIX g
+#endif
+
+/**
+ * @brief Specify the style used for symbol names.
+ *
+ * Possible options are
+ * - INCBIN_STYLE_CAMEL "CamelCase"
+ * - INCBIN_STYLE_SNAKE "snake_case"
+ *
+ * Default option is *INCBIN_STYLE_CAMEL* producing symbols of the form:
+ * @code
+ * #include "incbin.h"
+ * INCBIN(Foo, "foo.txt");
+ *
+ * // Now you have the following symbols:
+ * // const unsigned char <prefix>FooData[];
+ * // const unsigned char *const <prefix>FooEnd;
+ * // const unsigned int <prefix>FooSize;
+ * @endcode
+ *
+ * If however you specify a style before including: e.g:
+ * @code
+ * #define INCBIN_STYLE INCBIN_STYLE_SNAKE
+ * #include "incbin.h"
+ * INCBIN(foo, "foo.txt");
+ *
+ * // Now you have the following symbols:
+ * // const unsigned char <prefix>foo_data[];
+ * // const unsigned char *const <prefix>foo_end;
+ * // const unsigned int <prefix>foo_size;
+ * @endcode
+ */
+#if !defined(INCBIN_STYLE)
+#  define INCBIN_STYLE INCBIN_STYLE_CAMEL
+#endif
+
+/* Style lookup tables */
+#define INCBIN_STYLE_0_DATA Data
+#define INCBIN_STYLE_0_END End
+#define INCBIN_STYLE_0_SIZE Size
+#define INCBIN_STYLE_1_DATA _data
+#define INCBIN_STYLE_1_END _end
+#define INCBIN_STYLE_1_SIZE _size
+
+/* Style lookup: returning identifier */
+#define INCBIN_STYLE_IDENT(TYPE) \
+    INCBIN_CONCATENATE( \
+        INCBIN_STYLE_, \
+        INCBIN_CONCATENATE( \
+            INCBIN_EVAL(INCBIN_STYLE), \
+            INCBIN_CONCATENATE(_, TYPE)))
+
+/* Style lookup: returning string literal */
+#define INCBIN_STYLE_STRING(TYPE) \
+    INCBIN_STRINGIZE( \
+        INCBIN_STYLE_IDENT(TYPE)) \
+
+/* Generate the global labels by indirectly invoking the macro with our style
+ * type and concatenating the name against them. */
+#define INCBIN_GLOBAL_LABELS(NAME, TYPE) \
+    INCBIN_INVOKE( \
+        INCBIN_GLOBAL, \
+        INCBIN_CONCATENATE( \
+            NAME, \
+            INCBIN_INVOKE( \
+                INCBIN_STYLE_IDENT, \
+                TYPE))) \
+    INCBIN_INVOKE( \
+        INCBIN_TYPE, \
+        INCBIN_CONCATENATE( \
+            NAME, \
+            INCBIN_INVOKE( \
+                INCBIN_STYLE_IDENT, \
+                TYPE)))
+
+/**
+ * @brief Externally reference binary data included in another translation unit.
+ *
+ * Produces three external symbols that reference the binary data included in
+ * another translation unit.
+ *
+ * The symbol names are a concatenation of `INCBIN_PREFIX' before *NAME*; with
+ * "Data", as well as "End" and "Size" after. An example is provided below.
+ *
+ * @param NAME The name given for the binary data
+ *
+ * @code
+ * INCBIN_EXTERN(Foo);
+ *
+ * // Now you have the following symbols:
+ * // extern const unsigned char <prefix>FooData[];
+ * // extern const unsigned char *const <prefix>FooEnd;
+ * // extern const unsigned int <prefix>FooSize;
+ * @endcode
+ */
+#define INCBIN_EXTERN(NAME) \
+    INCBIN_EXTERNAL const INCBIN_ALIGN unsigned char \
+        INCBIN_CONCATENATE( \
+            INCBIN_CONCATENATE(INCBIN_PREFIX, NAME), \
+            INCBIN_STYLE_IDENT(DATA))[]; \
+    INCBIN_EXTERNAL const INCBIN_ALIGN unsigned char *const \
+    INCBIN_CONCATENATE( \
+        INCBIN_CONCATENATE(INCBIN_PREFIX, NAME), \
+        INCBIN_STYLE_IDENT(END)); \
+    INCBIN_EXTERNAL const unsigned int \
+        INCBIN_CONCATENATE( \
+            INCBIN_CONCATENATE(INCBIN_PREFIX, NAME), \
+            INCBIN_STYLE_IDENT(SIZE))
+
+/**
+ * @brief Include a binary file into the current translation unit.
+ *
+ * Includes a binary file into the current translation unit, producing three symbols
+ * for objects that encode the data and size respectively.
+ *
+ * The symbol names are a concatenation of `INCBIN_PREFIX' before *NAME*; with
+ * "Data", as well as "End" and "Size" after. An example is provided below.
+ *
+ * @param NAME The name to associate with this binary data (as an identifier.)
+ * @param FILENAME The file to include (as a string literal.)
+ *
+ * @code
+ * INCBIN(Icon, "icon.png");
+ *
+ * // Now you have the following symbols:
+ * // const unsigned char <prefix>IconData[];
+ * // const unsigned char *const <prefix>IconEnd;
+ * // const unsigned int <prefix>IconSize;
+ * @endcode
+ *
+ * @warning This must be used in global scope
+ * @warning The identifiers may be different if INCBIN_STYLE is not default
+ *
+ * To externally reference the data included by this in another translation unit
+ * please @see INCBIN_EXTERN.
+ */
+#ifdef _MSC_VER
+#define INCBIN(NAME, FILENAME) \
+    INCBIN_EXTERN(NAME)
+#else
+#define INCBIN(NAME, FILENAME) \
+    __asm__(INCBIN_SECTION \
+            INCBIN_GLOBAL_LABELS(NAME, DATA) \
+            INCBIN_ALIGN_HOST \
+            INCBIN_MANGLE INCBIN_STRINGIZE(INCBIN_PREFIX) #NAME INCBIN_STYLE_STRING(DATA) ":\n" \
+            INCBIN_MACRO " \"" FILENAME "\"\n" \
+            INCBIN_GLOBAL_LABELS(NAME, END) \
+            INCBIN_ALIGN_BYTE \
+            INCBIN_MANGLE INCBIN_STRINGIZE(INCBIN_PREFIX) #NAME INCBIN_STYLE_STRING(END) ":\n" \
+                INCBIN_BYTE "1\n" \
+            INCBIN_GLOBAL_LABELS(NAME, SIZE) \
+            INCBIN_ALIGN_HOST \
+            INCBIN_MANGLE INCBIN_STRINGIZE(INCBIN_PREFIX) #NAME INCBIN_STYLE_STRING(SIZE) ":\n" \
+                INCBIN_INT INCBIN_MANGLE INCBIN_STRINGIZE(INCBIN_PREFIX) #NAME INCBIN_STYLE_STRING(END) " - " \
+                           INCBIN_MANGLE INCBIN_STRINGIZE(INCBIN_PREFIX) #NAME INCBIN_STYLE_STRING(DATA) "\n" \
+            INCBIN_ALIGN_HOST \
+            ".text\n" \
+    ); \
+    INCBIN_EXTERN(NAME)
+
+#endif
+#endif

src/Stockfish/src/main.cpp

@@ -0,0 +1,53 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include <iostream>
+
+#include "bitboard.h"
+#include "endgame.h"
+#include "position.h"
+#include "psqt.h"
+#include "search.h"
+#include "syzygy/tbprobe.h"
+#include "thread.h"
+#include "tt.h"
+#include "uci.h"
+
+using namespace Stockfish;
+
+int main(int argc, char* argv[]) {
+
+  fakeout << engine_info() << fakeendl;
+
+  CommandLine::init(argc, argv);
+  UCI::init(Options);
+  Tune::init();
+  PSQT::init();
+  Bitboards::init();
+  Position::init();
+  Bitbases::init();
+  Endgames::init();
+  Threads.set(size_t(Options["Threads"]));
+  Search::clear(); // After threads are up
+  Eval::NNUE::init();
+
+  UCI::loop(argc, argv);
+
+  Threads.set(0);
+  return 0;
+}

src/Stockfish/src/material.cpp

@@ -0,0 +1,229 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include <cassert>
+#include <cstring>   // For std::memset
+
+#include "material.h"
+#include "thread.h"
+
+using namespace std;
+
+namespace Stockfish {
+
+namespace {
+  #define S(mg, eg) make_score(mg, eg)
+
+  // Polynomial material imbalance parameters
+
+  // One Score parameter for each pair (our piece, another of our pieces)
+  constexpr Score QuadraticOurs[][PIECE_TYPE_NB] = {
+    // OUR PIECE 2
+    // bishop pair    pawn         knight       bishop       rook           queen
+    {S(1419, 1455)                                                                  }, // Bishop pair
+    {S( 101,   28), S( 37,  39)                                                     }, // Pawn
+    {S(  57,   64), S(249, 187), S(-49, -62)                                        }, // Knight      OUR PIECE 1
+    {S(   0,    0), S(118, 137), S( 10,  27), S(  0,   0)                           }, // Bishop
+    {S( -63,  -68), S( -5,   3), S(100,  81), S(132, 118), S(-246, -244)            }, // Rook
+    {S(-210, -211), S( 37,  14), S(147, 141), S(161, 105), S(-158, -174), S(-9,-31) }  // Queen
+  };
+
+  // One Score parameter for each pair (our piece, their piece)
+  constexpr Score QuadraticTheirs[][PIECE_TYPE_NB] = {
+    // THEIR PIECE
+    // bishop pair   pawn         knight       bishop       rook         queen
+    {                                                                               }, // Bishop pair
+    {S(  33,  30)                                                                   }, // Pawn
+    {S(  46,  18), S(106,  84)                                                      }, // Knight      OUR PIECE
+    {S(  75,  35), S( 59,  44), S( 60,  15)                                         }, // Bishop
+    {S(  26,  35), S(  6,  22), S( 38,  39), S(-12,  -2)                            }, // Rook
+    {S(  97,  93), S(100, 163), S(-58, -91), S(112, 192), S(276, 225)               }  // Queen
+  };
+
+  #undef S
+
+  // Endgame evaluation and scaling functions are accessed directly and not through
+  // the function maps because they correspond to more than one material hash key.
+  Endgame<KXK>    EvaluateKXK[] = { Endgame<KXK>(WHITE),    Endgame<KXK>(BLACK) };
+
+  Endgame<KBPsK>  ScaleKBPsK[]  = { Endgame<KBPsK>(WHITE),  Endgame<KBPsK>(BLACK) };
+  Endgame<KQKRPs> ScaleKQKRPs[] = { Endgame<KQKRPs>(WHITE), Endgame<KQKRPs>(BLACK) };
+  Endgame<KPsK>   ScaleKPsK[]   = { Endgame<KPsK>(WHITE),   Endgame<KPsK>(BLACK) };
+  Endgame<KPKP>   ScaleKPKP[]   = { Endgame<KPKP>(WHITE),   Endgame<KPKP>(BLACK) };
+
+  // Helper used to detect a given material distribution
+  bool is_KXK(const Position& pos, Color us) {
+    return  !more_than_one(pos.pieces(~us))
+          && pos.non_pawn_material(us) >= RookValueMg;
+  }
+
+  bool is_KBPsK(const Position& pos, Color us) {
+    return   pos.non_pawn_material(us) == BishopValueMg
+          && pos.count<PAWN>(us) >= 1;
+  }
+
+  bool is_KQKRPs(const Position& pos, Color us) {
+    return  !pos.count<PAWN>(us)
+          && pos.non_pawn_material(us) == QueenValueMg
+          && pos.count<ROOK>(~us) == 1
+          && pos.count<PAWN>(~us) >= 1;
+  }
+
+
+  /// imbalance() calculates the imbalance by comparing the piece count of each
+  /// piece type for both colors.
+
+  template<Color Us>
+  Score imbalance(const int pieceCount[][PIECE_TYPE_NB]) {
+
+    constexpr Color Them = ~Us;
+
+    Score bonus = SCORE_ZERO;
+
+    // Second-degree polynomial material imbalance, by Tord Romstad
+    for (int pt1 = NO_PIECE_TYPE; pt1 <= QUEEN; ++pt1)
+    {
+        if (!pieceCount[Us][pt1])
+            continue;
+
+        int v = QuadraticOurs[pt1][pt1] * pieceCount[Us][pt1];
+
+        for (int pt2 = NO_PIECE_TYPE; pt2 < pt1; ++pt2)
+            v +=  QuadraticOurs[pt1][pt2] * pieceCount[Us][pt2]
+                + QuadraticTheirs[pt1][pt2] * pieceCount[Them][pt2];
+
+        bonus += pieceCount[Us][pt1] * v;
+    }
+
+    return bonus;
+  }
+
+} // namespace
+
+namespace Material {
+
+
+/// Material::probe() looks up the current position's material configuration in
+/// the material hash table. It returns a pointer to the Entry if the position
+/// is found. Otherwise a new Entry is computed and stored there, so we don't
+/// have to recompute all when the same material configuration occurs again.
+
+Entry* probe(const Position& pos) {
+
+  Key key = pos.material_key();
+  Entry* e = pos.this_thread()->materialTable[key];
+
+  if (e->key == key)
+      return e;
+
+  std::memset(e, 0, sizeof(Entry));
+  e->key = key;
+  e->factor[WHITE] = e->factor[BLACK] = (uint8_t)SCALE_FACTOR_NORMAL;
+
+  Value npm_w = pos.non_pawn_material(WHITE);
+  Value npm_b = pos.non_pawn_material(BLACK);
+  Value npm   = std::clamp(npm_w + npm_b, EndgameLimit, MidgameLimit);
+
+  // Map total non-pawn material into [PHASE_ENDGAME, PHASE_MIDGAME]
+  e->gamePhase = Phase(((npm - EndgameLimit) * PHASE_MIDGAME) / (MidgameLimit - EndgameLimit));
+
+  // Let's look if we have a specialized evaluation function for this particular
+  // material configuration. Firstly we look for a fixed configuration one, then
+  // for a generic one if the previous search failed.
+  if ((e->evaluationFunction = Endgames::probe<Value>(key)) != nullptr)
+      return e;
+
+  for (Color c : { WHITE, BLACK })
+      if (is_KXK(pos, c))
+      {
+          e->evaluationFunction = &EvaluateKXK[c];
+          return e;
+      }
+
+  // OK, we didn't find any special evaluation function for the current material
+  // configuration. Is there a suitable specialized scaling function?
+  const auto* sf = Endgames::probe<ScaleFactor>(key);
+
+  if (sf)
+  {
+      e->scalingFunction[sf->strongSide] = sf; // Only strong color assigned
+      return e;
+  }
+
+  // We didn't find any specialized scaling function, so fall back on generic
+  // ones that refer to more than one material distribution. Note that in this
+  // case we don't return after setting the function.
+  for (Color c : { WHITE, BLACK })
+  {
+    if (is_KBPsK(pos, c))
+        e->scalingFunction[c] = &ScaleKBPsK[c];
+
+    else if (is_KQKRPs(pos, c))
+        e->scalingFunction[c] = &ScaleKQKRPs[c];
+  }
+
+  if (npm_w + npm_b == VALUE_ZERO && pos.pieces(PAWN)) // Only pawns on the board
+  {
+      if (!pos.count<PAWN>(BLACK))
+      {
+          assert(pos.count<PAWN>(WHITE) >= 2);
+
+          e->scalingFunction[WHITE] = &ScaleKPsK[WHITE];
+      }
+      else if (!pos.count<PAWN>(WHITE))
+      {
+          assert(pos.count<PAWN>(BLACK) >= 2);
+
+          e->scalingFunction[BLACK] = &ScaleKPsK[BLACK];
+      }
+      else if (pos.count<PAWN>(WHITE) == 1 && pos.count<PAWN>(BLACK) == 1)
+      {
+          // This is a special case because we set scaling functions
+          // for both colors instead of only one.
+          e->scalingFunction[WHITE] = &ScaleKPKP[WHITE];
+          e->scalingFunction[BLACK] = &ScaleKPKP[BLACK];
+      }
+  }
+
+  // Zero or just one pawn makes it difficult to win, even with a small material
+  // advantage. This catches some trivial draws like KK, KBK and KNK and gives a
+  // drawish scale factor for cases such as KRKBP and KmmKm (except for KBBKN).
+  if (!pos.count<PAWN>(WHITE) && npm_w - npm_b <= BishopValueMg)
+      e->factor[WHITE] = uint8_t(npm_w <  RookValueMg   ? SCALE_FACTOR_DRAW :
+                                 npm_b <= BishopValueMg ? 4 : 14);
+
+  if (!pos.count<PAWN>(BLACK) && npm_b - npm_w <= BishopValueMg)
+      e->factor[BLACK] = uint8_t(npm_b <  RookValueMg   ? SCALE_FACTOR_DRAW :
+                                 npm_w <= BishopValueMg ? 4 : 14);
+
+  // Evaluate the material imbalance. We use PIECE_TYPE_NONE as a place holder
+  // for the bishop pair "extended piece", which allows us to be more flexible
+  // in defining bishop pair bonuses.
+  const int pieceCount[COLOR_NB][PIECE_TYPE_NB] = {
+  { pos.count<BISHOP>(WHITE) > 1, pos.count<PAWN>(WHITE), pos.count<KNIGHT>(WHITE),
+    pos.count<BISHOP>(WHITE)    , pos.count<ROOK>(WHITE), pos.count<QUEEN >(WHITE) },
+  { pos.count<BISHOP>(BLACK) > 1, pos.count<PAWN>(BLACK), pos.count<KNIGHT>(BLACK),
+    pos.count<BISHOP>(BLACK)    , pos.count<ROOK>(BLACK), pos.count<QUEEN >(BLACK) } };
+
+  e->score = (imbalance<WHITE>(pieceCount) - imbalance<BLACK>(pieceCount)) / 16;
+  return e;
+}
+
+} // namespace Material
+
+} // namespace Stockfish

src/Stockfish/src/material.h

@@ -0,0 +1,71 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef MATERIAL_H_INCLUDED
+#define MATERIAL_H_INCLUDED
+
+#include "endgame.h"
+#include "misc.h"
+#include "position.h"
+#include "types.h"
+
+namespace Stockfish::Material {
+
+/// Material::Entry contains various information about a material configuration.
+/// It contains a material imbalance evaluation, a function pointer to a special
+/// endgame evaluation function (which in most cases is nullptr, meaning that the
+/// standard evaluation function will be used), and scale factors.
+///
+/// The scale factors are used to scale the evaluation score up or down. For
+/// instance, in KRB vs KR endgames, the score is scaled down by a factor of 4,
+/// which will result in scores of absolute value less than one pawn.
+
+struct Entry {
+
+  Score imbalance() const { return score; }
+  Phase game_phase() const { return (Phase)gamePhase; }
+  bool specialized_eval_exists() const { return evaluationFunction != nullptr; }
+  Value evaluate(const Position& pos) const { return (*evaluationFunction)(pos); }
+
+  // scale_factor() takes a position and a color as input and returns a scale factor
+  // for the given color. We have to provide the position in addition to the color
+  // because the scale factor may also be a function which should be applied to
+  // the position. For instance, in KBP vs K endgames, the scaling function looks
+  // for rook pawns and wrong-colored bishops.
+  ScaleFactor scale_factor(const Position& pos, Color c) const {
+    ScaleFactor sf = scalingFunction[c] ? (*scalingFunction[c])(pos)
+                                        :  SCALE_FACTOR_NONE;
+    return sf != SCALE_FACTOR_NONE ? sf : ScaleFactor(factor[c]);
+  }
+
+  Key key;
+  const EndgameBase<Value>* evaluationFunction;
+  const EndgameBase<ScaleFactor>* scalingFunction[COLOR_NB]; // Could be one for each
+                                                             // side (e.g. KPKP, KBPsK)
+  Score score;
+  int16_t gamePhase;
+  uint8_t factor[COLOR_NB];
+};
+
+using Table = HashTable<Entry, 8192>;
+
+Entry* probe(const Position& pos);
+
+} // namespace Stockfish::Material
+
+#endif // #ifndef MATERIAL_H_INCLUDED

src/Stockfish/src/misc.cpp

@@ -0,0 +1,787 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifdef _WIN32
+#if _WIN32_WINNT < 0x0601
+#undef  _WIN32_WINNT
+#define _WIN32_WINNT 0x0601 // Force to include needed API prototypes
+#endif
+
+#ifndef NOMINMAX
+#define NOMINMAX
+#endif
+
+#include <windows.h>
+// The needed Windows API for processor groups could be missed from old Windows
+// versions, so instead of calling them directly (forcing the linker to resolve
+// the calls at compile time), try to load them at runtime. To do this we need
+// first to define the corresponding function pointers.
+extern "C" {
+using fun1_t = bool(*)(LOGICAL_PROCESSOR_RELATIONSHIP,
+                       PSYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX, PDWORD);
+using fun2_t = bool(*)(USHORT, PGROUP_AFFINITY);
+using fun3_t = bool(*)(HANDLE, CONST GROUP_AFFINITY*, PGROUP_AFFINITY);
+using fun4_t = bool(*)(USHORT, PGROUP_AFFINITY, USHORT, PUSHORT);
+using fun5_t = WORD(*)();
+using fun6_t = bool(*)(HANDLE, DWORD, PHANDLE);
+using fun7_t = bool(*)(LPCSTR, LPCSTR, PLUID);
+using fun8_t = bool(*)(HANDLE, BOOL, PTOKEN_PRIVILEGES, DWORD, PTOKEN_PRIVILEGES, PDWORD);
+}
+#endif
+
+#include <cmath>
+#include <cstdlib>
+#include <fstream>
+#include <iomanip>
+#include <iostream>
+#include <sstream>
+#include <string_view>
+#include <vector>
+
+#if defined(__linux__) && !defined(__ANDROID__)
+#include <stdlib.h>
+#include <sys/mman.h>
+#endif
+
+#if defined(__APPLE__) || defined(__ANDROID__) || defined(__OpenBSD__) || (defined(__GLIBCXX__) && !defined(_GLIBCXX_HAVE_ALIGNED_ALLOC) && !defined(_WIN32)) || defined(__e2k__)
+#define POSIXALIGNEDALLOC
+#include <stdlib.h>
+#endif
+
+#include "misc.h"
+#include "thread.h"
+
+using namespace std;
+
+namespace Stockfish {
+
+namespace {
+
+/// Version number or dev.
+constexpr string_view version = "16";
+
+/// Our fancy logging facility. The trick here is to replace cin.rdbuf() and
+/// cout.rdbuf() with two Tie objects that tie cin and cout to a file stream. We
+/// can toggle the logging of fakeout and std:cin at runtime whilst preserving
+/// usual I/O functionality, all without changing a single line of code!
+/// Idea from http://groups.google.com/group/comp.lang.c++/msg/1d941c0f26ea0d81
+
+struct Tie: public streambuf { // MSVC requires split streambuf for cin and cout
+
+  Tie(streambuf* b, streambuf* l) : buf(b), logBuf(l) {}
+
+  int sync() override { return logBuf->pubsync(), buf->pubsync(); }
+  int overflow(int c) override { return log(buf->sputc((char)c), "<< "); }
+  int underflow() override { return buf->sgetc(); }
+  int uflow() override { return log(buf->sbumpc(), ">> "); }
+
+  streambuf *buf, *logBuf;
+
+  int log(int c, const char* prefix) {
+
+    static int last = '\n'; // Single log file
+
+    if (last == '\n')
+        logBuf->sputn(prefix, 3);
+
+    return last = logBuf->sputc((char)c);
+  }
+};
+
+class Logger {
+
+  Logger() : in(cin.rdbuf(), file.rdbuf()), out(cout.rdbuf(), file.rdbuf()) {}
+ ~Logger() { start(""); }
+
+  ofstream file;
+  Tie in, out;
+
+public:
+  static void start(const std::string& fname) {
+
+    static Logger l;
+
+    if (l.file.is_open())
+    {
+        cout.rdbuf(l.out.buf);
+        cin.rdbuf(l.in.buf);
+        l.file.close();
+    }
+
+    if (!fname.empty())
+    {
+        l.file.open(fname, ifstream::out);
+
+        if (!l.file.is_open())
+        {
+            cerr << "Unable to open debug log file " << fname << endl;
+            exit(EXIT_FAILURE);
+        }
+
+        cin.rdbuf(&l.in);
+        cout.rdbuf(&l.out);
+    }
+  }
+};
+
+} // namespace
+
+
+/// engine_info() returns the full name of the current Stockfish version.
+/// For local dev compiles we try to append the commit sha and commit date
+/// from git if that fails only the local compilation date is set and "nogit" is specified:
+/// Stockfish dev-YYYYMMDD-SHA
+/// or
+/// Stockfish dev-YYYYMMDD-nogit
+///
+/// For releases (non dev builds) we only include the version number:
+/// Stockfish version
+
+string engine_info(bool to_uci) {
+  stringstream ss;
+  ss << "Stockfish " << version << setfill('0');
+
+  if constexpr (version == "dev")
+  {
+      ss << "-";
+      #ifdef GIT_DATE
+      ss << stringify(GIT_DATE);
+      #else
+      constexpr string_view months("Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec");
+      string month, day, year;
+      stringstream date(__DATE__); // From compiler, format is "Sep 21 2008"
+
+      date >> month >> day >> year;
+      ss << year << setw(2) << setfill('0') << (1 + months.find(month) / 4) << setw(2) << setfill('0') << day;
+      #endif
+
+      ss << "-";
+
+      #ifdef GIT_SHA
+      ss << stringify(GIT_SHA);
+      #else
+      ss << "nogit";
+      #endif
+  }
+
+  ss << (to_uci  ? "\nid author ": " by ")
+     << "the Stockfish developers (see AUTHORS file)";
+
+  return ss.str();
+}
+
+
+/// compiler_info() returns a string trying to describe the compiler we use
+
+std::string compiler_info() {
+
+  #define make_version_string(major, minor, patch) stringify(major) "." stringify(minor) "." stringify(patch)
+
+/// Predefined macros hell:
+///
+/// __GNUC__           Compiler is gcc, Clang or Intel on Linux
+/// __INTEL_COMPILER   Compiler is Intel
+/// _MSC_VER           Compiler is MSVC or Intel on Windows
+/// _WIN32             Building on Windows (any)
+/// _WIN64             Building on Windows 64 bit
+
+  std::string compiler = "\nCompiled by ";
+
+  #ifdef __clang__
+     compiler += "clang++ ";
+     compiler += make_version_string(__clang_major__, __clang_minor__, __clang_patchlevel__);
+  #elif __INTEL_COMPILER
+     compiler += "Intel compiler ";
+     compiler += "(version ";
+     compiler += stringify(__INTEL_COMPILER) " update " stringify(__INTEL_COMPILER_UPDATE);
+     compiler += ")";
+  #elif _MSC_VER
+     compiler += "MSVC ";
+     compiler += "(version ";
+     compiler += stringify(_MSC_FULL_VER) "." stringify(_MSC_BUILD);
+     compiler += ")";
+  #elif defined(__e2k__) && defined(__LCC__)
+    #define dot_ver2(n) \
+      compiler += (char)'.'; \
+      compiler += (char)('0' + (n) / 10); \
+      compiler += (char)('0' + (n) % 10);
+
+     compiler += "MCST LCC ";
+     compiler += "(version ";
+     compiler += std::to_string(__LCC__ / 100);
+     dot_ver2(__LCC__ % 100)
+     dot_ver2(__LCC_MINOR__)
+     compiler += ")";
+  #elif __GNUC__
+     compiler += "g++ (GNUC) ";
+     compiler += make_version_string(__GNUC__, __GNUC_MINOR__, __GNUC_PATCHLEVEL__);
+  #else
+     compiler += "Unknown compiler ";
+     compiler += "(unknown version)";
+  #endif
+
+  #if defined(__APPLE__)
+     compiler += " on Apple";
+  #elif defined(__CYGWIN__)
+     compiler += " on Cygwin";
+  #elif defined(__MINGW64__)
+     compiler += " on MinGW64";
+  #elif defined(__MINGW32__)
+     compiler += " on MinGW32";
+  #elif defined(__ANDROID__)
+     compiler += " on Android";
+  #elif defined(__linux__)
+     compiler += " on Linux";
+  #elif defined(_WIN64)
+     compiler += " on Microsoft Windows 64-bit";
+  #elif defined(_WIN32)
+     compiler += " on Microsoft Windows 32-bit";
+  #else
+     compiler += " on unknown system";
+  #endif
+
+  compiler += "\nCompilation settings include: ";
+  compiler += (Is64Bit ? " 64bit" : " 32bit");
+  #if defined(USE_VNNI)
+    compiler += " VNNI";
+  #endif
+  #if defined(USE_AVX512)
+    compiler += " AVX512";
+  #endif
+  compiler += (HasPext ? " BMI2" : "");
+  #if defined(USE_AVX2)
+    compiler += " AVX2";
+  #endif
+  #if defined(USE_SSE41)
+    compiler += " SSE41";
+  #endif
+  #if defined(USE_SSSE3)
+    compiler += " SSSE3";
+  #endif
+  #if defined(USE_SSE2)
+    compiler += " SSE2";
+  #endif
+  compiler += (HasPopCnt ? " POPCNT" : "");
+  #if defined(USE_MMX)
+    compiler += " MMX";
+  #endif
+  #if defined(USE_NEON)
+    compiler += " NEON";
+  #endif
+
+  #if !defined(NDEBUG)
+    compiler += " DEBUG";
+  #endif
+
+  compiler += "\n__VERSION__ macro expands to: ";
+  #ifdef __VERSION__
+     compiler += __VERSION__;
+  #else
+     compiler += "(undefined macro)";
+  #endif
+  compiler += "\n";
+
+  return compiler;
+}
+
+
+/// Debug functions used mainly to collect run-time statistics
+constexpr int MaxDebugSlots = 32;
+
+namespace {
+
+template<size_t N>
+struct DebugInfo {
+    std::atomic<int64_t> data[N] = { 0 };
+
+    constexpr inline std::atomic<int64_t>& operator[](int index) { return data[index]; }
+};
+
+DebugInfo<2> hit[MaxDebugSlots];
+DebugInfo<2> mean[MaxDebugSlots];
+DebugInfo<3> stdev[MaxDebugSlots];
+DebugInfo<6> correl[MaxDebugSlots];
+
+}  // namespace
+
+void dbg_hit_on(bool cond, int slot) {
+
+    ++hit[slot][0];
+    if (cond)
+        ++hit[slot][1];
+}
+
+void dbg_mean_of(int64_t value, int slot) {
+
+    ++mean[slot][0];
+    mean[slot][1] += value;
+}
+
+void dbg_stdev_of(int64_t value, int slot) {
+
+    ++stdev[slot][0];
+    stdev[slot][1] += value;
+    stdev[slot][2] += value * value;
+}
+
+void dbg_correl_of(int64_t value1, int64_t value2, int slot) {
+
+    ++correl[slot][0];
+    correl[slot][1] += value1;
+    correl[slot][2] += value1 * value1;
+    correl[slot][3] += value2;
+    correl[slot][4] += value2 * value2;
+    correl[slot][5] += value1 * value2;
+}
+
+void dbg_print() {
+
+    int64_t n;
+    auto E   = [&n](int64_t x) { return double(x) / n; };
+    auto sqr = [](double x) { return x * x; };
+
+    for (int i = 0; i < MaxDebugSlots; ++i)
+        if ((n = hit[i][0]))
+            std::cerr << "Hit #" << i
+                      << ": Total " << n << " Hits " << hit[i][1]
+                      << " Hit Rate (%) " << 100.0 * E(hit[i][1])
+                      << fakeendl;
+
+    for (int i = 0; i < MaxDebugSlots; ++i)
+        if ((n = mean[i][0]))
+        {
+            std::cerr << "Mean #" << i
+                      << ": Total " << n << " Mean " << E(mean[i][1])
+                      << fakeendl;
+        }
+
+    for (int i = 0; i < MaxDebugSlots; ++i)
+        if ((n = stdev[i][0]))
+        {
+            double r = sqrtl(E(stdev[i][2]) - sqr(E(stdev[i][1])));
+            std::cerr << "Stdev #" << i
+                      << ": Total " << n << " Stdev " << r
+                      << fakeendl;
+        }
+
+    for (int i = 0; i < MaxDebugSlots; ++i)
+        if ((n = correl[i][0]))
+        {
+            double r = (E(correl[i][5]) - E(correl[i][1]) * E(correl[i][3]))
+                       / (  sqrtl(E(correl[i][2]) - sqr(E(correl[i][1])))
+                          * sqrtl(E(correl[i][4]) - sqr(E(correl[i][3]))));
+            std::cerr << "Correl. #" << i
+                      << ": Total " << n << " Coefficient " << r
+                      << fakeendl;
+        }
+}
+
+
+/// Used to serialize access to fakeout to avoid multiple threads writing at
+/// the same time.
+
+std::ostream& operator<<(std::ostream& os, SyncCout sc) {
+
+  static std::mutex m;
+
+  if (sc == IO_LOCK)
+      m.lock();
+
+  if (sc == IO_UNLOCK)
+      m.unlock();
+
+  return os;
+}
+
+
+/// Trampoline helper to avoid moving Logger to misc.h
+void start_logger(const std::string& fname) { Logger::start(fname); }
+
+
+/// prefetch() preloads the given address in L1/L2 cache. This is a non-blocking
+/// function that doesn't stall the CPU waiting for data to be loaded from memory,
+/// which can be quite slow.
+#ifdef NO_PREFETCH
+
+void prefetch(void*) {}
+
+#else
+
+void prefetch(void* addr) {
+
+#  if defined(__INTEL_COMPILER)
+   // This hack prevents prefetches from being optimized away by
+   // Intel compiler. Both MSVC and gcc seem not be affected by this.
+   __asm__ ("");
+#  endif
+
+#  if defined(__INTEL_COMPILER) || defined(_MSC_VER)
+  _mm_prefetch((char*)addr, _MM_HINT_T0);
+#  else
+  __builtin_prefetch(addr);
+#  endif
+}
+
+#endif
+
+
+/// std_aligned_alloc() is our wrapper for systems where the c++17 implementation
+/// does not guarantee the availability of aligned_alloc(). Memory allocated with
+/// std_aligned_alloc() must be freed with std_aligned_free().
+
+void* std_aligned_alloc(size_t alignment, size_t size) {
+
+#if defined(POSIXALIGNEDALLOC)
+  void *mem;
+  return posix_memalign(&mem, alignment, size) ? nullptr : mem;
+#elif defined(_WIN32) && !defined(_M_ARM) && !defined(_M_ARM64)
+  return _mm_malloc(size, alignment);
+#elif defined(_WIN32)
+  return _aligned_malloc(size, alignment);
+#else
+  return std::aligned_alloc(alignment, size);
+#endif
+}
+
+void std_aligned_free(void* ptr) {
+
+#if defined(POSIXALIGNEDALLOC)
+  free(ptr);
+#elif defined(_WIN32) && !defined(_M_ARM) && !defined(_M_ARM64)
+  _mm_free(ptr);
+#elif defined(_WIN32)
+  _aligned_free(ptr);
+#else
+  free(ptr);
+#endif
+}
+
+/// aligned_large_pages_alloc() will return suitably aligned memory, if possible using large pages.
+
+#if defined(_WIN32)
+
+static void* aligned_large_pages_alloc_windows([[maybe_unused]] size_t allocSize) {
+
+  #if !defined(_WIN64)
+    return nullptr;
+  #else
+
+  HANDLE hProcessToken { };
+  LUID luid { };
+  void* mem = nullptr;
+
+  const size_t largePageSize = GetLargePageMinimum();
+  if (!largePageSize)
+      return nullptr;
+
+  // Dynamically link OpenProcessToken, LookupPrivilegeValue and AdjustTokenPrivileges
+
+  HMODULE hAdvapi32 = GetModuleHandle(TEXT("advapi32.dll"));
+
+  if (!hAdvapi32)
+      hAdvapi32 = LoadLibrary(TEXT("advapi32.dll"));
+
+  auto fun6 = (fun6_t)(void(*)())GetProcAddress(hAdvapi32, "OpenProcessToken");
+  if (!fun6)
+      return nullptr;
+  auto fun7 = (fun7_t)(void(*)())GetProcAddress(hAdvapi32, "LookupPrivilegeValueA");
+  if (!fun7)
+      return nullptr;
+  auto fun8 = (fun8_t)(void(*)())GetProcAddress(hAdvapi32, "AdjustTokenPrivileges");
+  if (!fun8)
+      return nullptr;
+
+  // We need SeLockMemoryPrivilege, so try to enable it for the process
+  if (!fun6( // OpenProcessToken()
+      GetCurrentProcess(), TOKEN_ADJUST_PRIVILEGES | TOKEN_QUERY, &hProcessToken))
+          return nullptr;
+
+  if (fun7( // LookupPrivilegeValue(nullptr, SE_LOCK_MEMORY_NAME, &luid)
+      nullptr, "SeLockMemoryPrivilege", &luid))
+  {
+      TOKEN_PRIVILEGES tp { };
+      TOKEN_PRIVILEGES prevTp { };
+      DWORD prevTpLen = 0;
+
+      tp.PrivilegeCount = 1;
+      tp.Privileges[0].Luid = luid;
+      tp.Privileges[0].Attributes = SE_PRIVILEGE_ENABLED;
+
+      // Try to enable SeLockMemoryPrivilege. Note that even if AdjustTokenPrivileges() succeeds,
+      // we still need to query GetLastError() to ensure that the privileges were actually obtained.
+      if (fun8( // AdjustTokenPrivileges()
+              hProcessToken, FALSE, &tp, sizeof(TOKEN_PRIVILEGES), &prevTp, &prevTpLen) &&
+          GetLastError() == ERROR_SUCCESS)
+      {
+          // Round up size to full pages and allocate
+          allocSize = (allocSize + largePageSize - 1) & ~size_t(largePageSize - 1);
+          mem = VirtualAlloc(
+              nullptr, allocSize, MEM_RESERVE | MEM_COMMIT | MEM_LARGE_PAGES, PAGE_READWRITE);
+
+          // Privilege no longer needed, restore previous state
+          fun8( // AdjustTokenPrivileges ()
+              hProcessToken, FALSE, &prevTp, 0, nullptr, nullptr);
+      }
+  }
+
+  CloseHandle(hProcessToken);
+
+  return mem;
+
+  #endif
+}
+
+void* aligned_large_pages_alloc(size_t allocSize) {
+
+  // Try to allocate large pages
+  void* mem = aligned_large_pages_alloc_windows(allocSize);
+
+  // Fall back to regular, page aligned, allocation if necessary
+  if (!mem)
+      mem = VirtualAlloc(nullptr, allocSize, MEM_RESERVE | MEM_COMMIT, PAGE_READWRITE);
+
+  return mem;
+}
+
+#else
+
+void* aligned_large_pages_alloc(size_t allocSize) {
+
+#if defined(__linux__)
+  constexpr size_t alignment = 2 * 1024 * 1024; // assumed 2MB page size
+#else
+  constexpr size_t alignment = 4096; // assumed small page size
+#endif
+
+  // round up to multiples of alignment
+  size_t size = ((allocSize + alignment - 1) / alignment) * alignment;
+  void *mem = std_aligned_alloc(alignment, size);
+#if defined(MADV_HUGEPAGE)
+  madvise(mem, size, MADV_HUGEPAGE);
+#endif
+  return mem;
+}
+
+#endif
+
+
+/// aligned_large_pages_free() will free the previously allocated ttmem
+
+#if defined(_WIN32)
+
+void aligned_large_pages_free(void* mem) {
+
+  if (mem && !VirtualFree(mem, 0, MEM_RELEASE))
+  {
+      DWORD err = GetLastError();
+      std::cerr << "Failed to free large page memory. Error code: 0x"
+                << std::hex << err
+                << std::dec << fakeendl;
+      exit(EXIT_FAILURE);
+  }
+}
+
+#else
+
+void aligned_large_pages_free(void *mem) {
+  std_aligned_free(mem);
+}
+
+#endif
+
+
+namespace WinProcGroup {
+
+#ifndef _WIN32
+
+void bindThisThread(size_t) {}
+
+#else
+
+/// best_node() retrieves logical processor information using Windows specific
+/// API and returns the best node id for the thread with index idx. Original
+/// code from Texel by Peter Österlund.
+
+static int best_node(size_t idx) {
+
+  int threads = 0;
+  int nodes = 0;
+  int cores = 0;
+  DWORD returnLength = 0;
+  DWORD byteOffset = 0;
+
+  // Early exit if the needed API is not available at runtime
+  HMODULE k32 = GetModuleHandle(TEXT("Kernel32.dll"));
+  auto fun1 = (fun1_t)(void(*)())GetProcAddress(k32, "GetLogicalProcessorInformationEx");
+  if (!fun1)
+      return -1;
+
+  // First call to GetLogicalProcessorInformationEx() to get returnLength.
+  // We expect the call to fail due to null buffer.
+  if (fun1(RelationAll, nullptr, &returnLength))
+      return -1;
+
+  // Once we know returnLength, allocate the buffer
+  SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX *buffer, *ptr;
+  ptr = buffer = (SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX*)malloc(returnLength);
+
+  // Second call to GetLogicalProcessorInformationEx(), now we expect to succeed
+  if (!fun1(RelationAll, buffer, &returnLength))
+  {
+      free(buffer);
+      return -1;
+  }
+
+  while (byteOffset < returnLength)
+  {
+      if (ptr->Relationship == RelationNumaNode)
+          nodes++;
+
+      else if (ptr->Relationship == RelationProcessorCore)
+      {
+          cores++;
+          threads += (ptr->Processor.Flags == LTP_PC_SMT) ? 2 : 1;
+      }
+
+      assert(ptr->Size);
+      byteOffset += ptr->Size;
+      ptr = (SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX*)(((char*)ptr) + ptr->Size);
+  }
+
+  free(buffer);
+
+  std::vector<int> groups;
+
+  // Run as many threads as possible on the same node until core limit is
+  // reached, then move on filling the next node.
+  for (int n = 0; n < nodes; n++)
+      for (int i = 0; i < cores / nodes; i++)
+          groups.push_back(n);
+
+  // In case a core has more than one logical processor (we assume 2) and we
+  // have still threads to allocate, then spread them evenly across available
+  // nodes.
+  for (int t = 0; t < threads - cores; t++)
+      groups.push_back(t % nodes);
+
+  // If we still have more threads than the total number of logical processors
+  // then return -1 and let the OS to decide what to do.
+  return idx < groups.size() ? groups[idx] : -1;
+}
+
+
+/// bindThisThread() set the group affinity of the current thread
+
+void bindThisThread(size_t idx) {
+
+  // Use only local variables to be thread-safe
+  int node = best_node(idx);
+
+  if (node == -1)
+      return;
+
+  // Early exit if the needed API are not available at runtime
+  HMODULE k32 = GetModuleHandle(TEXT("Kernel32.dll"));
+  auto fun2 = (fun2_t)(void(*)())GetProcAddress(k32, "GetNumaNodeProcessorMaskEx");
+  auto fun3 = (fun3_t)(void(*)())GetProcAddress(k32, "SetThreadGroupAffinity");
+  auto fun4 = (fun4_t)(void(*)())GetProcAddress(k32, "GetNumaNodeProcessorMask2");
+  auto fun5 = (fun5_t)(void(*)())GetProcAddress(k32, "GetMaximumProcessorGroupCount");
+
+  if (!fun2 || !fun3)
+      return;
+
+  if (!fun4 || !fun5)
+  {
+      GROUP_AFFINITY affinity;
+      if (fun2(node, &affinity))                                                 // GetNumaNodeProcessorMaskEx
+          fun3(GetCurrentThread(), &affinity, nullptr);                          // SetThreadGroupAffinity
+  }
+  else
+  {
+      // If a numa node has more than one processor group, we assume they are
+      // sized equal and we spread threads evenly across the groups.
+      USHORT elements, returnedElements;
+      elements = fun5();                                                         // GetMaximumProcessorGroupCount
+      GROUP_AFFINITY *affinity = (GROUP_AFFINITY*)malloc(elements * sizeof(GROUP_AFFINITY));
+      if (fun4(node, affinity, elements, &returnedElements))                     // GetNumaNodeProcessorMask2
+          fun3(GetCurrentThread(), &affinity[idx % returnedElements], nullptr);  // SetThreadGroupAffinity
+      free(affinity);
+  }
+}
+
+#endif
+
+} // namespace WinProcGroup
+
+#ifdef _WIN32
+#include <direct.h>
+#define GETCWD _getcwd
+#else
+#include <unistd.h>
+#define GETCWD getcwd
+#endif
+
+namespace CommandLine {
+
+string argv0;            // path+name of the executable binary, as given by argv[0]
+string binaryDirectory;  // path of the executable directory
+string workingDirectory; // path of the working directory
+
+void init([[maybe_unused]] int argc, char* argv[]) {
+    string pathSeparator;
+
+    // extract the path+name of the executable binary
+    argv0 = argv[0];
+
+#ifdef _WIN32
+    pathSeparator = "\\";
+  #ifdef _MSC_VER
+    // Under windows argv[0] may not have the extension. Also _get_pgmptr() had
+    // issues in some windows 10 versions, so check returned values carefully.
+    char* pgmptr = nullptr;
+    if (!_get_pgmptr(&pgmptr) && pgmptr != nullptr && *pgmptr)
+        argv0 = pgmptr;
+  #endif
+#else
+    pathSeparator = "/";
+#endif
+
+    // extract the working directory
+    workingDirectory = "";
+    char buff[40000];
+    char* cwd = GETCWD(buff, 40000);
+    if (cwd)
+        workingDirectory = cwd;
+
+    // extract the binary directory path from argv0
+    binaryDirectory = argv0;
+    size_t pos = binaryDirectory.find_last_of("\\/");
+    if (pos == std::string::npos)
+        binaryDirectory = "." + pathSeparator;
+    else
+        binaryDirectory.resize(pos + 1);
+
+    // pattern replacement: "./" at the start of path is replaced by the working directory
+    if (binaryDirectory.find("." + pathSeparator) == 0)
+        binaryDirectory.replace(0, 1, workingDirectory);
+}
+
+
+} // namespace CommandLine
+
+} // namespace Stockfish

src/Stockfish/src/misc.h

@@ -0,0 +1,176 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef MISC_H_INCLUDED
+#define MISC_H_INCLUDED
+
+#include <cassert>
+#include <chrono>
+#include <ostream>
+#include <string>
+#include <vector>
+#include <cstdint>
+
+#include "types.h"
+
+#define stringify2(x) #x
+#define stringify(x) stringify2(x)
+
+namespace Stockfish {
+
+std::string engine_info(bool to_uci = false);
+std::string compiler_info();
+void prefetch(void* addr);
+void start_logger(const std::string& fname);
+void* std_aligned_alloc(size_t alignment, size_t size);
+void std_aligned_free(void* ptr);
+void* aligned_large_pages_alloc(size_t size); // memory aligned by page size, min alignment: 4096 bytes
+void aligned_large_pages_free(void* mem); // nop if mem == nullptr
+
+void dbg_hit_on(bool cond, int slot = 0);
+void dbg_mean_of(int64_t value, int slot = 0);
+void dbg_stdev_of(int64_t value, int slot = 0);
+void dbg_correl_of(int64_t value1, int64_t value2, int slot = 0);
+void dbg_print();
+
+using TimePoint = std::chrono::milliseconds::rep; // A value in milliseconds
+static_assert(sizeof(TimePoint) == sizeof(int64_t), "TimePoint should be 64 bits");
+inline TimePoint now() {
+  return std::chrono::duration_cast<std::chrono::milliseconds>
+        (std::chrono::steady_clock::now().time_since_epoch()).count();
+}
+
+template<class Entry, int Size>
+struct HashTable {
+  Entry* operator[](Key key) { return &table[(uint32_t)key & (Size - 1)]; }
+
+private:
+  std::vector<Entry> table = std::vector<Entry>(Size); // Allocate on the heap
+};
+
+
+enum SyncCout { IO_LOCK, IO_UNLOCK };
+std::ostream& operator<<(std::ostream&, SyncCout);
+
+#define sync_cout fakeout << IO_LOCK
+#define sync_endl fakeendl << IO_UNLOCK
+
+
+// align_ptr_up() : get the first aligned element of an array.
+// ptr must point to an array of size at least `sizeof(T) * N + alignment` bytes,
+// where N is the number of elements in the array.
+template <uintptr_t Alignment, typename T>
+T* align_ptr_up(T* ptr)
+{
+  static_assert(alignof(T) < Alignment);
+
+  const uintptr_t ptrint = reinterpret_cast<uintptr_t>(reinterpret_cast<char*>(ptr));
+  return reinterpret_cast<T*>(reinterpret_cast<char*>((ptrint + (Alignment - 1)) / Alignment * Alignment));
+}
+
+
+// IsLittleEndian : true if and only if the binary is compiled on a little endian machine
+static inline const union { uint32_t i; char c[4]; } Le = { 0x01020304 };
+static inline const bool IsLittleEndian = (Le.c[0] == 4);
+
+
+template <typename T, std::size_t MaxSize>
+class ValueList {
+
+public:
+  std::size_t size() const { return size_; }
+  void push_back(const T& value) { values_[size_++] = value; }
+  const T* begin() const { return values_; }
+  const T* end() const { return values_ + size_; }
+
+private:
+  T values_[MaxSize];
+  std::size_t size_ = 0;
+};
+
+
+/// xorshift64star Pseudo-Random Number Generator
+/// This class is based on original code written and dedicated
+/// to the public domain by Sebastiano Vigna (2014).
+/// It has the following characteristics:
+///
+///  -  Outputs 64-bit numbers
+///  -  Passes Dieharder and SmallCrush test batteries
+///  -  Does not require warm-up, no zeroland to escape
+///  -  Internal state is a single 64-bit integer
+///  -  Period is 2^64 - 1
+///  -  Speed: 1.60 ns/call (Core i7 @3.40GHz)
+///
+/// For further analysis see
+///   <http://vigna.di.unimi.it/ftp/papers/xorshift.pdf>
+
+class PRNG {
+
+  uint64_t s;
+
+  uint64_t rand64() {
+
+    s ^= s >> 12, s ^= s << 25, s ^= s >> 27;
+    return s * 2685821657736338717LL;
+  }
+
+public:
+  PRNG(uint64_t seed) : s(seed) { assert(seed); }
+
+  template<typename T> T rand() { return T(rand64()); }
+
+  /// Special generator used to fast init magic numbers.
+  /// Output values only have 1/8th of their bits set on average.
+  template<typename T> T sparse_rand()
+  { return T(rand64() & rand64() & rand64()); }
+};
+
+inline uint64_t mul_hi64(uint64_t a, uint64_t b) {
+#if defined(__GNUC__) && defined(IS_64BIT)
+    __extension__ using uint128 = unsigned __int128;
+    return ((uint128)a * (uint128)b) >> 64;
+#else
+    uint64_t aL = (uint32_t)a, aH = a >> 32;
+    uint64_t bL = (uint32_t)b, bH = b >> 32;
+    uint64_t c1 = (aL * bL) >> 32;
+    uint64_t c2 = aH * bL + c1;
+    uint64_t c3 = aL * bH + (uint32_t)c2;
+    return aH * bH + (c2 >> 32) + (c3 >> 32);
+#endif
+}
+
+/// Under Windows it is not possible for a process to run on more than one
+/// logical processor group. This usually means to be limited to use max 64
+/// cores. To overcome this, some special platform specific API should be
+/// called to set group affinity for each thread. Original code from Texel by
+/// Peter Österlund.
+
+namespace WinProcGroup {
+  void bindThisThread(size_t idx);
+}
+
+namespace CommandLine {
+  void init(int argc, char* argv[]);
+
+  extern std::string binaryDirectory;  // path of the executable directory
+  extern std::string workingDirectory; // path of the working directory
+}
+
+} // namespace Stockfish
+
+#endif // #ifndef MISC_H_INCLUDED

src/Stockfish/src/movegen.cpp

@@ -0,0 +1,284 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include <cassert>
+
+#include "movegen.h"
+#include "position.h"
+
+namespace Stockfish {
+
+namespace {
+
+  template<GenType Type, Direction D, bool Enemy>
+  ExtMove* make_promotions(ExtMove* moveList, [[maybe_unused]] Square to) {
+
+    if constexpr (Type == CAPTURES || Type == EVASIONS || Type == NON_EVASIONS)
+    {
+        *moveList++ = make<PROMOTION>(to - D, to, QUEEN);
+        if constexpr (Enemy && Type == CAPTURES)
+        {
+            *moveList++ = make<PROMOTION>(to - D, to, ROOK);
+            *moveList++ = make<PROMOTION>(to - D, to, BISHOP);
+            *moveList++ = make<PROMOTION>(to - D, to, KNIGHT);
+        }
+    }
+
+    if constexpr ((Type == QUIETS && !Enemy) || Type == EVASIONS || Type == NON_EVASIONS)
+    {
+        *moveList++ = make<PROMOTION>(to - D, to, ROOK);
+        *moveList++ = make<PROMOTION>(to - D, to, BISHOP);
+        *moveList++ = make<PROMOTION>(to - D, to, KNIGHT);
+    }
+
+    return moveList;
+  }
+
+
+  template<Color Us, GenType Type>
+  ExtMove* generate_pawn_moves(const Position& pos, ExtMove* moveList, Bitboard target) {
+
+    constexpr Color     Them     = ~Us;
+    constexpr Bitboard  TRank7BB = (Us == WHITE ? Rank7BB    : Rank2BB);
+    constexpr Bitboard  TRank3BB = (Us == WHITE ? Rank3BB    : Rank6BB);
+    constexpr Direction Up       = pawn_push(Us);
+    constexpr Direction UpRight  = (Us == WHITE ? NORTH_EAST : SOUTH_WEST);
+    constexpr Direction UpLeft   = (Us == WHITE ? NORTH_WEST : SOUTH_EAST);
+
+    const Bitboard emptySquares = ~pos.pieces();
+    const Bitboard enemies      =  Type == EVASIONS ? pos.checkers()
+                                                    : pos.pieces(Them);
+
+    Bitboard pawnsOn7    = pos.pieces(Us, PAWN) &  TRank7BB;
+    Bitboard pawnsNotOn7 = pos.pieces(Us, PAWN) & ~TRank7BB;
+
+    // Single and double pawn pushes, no promotions
+    if constexpr (Type != CAPTURES)
+    {
+        Bitboard b1 = shift<Up>(pawnsNotOn7)   & emptySquares;
+        Bitboard b2 = shift<Up>(b1 & TRank3BB) & emptySquares;
+
+        if constexpr (Type == EVASIONS) // Consider only blocking squares
+        {
+            b1 &= target;
+            b2 &= target;
+        }
+
+        if constexpr (Type == QUIET_CHECKS)
+        {
+            // To make a quiet check, you either make a direct check by pushing a pawn
+            // or push a blocker pawn that is not on the same file as the enemy king.
+            // Discovered check promotion has been already generated amongst the captures.
+            Square ksq = pos.square<KING>(Them);
+            Bitboard dcCandidatePawns = pos.blockers_for_king(Them) & ~file_bb(ksq);
+            b1 &= pawn_attacks_bb(Them, ksq) | shift<   Up>(dcCandidatePawns);
+            b2 &= pawn_attacks_bb(Them, ksq) | shift<Up+Up>(dcCandidatePawns);
+        }
+
+        while (b1)
+        {
+            Square to = pop_lsb(b1);
+            *moveList++ = make_move(to - Up, to);
+        }
+
+        while (b2)
+        {
+            Square to = pop_lsb(b2);
+            *moveList++ = make_move(to - Up - Up, to);
+        }
+    }
+
+    // Promotions and underpromotions
+    if (pawnsOn7)
+    {
+        Bitboard b1 = shift<UpRight>(pawnsOn7) & enemies;
+        Bitboard b2 = shift<UpLeft >(pawnsOn7) & enemies;
+        Bitboard b3 = shift<Up     >(pawnsOn7) & emptySquares;
+
+        if constexpr (Type == EVASIONS)
+            b3 &= target;
+
+        while (b1)
+            moveList = make_promotions<Type, UpRight, true>(moveList, pop_lsb(b1));
+
+        while (b2)
+            moveList = make_promotions<Type, UpLeft, true>(moveList, pop_lsb(b2));
+
+        while (b3)
+            moveList = make_promotions<Type, Up,    false>(moveList, pop_lsb(b3));
+    }
+
+    // Standard and en passant captures
+    if constexpr (Type == CAPTURES || Type == EVASIONS || Type == NON_EVASIONS)
+    {
+        Bitboard b1 = shift<UpRight>(pawnsNotOn7) & enemies;
+        Bitboard b2 = shift<UpLeft >(pawnsNotOn7) & enemies;
+
+        while (b1)
+        {
+            Square to = pop_lsb(b1);
+            *moveList++ = make_move(to - UpRight, to);
+        }
+
+        while (b2)
+        {
+            Square to = pop_lsb(b2);
+            *moveList++ = make_move(to - UpLeft, to);
+        }
+
+        if (pos.ep_square() != SQ_NONE)
+        {
+            assert(rank_of(pos.ep_square()) == relative_rank(Us, RANK_6));
+
+            // An en passant capture cannot resolve a discovered check
+            if (Type == EVASIONS && (target & (pos.ep_square() + Up)))
+                return moveList;
+
+            b1 = pawnsNotOn7 & pawn_attacks_bb(Them, pos.ep_square());
+
+            assert(b1);
+
+            while (b1)
+                *moveList++ = make<EN_PASSANT>(pop_lsb(b1), pos.ep_square());
+        }
+    }
+
+    return moveList;
+  }
+
+
+  template<Color Us, PieceType Pt, bool Checks>
+  ExtMove* generate_moves(const Position& pos, ExtMove* moveList, Bitboard target) {
+
+    static_assert(Pt != KING && Pt != PAWN, "Unsupported piece type in generate_moves()");
+
+    Bitboard bb = pos.pieces(Us, Pt);
+
+    while (bb)
+    {
+        Square from = pop_lsb(bb);
+        Bitboard b = attacks_bb<Pt>(from, pos.pieces()) & target;
+
+        // To check, you either move freely a blocker or make a direct check.
+        if (Checks && (Pt == QUEEN || !(pos.blockers_for_king(~Us) & from)))
+            b &= pos.check_squares(Pt);
+
+        while (b)
+            *moveList++ = make_move(from, pop_lsb(b));
+    }
+
+    return moveList;
+  }
+
+
+  template<Color Us, GenType Type>
+  ExtMove* generate_all(const Position& pos, ExtMove* moveList) {
+
+    static_assert(Type != LEGAL, "Unsupported type in generate_all()");
+
+    constexpr bool Checks = Type == QUIET_CHECKS; // Reduce template instantiations
+    const Square ksq = pos.square<KING>(Us);
+    Bitboard target;
+
+    // Skip generating non-king moves when in double check
+    if (Type != EVASIONS || !more_than_one(pos.checkers()))
+    {
+        target = Type == EVASIONS     ?  between_bb(ksq, lsb(pos.checkers()))
+               : Type == NON_EVASIONS ? ~pos.pieces( Us)
+               : Type == CAPTURES     ?  pos.pieces(~Us)
+                                      : ~pos.pieces(   ); // QUIETS || QUIET_CHECKS
+
+        moveList = generate_pawn_moves<Us, Type>(pos, moveList, target);
+        moveList = generate_moves<Us, KNIGHT, Checks>(pos, moveList, target);
+        moveList = generate_moves<Us, BISHOP, Checks>(pos, moveList, target);
+        moveList = generate_moves<Us,   ROOK, Checks>(pos, moveList, target);
+        moveList = generate_moves<Us,  QUEEN, Checks>(pos, moveList, target);
+    }
+
+    if (!Checks || pos.blockers_for_king(~Us) & ksq)
+    {
+        Bitboard b = attacks_bb<KING>(ksq) & (Type == EVASIONS ? ~pos.pieces(Us) : target);
+        if (Checks)
+            b &= ~attacks_bb<QUEEN>(pos.square<KING>(~Us));
+
+        while (b)
+            *moveList++ = make_move(ksq, pop_lsb(b));
+
+        if ((Type == QUIETS || Type == NON_EVASIONS) && pos.can_castle(Us & ANY_CASTLING))
+            for (CastlingRights cr : { Us & KING_SIDE, Us & QUEEN_SIDE } )
+                if (!pos.castling_impeded(cr) && pos.can_castle(cr))
+                    *moveList++ = make<CASTLING>(ksq, pos.castling_rook_square(cr));
+    }
+
+    return moveList;
+  }
+
+} // namespace
+
+
+/// <CAPTURES>     Generates all pseudo-legal captures plus queen promotions
+/// <QUIETS>       Generates all pseudo-legal non-captures and underpromotions
+/// <EVASIONS>     Generates all pseudo-legal check evasions when the side to move is in check
+/// <QUIET_CHECKS> Generates all pseudo-legal non-captures giving check, except castling and promotions
+/// <NON_EVASIONS> Generates all pseudo-legal captures and non-captures
+///
+/// Returns a pointer to the end of the move list.
+
+template<GenType Type>
+ExtMove* generate(const Position& pos, ExtMove* moveList) {
+
+  static_assert(Type != LEGAL, "Unsupported type in generate()");
+  assert((Type == EVASIONS) == (bool)pos.checkers());
+
+  Color us = pos.side_to_move();
+
+  return us == WHITE ? generate_all<WHITE, Type>(pos, moveList)
+                     : generate_all<BLACK, Type>(pos, moveList);
+}
+
+// Explicit template instantiations
+template ExtMove* generate<CAPTURES>(const Position&, ExtMove*);
+template ExtMove* generate<QUIETS>(const Position&, ExtMove*);
+template ExtMove* generate<EVASIONS>(const Position&, ExtMove*);
+template ExtMove* generate<QUIET_CHECKS>(const Position&, ExtMove*);
+template ExtMove* generate<NON_EVASIONS>(const Position&, ExtMove*);
+
+
+/// generate<LEGAL> generates all the legal moves in the given position
+
+template<>
+ExtMove* generate<LEGAL>(const Position& pos, ExtMove* moveList) {
+
+  Color us = pos.side_to_move();
+  Bitboard pinned = pos.blockers_for_king(us) & pos.pieces(us);
+  Square ksq = pos.square<KING>(us);
+  ExtMove* cur = moveList;
+
+  moveList = pos.checkers() ? generate<EVASIONS    >(pos, moveList)
+                            : generate<NON_EVASIONS>(pos, moveList);
+  while (cur != moveList)
+      if (  ((pinned & from_sq(*cur)) || from_sq(*cur) == ksq || type_of(*cur) == EN_PASSANT)
+          && !pos.legal(*cur))
+          *cur = (--moveList)->move;
+      else
+          ++cur;
+
+  return moveList;
+}
+
+} // namespace Stockfish

src/Stockfish/src/movegen.h

@@ -0,0 +1,77 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef MOVEGEN_H_INCLUDED
+#define MOVEGEN_H_INCLUDED
+
+#include <algorithm>
+
+#include "types.h"
+
+namespace Stockfish {
+
+class Position;
+
+enum GenType {
+  CAPTURES,
+  QUIETS,
+  QUIET_CHECKS,
+  EVASIONS,
+  NON_EVASIONS,
+  LEGAL
+};
+
+struct ExtMove {
+  Move move;
+  int value;
+
+  operator Move() const { return move; }
+  void operator=(Move m) { move = m; }
+
+  // Inhibit unwanted implicit conversions to Move
+  // with an ambiguity that yields to a compile error.
+  operator float() const = delete;
+};
+
+inline bool operator<(const ExtMove& f, const ExtMove& s) {
+  return f.value < s.value;
+}
+
+template<GenType>
+ExtMove* generate(const Position& pos, ExtMove* moveList);
+
+/// The MoveList struct is a simple wrapper around generate(). It sometimes comes
+/// in handy to use this class instead of the low level generate() function.
+template<GenType T>
+struct MoveList {
+
+  explicit MoveList(const Position& pos) : last(generate<T>(pos, moveList)) {}
+  const ExtMove* begin() const { return moveList; }
+  const ExtMove* end() const { return last; }
+  size_t size() const { return last - moveList; }
+  bool contains(Move move) const {
+    return std::find(begin(), end(), move) != end();
+  }
+
+private:
+  ExtMove moveList[MAX_MOVES], *last;
+};
+
+} // namespace Stockfish
+
+#endif // #ifndef MOVEGEN_H_INCLUDED

src/Stockfish/src/movepick.cpp

@@ -0,0 +1,295 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include <cassert>
+
+#include "bitboard.h"
+#include "movepick.h"
+
+namespace Stockfish {
+
+namespace {
+
+  enum Stages {
+    MAIN_TT, CAPTURE_INIT, GOOD_CAPTURE, REFUTATION, QUIET_INIT, QUIET, BAD_CAPTURE,
+    EVASION_TT, EVASION_INIT, EVASION,
+    PROBCUT_TT, PROBCUT_INIT, PROBCUT,
+    QSEARCH_TT, QCAPTURE_INIT, QCAPTURE, QCHECK_INIT, QCHECK
+  };
+
+  // partial_insertion_sort() sorts moves in descending order up to and including
+  // a given limit. The order of moves smaller than the limit is left unspecified.
+  void partial_insertion_sort(ExtMove* begin, ExtMove* end, int limit) {
+
+    for (ExtMove *sortedEnd = begin, *p = begin + 1; p < end; ++p)
+        if (p->value >= limit)
+        {
+            ExtMove tmp = *p, *q;
+            *p = *++sortedEnd;
+            for (q = sortedEnd; q != begin && *(q - 1) < tmp; --q)
+                *q = *(q - 1);
+            *q = tmp;
+        }
+  }
+
+} // namespace
+
+
+/// Constructors of the MovePicker class. As arguments we pass information
+/// to help it to return the (presumably) good moves first, to decide which
+/// moves to return (in the quiescence search, for instance, we only want to
+/// search captures, promotions, and some checks) and how important good move
+/// ordering is at the current node.
+
+/// MovePicker constructor for the main search
+MovePicker::MovePicker(const Position& p, Move ttm, Depth d, const ButterflyHistory* mh,
+                                                             const CapturePieceToHistory* cph,
+                                                             const PieceToHistory** ch,
+                                                             Move cm,
+                                                             const Move* killers)
+           : pos(p), mainHistory(mh), captureHistory(cph), continuationHistory(ch),
+             ttMove(ttm), refutations{{killers[0], 0}, {killers[1], 0}, {cm, 0}}, depth(d)
+{
+  assert(d > 0);
+
+  stage = (pos.checkers() ? EVASION_TT : MAIN_TT) +
+          !(ttm && pos.pseudo_legal(ttm));
+}
+
+/// MovePicker constructor for quiescence search
+MovePicker::MovePicker(const Position& p, Move ttm, Depth d, const ButterflyHistory* mh,
+                                                             const CapturePieceToHistory* cph,
+                                                             const PieceToHistory** ch,
+                                                             Square rs)
+           : pos(p), mainHistory(mh), captureHistory(cph), continuationHistory(ch), ttMove(ttm), recaptureSquare(rs), depth(d)
+{
+  assert(d <= 0);
+
+  stage = (pos.checkers() ? EVASION_TT : QSEARCH_TT) +
+          !(   ttm
+            && pos.pseudo_legal(ttm));
+}
+
+/// MovePicker constructor for ProbCut: we generate captures with SEE greater
+/// than or equal to the given threshold.
+MovePicker::MovePicker(const Position& p, Move ttm, Value th, const CapturePieceToHistory* cph)
+           : pos(p), captureHistory(cph), ttMove(ttm), threshold(th)
+{
+  assert(!pos.checkers());
+
+  stage = PROBCUT_TT + !(ttm && pos.capture_stage(ttm)
+                             && pos.pseudo_legal(ttm)
+                             && pos.see_ge(ttm, threshold));
+}
+
+/// MovePicker::score() assigns a numerical value to each move in a list, used
+/// for sorting. Captures are ordered by Most Valuable Victim (MVV), preferring
+/// captures with a good history. Quiets moves are ordered using the history tables.
+template<GenType Type>
+void MovePicker::score() {
+
+  static_assert(Type == CAPTURES || Type == QUIETS || Type == EVASIONS, "Wrong type");
+
+  [[maybe_unused]] Bitboard threatenedByPawn, threatenedByMinor, threatenedByRook, threatenedPieces;
+  if constexpr (Type == QUIETS)
+  {
+      Color us = pos.side_to_move();
+
+      threatenedByPawn  = pos.attacks_by<PAWN>(~us);
+      threatenedByMinor = pos.attacks_by<KNIGHT>(~us) | pos.attacks_by<BISHOP>(~us) | threatenedByPawn;
+      threatenedByRook  = pos.attacks_by<ROOK>(~us) | threatenedByMinor;
+
+      // Pieces threatened by pieces of lesser material value
+      threatenedPieces = (pos.pieces(us, QUEEN) & threatenedByRook)
+                       | (pos.pieces(us, ROOK)  & threatenedByMinor)
+                       | (pos.pieces(us, KNIGHT, BISHOP) & threatenedByPawn);
+  }
+
+  for (auto& m : *this)
+      if constexpr (Type == CAPTURES)
+          m.value =  (7 * int(PieceValue[MG][pos.piece_on(to_sq(m))])
+                   +     (*captureHistory)[pos.moved_piece(m)][to_sq(m)][type_of(pos.piece_on(to_sq(m)))]) / 16;
+
+      else if constexpr (Type == QUIETS)
+          m.value =  2 * (*mainHistory)[pos.side_to_move()][from_to(m)]
+                   + 2 * (*continuationHistory[0])[pos.moved_piece(m)][to_sq(m)]
+                   +     (*continuationHistory[1])[pos.moved_piece(m)][to_sq(m)]
+                   +     (*continuationHistory[3])[pos.moved_piece(m)][to_sq(m)]
+                   +     (*continuationHistory[5])[pos.moved_piece(m)][to_sq(m)]
+                   +     (threatenedPieces & from_sq(m) ?
+                           (type_of(pos.moved_piece(m)) == QUEEN && !(to_sq(m) & threatenedByRook)  ? 50000
+                          : type_of(pos.moved_piece(m)) == ROOK  && !(to_sq(m) & threatenedByMinor) ? 25000
+                          :                                         !(to_sq(m) & threatenedByPawn)  ? 15000
+                          :                                                                           0)
+                          :                                                                           0)
+                   +     bool(pos.check_squares(type_of(pos.moved_piece(m))) & to_sq(m)) * 16384;
+      else // Type == EVASIONS
+      {
+          if (pos.capture_stage(m))
+              m.value =  PieceValue[MG][pos.piece_on(to_sq(m))]
+                       - Value(type_of(pos.moved_piece(m)))
+                       + (1 << 28);
+          else
+              m.value =  (*mainHistory)[pos.side_to_move()][from_to(m)]
+                       + (*continuationHistory[0])[pos.moved_piece(m)][to_sq(m)];
+      }
+}
+
+/// MovePicker::select() returns the next move satisfying a predicate function.
+/// It never returns the TT move.
+template<MovePicker::PickType T, typename Pred>
+Move MovePicker::select(Pred filter) {
+
+  while (cur < endMoves)
+  {
+      if constexpr (T == Best)
+          std::swap(*cur, *std::max_element(cur, endMoves));
+
+      if (*cur != ttMove && filter())
+          return *cur++;
+
+      cur++;
+  }
+  return MOVE_NONE;
+}
+
+/// MovePicker::next_move() is the most important method of the MovePicker class. It
+/// returns a new pseudo-legal move every time it is called until there are no more
+/// moves left, picking the move with the highest score from a list of generated moves.
+Move MovePicker::next_move(bool skipQuiets) {
+
+top:
+  switch (stage) {
+
+  case MAIN_TT:
+  case EVASION_TT:
+  case QSEARCH_TT:
+  case PROBCUT_TT:
+      ++stage;
+      return ttMove;
+
+  case CAPTURE_INIT:
+  case PROBCUT_INIT:
+  case QCAPTURE_INIT:
+      cur = endBadCaptures = moves;
+      endMoves = generate<CAPTURES>(pos, cur);
+
+      score<CAPTURES>();
+      partial_insertion_sort(cur, endMoves, std::numeric_limits<int>::min());
+      ++stage;
+      goto top;
+
+  case GOOD_CAPTURE:
+      if (select<Next>([&](){
+                       return pos.see_ge(*cur, Value(-cur->value)) ?
+                              // Move losing capture to endBadCaptures to be tried later
+                              true : (*endBadCaptures++ = *cur, false); }))
+          return *(cur - 1);
+
+      // Prepare the pointers to loop over the refutations array
+      cur = std::begin(refutations);
+      endMoves = std::end(refutations);
+
+      // If the countermove is the same as a killer, skip it
+      if (   refutations[0].move == refutations[2].move
+          || refutations[1].move == refutations[2].move)
+          --endMoves;
+
+      ++stage;
+      [[fallthrough]];
+
+  case REFUTATION:
+      if (select<Next>([&](){ return    *cur != MOVE_NONE
+                                    && !pos.capture_stage(*cur)
+                                    &&  pos.pseudo_legal(*cur); }))
+          return *(cur - 1);
+      ++stage;
+      [[fallthrough]];
+
+  case QUIET_INIT:
+      if (!skipQuiets)
+      {
+          cur = endBadCaptures;
+          endMoves = generate<QUIETS>(pos, cur);
+
+          score<QUIETS>();
+          partial_insertion_sort(cur, endMoves, -3000 * depth);
+      }
+
+      ++stage;
+      [[fallthrough]];
+
+  case QUIET:
+      if (   !skipQuiets
+          && select<Next>([&](){return   *cur != refutations[0].move
+                                      && *cur != refutations[1].move
+                                      && *cur != refutations[2].move;}))
+          return *(cur - 1);
+
+      // Prepare the pointers to loop over the bad captures
+      cur = moves;
+      endMoves = endBadCaptures;
+
+      ++stage;
+      [[fallthrough]];
+
+  case BAD_CAPTURE:
+      return select<Next>([](){ return true; });
+
+  case EVASION_INIT:
+      cur = moves;
+      endMoves = generate<EVASIONS>(pos, cur);
+
+      score<EVASIONS>();
+      ++stage;
+      [[fallthrough]];
+
+  case EVASION:
+      return select<Best>([](){ return true; });
+
+  case PROBCUT:
+      return select<Next>([&](){ return pos.see_ge(*cur, threshold); });
+
+  case QCAPTURE:
+      if (select<Next>([&](){ return   depth > DEPTH_QS_RECAPTURES
+                                    || to_sq(*cur) == recaptureSquare; }))
+          return *(cur - 1);
+
+      // If we did not find any move and we do not try checks, we have finished
+      if (depth != DEPTH_QS_CHECKS)
+          return MOVE_NONE;
+
+      ++stage;
+      [[fallthrough]];
+
+  case QCHECK_INIT:
+      cur = moves;
+      endMoves = generate<QUIET_CHECKS>(pos, cur);
+
+      ++stage;
+      [[fallthrough]];
+
+  case QCHECK:
+      return select<Next>([](){ return true; });
+  }
+
+  assert(false);
+  return MOVE_NONE; // Silence warning
+}
+
+} // namespace Stockfish

src/Stockfish/src/movepick.h

@@ -0,0 +1,155 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef MOVEPICK_H_INCLUDED
+#define MOVEPICK_H_INCLUDED
+
+#include <array>
+#include <limits>
+#include <type_traits>
+
+#include "movegen.h"
+#include "position.h"
+#include "types.h"
+
+namespace Stockfish {
+
+/// StatsEntry stores the stat table value. It is usually a number but could
+/// be a move or even a nested history. We use a class instead of naked value
+/// to directly call history update operator<<() on the entry so to use stats
+/// tables at caller sites as simple multi-dim arrays.
+template<typename T, int D>
+class StatsEntry {
+
+  T entry;
+
+public:
+  void operator=(const T& v) { entry = v; }
+  T* operator&() { return &entry; }
+  T* operator->() { return &entry; }
+  operator const T&() const { return entry; }
+
+  void operator<<(int bonus) {
+    assert(abs(bonus) <= D); // Ensure range is [-D, D]
+    static_assert(D <= std::numeric_limits<T>::max(), "D overflows T");
+
+    entry += bonus - entry * abs(bonus) / D;
+
+    assert(abs(entry) <= D);
+  }
+};
+
+/// Stats is a generic N-dimensional array used to store various statistics.
+/// The first template parameter T is the base type of the array, the second
+/// template parameter D limits the range of updates in [-D, D] when we update
+/// values with the << operator, while the last parameters (Size and Sizes)
+/// encode the dimensions of the array.
+template <typename T, int D, int Size, int... Sizes>
+struct Stats : public std::array<Stats<T, D, Sizes...>, Size>
+{
+  using stats = Stats<T, D, Size, Sizes...>;
+
+  void fill(const T& v) {
+
+    // For standard-layout 'this' points to first struct member
+    assert(std::is_standard_layout<stats>::value);
+
+    using entry = StatsEntry<T, D>;
+    entry* p = reinterpret_cast<entry*>(this);
+    std::fill(p, p + sizeof(*this) / sizeof(entry), v);
+  }
+};
+
+template <typename T, int D, int Size>
+struct Stats<T, D, Size> : public std::array<StatsEntry<T, D>, Size> {};
+
+/// In stats table, D=0 means that the template parameter is not used
+enum StatsParams { NOT_USED = 0 };
+enum StatsType { NoCaptures, Captures };
+
+/// ButterflyHistory records how often quiet moves have been successful or
+/// unsuccessful during the current search, and is used for reduction and move
+/// ordering decisions. It uses 2 tables (one for each color) indexed by
+/// the move's from and to squares, see www.chessprogramming.org/Butterfly_Boards
+/// (~11 elo)
+using ButterflyHistory = Stats<int16_t, 7183, COLOR_NB, int(SQUARE_NB) * int(SQUARE_NB)>;
+
+/// CounterMoveHistory stores counter moves indexed by [piece][to] of the previous
+/// move, see www.chessprogramming.org/Countermove_Heuristic
+using CounterMoveHistory = Stats<Move, NOT_USED, PIECE_NB, SQUARE_NB>;
+
+/// CapturePieceToHistory is addressed by a move's [piece][to][captured piece type]
+using CapturePieceToHistory = Stats<int16_t, 10692, PIECE_NB, SQUARE_NB, PIECE_TYPE_NB>;
+
+/// PieceToHistory is like ButterflyHistory but is addressed by a move's [piece][to]
+using PieceToHistory = Stats<int16_t, 29952, PIECE_NB, SQUARE_NB>;
+
+/// ContinuationHistory is the combined history of a given pair of moves, usually
+/// the current one given a previous one. The nested history table is based on
+/// PieceToHistory instead of ButterflyBoards.
+/// (~63 elo)
+using ContinuationHistory = Stats<PieceToHistory, NOT_USED, PIECE_NB, SQUARE_NB>;
+
+
+/// MovePicker class is used to pick one pseudo-legal move at a time from the
+/// current position. The most important method is next_move(), which returns a
+/// new pseudo-legal move each time it is called, until there are no moves left,
+/// when MOVE_NONE is returned. In order to improve the efficiency of the
+/// alpha-beta algorithm, MovePicker attempts to return the moves which are most
+/// likely to get a cut-off first.
+class MovePicker {
+
+  enum PickType { Next, Best };
+
+public:
+  MovePicker(const MovePicker&) = delete;
+  MovePicker& operator=(const MovePicker&) = delete;
+  MovePicker(const Position&, Move, Depth, const ButterflyHistory*,
+                                           const CapturePieceToHistory*,
+                                           const PieceToHistory**,
+                                           Move,
+                                           const Move*);
+  MovePicker(const Position&, Move, Depth, const ButterflyHistory*,
+                                           const CapturePieceToHistory*,
+                                           const PieceToHistory**,
+                                           Square);
+  MovePicker(const Position&, Move, Value, const CapturePieceToHistory*);
+  Move next_move(bool skipQuiets = false);
+
+private:
+  template<PickType T, typename Pred> Move select(Pred);
+  template<GenType> void score();
+  ExtMove* begin() { return cur; }
+  ExtMove* end() { return endMoves; }
+
+  const Position& pos;
+  const ButterflyHistory* mainHistory;
+  const CapturePieceToHistory* captureHistory;
+  const PieceToHistory** continuationHistory;
+  Move ttMove;
+  ExtMove refutations[3], *cur, *endMoves, *endBadCaptures;
+  int stage;
+  Square recaptureSquare;
+  Value threshold;
+  Depth depth;
+  ExtMove moves[MAX_MOVES];
+};
+
+} // namespace Stockfish
+
+#endif // #ifndef MOVEPICK_H_INCLUDED

src/Stockfish/src/nnue/evaluate_nnue.cpp

@@ -0,0 +1,405 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+// Code for calculating NNUE evaluation function
+
+#include <fstream>
+#include <iomanip>
+#include <iostream>
+#include <set>
+#include <sstream>
+#include <string_view>
+
+#include "../evaluate.h"
+#include "../position.h"
+#include "../uci.h"
+#include "../types.h"
+
+#include "evaluate_nnue.h"
+
+namespace Stockfish::Eval::NNUE {
+
+  // Input feature converter
+  LargePagePtr<FeatureTransformer> featureTransformer;
+
+  // Evaluation function
+  AlignedPtr<Network> network[LayerStacks];
+
+  // Evaluation function file name
+  std::string fileName;
+  std::string netDescription;
+
+  namespace Detail {
+
+  // Initialize the evaluation function parameters
+  template <typename T>
+  void initialize(AlignedPtr<T>& pointer) {
+
+    pointer.reset(reinterpret_cast<T*>(std_aligned_alloc(alignof(T), sizeof(T))));
+    std::memset(pointer.get(), 0, sizeof(T));
+  }
+
+  template <typename T>
+  void initialize(LargePagePtr<T>& pointer) {
+
+    static_assert(alignof(T) <= 4096, "aligned_large_pages_alloc() may fail for such a big alignment requirement of T");
+    pointer.reset(reinterpret_cast<T*>(aligned_large_pages_alloc(sizeof(T))));
+    std::memset(pointer.get(), 0, sizeof(T));
+  }
+
+  // Read evaluation function parameters
+  template <typename T>
+  bool read_parameters(std::istream& stream, T& reference) {
+
+    std::uint32_t header;
+    header = read_little_endian<std::uint32_t>(stream);
+    if (!stream || header != T::get_hash_value()) return false;
+    return reference.read_parameters(stream);
+  }
+
+  // Write evaluation function parameters
+  template <typename T>
+  bool write_parameters(std::ostream& stream, const T& reference) {
+
+    write_little_endian<std::uint32_t>(stream, T::get_hash_value());
+    return reference.write_parameters(stream);
+  }
+
+  }  // namespace Detail
+
+  // Initialize the evaluation function parameters
+  static void initialize() {
+
+    Detail::initialize(featureTransformer);
+    for (std::size_t i = 0; i < LayerStacks; ++i)
+      Detail::initialize(network[i]);
+  }
+
+  // Read network header
+  static bool read_header(std::istream& stream, std::uint32_t* hashValue, std::string* desc)
+  {
+    std::uint32_t version, size;
+
+    version     = read_little_endian<std::uint32_t>(stream);
+    *hashValue  = read_little_endian<std::uint32_t>(stream);
+    size        = read_little_endian<std::uint32_t>(stream);
+    if (!stream || version != Version) return false;
+    desc->resize(size);
+    stream.read(&(*desc)[0], size);
+    return !stream.fail();
+  }
+
+  // Write network header
+  static bool write_header(std::ostream& stream, std::uint32_t hashValue, const std::string& desc)
+  {
+    write_little_endian<std::uint32_t>(stream, Version);
+    write_little_endian<std::uint32_t>(stream, hashValue);
+    write_little_endian<std::uint32_t>(stream, (std::uint32_t)desc.size());
+    stream.write(&desc[0], desc.size());
+    return !stream.fail();
+  }
+
+  // Read network parameters
+  static bool read_parameters(std::istream& stream) {
+
+    std::uint32_t hashValue;
+    if (!read_header(stream, &hashValue, &netDescription)) return false;
+    if (hashValue != HashValue) return false;
+    if (!Detail::read_parameters(stream, *featureTransformer)) return false;
+    for (std::size_t i = 0; i < LayerStacks; ++i)
+      if (!Detail::read_parameters(stream, *(network[i]))) return false;
+    return stream && stream.peek() == std::ios::traits_type::eof();
+  }
+
+  // Write network parameters
+  static bool write_parameters(std::ostream& stream) {
+
+    if (!write_header(stream, HashValue, netDescription)) return false;
+    if (!Detail::write_parameters(stream, *featureTransformer)) return false;
+    for (std::size_t i = 0; i < LayerStacks; ++i)
+      if (!Detail::write_parameters(stream, *(network[i]))) return false;
+    return (bool)stream;
+  }
+
+  void hint_common_parent_position(const Position& pos) {
+    if (Eval::useNNUE)
+        featureTransformer->hint_common_access(pos);
+  }
+
+  // Evaluation function. Perform differential calculation.
+  Value evaluate(const Position& pos, bool adjusted, int* complexity) {
+
+    // We manually align the arrays on the stack because with gcc < 9.3
+    // overaligning stack variables with alignas() doesn't work correctly.
+
+    constexpr uint64_t alignment = CacheLineSize;
+    constexpr int delta = 24;
+
+#if defined(ALIGNAS_ON_STACK_VARIABLES_BROKEN)
+    TransformedFeatureType transformedFeaturesUnaligned[
+      FeatureTransformer::BufferSize + alignment / sizeof(TransformedFeatureType)];
+
+    auto* transformedFeatures = align_ptr_up<alignment>(&transformedFeaturesUnaligned[0]);
+#else
+    alignas(alignment)
+      TransformedFeatureType transformedFeatures[FeatureTransformer::BufferSize];
+#endif
+
+    ASSERT_ALIGNED(transformedFeatures, alignment);
+
+    const int bucket = (pos.count<ALL_PIECES>() - 1) / 4;
+    const auto psqt = featureTransformer->transform(pos, transformedFeatures, bucket);
+    const auto positional = network[bucket]->propagate(transformedFeatures);
+
+    if (complexity)
+        *complexity = abs(psqt - positional) / OutputScale;
+
+    // Give more value to positional evaluation when adjusted flag is set
+    if (adjusted)
+        return static_cast<Value>(((1024 - delta) * psqt + (1024 + delta) * positional) / (1024 * OutputScale));
+    else
+        return static_cast<Value>((psqt + positional) / OutputScale);
+  }
+
+  struct NnueEvalTrace {
+    static_assert(LayerStacks == PSQTBuckets);
+
+    Value psqt[LayerStacks];
+    Value positional[LayerStacks];
+    std::size_t correctBucket;
+  };
+
+  static NnueEvalTrace trace_evaluate(const Position& pos) {
+
+    // We manually align the arrays on the stack because with gcc < 9.3
+    // overaligning stack variables with alignas() doesn't work correctly.
+
+    constexpr uint64_t alignment = CacheLineSize;
+
+#if defined(ALIGNAS_ON_STACK_VARIABLES_BROKEN)
+    TransformedFeatureType transformedFeaturesUnaligned[
+      FeatureTransformer::BufferSize + alignment / sizeof(TransformedFeatureType)];
+
+    auto* transformedFeatures = align_ptr_up<alignment>(&transformedFeaturesUnaligned[0]);
+#else
+    alignas(alignment)
+      TransformedFeatureType transformedFeatures[FeatureTransformer::BufferSize];
+#endif
+
+    ASSERT_ALIGNED(transformedFeatures, alignment);
+
+    NnueEvalTrace t{};
+    t.correctBucket = (pos.count<ALL_PIECES>() - 1) / 4;
+    for (IndexType bucket = 0; bucket < LayerStacks; ++bucket) {
+      const auto materialist = featureTransformer->transform(pos, transformedFeatures, bucket);
+      const auto positional = network[bucket]->propagate(transformedFeatures);
+
+      t.psqt[bucket] = static_cast<Value>( materialist / OutputScale );
+      t.positional[bucket] = static_cast<Value>( positional / OutputScale );
+    }
+
+    return t;
+  }
+
+  constexpr std::string_view PieceToChar(" PNBRQK  pnbrqk");
+
+
+  // format_cp_compact() converts a Value into (centi)pawns and writes it in a buffer.
+  // The buffer must have capacity for at least 5 chars.
+  static void format_cp_compact(Value v, char* buffer) {
+
+    buffer[0] = (v < 0 ? '-' : v > 0 ? '+' : ' ');
+
+    int cp = std::abs(100 * v / UCI::NormalizeToPawnValue);
+    if (cp >= 10000)
+    {
+        buffer[1] = '0' + cp / 10000; cp %= 10000;
+        buffer[2] = '0' + cp / 1000; cp %= 1000;
+        buffer[3] = '0' + cp / 100;
+        buffer[4] = ' ';
+    }
+    else if (cp >= 1000)
+    {
+        buffer[1] = '0' + cp / 1000; cp %= 1000;
+        buffer[2] = '0' + cp / 100; cp %= 100;
+        buffer[3] = '.';
+        buffer[4] = '0' + cp / 10;
+    }
+    else
+    {
+        buffer[1] = '0' + cp / 100; cp %= 100;
+        buffer[2] = '.';
+        buffer[3] = '0' + cp / 10; cp %= 10;
+        buffer[4] = '0' + cp / 1;
+    }
+  }
+
+
+  // format_cp_aligned_dot() converts a Value into (centi)pawns, always keeping two decimals.
+  static void format_cp_aligned_dot(Value v, std::stringstream &stream) {
+    const double cp = 1.0 * std::abs(int(v)) / UCI::NormalizeToPawnValue;
+
+    stream << (v < 0 ? '-' : v > 0 ? '+' : ' ')
+           << std::setiosflags(std::ios::fixed)
+           << std::setw(6)
+           << std::setprecision(2)
+           << cp;
+  }
+
+
+  // trace() returns a string with the value of each piece on a board,
+  // and a table for (PSQT, Layers) values bucket by bucket.
+
+  std::string trace(Position& pos) {
+
+    std::stringstream ss;
+
+    char board[3*8+1][8*8+2];
+    std::memset(board, ' ', sizeof(board));
+    for (int row = 0; row < 3*8+1; ++row)
+      board[row][8*8+1] = '\0';
+
+    // A lambda to output one box of the board
+    auto writeSquare = [&board](File file, Rank rank, Piece pc, Value value) {
+
+      const int x = ((int)file) * 8;
+      const int y = (7 - (int)rank) * 3;
+      for (int i = 1; i < 8; ++i)
+         board[y][x+i] = board[y+3][x+i] = '-';
+      for (int i = 1; i < 3; ++i)
+         board[y+i][x] = board[y+i][x+8] = '|';
+      board[y][x] = board[y][x+8] = board[y+3][x+8] = board[y+3][x] = '+';
+      if (pc != NO_PIECE)
+        board[y+1][x+4] = PieceToChar[pc];
+      if (value != VALUE_NONE)
+        format_cp_compact(value, &board[y+2][x+2]);
+    };
+
+    // We estimate the value of each piece by doing a differential evaluation from
+    // the current base eval, simulating the removal of the piece from its square.
+    Value base = evaluate(pos);
+    base = pos.side_to_move() == WHITE ? base : -base;
+
+    for (File f = FILE_A; f <= FILE_H; ++f)
+      for (Rank r = RANK_1; r <= RANK_8; ++r)
+      {
+        Square sq = make_square(f, r);
+        Piece pc = pos.piece_on(sq);
+        Value v = VALUE_NONE;
+
+        if (pc != NO_PIECE && type_of(pc) != KING)
+        {
+          auto st = pos.state();
+
+          pos.remove_piece(sq);
+          st->accumulator.computed[WHITE] = false;
+          st->accumulator.computed[BLACK] = false;
+
+          Value eval = evaluate(pos);
+          eval = pos.side_to_move() == WHITE ? eval : -eval;
+          v = base - eval;
+
+          pos.put_piece(pc, sq);
+          st->accumulator.computed[WHITE] = false;
+          st->accumulator.computed[BLACK] = false;
+        }
+
+        writeSquare(f, r, pc, v);
+      }
+
+    ss << " NNUE derived piece values:\n";
+    for (int row = 0; row < 3*8+1; ++row)
+        ss << board[row] << '\n';
+    ss << '\n';
+
+    auto t = trace_evaluate(pos);
+
+    ss << " NNUE network contributions "
+       << (pos.side_to_move() == WHITE ? "(White to move)" : "(Black to move)") << fakeendl
+       << "+------------+------------+------------+------------+\n"
+       << "|   Bucket   |  Material  | Positional |   Total    |\n"
+       << "|            |   (PSQT)   |  (Layers)  |            |\n"
+       << "+------------+------------+------------+------------+\n";
+
+    for (std::size_t bucket = 0; bucket < LayerStacks; ++bucket)
+    {
+      ss <<  "|  " << bucket    << "        ";
+      ss << " |  "; format_cp_aligned_dot(t.psqt[bucket], ss); ss << "  "
+         << " |  "; format_cp_aligned_dot(t.positional[bucket], ss); ss << "  "
+         << " |  "; format_cp_aligned_dot(t.psqt[bucket] + t.positional[bucket], ss); ss << "  "
+         << " |";
+      if (bucket == t.correctBucket)
+          ss << " <-- this bucket is used";
+      ss << '\n';
+    }
+
+    ss << "+------------+------------+------------+------------+\n";
+
+    return ss.str();
+  }
+
+
+  // Load eval, from a file stream or a memory stream
+  bool load_eval(std::string name, std::istream& stream) {
+
+    initialize();
+    fileName = name;
+    return read_parameters(stream);
+  }
+
+  // Save eval, to a file stream or a memory stream
+  bool save_eval(std::ostream& stream) {
+
+    if (fileName.empty())
+      return false;
+
+    return write_parameters(stream);
+  }
+
+  /// Save eval, to a file given by its name
+  bool save_eval(const std::optional<std::string>& filename) {
+
+    std::string actualFilename;
+    std::string msg;
+
+    if (filename.has_value())
+        actualFilename = filename.value();
+    else
+    {
+        if (currentEvalFileName != EvalFileDefaultName)
+        {
+             msg = "Failed to export a net. A non-embedded net can only be saved if the filename is specified";
+
+             sync_cout << msg << sync_endl;
+             return false;
+        }
+        actualFilename = EvalFileDefaultName;
+    }
+
+    std::ofstream stream(actualFilename, std::ios_base::binary);
+    bool saved = save_eval(stream);
+
+    msg = saved ? "Network saved successfully to " + actualFilename
+                : "Failed to export a net";
+
+    sync_cout << msg << sync_endl;
+    return saved;
+  }
+
+
+} // namespace Stockfish::Eval::NNUE

src/Stockfish/src/nnue/evaluate_nnue.h

@@ -0,0 +1,68 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+// header used in NNUE evaluation function
+
+#ifndef NNUE_EVALUATE_NNUE_H_INCLUDED
+#define NNUE_EVALUATE_NNUE_H_INCLUDED
+
+#include "nnue_feature_transformer.h"
+
+#include <memory>
+
+namespace Stockfish::Eval::NNUE {
+
+  // Hash value of evaluation function structure
+  constexpr std::uint32_t HashValue =
+      FeatureTransformer::get_hash_value() ^ Network::get_hash_value();
+
+
+  // Deleter for automating release of memory area
+  template <typename T>
+  struct AlignedDeleter {
+    void operator()(T* ptr) const {
+      ptr->~T();
+      std_aligned_free(ptr);
+    }
+  };
+
+  template <typename T>
+  struct LargePageDeleter {
+    void operator()(T* ptr) const {
+      ptr->~T();
+      aligned_large_pages_free(ptr);
+    }
+  };
+
+  template <typename T>
+  using AlignedPtr = std::unique_ptr<T, AlignedDeleter<T>>;
+
+  template <typename T>
+  using LargePagePtr = std::unique_ptr<T, LargePageDeleter<T>>;
+
+  std::string trace(Position& pos);
+  Value evaluate(const Position& pos, bool adjusted = false, int* complexity = nullptr);
+  void hint_common_parent_position(const Position& pos);
+
+  bool load_eval(std::string name, std::istream& stream);
+  bool save_eval(std::ostream& stream);
+  bool save_eval(const std::optional<std::string>& filename);
+
+}  // namespace Stockfish::Eval::NNUE
+
+#endif // #ifndef NNUE_EVALUATE_NNUE_H_INCLUDED

src/Stockfish/src/nnue/features/half_ka_v2_hm.cpp

@@ -0,0 +1,84 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+//Definition of input features HalfKAv2_hm of NNUE evaluation function
+
+#include "half_ka_v2_hm.h"
+
+#include "../../position.h"
+
+namespace Stockfish::Eval::NNUE::Features {
+
+  // Index of a feature for a given king position and another piece on some square
+  template<Color Perspective>
+  inline IndexType HalfKAv2_hm::make_index(Square s, Piece pc, Square ksq) {
+    return IndexType((int(s) ^ OrientTBL[Perspective][ksq]) + PieceSquareIndex[Perspective][pc] + KingBuckets[Perspective][ksq]);
+  }
+
+  // Get a list of indices for active features
+  template<Color Perspective>
+  void HalfKAv2_hm::append_active_indices(
+    const Position& pos,
+    IndexList& active
+  ) {
+    Square ksq = pos.square<KING>(Perspective);
+    Bitboard bb = pos.pieces();
+    while (bb)
+    {
+      Square s = pop_lsb(bb);
+      active.push_back(make_index<Perspective>(s, pos.piece_on(s), ksq));
+    }
+  }
+
+  // Explicit template instantiations
+  template void HalfKAv2_hm::append_active_indices<WHITE>(const Position& pos, IndexList& active);
+  template void HalfKAv2_hm::append_active_indices<BLACK>(const Position& pos, IndexList& active);
+
+  // append_changed_indices() : get a list of indices for recently changed features
+  template<Color Perspective>
+  void HalfKAv2_hm::append_changed_indices(
+    Square ksq,
+    const DirtyPiece& dp,
+    IndexList& removed,
+    IndexList& added
+  ) {
+    for (int i = 0; i < dp.dirty_num; ++i) {
+      if (dp.from[i] != SQ_NONE)
+        removed.push_back(make_index<Perspective>(dp.from[i], dp.piece[i], ksq));
+      if (dp.to[i] != SQ_NONE)
+        added.push_back(make_index<Perspective>(dp.to[i], dp.piece[i], ksq));
+    }
+  }
+
+  // Explicit template instantiations
+  template void HalfKAv2_hm::append_changed_indices<WHITE>(Square ksq, const DirtyPiece& dp, IndexList& removed, IndexList& added);
+  template void HalfKAv2_hm::append_changed_indices<BLACK>(Square ksq, const DirtyPiece& dp, IndexList& removed, IndexList& added);
+
+  int HalfKAv2_hm::update_cost(const StateInfo* st) {
+    return st->dirtyPiece.dirty_num;
+  }
+
+  int HalfKAv2_hm::refresh_cost(const Position& pos) {
+    return pos.count<ALL_PIECES>();
+  }
+
+  bool HalfKAv2_hm::requires_refresh(const StateInfo* st, Color perspective) {
+    return st->dirtyPiece.piece[0] == make_piece(perspective, KING);
+  }
+
+}  // namespace Stockfish::Eval::NNUE::Features

src/Stockfish/src/nnue/features/half_ka_v2_hm.h

@@ -0,0 +1,152 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+//Definition of input features HalfKP of NNUE evaluation function
+
+#ifndef NNUE_FEATURES_HALF_KA_V2_HM_H_INCLUDED
+#define NNUE_FEATURES_HALF_KA_V2_HM_H_INCLUDED
+
+#include "../nnue_common.h"
+
+#include "../../evaluate.h"
+#include "../../misc.h"
+
+namespace Stockfish {
+  struct StateInfo;
+}
+
+namespace Stockfish::Eval::NNUE::Features {
+
+  // Feature HalfKAv2_hm: Combination of the position of own king
+  // and the position of pieces. Position mirrored such that king always on e..h files.
+  class HalfKAv2_hm {
+
+    // unique number for each piece type on each square
+    enum {
+      PS_NONE     =  0,
+      PS_W_PAWN   =  0,
+      PS_B_PAWN   =  1 * SQUARE_NB,
+      PS_W_KNIGHT =  2 * SQUARE_NB,
+      PS_B_KNIGHT =  3 * SQUARE_NB,
+      PS_W_BISHOP =  4 * SQUARE_NB,
+      PS_B_BISHOP =  5 * SQUARE_NB,
+      PS_W_ROOK   =  6 * SQUARE_NB,
+      PS_B_ROOK   =  7 * SQUARE_NB,
+      PS_W_QUEEN  =  8 * SQUARE_NB,
+      PS_B_QUEEN  =  9 * SQUARE_NB,
+      PS_KING     = 10 * SQUARE_NB,
+      PS_NB       = 11 * SQUARE_NB
+    };
+
+    static constexpr IndexType PieceSquareIndex[COLOR_NB][PIECE_NB] = {
+      // convention: W - us, B - them
+      // viewed from other side, W and B are reversed
+      { PS_NONE, PS_W_PAWN, PS_W_KNIGHT, PS_W_BISHOP, PS_W_ROOK, PS_W_QUEEN, PS_KING, PS_NONE,
+        PS_NONE, PS_B_PAWN, PS_B_KNIGHT, PS_B_BISHOP, PS_B_ROOK, PS_B_QUEEN, PS_KING, PS_NONE },
+      { PS_NONE, PS_B_PAWN, PS_B_KNIGHT, PS_B_BISHOP, PS_B_ROOK, PS_B_QUEEN, PS_KING, PS_NONE,
+        PS_NONE, PS_W_PAWN, PS_W_KNIGHT, PS_W_BISHOP, PS_W_ROOK, PS_W_QUEEN, PS_KING, PS_NONE }
+    };
+
+    // Index of a feature for a given king position and another piece on some square
+    template<Color Perspective>
+    static IndexType make_index(Square s, Piece pc, Square ksq);
+
+   public:
+    // Feature name
+    static constexpr const char* Name = "HalfKAv2_hm(Friend)";
+
+    // Hash value embedded in the evaluation file
+    static constexpr std::uint32_t HashValue = 0x7f234cb8u;
+
+    // Number of feature dimensions
+    static constexpr IndexType Dimensions =
+        static_cast<IndexType>(SQUARE_NB) * static_cast<IndexType>(PS_NB) / 2;
+
+#define B(v) (v * PS_NB)
+    static constexpr int KingBuckets[COLOR_NB][SQUARE_NB] = {
+      { B(28), B(29), B(30), B(31), B(31), B(30), B(29), B(28),
+        B(24), B(25), B(26), B(27), B(27), B(26), B(25), B(24),
+        B(20), B(21), B(22), B(23), B(23), B(22), B(21), B(20),
+        B(16), B(17), B(18), B(19), B(19), B(18), B(17), B(16),
+        B(12), B(13), B(14), B(15), B(15), B(14), B(13), B(12),
+        B( 8), B( 9), B(10), B(11), B(11), B(10), B( 9), B( 8),
+        B( 4), B( 5), B( 6), B( 7), B( 7), B( 6), B( 5), B( 4),
+        B( 0), B( 1), B( 2), B( 3), B( 3), B( 2), B( 1), B( 0) },
+      { B( 0), B( 1), B( 2), B( 3), B( 3), B( 2), B( 1), B( 0),
+        B( 4), B( 5), B( 6), B( 7), B( 7), B( 6), B( 5), B( 4),
+        B( 8), B( 9), B(10), B(11), B(11), B(10), B( 9), B( 8),
+        B(12), B(13), B(14), B(15), B(15), B(14), B(13), B(12),
+        B(16), B(17), B(18), B(19), B(19), B(18), B(17), B(16),
+        B(20), B(21), B(22), B(23), B(23), B(22), B(21), B(20),
+        B(24), B(25), B(26), B(27), B(27), B(26), B(25), B(24),
+        B(28), B(29), B(30), B(31), B(31), B(30), B(29), B(28) }
+    };
+#undef B
+
+    // Orient a square according to perspective (rotates by 180 for black)
+    static constexpr int OrientTBL[COLOR_NB][SQUARE_NB] = {
+      { SQ_H1, SQ_H1, SQ_H1, SQ_H1, SQ_A1, SQ_A1, SQ_A1, SQ_A1,
+        SQ_H1, SQ_H1, SQ_H1, SQ_H1, SQ_A1, SQ_A1, SQ_A1, SQ_A1,
+        SQ_H1, SQ_H1, SQ_H1, SQ_H1, SQ_A1, SQ_A1, SQ_A1, SQ_A1,
+        SQ_H1, SQ_H1, SQ_H1, SQ_H1, SQ_A1, SQ_A1, SQ_A1, SQ_A1,
+        SQ_H1, SQ_H1, SQ_H1, SQ_H1, SQ_A1, SQ_A1, SQ_A1, SQ_A1,
+        SQ_H1, SQ_H1, SQ_H1, SQ_H1, SQ_A1, SQ_A1, SQ_A1, SQ_A1,
+        SQ_H1, SQ_H1, SQ_H1, SQ_H1, SQ_A1, SQ_A1, SQ_A1, SQ_A1,
+        SQ_H1, SQ_H1, SQ_H1, SQ_H1, SQ_A1, SQ_A1, SQ_A1, SQ_A1 },
+      { SQ_H8, SQ_H8, SQ_H8, SQ_H8, SQ_A8, SQ_A8, SQ_A8, SQ_A8,
+        SQ_H8, SQ_H8, SQ_H8, SQ_H8, SQ_A8, SQ_A8, SQ_A8, SQ_A8,
+        SQ_H8, SQ_H8, SQ_H8, SQ_H8, SQ_A8, SQ_A8, SQ_A8, SQ_A8,
+        SQ_H8, SQ_H8, SQ_H8, SQ_H8, SQ_A8, SQ_A8, SQ_A8, SQ_A8,
+        SQ_H8, SQ_H8, SQ_H8, SQ_H8, SQ_A8, SQ_A8, SQ_A8, SQ_A8,
+        SQ_H8, SQ_H8, SQ_H8, SQ_H8, SQ_A8, SQ_A8, SQ_A8, SQ_A8,
+        SQ_H8, SQ_H8, SQ_H8, SQ_H8, SQ_A8, SQ_A8, SQ_A8, SQ_A8,
+        SQ_H8, SQ_H8, SQ_H8, SQ_H8, SQ_A8, SQ_A8, SQ_A8, SQ_A8 }
+    };
+
+    // Maximum number of simultaneously active features.
+    static constexpr IndexType MaxActiveDimensions = 32;
+    using IndexList = ValueList<IndexType, MaxActiveDimensions>;
+
+    // Get a list of indices for active features
+    template<Color Perspective>
+    static void append_active_indices(
+      const Position& pos,
+      IndexList& active);
+
+    // Get a list of indices for recently changed features
+    template<Color Perspective>
+    static void append_changed_indices(
+      Square ksq,
+      const DirtyPiece& dp,
+      IndexList& removed,
+      IndexList& added
+    );
+
+    // Returns the cost of updating one perspective, the most costly one.
+    // Assumes no refresh needed.
+    static int update_cost(const StateInfo* st);
+    static int refresh_cost(const Position& pos);
+
+    // Returns whether the change stored in this StateInfo means that
+    // a full accumulator refresh is required.
+    static bool requires_refresh(const StateInfo* st, Color perspective);
+  };
+
+}  // namespace Stockfish::Eval::NNUE::Features
+
+#endif // #ifndef NNUE_FEATURES_HALF_KA_V2_HM_H_INCLUDED

src/Stockfish/src/nnue/layers/affine_transform.h

@@ -0,0 +1,563 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+// Definition of layer AffineTransform of NNUE evaluation function
+
+#ifndef NNUE_LAYERS_AFFINE_TRANSFORM_H_INCLUDED
+#define NNUE_LAYERS_AFFINE_TRANSFORM_H_INCLUDED
+
+#include <iostream>
+#include <algorithm>
+#include <type_traits>
+#include "../nnue_common.h"
+#include "simd.h"
+
+/*
+  This file contains the definition for a fully connected layer (aka affine transform).
+  Two approaches are employed, depending on the sizes of the transform.
+
+  Approach 1 (a specialization for large inputs):
+    - used when the PaddedInputDimensions >= 128
+    - uses AVX512 if possible
+    - processes inputs in batches of 2*InputSimdWidth
+      - so in batches of 128 for AVX512
+    - the weight blocks of size InputSimdWidth are transposed such that
+      access is sequential
+    - N columns of the weight matrix are processed a time, where N
+      depends on the architecture (the amount of registers)
+    - accumulate + hadd is used
+
+  Approach 2 (a specialization for small inputs):
+    - used when the PaddedInputDimensions < 128
+    - expected use-case is for when PaddedInputDimensions == 32 and InputDimensions <= 32.
+      - that's why AVX512 is hard to implement
+    - expected use-case is small layers
+      - not optimized as well as the approach 1
+    - inputs are processed in chunks of 4, weights are respectively transposed
+    - accumulation happens directly to int32s
+*/
+
+namespace Stockfish::Eval::NNUE::Layers {
+
+// Fallback implementation for older/other architectures.
+// Identical for both approaches. Requires the input to be padded to at least 16 values.
+#if !defined(USE_SSSE3)
+  template <IndexType InputDimensions, IndexType PaddedInputDimensions, IndexType OutputDimensions>
+  static void affine_transform_non_ssse3(std::int32_t* output, const std::int8_t* weights, const std::int32_t* biases, const std::uint8_t* input)
+  {
+# if defined(USE_SSE2)
+    // At least a multiple of 16, with SSE2.
+    constexpr IndexType NumChunks = ceil_to_multiple<IndexType>(InputDimensions, 16) / 16;
+    const __m128i Zeros = _mm_setzero_si128();
+    const auto inputVector = reinterpret_cast<const __m128i*>(input);
+
+# elif defined(USE_MMX)
+    constexpr IndexType NumChunks = ceil_to_multiple<IndexType>(InputDimensions, 8) / 8;
+    const __m64 Zeros = _mm_setzero_si64();
+    const auto inputVector = reinterpret_cast<const __m64*>(input);
+
+# elif defined(USE_NEON_DOTPROD)
+    constexpr IndexType NumChunks = ceil_to_multiple<IndexType>(InputDimensions, 16) / 16;
+    const auto inputVector = reinterpret_cast<const int8x16_t*>(input);
+
+# elif defined(USE_NEON)
+    constexpr IndexType NumChunks = ceil_to_multiple<IndexType>(InputDimensions, 16) / 16;
+    const auto inputVector = reinterpret_cast<const int8x8_t*>(input);
+# endif
+
+    for (IndexType i = 0; i < OutputDimensions; ++i) {
+      const IndexType offset = i * PaddedInputDimensions;
+
+# if defined(USE_SSE2)
+      __m128i sumLo = _mm_cvtsi32_si128(biases[i]);
+      __m128i sumHi = Zeros;
+      const auto row = reinterpret_cast<const __m128i*>(&weights[offset]);
+      for (IndexType j = 0; j < NumChunks; ++j) {
+        __m128i row_j = _mm_load_si128(&row[j]);
+        __m128i input_j = _mm_load_si128(&inputVector[j]);
+        __m128i extendedRowLo = _mm_srai_epi16(_mm_unpacklo_epi8(row_j, row_j), 8);
+        __m128i extendedRowHi = _mm_srai_epi16(_mm_unpackhi_epi8(row_j, row_j), 8);
+        __m128i extendedInputLo = _mm_unpacklo_epi8(input_j, Zeros);
+        __m128i extendedInputHi = _mm_unpackhi_epi8(input_j, Zeros);
+        __m128i productLo = _mm_madd_epi16(extendedRowLo, extendedInputLo);
+        __m128i productHi = _mm_madd_epi16(extendedRowHi, extendedInputHi);
+        sumLo = _mm_add_epi32(sumLo, productLo);
+        sumHi = _mm_add_epi32(sumHi, productHi);
+      }
+      __m128i sum = _mm_add_epi32(sumLo, sumHi);
+      __m128i sumHigh_64 = _mm_shuffle_epi32(sum, _MM_SHUFFLE(1, 0, 3, 2));
+      sum = _mm_add_epi32(sum, sumHigh_64);
+      __m128i sum_second_32 = _mm_shufflelo_epi16(sum, _MM_SHUFFLE(1, 0, 3, 2));
+      sum = _mm_add_epi32(sum, sum_second_32);
+      output[i] = _mm_cvtsi128_si32(sum);
+
+# elif defined(USE_MMX)
+      __m64 sumLo = _mm_cvtsi32_si64(biases[i]);
+      __m64 sumHi = Zeros;
+      const auto row = reinterpret_cast<const __m64*>(&weights[offset]);
+      for (IndexType j = 0; j < NumChunks; ++j) {
+        __m64 row_j = row[j];
+        __m64 input_j = inputVector[j];
+        __m64 extendedRowLo = _mm_srai_pi16(_mm_unpacklo_pi8(row_j, row_j), 8);
+        __m64 extendedRowHi = _mm_srai_pi16(_mm_unpackhi_pi8(row_j, row_j), 8);
+        __m64 extendedInputLo = _mm_unpacklo_pi8(input_j, Zeros);
+        __m64 extendedInputHi = _mm_unpackhi_pi8(input_j, Zeros);
+        __m64 productLo = _mm_madd_pi16(extendedRowLo, extendedInputLo);
+        __m64 productHi = _mm_madd_pi16(extendedRowHi, extendedInputHi);
+        sumLo = _mm_add_pi32(sumLo, productLo);
+        sumHi = _mm_add_pi32(sumHi, productHi);
+      }
+      __m64 sum = _mm_add_pi32(sumLo, sumHi);
+      sum = _mm_add_pi32(sum, _mm_unpackhi_pi32(sum, sum));
+      output[i] = _mm_cvtsi64_si32(sum);
+
+# elif defined(USE_NEON_DOTPROD)
+      int32x4_t sum = {biases[i]};
+      const auto row = reinterpret_cast<const int8x16_t*>(&weights[offset]);
+      for (IndexType j = 0; j < NumChunks; ++j) {
+        sum = vdotq_s32(sum, inputVector[j], row[j]);
+      }
+      output[i] = vaddvq_s32(sum);
+
+# elif defined(USE_NEON)
+      int32x4_t sum = {biases[i]};
+      const auto row = reinterpret_cast<const int8x8_t*>(&weights[offset]);
+      for (IndexType j = 0; j < NumChunks; ++j) {
+        int16x8_t product = vmull_s8(inputVector[j * 2], row[j * 2]);
+        product = vmlal_s8(product, inputVector[j * 2 + 1], row[j * 2 + 1]);
+        sum = vpadalq_s16(sum, product);
+      }
+      output[i] = sum[0] + sum[1] + sum[2] + sum[3];
+
+# else
+      std::int32_t sum = biases[i];
+      for (IndexType j = 0; j < InputDimensions; ++j) {
+        sum += weights[offset + j] * input[j];
+      }
+      output[i] = sum;
+# endif
+    }
+
+# if defined(USE_MMX)
+    _mm_empty();
+# endif
+  }
+#endif
+
+  template <IndexType InDims, IndexType OutDims, typename Enabled = void>
+  class AffineTransform;
+
+#if defined (USE_AVX512)
+  constexpr IndexType LargeInputSize = 2 * 64;
+#else
+  constexpr IndexType LargeInputSize = std::numeric_limits<IndexType>::max();
+#endif
+
+  // A specialization for large inputs
+  template <IndexType InDims, IndexType OutDims>
+  class AffineTransform<InDims, OutDims, std::enable_if_t<(ceil_to_multiple<IndexType>(InDims, MaxSimdWidth) >= LargeInputSize)>> {
+   public:
+    // Input/output type
+    using InputType = std::uint8_t;
+    using OutputType = std::int32_t;
+
+    // Number of input/output dimensions
+    static constexpr IndexType InputDimensions = InDims;
+    static constexpr IndexType OutputDimensions = OutDims;
+
+    static constexpr IndexType PaddedInputDimensions =
+      ceil_to_multiple<IndexType>(InputDimensions, MaxSimdWidth);
+    static constexpr IndexType PaddedOutputDimensions =
+      ceil_to_multiple<IndexType>(OutputDimensions, MaxSimdWidth);
+
+    using OutputBuffer = OutputType[PaddedOutputDimensions];
+
+    static_assert(PaddedInputDimensions >= LargeInputSize, "Something went wrong. This specialization (for large inputs) should not have been chosen.");
+
+#if defined (USE_AVX512)
+    static constexpr IndexType InputSimdWidth = 64;
+    static constexpr IndexType MaxNumOutputRegs = 16;
+#elif defined (USE_AVX2)
+    static constexpr IndexType InputSimdWidth = 32;
+    static constexpr IndexType MaxNumOutputRegs = 8;
+#elif defined (USE_SSSE3)
+    static constexpr IndexType InputSimdWidth = 16;
+    static constexpr IndexType MaxNumOutputRegs = 8;
+#elif defined (USE_NEON_DOTPROD)
+    static constexpr IndexType InputSimdWidth = 16;
+    static constexpr IndexType MaxNumOutputRegs = 8;
+#elif defined (USE_NEON)
+    static constexpr IndexType InputSimdWidth = 8;
+    static constexpr IndexType MaxNumOutputRegs = 8;
+#else
+    // The fallback implementation will not have permuted weights.
+    // We define these to avoid a lot of ifdefs later.
+    static constexpr IndexType InputSimdWidth = 1;
+    static constexpr IndexType MaxNumOutputRegs = 1;
+#endif
+
+    // A big block is a region in the weight matrix of the size [PaddedInputDimensions, NumOutputRegs].
+    // A small block is a region of size [InputSimdWidth, 1]
+
+    static constexpr IndexType NumOutputRegs = std::min(MaxNumOutputRegs, OutputDimensions);
+    static constexpr IndexType SmallBlockSize = InputSimdWidth;
+    static constexpr IndexType BigBlockSize = NumOutputRegs * PaddedInputDimensions;
+    static constexpr IndexType NumSmallBlocksInBigBlock = BigBlockSize / SmallBlockSize;
+    static constexpr IndexType NumSmallBlocksPerOutput = PaddedInputDimensions / SmallBlockSize;
+    static constexpr IndexType NumBigBlocks = OutputDimensions / NumOutputRegs;
+
+    static_assert(OutputDimensions % NumOutputRegs == 0);
+
+    // Hash value embedded in the evaluation file
+    static constexpr std::uint32_t get_hash_value(std::uint32_t prevHash) {
+      std::uint32_t hashValue = 0xCC03DAE4u;
+      hashValue += OutputDimensions;
+      hashValue ^= prevHash >> 1;
+      hashValue ^= prevHash << 31;
+      return hashValue;
+    }
+
+    /*
+      Transposes the small blocks within a block.
+      Effectively means that weights can be traversed sequentially during inference.
+    */
+    static IndexType get_weight_index(IndexType i)
+    {
+      const IndexType smallBlock = (i / SmallBlockSize) % NumSmallBlocksInBigBlock;
+      const IndexType smallBlockCol = smallBlock / NumSmallBlocksPerOutput;
+      const IndexType smallBlockRow = smallBlock % NumSmallBlocksPerOutput;
+      const IndexType bigBlock   = i / BigBlockSize;
+      const IndexType rest       = i % SmallBlockSize;
+
+      const IndexType idx =
+          bigBlock * BigBlockSize
+        + smallBlockRow * SmallBlockSize * NumOutputRegs
+        + smallBlockCol * SmallBlockSize
+        + rest;
+
+      return idx;
+    }
+
+    // Read network parameters
+    bool read_parameters(std::istream& stream) {
+      read_little_endian<BiasType>(stream, biases, OutputDimensions);
+
+      for (IndexType i = 0; i < OutputDimensions * PaddedInputDimensions; ++i)
+        weights[get_weight_index(i)] = read_little_endian<WeightType>(stream);
+
+      return !stream.fail();
+    }
+
+    // Write network parameters
+    bool write_parameters(std::ostream& stream) const {
+      write_little_endian<BiasType>(stream, biases, OutputDimensions);
+
+      for (IndexType i = 0; i < OutputDimensions * PaddedInputDimensions; ++i)
+        write_little_endian<WeightType>(stream, weights[get_weight_index(i)]);
+
+      return !stream.fail();
+    }
+
+    // Forward propagation
+    const OutputType* propagate(
+        const InputType* input, OutputType* output) const {
+
+#if defined (USE_AVX512)
+      using acc_vec_t = __m512i;
+      using bias_vec_t = __m128i;
+      using weight_vec_t = __m512i;
+      using in_vec_t = __m512i;
+      #define vec_zero _mm512_setzero_si512()
+      #define vec_add_dpbusd_32x2 Simd::m512_add_dpbusd_epi32x2
+      #define vec_hadd Simd::m512_hadd
+      #define vec_haddx4 Simd::m512_haddx4
+#elif defined (USE_AVX2)
+      using acc_vec_t = __m256i;
+      using bias_vec_t = __m128i;
+      using weight_vec_t = __m256i;
+      using in_vec_t = __m256i;
+      #define vec_zero _mm256_setzero_si256()
+      #define vec_add_dpbusd_32x2 Simd::m256_add_dpbusd_epi32x2
+      #define vec_hadd Simd::m256_hadd
+      #define vec_haddx4 Simd::m256_haddx4
+#elif defined (USE_SSSE3)
+      using acc_vec_t = __m128i;
+      using bias_vec_t = __m128i;
+      using weight_vec_t = __m128i;
+      using in_vec_t = __m128i;
+      #define vec_zero _mm_setzero_si128()
+      #define vec_add_dpbusd_32x2 Simd::m128_add_dpbusd_epi32x2
+      #define vec_hadd Simd::m128_hadd
+      #define vec_haddx4 Simd::m128_haddx4
+#elif defined (USE_NEON_DOTPROD)
+      using acc_vec_t = int32x4_t;
+      using bias_vec_t = int32x4_t;
+      using weight_vec_t = int8x16_t;
+      using in_vec_t = int8x16_t;
+      #define vec_zero {0}
+      #define vec_add_dpbusd_32x2 Simd::dotprod_m128_add_dpbusd_epi32x2
+      #define vec_hadd Simd::neon_m128_hadd
+      #define vec_haddx4 Simd::neon_m128_haddx4
+#elif defined (USE_NEON)
+      using acc_vec_t = int32x4_t;
+      using bias_vec_t = int32x4_t;
+      using weight_vec_t = int8x8_t;
+      using in_vec_t = int8x8_t;
+      #define vec_zero {0}
+      #define vec_add_dpbusd_32x2 Simd::neon_m128_add_dpbusd_epi32x2
+      #define vec_hadd Simd::neon_m128_hadd
+      #define vec_haddx4 Simd::neon_m128_haddx4
+#endif
+
+#if defined (USE_SSSE3) || defined (USE_NEON)
+      const in_vec_t* invec = reinterpret_cast<const in_vec_t*>(input);
+
+      // Perform accumulation to registers for each big block
+      for (IndexType bigBlock = 0; bigBlock < NumBigBlocks; ++bigBlock)
+      {
+        acc_vec_t acc[NumOutputRegs] = { vec_zero };
+
+        // Each big block has NumOutputRegs small blocks in each "row", one per register.
+        // We process two small blocks at a time to save on one addition without VNNI.
+        for (IndexType smallBlock = 0; smallBlock < NumSmallBlocksPerOutput; smallBlock += 2)
+        {
+          const weight_vec_t* weightvec =
+            reinterpret_cast<const weight_vec_t*>(
+                weights
+              + bigBlock * BigBlockSize
+              + smallBlock * SmallBlockSize * NumOutputRegs);
+
+          const in_vec_t in0 = invec[smallBlock + 0];
+          const in_vec_t in1 = invec[smallBlock + 1];
+
+          for (IndexType k = 0; k < NumOutputRegs; ++k)
+            vec_add_dpbusd_32x2(acc[k], in0, weightvec[k], in1, weightvec[k + NumOutputRegs]);
+        }
+
+        // Horizontally add all accumulators.
+        if constexpr (NumOutputRegs % 4 == 0)
+        {
+          bias_vec_t* outputvec = reinterpret_cast<bias_vec_t*>(output);
+          const bias_vec_t* biasvec = reinterpret_cast<const bias_vec_t*>(biases);
+
+          for (IndexType k = 0; k < NumOutputRegs; k += 4)
+          {
+            const IndexType idx = (bigBlock * NumOutputRegs + k) / 4;
+            outputvec[idx] = vec_haddx4(acc[k+0], acc[k+1], acc[k+2], acc[k+3], biasvec[idx]);
+          }
+        }
+        else
+        {
+          for (IndexType k = 0; k < NumOutputRegs; ++k)
+          {
+            const IndexType idx = (bigBlock * NumOutputRegs + k);
+            output[idx] = vec_hadd(acc[k], biases[idx]);
+          }
+        }
+      }
+
+# undef vec_zero
+# undef vec_add_dpbusd_32x2
+# undef vec_hadd
+# undef vec_haddx4
+#else
+      // Use old implementation for the other architectures.
+      affine_transform_non_ssse3<
+        InputDimensions,
+        PaddedInputDimensions,
+        OutputDimensions>(output, weights, biases, input);
+
+#endif
+
+      return output;
+    }
+
+   private:
+    using BiasType = OutputType;
+    using WeightType = std::int8_t;
+
+    alignas(CacheLineSize) BiasType biases[OutputDimensions];
+    alignas(CacheLineSize) WeightType weights[OutputDimensions * PaddedInputDimensions];
+  };
+
+  // A specialization for small inputs
+  template <IndexType InDims, IndexType OutDims>
+  class AffineTransform<InDims, OutDims, std::enable_if_t<(ceil_to_multiple<IndexType>(InDims, MaxSimdWidth) < LargeInputSize)>> {
+   public:
+    // Input/output type
+    // Input/output type
+    using InputType = std::uint8_t;
+    using OutputType = std::int32_t;
+
+    // Number of input/output dimensions
+    static constexpr IndexType InputDimensions = InDims;
+    static constexpr IndexType OutputDimensions = OutDims;
+
+    static constexpr IndexType PaddedInputDimensions =
+      ceil_to_multiple<IndexType>(InputDimensions, MaxSimdWidth);
+    static constexpr IndexType PaddedOutputDimensions =
+      ceil_to_multiple<IndexType>(OutputDimensions, MaxSimdWidth);
+
+    using OutputBuffer = OutputType[PaddedOutputDimensions];
+
+    static_assert(PaddedInputDimensions < LargeInputSize, "Something went wrong. This specialization (for small inputs) should not have been chosen.");
+
+    // Hash value embedded in the evaluation file
+    static constexpr std::uint32_t get_hash_value(std::uint32_t prevHash) {
+      std::uint32_t hashValue = 0xCC03DAE4u;
+      hashValue += OutputDimensions;
+      hashValue ^= prevHash >> 1;
+      hashValue ^= prevHash << 31;
+      return hashValue;
+    }
+
+    static IndexType get_weight_index_scrambled(IndexType i)
+    {
+      return
+        (i / 4) % (PaddedInputDimensions / 4) * OutputDimensions * 4 +
+        i / PaddedInputDimensions * 4 +
+        i % 4;
+    }
+
+    static IndexType get_weight_index(IndexType i)
+    {
+#if defined (USE_SSSE3)
+      return get_weight_index_scrambled(i);
+#else
+      return i;
+#endif
+    }
+
+    // Read network parameters
+    bool read_parameters(std::istream& stream) {
+      read_little_endian<BiasType>(stream, biases, OutputDimensions);
+      for (IndexType i = 0; i < OutputDimensions * PaddedInputDimensions; ++i)
+        weights[get_weight_index(i)] = read_little_endian<WeightType>(stream);
+
+      return !stream.fail();
+    }
+
+    // Write network parameters
+    bool write_parameters(std::ostream& stream) const {
+      write_little_endian<BiasType>(stream, biases, OutputDimensions);
+
+      for (IndexType i = 0; i < OutputDimensions * PaddedInputDimensions; ++i)
+        write_little_endian<WeightType>(stream, weights[get_weight_index(i)]);
+
+      return !stream.fail();
+    }
+    // Forward propagation
+    const OutputType* propagate(
+        const InputType* input, OutputType* output) const {
+
+#if defined (USE_AVX512)
+      using vec_t = __m512i;
+      #define vec_setzero _mm512_setzero_si512
+      #define vec_set_32 _mm512_set1_epi32
+      #define vec_add_dpbusd_32 Simd::m512_add_dpbusd_epi32
+      #define vec_add_dpbusd_32x2 Simd::m512_add_dpbusd_epi32x2
+      #define vec_hadd Simd::m512_hadd
+#elif defined (USE_AVX2)
+      using vec_t = __m256i;
+      #define vec_setzero _mm256_setzero_si256
+      #define vec_set_32 _mm256_set1_epi32
+      #define vec_add_dpbusd_32 Simd::m256_add_dpbusd_epi32
+      #define vec_add_dpbusd_32x2 Simd::m256_add_dpbusd_epi32x2
+      #define vec_hadd Simd::m256_hadd
+#elif defined (USE_SSSE3)
+      using vec_t = __m128i;
+      #define vec_setzero _mm_setzero_si128
+      #define vec_set_32 _mm_set1_epi32
+      #define vec_add_dpbusd_32 Simd::m128_add_dpbusd_epi32
+      #define vec_add_dpbusd_32x2 Simd::m128_add_dpbusd_epi32x2
+      #define vec_hadd Simd::m128_hadd
+#endif
+
+#if defined (USE_SSSE3)
+      const auto inputVector = reinterpret_cast<const vec_t*>(input);
+
+      static constexpr IndexType OutputSimdWidth = sizeof(vec_t) / sizeof(OutputType);
+
+      static_assert(OutputDimensions % OutputSimdWidth == 0 || OutputDimensions == 1);
+
+      if constexpr (OutputDimensions % OutputSimdWidth == 0)
+      {
+        constexpr IndexType NumChunks = ceil_to_multiple<IndexType>(InputDimensions, 8) / 4;
+        constexpr IndexType NumRegs = OutputDimensions / OutputSimdWidth;
+
+        const auto input32 = reinterpret_cast<const std::int32_t*>(input);
+        const vec_t* biasvec = reinterpret_cast<const vec_t*>(biases);
+        vec_t acc[NumRegs];
+        for (IndexType k = 0; k < NumRegs; ++k)
+          acc[k] = biasvec[k];
+
+        for (IndexType i = 0; i < NumChunks; i += 2)
+        {
+          const vec_t in0 = vec_set_32(input32[i + 0]);
+          const vec_t in1 = vec_set_32(input32[i + 1]);
+          const auto col0 = reinterpret_cast<const vec_t*>(&weights[(i + 0) * OutputDimensions * 4]);
+          const auto col1 = reinterpret_cast<const vec_t*>(&weights[(i + 1) * OutputDimensions * 4]);
+          for (IndexType k = 0; k < NumRegs; ++k)
+            vec_add_dpbusd_32x2(acc[k], in0, col0[k], in1, col1[k]);
+        }
+
+        vec_t* outptr = reinterpret_cast<vec_t*>(output);
+        for (IndexType k = 0; k < NumRegs; ++k)
+          outptr[k] = acc[k];
+      }
+      else if constexpr (OutputDimensions == 1)
+      {
+        constexpr IndexType NumChunks = PaddedInputDimensions / SimdWidth;
+        vec_t sum0 = vec_setzero();
+        const auto row0 = reinterpret_cast<const vec_t*>(&weights[0]);
+
+        for (int j = 0; j < (int)NumChunks; ++j)
+        {
+          const vec_t in = inputVector[j];
+          vec_add_dpbusd_32(sum0, in, row0[j]);
+        }
+        output[0] = vec_hadd(sum0, biases[0]);
+      }
+
+# undef vec_setzero
+# undef vec_set_32
+# undef vec_add_dpbusd_32
+# undef vec_add_dpbusd_32x2
+# undef vec_hadd
+#else
+      // Use old implementation for the other architectures.
+      affine_transform_non_ssse3<
+        InputDimensions,
+        PaddedInputDimensions,
+        OutputDimensions>(output, weights, biases, input);
+#endif
+
+      return output;
+    }
+
+   private:
+    using BiasType = OutputType;
+    using WeightType = std::int8_t;
+
+    alignas(CacheLineSize) BiasType biases[OutputDimensions];
+    alignas(CacheLineSize) WeightType weights[OutputDimensions * PaddedInputDimensions];
+  };
+
+}  // namespace Stockfish::Eval::NNUE::Layers
+
+#endif // #ifndef NNUE_LAYERS_AFFINE_TRANSFORM_H_INCLUDED

src/Stockfish/src/nnue/layers/affine_transform_sparse_input.h

@@ -0,0 +1,286 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+// Definition of layer AffineTransformSparseInput of NNUE evaluation function
+
+#ifndef NNUE_LAYERS_AFFINE_TRANSFORM_SPARSE_INPUT_H_INCLUDED
+#define NNUE_LAYERS_AFFINE_TRANSFORM_SPARSE_INPUT_H_INCLUDED
+
+#include <iostream>
+#include <algorithm>
+#include <array>
+#include <type_traits>
+#include "../nnue_common.h"
+#include "affine_transform.h"
+#include "simd.h"
+
+/*
+  This file contains the definition for a fully connected layer (aka affine transform) with block sparse input.
+*/
+
+namespace Stockfish::Eval::NNUE::Layers {
+#if defined(__GNUC__)  // GCC, Clang, ICC
+
+  static inline IndexType lsb_(std::uint32_t b) {
+    assert(b);
+    return IndexType(__builtin_ctzl(b));
+  }
+
+#elif defined(_MSC_VER)  // MSVC
+
+  static inline IndexType lsb_(std::uint32_t b) {
+    assert(b);
+    unsigned long idx;
+    _BitScanForward(&idx, b);
+    return (IndexType) idx;
+  }
+
+#else  // Compiler is neither GCC nor MSVC compatible
+
+#error "Compiler not supported."
+
+#endif
+
+
+#if defined(USE_SSSE3)
+  alignas(CacheLineSize) static inline const std::array<std::array<std::uint16_t, 8>, 256> lookup_indices = [](){
+    std::array<std::array<std::uint16_t, 8>, 256> v{};
+    for (int i = 0; i < 256; ++i)
+    {
+      int j = i;
+      int k = 0;
+      while(j)
+      {
+        const IndexType lsbIndex = lsb_(std::uint32_t(j));
+        j &= j - 1;
+        v[i][k] = lsbIndex;
+        ++k;
+      }
+    }
+    return v;
+  }();
+  alignas(CacheLineSize) static inline const std::array<unsigned, 256> lookup_count = [](){
+    std::array<unsigned, 256> v;
+    for (int i = 0; i < 256; ++i)
+    {
+      int j = i;
+      int k = 0;
+      while(j)
+      {
+        j &= j - 1;
+        ++k;
+      }
+      v[i] = k;
+    }
+    return v;
+  }();
+
+  // Find indices of nonzero numbers in an int32_t array
+  template<const IndexType InputDimensions>
+  void find_nnz(const std::int32_t* input, std::uint16_t* out, IndexType& count_out) {
+#if defined (USE_AVX512)
+    using vec_t = __m512i;
+    #define vec_nnz(a) _mm512_cmpgt_epi32_mask(a, _mm512_setzero_si512())
+#elif defined (USE_AVX2)
+    using vec_t = __m256i;
+    #define vec_nnz(a) _mm256_movemask_ps(_mm256_castsi256_ps(_mm256_cmpgt_epi32(a, _mm256_setzero_si256())))
+#elif defined (USE_SSSE3)
+    using vec_t = __m128i;
+    #define vec_nnz(a) _mm_movemask_ps(_mm_castsi128_ps(_mm_cmpgt_epi32(a, _mm_setzero_si128())))
+#endif
+    constexpr IndexType InputSimdWidth = sizeof(vec_t) / sizeof(std::int32_t);
+    // Inputs are processed InputSimdWidth at a time and outputs are processed 8 at a time so we process in chunks of max(InputSimdWidth, 8)
+    constexpr IndexType ChunkSize = std::max<IndexType>(InputSimdWidth, 8);
+    constexpr IndexType NumChunks = InputDimensions / ChunkSize;
+    constexpr IndexType InputsPerChunk = ChunkSize / InputSimdWidth;
+    constexpr IndexType OutputsPerChunk = ChunkSize / 8;
+
+    const auto inputVector = reinterpret_cast<const vec_t*>(input);
+    IndexType count = 0;
+    __m128i base = _mm_set1_epi16(0);
+    __m128i increment = _mm_set1_epi16(8);
+    for (IndexType i = 0; i < NumChunks; ++i)
+    {
+      // bitmask of nonzero values in this chunk
+      unsigned nnz = 0;
+      for (IndexType j = 0; j < InputsPerChunk; ++j)
+      {
+        const vec_t inputChunk = inputVector[i * InputsPerChunk + j];
+        nnz |= (unsigned)vec_nnz(inputChunk) << (j * InputSimdWidth);
+      }
+      for (IndexType j = 0; j < OutputsPerChunk; ++j)
+      {
+        const auto lookup = (nnz >> (j * 8)) & 0xFF;
+        const auto offsets = _mm_loadu_si128(reinterpret_cast<const __m128i*>(&lookup_indices[lookup]));
+        _mm_storeu_si128(reinterpret_cast<__m128i*>(out + count), _mm_add_epi16(base, offsets));
+        count += lookup_count[lookup];
+        base = _mm_add_epi16(base, increment);
+      }
+    }
+    count_out = count;
+  }
+# undef vec_nnz
+#endif
+
+  // Sparse input implementation
+  template <IndexType InDims, IndexType OutDims>
+  class AffineTransformSparseInput {
+   public:
+    // Input/output type
+    // Input/output type
+    using InputType = std::uint8_t;
+    using OutputType = std::int32_t;
+
+    // Number of input/output dimensions
+    static constexpr IndexType InputDimensions = InDims;
+    static constexpr IndexType OutputDimensions = OutDims;
+
+    static_assert(OutputDimensions % 16 == 0, "Only implemented for OutputDimensions divisible by 16.");
+
+    static constexpr IndexType PaddedInputDimensions =
+      ceil_to_multiple<IndexType>(InputDimensions, MaxSimdWidth);
+    static constexpr IndexType PaddedOutputDimensions =
+      ceil_to_multiple<IndexType>(OutputDimensions, MaxSimdWidth);
+
+#if defined (USE_SSSE3)
+    static constexpr IndexType ChunkSize = 4;
+#else
+    static constexpr IndexType ChunkSize = 1;
+#endif
+
+    using OutputBuffer = OutputType[PaddedOutputDimensions];
+
+    // Hash value embedded in the evaluation file
+    static constexpr std::uint32_t get_hash_value(std::uint32_t prevHash) {
+      std::uint32_t hashValue = 0xCC03DAE4u;
+      hashValue += OutputDimensions;
+      hashValue ^= prevHash >> 1;
+      hashValue ^= prevHash << 31;
+      return hashValue;
+    }
+
+    static IndexType get_weight_index_scrambled(IndexType i)
+    {
+      return
+        (i / ChunkSize) % (PaddedInputDimensions / ChunkSize) * OutputDimensions * ChunkSize +
+        i / PaddedInputDimensions * ChunkSize +
+        i % ChunkSize;
+    }
+
+    static IndexType get_weight_index(IndexType i)
+    {
+#if defined (USE_SSSE3)
+      return get_weight_index_scrambled(i);
+#else
+      return i;
+#endif
+    }
+
+    // Read network parameters
+    bool read_parameters(std::istream& stream) {
+      read_little_endian<BiasType>(stream, biases, OutputDimensions);
+      for (IndexType i = 0; i < OutputDimensions * PaddedInputDimensions; ++i)
+        weights[get_weight_index(i)] = read_little_endian<WeightType>(stream);
+
+      return !stream.fail();
+    }
+
+    // Write network parameters
+    bool write_parameters(std::ostream& stream) const {
+      write_little_endian<BiasType>(stream, biases, OutputDimensions);
+
+      for (IndexType i = 0; i < OutputDimensions * PaddedInputDimensions; ++i)
+        write_little_endian<WeightType>(stream, weights[get_weight_index(i)]);
+
+      return !stream.fail();
+    }
+    // Forward propagation
+    const OutputType* propagate(
+        const InputType* input, OutputType* output) const {
+
+#if defined (USE_SSSE3)
+#if defined (USE_AVX512)
+      using vec_t = __m512i;
+      #define vec_setzero _mm512_setzero_si512
+      #define vec_set_32 _mm512_set1_epi32
+      #define vec_add_dpbusd_32 Simd::m512_add_dpbusd_epi32
+#elif defined (USE_AVX2)
+      using vec_t = __m256i;
+      #define vec_setzero _mm256_setzero_si256
+      #define vec_set_32 _mm256_set1_epi32
+      #define vec_add_dpbusd_32 Simd::m256_add_dpbusd_epi32
+#elif defined (USE_SSSE3)
+      using vec_t = __m128i;
+      #define vec_setzero _mm_setzero_si128
+      #define vec_set_32 _mm_set1_epi32
+      #define vec_add_dpbusd_32 Simd::m128_add_dpbusd_epi32
+#endif
+      static constexpr IndexType OutputSimdWidth = sizeof(vec_t) / sizeof(OutputType);
+
+      constexpr IndexType NumChunks = ceil_to_multiple<IndexType>(InputDimensions, 8) / ChunkSize;
+      constexpr IndexType NumRegs = OutputDimensions / OutputSimdWidth;
+      std::uint16_t nnz[NumChunks];
+      IndexType count;
+
+      const auto input32 = reinterpret_cast<const std::int32_t*>(input);
+
+      // Find indices of nonzero 32bit blocks
+      find_nnz<NumChunks>(input32, nnz, count);
+
+      const vec_t* biasvec = reinterpret_cast<const vec_t*>(biases);
+      vec_t acc[NumRegs];
+      for (IndexType k = 0; k < NumRegs; ++k)
+        acc[k] = biasvec[k];
+
+      for (IndexType j = 0; j < count; ++j)
+      {
+        const auto i = nnz[j];
+        const vec_t in = vec_set_32(input32[i]);
+        const auto col = reinterpret_cast<const vec_t*>(&weights[i * OutputDimensions * ChunkSize]);
+        for (IndexType k = 0; k < NumRegs; ++k)
+          vec_add_dpbusd_32(acc[k], in, col[k]);
+      }
+
+      vec_t* outptr = reinterpret_cast<vec_t*>(output);
+      for (IndexType k = 0; k < NumRegs; ++k)
+        outptr[k] = acc[k];
+# undef vec_setzero
+# undef vec_set_32
+# undef vec_add_dpbusd_32
+#else
+      // Use dense implementation for the other architectures.
+      affine_transform_non_ssse3<
+        InputDimensions,
+        PaddedInputDimensions,
+        OutputDimensions>(output, weights, biases, input);
+#endif
+
+      return output;
+    }
+
+   private:
+    using BiasType = OutputType;
+    using WeightType = std::int8_t;
+
+    alignas(CacheLineSize) BiasType biases[OutputDimensions];
+    alignas(CacheLineSize) WeightType weights[OutputDimensions * PaddedInputDimensions];
+  };
+
+}  // namespace Stockfish::Eval::NNUE::Layers
+
+#endif // #ifndef NNUE_LAYERS_AFFINE_TRANSFORM_SPARSE_INPUT_H_INCLUDED

src/Stockfish/src/nnue/layers/clipped_relu.h

@@ -0,0 +1,180 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+// Definition of layer ClippedReLU of NNUE evaluation function
+
+#ifndef NNUE_LAYERS_CLIPPED_RELU_H_INCLUDED
+#define NNUE_LAYERS_CLIPPED_RELU_H_INCLUDED
+
+#include "../nnue_common.h"
+
+namespace Stockfish::Eval::NNUE::Layers {
+
+  // Clipped ReLU
+  template <IndexType InDims>
+  class ClippedReLU {
+   public:
+    // Input/output type
+    using InputType = std::int32_t;
+    using OutputType = std::uint8_t;
+
+    // Number of input/output dimensions
+    static constexpr IndexType InputDimensions = InDims;
+    static constexpr IndexType OutputDimensions = InputDimensions;
+    static constexpr IndexType PaddedOutputDimensions =
+        ceil_to_multiple<IndexType>(OutputDimensions, 32);
+
+    using OutputBuffer = OutputType[PaddedOutputDimensions];
+
+    // Hash value embedded in the evaluation file
+    static constexpr std::uint32_t get_hash_value(std::uint32_t prevHash) {
+      std::uint32_t hashValue = 0x538D24C7u;
+      hashValue += prevHash;
+      return hashValue;
+    }
+
+    // Read network parameters
+    bool read_parameters(std::istream&) {
+      return true;
+    }
+
+    // Write network parameters
+    bool write_parameters(std::ostream&) const {
+      return true;
+    }
+
+    // Forward propagation
+    const OutputType* propagate(
+        const InputType* input, OutputType* output) const {
+
+  #if defined(USE_AVX2)
+      if constexpr (InputDimensions % SimdWidth == 0) {
+        constexpr IndexType NumChunks = InputDimensions / SimdWidth;
+        const __m256i Zero = _mm256_setzero_si256();
+        const __m256i Offsets = _mm256_set_epi32(7, 3, 6, 2, 5, 1, 4, 0);
+        const auto in = reinterpret_cast<const __m256i*>(input);
+        const auto out = reinterpret_cast<__m256i*>(output);
+        for (IndexType i = 0; i < NumChunks; ++i) {
+          const __m256i words0 = _mm256_srai_epi16(_mm256_packs_epi32(
+              _mm256_load_si256(&in[i * 4 + 0]),
+              _mm256_load_si256(&in[i * 4 + 1])), WeightScaleBits);
+          const __m256i words1 = _mm256_srai_epi16(_mm256_packs_epi32(
+              _mm256_load_si256(&in[i * 4 + 2]),
+              _mm256_load_si256(&in[i * 4 + 3])), WeightScaleBits);
+          _mm256_store_si256(&out[i], _mm256_permutevar8x32_epi32(_mm256_max_epi8(
+              _mm256_packs_epi16(words0, words1), Zero), Offsets));
+        }
+      } else {
+        constexpr IndexType NumChunks = InputDimensions / (SimdWidth / 2);
+        const __m128i Zero = _mm_setzero_si128();
+        const auto in = reinterpret_cast<const __m128i*>(input);
+        const auto out = reinterpret_cast<__m128i*>(output);
+        for (IndexType i = 0; i < NumChunks; ++i) {
+          const __m128i words0 = _mm_srai_epi16(_mm_packs_epi32(
+              _mm_load_si128(&in[i * 4 + 0]),
+              _mm_load_si128(&in[i * 4 + 1])), WeightScaleBits);
+          const __m128i words1 = _mm_srai_epi16(_mm_packs_epi32(
+              _mm_load_si128(&in[i * 4 + 2]),
+              _mm_load_si128(&in[i * 4 + 3])), WeightScaleBits);
+          const __m128i packedbytes = _mm_packs_epi16(words0, words1);
+          _mm_store_si128(&out[i], _mm_max_epi8(packedbytes, Zero));
+        }
+      }
+      constexpr IndexType Start =
+        InputDimensions % SimdWidth == 0
+        ? InputDimensions / SimdWidth * SimdWidth
+        : InputDimensions / (SimdWidth / 2) * (SimdWidth / 2);
+
+  #elif defined(USE_SSE2)
+      constexpr IndexType NumChunks = InputDimensions / SimdWidth;
+
+  #ifdef USE_SSE41
+      const __m128i Zero = _mm_setzero_si128();
+  #else
+      const __m128i k0x80s = _mm_set1_epi8(-128);
+  #endif
+
+      const auto in = reinterpret_cast<const __m128i*>(input);
+      const auto out = reinterpret_cast<__m128i*>(output);
+      for (IndexType i = 0; i < NumChunks; ++i) {
+        const __m128i words0 = _mm_srai_epi16(_mm_packs_epi32(
+            _mm_load_si128(&in[i * 4 + 0]),
+            _mm_load_si128(&in[i * 4 + 1])), WeightScaleBits);
+        const __m128i words1 = _mm_srai_epi16(_mm_packs_epi32(
+            _mm_load_si128(&in[i * 4 + 2]),
+            _mm_load_si128(&in[i * 4 + 3])), WeightScaleBits);
+        const __m128i packedbytes = _mm_packs_epi16(words0, words1);
+        _mm_store_si128(&out[i],
+
+  #ifdef USE_SSE41
+          _mm_max_epi8(packedbytes, Zero)
+  #else
+          _mm_subs_epi8(_mm_adds_epi8(packedbytes, k0x80s), k0x80s)
+  #endif
+
+        );
+      }
+      constexpr IndexType Start = NumChunks * SimdWidth;
+
+  #elif defined(USE_MMX)
+      constexpr IndexType NumChunks = InputDimensions / SimdWidth;
+      const __m64 k0x80s = _mm_set1_pi8(-128);
+      const auto in = reinterpret_cast<const __m64*>(input);
+      const auto out = reinterpret_cast<__m64*>(output);
+      for (IndexType i = 0; i < NumChunks; ++i) {
+        const __m64 words0 = _mm_srai_pi16(
+            _mm_packs_pi32(in[i * 4 + 0], in[i * 4 + 1]),
+            WeightScaleBits);
+        const __m64 words1 = _mm_srai_pi16(
+            _mm_packs_pi32(in[i * 4 + 2], in[i * 4 + 3]),
+            WeightScaleBits);
+        const __m64 packedbytes = _mm_packs_pi16(words0, words1);
+        out[i] = _mm_subs_pi8(_mm_adds_pi8(packedbytes, k0x80s), k0x80s);
+      }
+      _mm_empty();
+      constexpr IndexType Start = NumChunks * SimdWidth;
+
+  #elif defined(USE_NEON)
+      constexpr IndexType NumChunks = InputDimensions / (SimdWidth / 2);
+      const int8x8_t Zero = {0};
+      const auto in = reinterpret_cast<const int32x4_t*>(input);
+      const auto out = reinterpret_cast<int8x8_t*>(output);
+      for (IndexType i = 0; i < NumChunks; ++i) {
+        int16x8_t shifted;
+        const auto pack = reinterpret_cast<int16x4_t*>(&shifted);
+        pack[0] = vqshrn_n_s32(in[i * 2 + 0], WeightScaleBits);
+        pack[1] = vqshrn_n_s32(in[i * 2 + 1], WeightScaleBits);
+        out[i] = vmax_s8(vqmovn_s16(shifted), Zero);
+      }
+      constexpr IndexType Start = NumChunks * (SimdWidth / 2);
+  #else
+      constexpr IndexType Start = 0;
+  #endif
+
+      for (IndexType i = Start; i < InputDimensions; ++i) {
+        output[i] = static_cast<OutputType>(
+            std::max(0, std::min(127, input[i] >> WeightScaleBits)));
+      }
+
+      return output;
+    }
+  };
+
+}  // namespace Stockfish::Eval::NNUE::Layers
+
+#endif // NNUE_LAYERS_CLIPPED_RELU_H_INCLUDED

src/Stockfish/src/nnue/layers/simd.h

@@ -0,0 +1,403 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef STOCKFISH_SIMD_H_INCLUDED
+#define STOCKFISH_SIMD_H_INCLUDED
+
+#if defined(USE_AVX2)
+# include <immintrin.h>
+
+#elif defined(USE_SSE41)
+# include <smmintrin.h>
+
+#elif defined(USE_SSSE3)
+# include <tmmintrin.h>
+
+#elif defined(USE_SSE2)
+# include <emmintrin.h>
+
+#elif defined(USE_MMX)
+# include <mmintrin.h>
+
+#elif defined(USE_NEON)
+# include <arm_neon.h>
+#endif
+
+// The inline asm is only safe for GCC, where it is necessary to get good codegen.
+// See https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101693
+// Clang does fine without it.
+// Play around here: https://godbolt.org/z/7EWqrYq51
+#if (defined(__GNUC__) && !defined(__clang__) && !defined(__INTEL_COMPILER))
+#define USE_INLINE_ASM
+#endif
+
+// Use either the AVX512 or AVX-VNNI version of the VNNI instructions.
+#if defined(USE_AVXVNNI)
+#define VNNI_PREFIX "%{vex%} "
+#else
+#define VNNI_PREFIX ""
+#endif
+
+namespace Stockfish::Simd {
+
+#if defined (USE_AVX512)
+
+    [[maybe_unused]] static int m512_hadd(__m512i sum, int bias) {
+      return _mm512_reduce_add_epi32(sum) + bias;
+    }
+
+    /*
+      Parameters:
+        sum0 = [zmm0.i128[0], zmm0.i128[1], zmm0.i128[2], zmm0.i128[3]]
+        sum1 = [zmm1.i128[0], zmm1.i128[1], zmm1.i128[2], zmm1.i128[3]]
+        sum2 = [zmm2.i128[0], zmm2.i128[1], zmm2.i128[2], zmm2.i128[3]]
+        sum3 = [zmm3.i128[0], zmm3.i128[1], zmm3.i128[2], zmm3.i128[3]]
+
+      Returns:
+        ret = [
+          reduce_add_epi32(zmm0.i128[0]), reduce_add_epi32(zmm1.i128[0]), reduce_add_epi32(zmm2.i128[0]), reduce_add_epi32(zmm3.i128[0]),
+          reduce_add_epi32(zmm0.i128[1]), reduce_add_epi32(zmm1.i128[1]), reduce_add_epi32(zmm2.i128[1]), reduce_add_epi32(zmm3.i128[1]),
+          reduce_add_epi32(zmm0.i128[2]), reduce_add_epi32(zmm1.i128[2]), reduce_add_epi32(zmm2.i128[2]), reduce_add_epi32(zmm3.i128[2]),
+          reduce_add_epi32(zmm0.i128[3]), reduce_add_epi32(zmm1.i128[3]), reduce_add_epi32(zmm2.i128[3]), reduce_add_epi32(zmm3.i128[3])
+        ]
+    */
+    [[maybe_unused]] static __m512i m512_hadd128x16_interleave(
+        __m512i sum0, __m512i sum1, __m512i sum2, __m512i sum3) {
+
+      __m512i sum01a = _mm512_unpacklo_epi32(sum0, sum1);
+      __m512i sum01b = _mm512_unpackhi_epi32(sum0, sum1);
+
+      __m512i sum23a = _mm512_unpacklo_epi32(sum2, sum3);
+      __m512i sum23b = _mm512_unpackhi_epi32(sum2, sum3);
+
+      __m512i sum01 = _mm512_add_epi32(sum01a, sum01b);
+      __m512i sum23 = _mm512_add_epi32(sum23a, sum23b);
+
+      __m512i sum0123a = _mm512_unpacklo_epi64(sum01, sum23);
+      __m512i sum0123b = _mm512_unpackhi_epi64(sum01, sum23);
+
+      return _mm512_add_epi32(sum0123a, sum0123b);
+    }
+
+    [[maybe_unused]] static __m128i m512_haddx4(
+        __m512i sum0, __m512i sum1, __m512i sum2, __m512i sum3,
+        __m128i bias) {
+
+      __m512i sum = m512_hadd128x16_interleave(sum0, sum1, sum2, sum3);
+
+      __m256i sum256lo = _mm512_castsi512_si256(sum);
+      __m256i sum256hi = _mm512_extracti64x4_epi64(sum, 1);
+
+      sum256lo = _mm256_add_epi32(sum256lo, sum256hi);
+
+      __m128i sum128lo = _mm256_castsi256_si128(sum256lo);
+      __m128i sum128hi = _mm256_extracti128_si256(sum256lo, 1);
+
+      return _mm_add_epi32(_mm_add_epi32(sum128lo, sum128hi), bias);
+    }
+
+    [[maybe_unused]] static void m512_add_dpbusd_epi32(
+        __m512i& acc,
+        __m512i a,
+        __m512i b) {
+
+# if defined (USE_VNNI)
+#   if defined (USE_INLINE_ASM)
+      asm(
+        "vpdpbusd %[b], %[a], %[acc]\n\t"
+        : [acc]"+v"(acc)
+        : [a]"v"(a), [b]"vm"(b)
+      );
+#   else
+      acc = _mm512_dpbusd_epi32(acc, a, b);
+#   endif
+# else
+#   if defined (USE_INLINE_ASM)
+      __m512i tmp = _mm512_maddubs_epi16(a, b);
+      asm(
+          "vpmaddwd    %[tmp], %[ones], %[tmp]\n\t"
+          "vpaddd      %[acc], %[tmp], %[acc]\n\t"
+          : [acc]"+v"(acc), [tmp]"+&v"(tmp)
+          : [ones]"v"(_mm512_set1_epi16(1))
+      );
+#   else
+      __m512i product0 = _mm512_maddubs_epi16(a, b);
+      product0 = _mm512_madd_epi16(product0, _mm512_set1_epi16(1));
+      acc = _mm512_add_epi32(acc, product0);
+#   endif
+# endif
+    }
+
+    [[maybe_unused]] static void m512_add_dpbusd_epi32x2(
+        __m512i& acc,
+        __m512i a0, __m512i b0,
+        __m512i a1, __m512i b1) {
+
+# if defined (USE_VNNI)
+#   if defined (USE_INLINE_ASM)
+      asm(
+        "vpdpbusd %[b0], %[a0], %[acc]\n\t"
+        "vpdpbusd %[b1], %[a1], %[acc]\n\t"
+        : [acc]"+&v"(acc)
+        : [a0]"v"(a0), [b0]"vm"(b0), [a1]"v"(a1), [b1]"vm"(b1)
+      );
+#   else
+      acc = _mm512_dpbusd_epi32(acc, a0, b0);
+      acc = _mm512_dpbusd_epi32(acc, a1, b1);
+#   endif
+# else
+#   if defined (USE_INLINE_ASM)
+      __m512i tmp0 = _mm512_maddubs_epi16(a0, b0);
+      __m512i tmp1 = _mm512_maddubs_epi16(a1, b1);
+      asm(
+          "vpmaddwd    %[tmp0], %[ones], %[tmp0]\n\t"
+          "vpmaddwd    %[tmp1], %[ones], %[tmp1]\n\t"
+          "vpaddd      %[tmp0], %[tmp1], %[tmp0]\n\t"
+          "vpaddd      %[acc], %[tmp0], %[acc]\n\t"
+          : [acc]"+v"(acc), [tmp0]"+&v"(tmp0), [tmp1]"+&v"(tmp1)
+          : [ones]"v"(_mm512_set1_epi16(1))
+      );
+#   else
+      __m512i product0 = _mm512_maddubs_epi16(a0, b0);
+      __m512i product1 = _mm512_maddubs_epi16(a1, b1);
+      product0 = _mm512_madd_epi16(product0, _mm512_set1_epi16(1));
+      product1 = _mm512_madd_epi16(product1, _mm512_set1_epi16(1));
+      acc = _mm512_add_epi32(acc, _mm512_add_epi32(product0, product1));
+#   endif
+# endif
+    }
+
+#endif
+
+#if defined (USE_AVX2)
+
+    [[maybe_unused]] static int m256_hadd(__m256i sum, int bias) {
+      __m128i sum128 = _mm_add_epi32(_mm256_castsi256_si128(sum), _mm256_extracti128_si256(sum, 1));
+      sum128 = _mm_add_epi32(sum128, _mm_shuffle_epi32(sum128, _MM_PERM_BADC));
+      sum128 = _mm_add_epi32(sum128, _mm_shuffle_epi32(sum128, _MM_PERM_CDAB));
+      return _mm_cvtsi128_si32(sum128) + bias;
+    }
+
+    [[maybe_unused]] static __m128i m256_haddx4(
+        __m256i sum0, __m256i sum1, __m256i sum2, __m256i sum3,
+        __m128i bias) {
+
+      sum0 = _mm256_hadd_epi32(sum0, sum1);
+      sum2 = _mm256_hadd_epi32(sum2, sum3);
+
+      sum0 = _mm256_hadd_epi32(sum0, sum2);
+
+      __m128i sum128lo = _mm256_castsi256_si128(sum0);
+      __m128i sum128hi = _mm256_extracti128_si256(sum0, 1);
+
+      return _mm_add_epi32(_mm_add_epi32(sum128lo, sum128hi), bias);
+    }
+
+    [[maybe_unused]] static void m256_add_dpbusd_epi32(
+        __m256i& acc,
+        __m256i a,
+        __m256i b) {
+
+# if defined (USE_VNNI)
+#   if defined (USE_INLINE_ASM)
+      asm(
+        VNNI_PREFIX "vpdpbusd %[b], %[a], %[acc]\n\t"
+        : [acc]"+v"(acc)
+        : [a]"v"(a), [b]"vm"(b)
+      );
+#   else
+      acc = _mm256_dpbusd_epi32(acc, a, b);
+#   endif
+# else
+#   if defined (USE_INLINE_ASM)
+      __m256i tmp = _mm256_maddubs_epi16(a, b);
+      asm(
+          "vpmaddwd    %[tmp], %[ones], %[tmp]\n\t"
+          "vpaddd      %[acc], %[tmp], %[acc]\n\t"
+          : [acc]"+v"(acc), [tmp]"+&v"(tmp)
+          : [ones]"v"(_mm256_set1_epi16(1))
+      );
+#   else
+      __m256i product0 = _mm256_maddubs_epi16(a, b);
+      product0 = _mm256_madd_epi16(product0, _mm256_set1_epi16(1));
+      acc = _mm256_add_epi32(acc, product0);
+#   endif
+# endif
+    }
+
+    [[maybe_unused]] static void m256_add_dpbusd_epi32x2(
+        __m256i& acc,
+        __m256i a0, __m256i b0,
+        __m256i a1, __m256i b1) {
+
+# if defined (USE_VNNI)
+#   if defined (USE_INLINE_ASM)
+      asm(
+        VNNI_PREFIX "vpdpbusd %[b0], %[a0], %[acc]\n\t"
+        VNNI_PREFIX "vpdpbusd %[b1], %[a1], %[acc]\n\t"
+        : [acc]"+&v"(acc)
+        : [a0]"v"(a0), [b0]"vm"(b0), [a1]"v"(a1), [b1]"vm"(b1)
+      );
+#   else
+      acc = _mm256_dpbusd_epi32(acc, a0, b0);
+      acc = _mm256_dpbusd_epi32(acc, a1, b1);
+#   endif
+# else
+#   if defined (USE_INLINE_ASM)
+      __m256i tmp0 = _mm256_maddubs_epi16(a0, b0);
+      __m256i tmp1 = _mm256_maddubs_epi16(a1, b1);
+      asm(
+          "vpmaddwd    %[tmp0], %[ones], %[tmp0]\n\t"
+          "vpmaddwd    %[tmp1], %[ones], %[tmp1]\n\t"
+          "vpaddd      %[tmp0], %[tmp1], %[tmp0]\n\t"
+          "vpaddd      %[acc], %[tmp0], %[acc]\n\t"
+          : [acc]"+v"(acc), [tmp0]"+&v"(tmp0), [tmp1]"+&v"(tmp1)
+          : [ones]"v"(_mm256_set1_epi16(1))
+      );
+#   else
+      __m256i product0 = _mm256_maddubs_epi16(a0, b0);
+      __m256i product1 = _mm256_maddubs_epi16(a1, b1);
+      product0 = _mm256_madd_epi16(product0, _mm256_set1_epi16(1));
+      product1 = _mm256_madd_epi16(product1, _mm256_set1_epi16(1));
+      acc = _mm256_add_epi32(acc, _mm256_add_epi32(product0, product1));
+#   endif
+# endif
+    }
+
+#endif
+
+#if defined (USE_SSSE3)
+
+    [[maybe_unused]] static int m128_hadd(__m128i sum, int bias) {
+      sum = _mm_add_epi32(sum, _mm_shuffle_epi32(sum, 0x4E)); //_MM_PERM_BADC
+      sum = _mm_add_epi32(sum, _mm_shuffle_epi32(sum, 0xB1)); //_MM_PERM_CDAB
+      return _mm_cvtsi128_si32(sum) + bias;
+    }
+
+    [[maybe_unused]] static __m128i m128_haddx4(
+        __m128i sum0, __m128i sum1, __m128i sum2, __m128i sum3,
+        __m128i bias) {
+
+      sum0 = _mm_hadd_epi32(sum0, sum1);
+      sum2 = _mm_hadd_epi32(sum2, sum3);
+      sum0 = _mm_hadd_epi32(sum0, sum2);
+      return _mm_add_epi32(sum0, bias);
+    }
+
+    [[maybe_unused]] static void m128_add_dpbusd_epi32(
+        __m128i& acc,
+        __m128i a,
+        __m128i b) {
+
+#   if defined (USE_INLINE_ASM)
+      __m128i tmp = _mm_maddubs_epi16(a, b);
+      asm(
+          "pmaddwd    %[ones], %[tmp]\n\t"
+          "paddd      %[tmp], %[acc]\n\t"
+          : [acc]"+v"(acc), [tmp]"+&v"(tmp)
+          : [ones]"v"(_mm_set1_epi16(1))
+      );
+#   else
+      __m128i product0 = _mm_maddubs_epi16(a, b);
+      product0 = _mm_madd_epi16(product0, _mm_set1_epi16(1));
+      acc = _mm_add_epi32(acc, product0);
+#   endif
+    }
+
+    [[maybe_unused]] static void m128_add_dpbusd_epi32x2(
+        __m128i& acc,
+        __m128i a0, __m128i b0,
+        __m128i a1, __m128i b1) {
+
+#   if defined (USE_INLINE_ASM)
+      __m128i tmp0 = _mm_maddubs_epi16(a0, b0);
+      __m128i tmp1 = _mm_maddubs_epi16(a1, b1);
+      asm(
+          "pmaddwd    %[ones], %[tmp0]\n\t"
+          "pmaddwd    %[ones], %[tmp1]\n\t"
+          "paddd      %[tmp1], %[tmp0]\n\t"
+          "paddd      %[tmp0], %[acc]\n\t"
+          : [acc]"+v"(acc), [tmp0]"+&v"(tmp0), [tmp1]"+&v"(tmp1)
+          : [ones]"v"(_mm_set1_epi16(1))
+      );
+#   else
+      __m128i product0 = _mm_maddubs_epi16(a0, b0);
+      __m128i product1 = _mm_maddubs_epi16(a1, b1);
+      product0 = _mm_madd_epi16(product0, _mm_set1_epi16(1));
+      product1 = _mm_madd_epi16(product1, _mm_set1_epi16(1));
+      acc = _mm_add_epi32(acc, _mm_add_epi32(product0, product1));
+#   endif
+    }
+
+#endif
+
+#if defined (USE_NEON_DOTPROD)
+
+    [[maybe_unused]] static void dotprod_m128_add_dpbusd_epi32x2(
+        int32x4_t& acc,
+        int8x16_t a0, int8x16_t b0,
+        int8x16_t a1, int8x16_t b1) {
+
+        acc = vdotq_s32(acc, a0, b0);
+        acc = vdotq_s32(acc, a1, b1);
+    }
+
+#endif
+
+#if defined (USE_NEON)
+
+    [[maybe_unused]] static int neon_m128_reduce_add_epi32(int32x4_t s) {
+#   if USE_NEON >= 8
+      return vaddvq_s32(s);
+#   else
+      return s[0] + s[1] + s[2] + s[3];
+#   endif
+    }
+
+    [[maybe_unused]] static int neon_m128_hadd(int32x4_t sum, int bias) {
+      return neon_m128_reduce_add_epi32(sum) + bias;
+    }
+
+    [[maybe_unused]] static int32x4_t neon_m128_haddx4(
+        int32x4_t sum0, int32x4_t sum1, int32x4_t sum2, int32x4_t sum3,
+        int32x4_t bias) {
+
+      int32x4_t hsums {
+        neon_m128_reduce_add_epi32(sum0),
+        neon_m128_reduce_add_epi32(sum1),
+        neon_m128_reduce_add_epi32(sum2),
+        neon_m128_reduce_add_epi32(sum3)
+      };
+      return vaddq_s32(hsums, bias);
+    }
+
+    [[maybe_unused]] static void neon_m128_add_dpbusd_epi32x2(
+        int32x4_t& acc,
+        int8x8_t a0, int8x8_t b0,
+        int8x8_t a1, int8x8_t b1) {
+
+      int16x8_t product = vmull_s8(a0, b0);
+      product = vmlal_s8(product, a1, b1);
+      acc = vpadalq_s16(acc, product);
+    }
+
+#endif
+
+}
+
+#endif // STOCKFISH_SIMD_H_INCLUDED

src/Stockfish/src/nnue/layers/sqr_clipped_relu.h

@@ -0,0 +1,120 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+// Definition of layer ClippedReLU of NNUE evaluation function
+
+#ifndef NNUE_LAYERS_SQR_CLIPPED_RELU_H_INCLUDED
+#define NNUE_LAYERS_SQR_CLIPPED_RELU_H_INCLUDED
+
+#include "../nnue_common.h"
+
+namespace Stockfish::Eval::NNUE::Layers {
+
+  // Clipped ReLU
+  template <IndexType InDims>
+  class SqrClippedReLU {
+   public:
+    // Input/output type
+    using InputType = std::int32_t;
+    using OutputType = std::uint8_t;
+
+    // Number of input/output dimensions
+    static constexpr IndexType InputDimensions = InDims;
+    static constexpr IndexType OutputDimensions = InputDimensions;
+    static constexpr IndexType PaddedOutputDimensions =
+        ceil_to_multiple<IndexType>(OutputDimensions, 32);
+
+    using OutputBuffer = OutputType[PaddedOutputDimensions];
+
+    // Hash value embedded in the evaluation file
+    static constexpr std::uint32_t get_hash_value(std::uint32_t prevHash) {
+      std::uint32_t hashValue = 0x538D24C7u;
+      hashValue += prevHash;
+      return hashValue;
+    }
+
+    // Read network parameters
+    bool read_parameters(std::istream&) {
+      return true;
+    }
+
+    // Write network parameters
+    bool write_parameters(std::ostream&) const {
+      return true;
+    }
+
+    // Forward propagation
+    const OutputType* propagate(
+        const InputType* input, OutputType* output) const {
+
+  #if defined(USE_SSE2)
+      constexpr IndexType NumChunks = InputDimensions / 16;
+
+  #ifdef USE_SSE41
+      const __m128i Zero = _mm_setzero_si128();
+  #else
+      const __m128i k0x80s = _mm_set1_epi8(-128);
+  #endif
+
+      static_assert(WeightScaleBits == 6);
+      const auto in = reinterpret_cast<const __m128i*>(input);
+      const auto out = reinterpret_cast<__m128i*>(output);
+      for (IndexType i = 0; i < NumChunks; ++i) {
+        __m128i words0 = _mm_packs_epi32(
+            _mm_load_si128(&in[i * 4 + 0]),
+            _mm_load_si128(&in[i * 4 + 1]));
+        __m128i words1 = _mm_packs_epi32(
+            _mm_load_si128(&in[i * 4 + 2]),
+            _mm_load_si128(&in[i * 4 + 3]));
+
+        // Not sure if
+        words0 = _mm_srli_epi16(_mm_mulhi_epi16(words0, words0), 3);
+        words1 = _mm_srli_epi16(_mm_mulhi_epi16(words1, words1), 3);
+
+        const __m128i packedbytes = _mm_packs_epi16(words0, words1);
+
+        _mm_store_si128(&out[i],
+
+  #ifdef USE_SSE41
+          _mm_max_epi8(packedbytes, Zero)
+  #else
+          _mm_subs_epi8(_mm_adds_epi8(packedbytes, k0x80s), k0x80s)
+  #endif
+
+        );
+      }
+      constexpr IndexType Start = NumChunks * 16;
+
+  #else
+      constexpr IndexType Start = 0;
+  #endif
+
+      for (IndexType i = Start; i < InputDimensions; ++i) {
+        output[i] = static_cast<OutputType>(
+            // really should be /127 but we need to make it fast
+            // needs to be accounted for in the trainer
+            std::max(0ll, std::min(127ll, (((long long)input[i] * input[i]) >> (2 * WeightScaleBits)) / 128)));
+      }
+
+      return output;
+    }
+  };
+
+}  // namespace Stockfish::Eval::NNUE::Layers
+
+#endif // NNUE_LAYERS_SQR_CLIPPED_RELU_H_INCLUDED

src/Stockfish/src/nnue/nnue_accumulator.h

@@ -0,0 +1,37 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+// Class for difference calculation of NNUE evaluation function
+
+#ifndef NNUE_ACCUMULATOR_H_INCLUDED
+#define NNUE_ACCUMULATOR_H_INCLUDED
+
+#include "nnue_architecture.h"
+
+namespace Stockfish::Eval::NNUE {
+
+  // Class that holds the result of affine transformation of input features
+  struct alignas(CacheLineSize) Accumulator {
+    std::int16_t accumulation[2][TransformedFeatureDimensions];
+    std::int32_t psqtAccumulation[2][PSQTBuckets];
+    bool computed[2];
+  };
+
+}  // namespace Stockfish::Eval::NNUE
+
+#endif // NNUE_ACCUMULATOR_H_INCLUDED

src/Stockfish/src/nnue/nnue_architecture.h

@@ -0,0 +1,137 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+// Input features and network structure used in NNUE evaluation function
+
+#ifndef NNUE_ARCHITECTURE_H_INCLUDED
+#define NNUE_ARCHITECTURE_H_INCLUDED
+
+#include <memory>
+
+#include "nnue_common.h"
+
+#include "features/half_ka_v2_hm.h"
+
+#include "layers/affine_transform_sparse_input.h"
+#include "layers/affine_transform.h"
+#include "layers/clipped_relu.h"
+#include "layers/sqr_clipped_relu.h"
+
+#include "../misc.h"
+
+namespace Stockfish::Eval::NNUE {
+
+// Input features used in evaluation function
+using FeatureSet = Features::HalfKAv2_hm;
+
+// Number of input feature dimensions after conversion
+constexpr IndexType TransformedFeatureDimensions = 1536;
+constexpr IndexType PSQTBuckets = 8;
+constexpr IndexType LayerStacks = 8;
+
+struct Network
+{
+  static constexpr int FC_0_OUTPUTS = 15;
+  static constexpr int FC_1_OUTPUTS = 32;
+
+  Layers::AffineTransformSparseInput<TransformedFeatureDimensions, FC_0_OUTPUTS + 1> fc_0;
+  Layers::SqrClippedReLU<FC_0_OUTPUTS + 1> ac_sqr_0;
+  Layers::ClippedReLU<FC_0_OUTPUTS + 1> ac_0;
+  Layers::AffineTransform<FC_0_OUTPUTS * 2, FC_1_OUTPUTS> fc_1;
+  Layers::ClippedReLU<FC_1_OUTPUTS> ac_1;
+  Layers::AffineTransform<FC_1_OUTPUTS, 1> fc_2;
+
+  // Hash value embedded in the evaluation file
+  static constexpr std::uint32_t get_hash_value() {
+    // input slice hash
+    std::uint32_t hashValue = 0xEC42E90Du;
+    hashValue ^= TransformedFeatureDimensions * 2;
+
+    hashValue = decltype(fc_0)::get_hash_value(hashValue);
+    hashValue = decltype(ac_0)::get_hash_value(hashValue);
+    hashValue = decltype(fc_1)::get_hash_value(hashValue);
+    hashValue = decltype(ac_1)::get_hash_value(hashValue);
+    hashValue = decltype(fc_2)::get_hash_value(hashValue);
+
+    return hashValue;
+  }
+
+  // Read network parameters
+  bool read_parameters(std::istream& stream) {
+    return   fc_0.read_parameters(stream)
+          && ac_0.read_parameters(stream)
+          && fc_1.read_parameters(stream)
+          && ac_1.read_parameters(stream)
+          && fc_2.read_parameters(stream);
+  }
+
+  // Write network parameters
+  bool write_parameters(std::ostream& stream) const {
+    return   fc_0.write_parameters(stream)
+          && ac_0.write_parameters(stream)
+          && fc_1.write_parameters(stream)
+          && ac_1.write_parameters(stream)
+          && fc_2.write_parameters(stream);
+  }
+
+  std::int32_t propagate(const TransformedFeatureType* transformedFeatures)
+  {
+    struct alignas(CacheLineSize) Buffer
+    {
+      alignas(CacheLineSize) decltype(fc_0)::OutputBuffer fc_0_out;
+      alignas(CacheLineSize) decltype(ac_sqr_0)::OutputType ac_sqr_0_out[ceil_to_multiple<IndexType>(FC_0_OUTPUTS * 2, 32)];
+      alignas(CacheLineSize) decltype(ac_0)::OutputBuffer ac_0_out;
+      alignas(CacheLineSize) decltype(fc_1)::OutputBuffer fc_1_out;
+      alignas(CacheLineSize) decltype(ac_1)::OutputBuffer ac_1_out;
+      alignas(CacheLineSize) decltype(fc_2)::OutputBuffer fc_2_out;
+
+      Buffer()
+      {
+          std::memset(this, 0, sizeof(*this));
+      }
+    };
+
+#if defined(__clang__) && (__APPLE__)
+    // workaround for a bug reported with xcode 12
+    static thread_local auto tlsBuffer = std::make_unique<Buffer>();
+    // Access TLS only once, cache result.
+    Buffer& buffer = *tlsBuffer;
+#else
+    alignas(CacheLineSize) static thread_local Buffer buffer;
+#endif
+
+    fc_0.propagate(transformedFeatures, buffer.fc_0_out);
+    ac_sqr_0.propagate(buffer.fc_0_out, buffer.ac_sqr_0_out);
+    ac_0.propagate(buffer.fc_0_out, buffer.ac_0_out);
+    std::memcpy(buffer.ac_sqr_0_out + FC_0_OUTPUTS, buffer.ac_0_out, FC_0_OUTPUTS * sizeof(decltype(ac_0)::OutputType));
+    fc_1.propagate(buffer.ac_sqr_0_out, buffer.fc_1_out);
+    ac_1.propagate(buffer.fc_1_out, buffer.ac_1_out);
+    fc_2.propagate(buffer.ac_1_out, buffer.fc_2_out);
+
+    // buffer.fc_0_out[FC_0_OUTPUTS] is such that 1.0 is equal to 127*(1<<WeightScaleBits) in quantized form
+    // but we want 1.0 to be equal to 600*OutputScale
+    std::int32_t fwdOut = int(buffer.fc_0_out[FC_0_OUTPUTS]) * (600*OutputScale) / (127*(1<<WeightScaleBits));
+    std::int32_t outputValue = buffer.fc_2_out[0] + fwdOut;
+
+    return outputValue;
+  }
+};
+
+}  // namespace Stockfish::Eval::NNUE
+
+#endif // #ifndef NNUE_ARCHITECTURE_H_INCLUDED

src/Stockfish/src/nnue/nnue_common.h

@@ -0,0 +1,241 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+// Constants used in NNUE evaluation function
+
+#ifndef NNUE_COMMON_H_INCLUDED
+#define NNUE_COMMON_H_INCLUDED
+
+#include <cstring>
+#include <iostream>
+
+#include "../misc.h"  // for IsLittleEndian
+
+#if defined(USE_AVX2)
+#include <immintrin.h>
+
+#elif defined(USE_SSE41)
+#include <smmintrin.h>
+
+#elif defined(USE_SSSE3)
+#include <tmmintrin.h>
+
+#elif defined(USE_SSE2)
+#include <emmintrin.h>
+
+#elif defined(USE_MMX)
+#include <mmintrin.h>
+
+#elif defined(USE_NEON)
+#include <arm_neon.h>
+#endif
+
+namespace Stockfish::Eval::NNUE {
+
+  // Version of the evaluation file
+  constexpr std::uint32_t Version = 0x7AF32F20u;
+
+  // Constant used in evaluation value calculation
+  constexpr int OutputScale = 16;
+  constexpr int WeightScaleBits = 6;
+
+  // Size of cache line (in bytes)
+  constexpr std::size_t CacheLineSize = 64;
+
+  constexpr const char Leb128MagicString[] = "COMPRESSED_LEB128";
+  constexpr const std::size_t Leb128MagicStringSize = sizeof(Leb128MagicString) - 1;
+
+  // SIMD width (in bytes)
+  #if defined(USE_AVX2)
+  constexpr std::size_t SimdWidth = 32;
+
+  #elif defined(USE_SSE2)
+  constexpr std::size_t SimdWidth = 16;
+
+  #elif defined(USE_MMX)
+  constexpr std::size_t SimdWidth = 8;
+
+  #elif defined(USE_NEON)
+  constexpr std::size_t SimdWidth = 16;
+  #endif
+
+  constexpr std::size_t MaxSimdWidth = 32;
+
+  // Type of input feature after conversion
+  using TransformedFeatureType = std::uint8_t;
+  using IndexType = std::uint32_t;
+
+  // Round n up to be a multiple of base
+  template <typename IntType>
+  constexpr IntType ceil_to_multiple(IntType n, IntType base) {
+      return (n + base - 1) / base * base;
+  }
+
+  // read_little_endian() is our utility to read an integer (signed or unsigned, any size)
+  // from a stream in little-endian order. We swap the byte order after the read if
+  // necessary to return a result with the byte ordering of the compiling machine.
+  template <typename IntType>
+  inline IntType read_little_endian(std::istream& stream) {
+      IntType result;
+
+      if (IsLittleEndian)
+          stream.read(reinterpret_cast<char*>(&result), sizeof(IntType));
+      else
+      {
+          std::uint8_t u[sizeof(IntType)];
+          typename std::make_unsigned<IntType>::type v = 0;
+
+          stream.read(reinterpret_cast<char*>(u), sizeof(IntType));
+          for (std::size_t i = 0; i < sizeof(IntType); ++i)
+              v = (v << 8) | u[sizeof(IntType) - i - 1];
+
+          std::memcpy(&result, &v, sizeof(IntType));
+      }
+
+      return result;
+  }
+
+  // write_little_endian() is our utility to write an integer (signed or unsigned, any size)
+  // to a stream in little-endian order. We swap the byte order before the write if
+  // necessary to always write in little endian order, independently of the byte
+  // ordering of the compiling machine.
+  template <typename IntType>
+  inline void write_little_endian(std::ostream& stream, IntType value) {
+
+      if (IsLittleEndian)
+          stream.write(reinterpret_cast<const char*>(&value), sizeof(IntType));
+      else
+      {
+          std::uint8_t u[sizeof(IntType)];
+          typename std::make_unsigned<IntType>::type v = value;
+
+          std::size_t i = 0;
+          // if constexpr to silence the warning about shift by 8
+          if constexpr (sizeof(IntType) > 1)
+          {
+            for (; i + 1 < sizeof(IntType); ++i)
+            {
+                u[i] = (std::uint8_t)v;
+                v >>= 8;
+            }
+          }
+          u[i] = (std::uint8_t)v;
+
+          stream.write(reinterpret_cast<char*>(u), sizeof(IntType));
+      }
+  }
+
+  // read_little_endian(s, out, N) : read integers in bulk from a little indian stream.
+  // This reads N integers from stream s and put them in array out.
+  template <typename IntType>
+  inline void read_little_endian(std::istream& stream, IntType* out, std::size_t count) {
+      if (IsLittleEndian)
+          stream.read(reinterpret_cast<char*>(out), sizeof(IntType) * count);
+      else
+          for (std::size_t i = 0; i < count; ++i)
+              out[i] = read_little_endian<IntType>(stream);
+  }
+
+  // write_little_endian(s, values, N) : write integers in bulk to a little indian stream.
+  // This takes N integers from array values and writes them on stream s.
+  template <typename IntType>
+  inline void write_little_endian(std::ostream& stream, const IntType* values, std::size_t count) {
+      if (IsLittleEndian)
+          stream.write(reinterpret_cast<const char*>(values), sizeof(IntType) * count);
+      else
+          for (std::size_t i = 0; i < count; ++i)
+              write_little_endian<IntType>(stream, values[i]);
+  }
+
+  template <typename IntType>
+  inline void read_leb_128(std::istream& stream, IntType* out, std::size_t count) {
+      static_assert(std::is_signed_v<IntType>, "Not implemented for unsigned types");
+      char leb128MagicString[Leb128MagicStringSize];
+      stream.read(leb128MagicString, Leb128MagicStringSize);
+      assert(strncmp(Leb128MagicString, leb128MagicString, Leb128MagicStringSize) == 0);
+      const std::uint32_t BUF_SIZE = 4096;
+      std::uint8_t buf[BUF_SIZE];
+      auto bytes_left = read_little_endian<std::uint32_t>(stream);
+      std::uint32_t buf_pos = BUF_SIZE;
+      for (std::size_t i = 0; i < count; ++i) {
+          IntType result = 0;
+          size_t shift = 0;
+          do {
+              if (buf_pos == BUF_SIZE) {
+                  stream.read(reinterpret_cast<char*>(buf), std::min(bytes_left, BUF_SIZE));
+                  buf_pos = 0;
+              }
+              std::uint8_t byte = buf[buf_pos++];
+              --bytes_left;
+              result |= (byte & 0x7f) << shift;
+              shift += 7;
+              if ((byte & 0x80) == 0) {
+                  out[i] = sizeof(IntType) * 8 <= shift || (byte & 0x40) == 0 ? result : result | ~((1 << shift) - 1);
+                  break;
+              }
+          } while (shift < sizeof(IntType) * 8);
+      }
+      assert(bytes_left == 0);
+  }
+
+  template <typename IntType>
+  inline void write_leb_128(std::ostream& stream, const IntType* values, std::size_t count) {
+      static_assert(std::is_signed_v<IntType>, "Not implemented for unsigned types");
+      stream.write(Leb128MagicString, Leb128MagicStringSize);
+      std::uint32_t byte_count = 0;
+      for (std::size_t i = 0; i < count; ++i) {
+          IntType value = values[i];
+          std::uint8_t byte;
+          do {
+              byte = value & 0x7f;
+              value >>= 7;
+              ++byte_count;
+          } while ((byte & 0x40) == 0 ? value != 0 : value != -1);
+      }
+      write_little_endian(stream, byte_count);
+      const std::uint32_t BUF_SIZE = 4096;
+      std::uint8_t buf[BUF_SIZE];
+      std::uint32_t buf_pos = 0;
+      auto flush = [&]() {
+          if (buf_pos > 0) {
+              stream.write(reinterpret_cast<char*>(buf), buf_pos);
+              buf_pos = 0;
+          }
+      };
+      auto write = [&](std::uint8_t byte) {
+          buf[buf_pos++] = byte;
+          if (buf_pos == BUF_SIZE) flush();
+      };
+      for (std::size_t i = 0; i < count; ++i) {
+          IntType value = values[i];
+          while (true) {
+              std::uint8_t byte = value & 0x7f;
+              value >>= 7;
+              if ((byte & 0x40) == 0 ? value == 0 : value == -1) {
+                  write(byte);
+                  break;
+              }
+              write(byte | 0x80);
+          }
+      }
+      flush();
+  }
+
+}  // namespace Stockfish::Eval::NNUE
+
+#endif // #ifndef NNUE_COMMON_H_INCLUDED

src/Stockfish/src/nnue/nnue_feature_transformer.h

@@ -0,0 +1,674 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+// A class that converts the input features of the NNUE evaluation function
+
+#ifndef NNUE_FEATURE_TRANSFORMER_H_INCLUDED
+#define NNUE_FEATURE_TRANSFORMER_H_INCLUDED
+
+#include "nnue_common.h"
+#include "nnue_architecture.h"
+
+#include <cstring> // std::memset()
+#include <utility> // std::pair
+
+namespace Stockfish::Eval::NNUE {
+
+  using BiasType       = std::int16_t;
+  using WeightType     = std::int16_t;
+  using PSQTWeightType = std::int32_t;
+
+  // If vector instructions are enabled, we update and refresh the
+  // accumulator tile by tile such that each tile fits in the CPU's
+  // vector registers.
+  #define VECTOR
+
+  static_assert(PSQTBuckets % 8 == 0,
+    "Per feature PSQT values cannot be processed at granularity lower than 8 at a time.");
+
+  #ifdef USE_AVX512
+  using vec_t = __m512i;
+  using psqt_vec_t = __m256i;
+  #define vec_load(a) _mm512_load_si512(a)
+  #define vec_store(a,b) _mm512_store_si512(a,b)
+  #define vec_add_16(a,b) _mm512_add_epi16(a,b)
+  #define vec_sub_16(a,b) _mm512_sub_epi16(a,b)
+  #define vec_mul_16(a,b) _mm512_mullo_epi16(a,b)
+  #define vec_zero() _mm512_setzero_epi32()
+  #define vec_set_16(a) _mm512_set1_epi16(a)
+  #define vec_max_16(a,b) _mm512_max_epi16(a,b)
+  #define vec_min_16(a,b) _mm512_min_epi16(a,b)
+  inline vec_t vec_msb_pack_16(vec_t a, vec_t b){
+    vec_t compacted = _mm512_packs_epi16(_mm512_srli_epi16(a,7),_mm512_srli_epi16(b,7));
+    return _mm512_permutexvar_epi64(_mm512_setr_epi64(0, 2, 4, 6, 1, 3, 5, 7), compacted);
+  }
+  #define vec_load_psqt(a) _mm256_load_si256(a)
+  #define vec_store_psqt(a,b) _mm256_store_si256(a,b)
+  #define vec_add_psqt_32(a,b) _mm256_add_epi32(a,b)
+  #define vec_sub_psqt_32(a,b) _mm256_sub_epi32(a,b)
+  #define vec_zero_psqt() _mm256_setzero_si256()
+  #define NumRegistersSIMD 32
+  #define MaxChunkSize 64
+
+  #elif USE_AVX2
+  using vec_t = __m256i;
+  using psqt_vec_t = __m256i;
+  #define vec_load(a) _mm256_load_si256(a)
+  #define vec_store(a,b) _mm256_store_si256(a,b)
+  #define vec_add_16(a,b) _mm256_add_epi16(a,b)
+  #define vec_sub_16(a,b) _mm256_sub_epi16(a,b)
+  #define vec_mul_16(a,b) _mm256_mullo_epi16(a,b)
+  #define vec_zero() _mm256_setzero_si256()
+  #define vec_set_16(a) _mm256_set1_epi16(a)
+  #define vec_max_16(a,b) _mm256_max_epi16(a,b)
+  #define vec_min_16(a,b) _mm256_min_epi16(a,b)
+  inline vec_t vec_msb_pack_16(vec_t a, vec_t b){
+    vec_t compacted = _mm256_packs_epi16(_mm256_srli_epi16(a,7), _mm256_srli_epi16(b,7));
+    return _mm256_permute4x64_epi64(compacted, 0b11011000);
+  }
+  #define vec_load_psqt(a) _mm256_load_si256(a)
+  #define vec_store_psqt(a,b) _mm256_store_si256(a,b)
+  #define vec_add_psqt_32(a,b) _mm256_add_epi32(a,b)
+  #define vec_sub_psqt_32(a,b) _mm256_sub_epi32(a,b)
+  #define vec_zero_psqt() _mm256_setzero_si256()
+  #define NumRegistersSIMD 16
+  #define MaxChunkSize 32
+
+  #elif USE_SSE2
+  using vec_t = __m128i;
+  using psqt_vec_t = __m128i;
+  #define vec_load(a) (*(a))
+  #define vec_store(a,b) *(a)=(b)
+  #define vec_add_16(a,b) _mm_add_epi16(a,b)
+  #define vec_sub_16(a,b) _mm_sub_epi16(a,b)
+  #define vec_mul_16(a,b) _mm_mullo_epi16(a,b)
+  #define vec_zero() _mm_setzero_si128()
+  #define vec_set_16(a) _mm_set1_epi16(a)
+  #define vec_max_16(a,b) _mm_max_epi16(a,b)
+  #define vec_min_16(a,b) _mm_min_epi16(a,b)
+  #define vec_msb_pack_16(a,b) _mm_packs_epi16(_mm_srli_epi16(a,7),_mm_srli_epi16(b,7))
+  #define vec_load_psqt(a) (*(a))
+  #define vec_store_psqt(a,b) *(a)=(b)
+  #define vec_add_psqt_32(a,b) _mm_add_epi32(a,b)
+  #define vec_sub_psqt_32(a,b) _mm_sub_epi32(a,b)
+  #define vec_zero_psqt() _mm_setzero_si128()
+  #define NumRegistersSIMD (Is64Bit ? 16 : 8)
+  #define MaxChunkSize 16
+
+  #elif USE_MMX
+  using vec_t = __m64;
+  using psqt_vec_t = __m64;
+  #define vec_load(a) (*(a))
+  #define vec_store(a,b) *(a)=(b)
+  #define vec_add_16(a,b) _mm_add_pi16(a,b)
+  #define vec_sub_16(a,b) _mm_sub_pi16(a,b)
+  #define vec_mul_16(a,b) _mm_mullo_pi16(a,b)
+  #define vec_zero() _mm_setzero_si64()
+  #define vec_set_16(a) _mm_set1_pi16(a)
+  inline vec_t vec_max_16(vec_t a,vec_t b){
+    vec_t comparison = _mm_cmpgt_pi16(a,b);
+    return _mm_or_si64(_mm_and_si64(comparison, a), _mm_andnot_si64(comparison, b));
+  }
+  inline vec_t vec_min_16(vec_t a,vec_t b){
+    vec_t comparison = _mm_cmpgt_pi16(a,b);
+    return _mm_or_si64(_mm_and_si64(comparison, b), _mm_andnot_si64(comparison, a));
+  }
+  #define vec_msb_pack_16(a,b) _mm_packs_pi16(_mm_srli_pi16(a,7),_mm_srli_pi16(b,7))
+  #define vec_load_psqt(a) (*(a))
+  #define vec_store_psqt(a,b) *(a)=(b)
+  #define vec_add_psqt_32(a,b) _mm_add_pi32(a,b)
+  #define vec_sub_psqt_32(a,b) _mm_sub_pi32(a,b)
+  #define vec_zero_psqt() _mm_setzero_si64()
+  #define vec_cleanup() _mm_empty()
+  #define NumRegistersSIMD 8
+  #define MaxChunkSize 8
+
+  #elif USE_NEON
+  using vec_t = int16x8_t;
+  using psqt_vec_t = int32x4_t;
+  #define vec_load(a) (*(a))
+  #define vec_store(a,b) *(a)=(b)
+  #define vec_add_16(a,b) vaddq_s16(a,b)
+  #define vec_sub_16(a,b) vsubq_s16(a,b)
+  #define vec_mul_16(a,b) vmulq_s16(a,b)
+  #define vec_zero() vec_t{0}
+  #define vec_set_16(a) vdupq_n_s16(a)
+  #define vec_max_16(a,b) vmaxq_s16(a,b)
+  #define vec_min_16(a,b) vminq_s16(a,b)
+  inline vec_t vec_msb_pack_16(vec_t a, vec_t b){
+    const int8x8_t shifta = vshrn_n_s16(a, 7);
+    const int8x8_t shiftb = vshrn_n_s16(b, 7);
+    const int8x16_t compacted = vcombine_s8(shifta,shiftb);
+    return *reinterpret_cast<const vec_t*> (&compacted);
+  }
+  #define vec_load_psqt(a) (*(a))
+  #define vec_store_psqt(a,b) *(a)=(b)
+  #define vec_add_psqt_32(a,b) vaddq_s32(a,b)
+  #define vec_sub_psqt_32(a,b) vsubq_s32(a,b)
+  #define vec_zero_psqt() psqt_vec_t{0}
+  #define NumRegistersSIMD 16
+  #define MaxChunkSize 16
+
+  #else
+  #undef VECTOR
+
+  #endif
+
+
+  #ifdef VECTOR
+
+      // Compute optimal SIMD register count for feature transformer accumulation.
+
+      // We use __m* types as template arguments, which causes GCC to emit warnings
+      // about losing some attribute information. This is irrelevant to us as we
+      // only take their size, so the following pragma are harmless.
+      #if defined(__GNUC__)
+      #pragma GCC diagnostic push
+      #pragma GCC diagnostic ignored "-Wignored-attributes"
+      #endif
+
+      template <typename SIMDRegisterType,
+                typename LaneType,
+                int      NumLanes,
+                int      MaxRegisters>
+      static constexpr int BestRegisterCount()
+      {
+          #define RegisterSize  sizeof(SIMDRegisterType)
+          #define LaneSize      sizeof(LaneType)
+
+          static_assert(RegisterSize >= LaneSize);
+          static_assert(MaxRegisters <= NumRegistersSIMD);
+          static_assert(MaxRegisters > 0);
+          static_assert(NumRegistersSIMD > 0);
+          static_assert(RegisterSize % LaneSize == 0);
+          static_assert((NumLanes * LaneSize) % RegisterSize == 0);
+
+          const int ideal = (NumLanes * LaneSize) / RegisterSize;
+          if (ideal <= MaxRegisters)
+            return ideal;
+
+          // Look for the largest divisor of the ideal register count that is smaller than MaxRegisters
+          for (int divisor = MaxRegisters; divisor > 1; --divisor)
+            if (ideal % divisor == 0)
+              return divisor;
+
+          return 1;
+      }
+
+      static constexpr int NumRegs     = BestRegisterCount<vec_t, WeightType, TransformedFeatureDimensions, NumRegistersSIMD>();
+      static constexpr int NumPsqtRegs = BestRegisterCount<psqt_vec_t, PSQTWeightType, PSQTBuckets, NumRegistersSIMD>();
+      #if defined(__GNUC__)
+      #pragma GCC diagnostic pop
+      #endif
+  #endif
+
+
+
+  // Input feature converter
+  class FeatureTransformer {
+
+   private:
+    // Number of output dimensions for one side
+    static constexpr IndexType HalfDimensions = TransformedFeatureDimensions;
+
+    #ifdef VECTOR
+    static constexpr IndexType TileHeight = NumRegs * sizeof(vec_t) / 2;
+    static constexpr IndexType PsqtTileHeight = NumPsqtRegs * sizeof(psqt_vec_t) / 4;
+    static_assert(HalfDimensions % TileHeight == 0, "TileHeight must divide HalfDimensions");
+    static_assert(PSQTBuckets % PsqtTileHeight == 0, "PsqtTileHeight must divide PSQTBuckets");
+    #endif
+
+   public:
+    // Output type
+    using OutputType = TransformedFeatureType;
+
+    // Number of input/output dimensions
+    static constexpr IndexType InputDimensions = FeatureSet::Dimensions;
+    static constexpr IndexType OutputDimensions = HalfDimensions;
+
+    // Size of forward propagation buffer
+    static constexpr std::size_t BufferSize =
+        OutputDimensions * sizeof(OutputType);
+
+    // Hash value embedded in the evaluation file
+    static constexpr std::uint32_t get_hash_value() {
+      return FeatureSet::HashValue ^ (OutputDimensions * 2);
+    }
+
+    // Read network parameters
+    bool read_parameters(std::istream& stream) {
+
+      read_leb_128<BiasType      >(stream, biases     , HalfDimensions                  );
+      read_leb_128<WeightType    >(stream, weights    , HalfDimensions * InputDimensions);
+      read_leb_128<PSQTWeightType>(stream, psqtWeights, PSQTBuckets    * InputDimensions);
+
+      return !stream.fail();
+    }
+
+    // Write network parameters
+    bool write_parameters(std::ostream& stream) const {
+
+      write_leb_128<BiasType      >(stream, biases     , HalfDimensions                  );
+      write_leb_128<WeightType    >(stream, weights    , HalfDimensions * InputDimensions);
+      write_leb_128<PSQTWeightType>(stream, psqtWeights, PSQTBuckets    * InputDimensions);
+
+      return !stream.fail();
+    }
+
+    // Convert input features
+    std::int32_t transform(const Position& pos, OutputType* output, int bucket) const {
+      update_accumulator<WHITE>(pos);
+      update_accumulator<BLACK>(pos);
+
+      const Color perspectives[2] = {pos.side_to_move(), ~pos.side_to_move()};
+      const auto& accumulation = pos.state()->accumulator.accumulation;
+      const auto& psqtAccumulation = pos.state()->accumulator.psqtAccumulation;
+
+      const auto psqt = (
+            psqtAccumulation[perspectives[0]][bucket]
+          - psqtAccumulation[perspectives[1]][bucket]
+        ) / 2;
+
+
+      for (IndexType p = 0; p < 2; ++p)
+      {
+          const IndexType offset = (HalfDimensions / 2) * p;
+
+#if defined(VECTOR)
+
+          constexpr IndexType OutputChunkSize = MaxChunkSize;
+          static_assert((HalfDimensions / 2) % OutputChunkSize == 0);
+          constexpr IndexType NumOutputChunks = HalfDimensions / 2 / OutputChunkSize;
+
+          vec_t Zero = vec_zero();
+          vec_t One = vec_set_16(127);
+
+          const vec_t* in0 = reinterpret_cast<const vec_t*>(&(accumulation[perspectives[p]][0]));
+          const vec_t* in1 = reinterpret_cast<const vec_t*>(&(accumulation[perspectives[p]][HalfDimensions / 2]));
+                vec_t* out = reinterpret_cast<      vec_t*>(output + offset);
+
+          for (IndexType j = 0; j < NumOutputChunks; j += 1)
+          {
+              const vec_t sum0a = vec_max_16(vec_min_16(in0[j * 2 + 0], One), Zero);
+              const vec_t sum0b = vec_max_16(vec_min_16(in0[j * 2 + 1], One), Zero);
+              const vec_t sum1a = vec_max_16(vec_min_16(in1[j * 2 + 0], One), Zero);
+              const vec_t sum1b = vec_max_16(vec_min_16(in1[j * 2 + 1], One), Zero);
+
+              const vec_t pa = vec_mul_16(sum0a, sum1a);
+              const vec_t pb = vec_mul_16(sum0b, sum1b);
+
+              out[j] = vec_msb_pack_16(pa, pb);
+          }
+
+#else
+
+          for (IndexType j = 0; j < HalfDimensions / 2; ++j) {
+              BiasType sum0 = accumulation[static_cast<int>(perspectives[p])][j + 0];
+              BiasType sum1 = accumulation[static_cast<int>(perspectives[p])][j + HalfDimensions / 2];
+              sum0 = std::max<int>(0, std::min<int>(127, sum0));
+              sum1 = std::max<int>(0, std::min<int>(127, sum1));
+              output[offset + j] = static_cast<OutputType>(sum0 * sum1 / 128);
+          }
+
+#endif
+      }
+
+#if defined(vec_cleanup)
+      vec_cleanup();
+#endif
+
+      return psqt;
+    } // end of function transform()
+
+    void hint_common_access(const Position& pos) const {
+      hint_common_access_for_perspective<WHITE>(pos);
+      hint_common_access_for_perspective<BLACK>(pos);
+    }
+
+   private:
+    template<Color Perspective>
+    [[nodiscard]] std::pair<StateInfo*, StateInfo*> try_find_computed_accumulator(const Position& pos) const {
+      // Look for a usable accumulator of an earlier position. We keep track
+      // of the estimated gain in terms of features to be added/subtracted.
+      StateInfo *st = pos.state(), *next = nullptr;
+      int gain = FeatureSet::refresh_cost(pos);
+      while (st->previous && !st->accumulator.computed[Perspective])
+      {
+        // This governs when a full feature refresh is needed and how many
+        // updates are better than just one full refresh.
+        if (   FeatureSet::requires_refresh(st, Perspective)
+            || (gain -= FeatureSet::update_cost(st) + 1) < 0)
+          break;
+        next = st;
+        st = st->previous;
+      }
+      return { st, next };
+    }
+
+    // NOTE: The parameter states_to_update is an array of position states, ending with nullptr.
+    //       All states must be sequential, that is states_to_update[i] must either be reachable
+    //       by repeatedly applying ->previous from states_to_update[i+1] or states_to_update[i] == nullptr.
+    //       computed_st must be reachable by repeatedly applying ->previous on states_to_update[0], if not nullptr.
+    template<Color Perspective, size_t N>
+    void update_accumulator_incremental(const Position& pos, StateInfo* computed_st, StateInfo* states_to_update[N]) const {
+      static_assert(N > 0);
+      assert(states_to_update[N-1] == nullptr);
+
+  #ifdef VECTOR
+      // Gcc-10.2 unnecessarily spills AVX2 registers if this array
+      // is defined in the VECTOR code below, once in each branch
+      vec_t acc[NumRegs];
+      psqt_vec_t psqt[NumPsqtRegs];
+  #endif
+
+      if (states_to_update[0] == nullptr)
+        return;
+
+      // Update incrementally going back through states_to_update.
+
+      // Gather all features to be updated.
+      const Square ksq = pos.square<KING>(Perspective);
+
+      // The size must be enough to contain the largest possible update.
+      // That might depend on the feature set and generally relies on the
+      // feature set's update cost calculation to be correct and never
+      // allow updates with more added/removed features than MaxActiveDimensions.
+      FeatureSet::IndexList removed[N-1], added[N-1];
+
+      {
+        int i = N-2; // last potential state to update. Skip last element because it must be nullptr.
+        while (states_to_update[i] == nullptr)
+          --i;
+
+        StateInfo *st2 = states_to_update[i];
+
+        for (; i >= 0; --i)
+        {
+          states_to_update[i]->accumulator.computed[Perspective] = true;
+
+          StateInfo* end_state = i == 0 ? computed_st : states_to_update[i - 1];
+
+          for (; st2 != end_state; st2 = st2->previous)
+            FeatureSet::append_changed_indices<Perspective>(
+              ksq, st2->dirtyPiece, removed[i], added[i]);
+        }
+      }
+
+      StateInfo* st = computed_st;
+
+      // Now update the accumulators listed in states_to_update[], where the last element is a sentinel.
+#ifdef VECTOR
+      for (IndexType j = 0; j < HalfDimensions / TileHeight; ++j)
+      {
+        // Load accumulator
+        auto accTile = reinterpret_cast<vec_t*>(
+          &st->accumulator.accumulation[Perspective][j * TileHeight]);
+        for (IndexType k = 0; k < NumRegs; ++k)
+          acc[k] = vec_load(&accTile[k]);
+
+        for (IndexType i = 0; states_to_update[i]; ++i)
+        {
+          // Difference calculation for the deactivated features
+          for (const auto index : removed[i])
+          {
+            const IndexType offset = HalfDimensions * index + j * TileHeight;
+            auto column = reinterpret_cast<const vec_t*>(&weights[offset]);
+            for (IndexType k = 0; k < NumRegs; ++k)
+              acc[k] = vec_sub_16(acc[k], column[k]);
+          }
+
+          // Difference calculation for the activated features
+          for (const auto index : added[i])
+          {
+            const IndexType offset = HalfDimensions * index + j * TileHeight;
+            auto column = reinterpret_cast<const vec_t*>(&weights[offset]);
+            for (IndexType k = 0; k < NumRegs; ++k)
+              acc[k] = vec_add_16(acc[k], column[k]);
+          }
+
+          // Store accumulator
+          accTile = reinterpret_cast<vec_t*>(
+            &states_to_update[i]->accumulator.accumulation[Perspective][j * TileHeight]);
+          for (IndexType k = 0; k < NumRegs; ++k)
+            vec_store(&accTile[k], acc[k]);
+        }
+      }
+
+      for (IndexType j = 0; j < PSQTBuckets / PsqtTileHeight; ++j)
+      {
+        // Load accumulator
+        auto accTilePsqt = reinterpret_cast<psqt_vec_t*>(
+          &st->accumulator.psqtAccumulation[Perspective][j * PsqtTileHeight]);
+        for (std::size_t k = 0; k < NumPsqtRegs; ++k)
+          psqt[k] = vec_load_psqt(&accTilePsqt[k]);
+
+        for (IndexType i = 0; states_to_update[i]; ++i)
+        {
+          // Difference calculation for the deactivated features
+          for (const auto index : removed[i])
+          {
+            const IndexType offset = PSQTBuckets * index + j * PsqtTileHeight;
+            auto columnPsqt = reinterpret_cast<const psqt_vec_t*>(&psqtWeights[offset]);
+            for (std::size_t k = 0; k < NumPsqtRegs; ++k)
+              psqt[k] = vec_sub_psqt_32(psqt[k], columnPsqt[k]);
+          }
+
+          // Difference calculation for the activated features
+          for (const auto index : added[i])
+          {
+            const IndexType offset = PSQTBuckets * index + j * PsqtTileHeight;
+            auto columnPsqt = reinterpret_cast<const psqt_vec_t*>(&psqtWeights[offset]);
+            for (std::size_t k = 0; k < NumPsqtRegs; ++k)
+              psqt[k] = vec_add_psqt_32(psqt[k], columnPsqt[k]);
+          }
+
+          // Store accumulator
+          accTilePsqt = reinterpret_cast<psqt_vec_t*>(
+            &states_to_update[i]->accumulator.psqtAccumulation[Perspective][j * PsqtTileHeight]);
+          for (std::size_t k = 0; k < NumPsqtRegs; ++k)
+            vec_store_psqt(&accTilePsqt[k], psqt[k]);
+        }
+      }
+
+#else
+      for (IndexType i = 0; states_to_update[i]; ++i)
+      {
+        std::memcpy(states_to_update[i]->accumulator.accumulation[Perspective],
+            st->accumulator.accumulation[Perspective],
+            HalfDimensions * sizeof(BiasType));
+
+        for (std::size_t k = 0; k < PSQTBuckets; ++k)
+          states_to_update[i]->accumulator.psqtAccumulation[Perspective][k] = st->accumulator.psqtAccumulation[Perspective][k];
+
+        st = states_to_update[i];
+
+        // Difference calculation for the deactivated features
+        for (const auto index : removed[i])
+        {
+          const IndexType offset = HalfDimensions * index;
+
+          for (IndexType j = 0; j < HalfDimensions; ++j)
+            st->accumulator.accumulation[Perspective][j] -= weights[offset + j];
+
+          for (std::size_t k = 0; k < PSQTBuckets; ++k)
+            st->accumulator.psqtAccumulation[Perspective][k] -= psqtWeights[index * PSQTBuckets + k];
+        }
+
+        // Difference calculation for the activated features
+        for (const auto index : added[i])
+        {
+          const IndexType offset = HalfDimensions * index;
+
+          for (IndexType j = 0; j < HalfDimensions; ++j)
+            st->accumulator.accumulation[Perspective][j] += weights[offset + j];
+
+          for (std::size_t k = 0; k < PSQTBuckets; ++k)
+            st->accumulator.psqtAccumulation[Perspective][k] += psqtWeights[index * PSQTBuckets + k];
+        }
+      }
+#endif
+
+  #if defined(USE_MMX)
+      _mm_empty();
+  #endif
+    }
+
+    template<Color Perspective>
+    void update_accumulator_refresh(const Position& pos) const {
+  #ifdef VECTOR
+      // Gcc-10.2 unnecessarily spills AVX2 registers if this array
+      // is defined in the VECTOR code below, once in each branch
+      vec_t acc[NumRegs];
+      psqt_vec_t psqt[NumPsqtRegs];
+  #endif
+
+      // Refresh the accumulator
+      // Could be extracted to a separate function because it's done in 2 places,
+      // but it's unclear if compilers would correctly handle register allocation.
+      auto& accumulator = pos.state()->accumulator;
+      accumulator.computed[Perspective] = true;
+      FeatureSet::IndexList active;
+      FeatureSet::append_active_indices<Perspective>(pos, active);
+
+#ifdef VECTOR
+      for (IndexType j = 0; j < HalfDimensions / TileHeight; ++j)
+      {
+        auto biasesTile = reinterpret_cast<const vec_t*>(
+            &biases[j * TileHeight]);
+        for (IndexType k = 0; k < NumRegs; ++k)
+          acc[k] = biasesTile[k];
+
+        for (const auto index : active)
+        {
+          const IndexType offset = HalfDimensions * index + j * TileHeight;
+          auto column = reinterpret_cast<const vec_t*>(&weights[offset]);
+
+          for (unsigned k = 0; k < NumRegs; ++k)
+            acc[k] = vec_add_16(acc[k], column[k]);
+        }
+
+        auto accTile = reinterpret_cast<vec_t*>(
+            &accumulator.accumulation[Perspective][j * TileHeight]);
+        for (unsigned k = 0; k < NumRegs; k++)
+          vec_store(&accTile[k], acc[k]);
+      }
+
+      for (IndexType j = 0; j < PSQTBuckets / PsqtTileHeight; ++j)
+      {
+        for (std::size_t k = 0; k < NumPsqtRegs; ++k)
+          psqt[k] = vec_zero_psqt();
+
+        for (const auto index : active)
+        {
+          const IndexType offset = PSQTBuckets * index + j * PsqtTileHeight;
+          auto columnPsqt = reinterpret_cast<const psqt_vec_t*>(&psqtWeights[offset]);
+
+          for (std::size_t k = 0; k < NumPsqtRegs; ++k)
+            psqt[k] = vec_add_psqt_32(psqt[k], columnPsqt[k]);
+        }
+
+        auto accTilePsqt = reinterpret_cast<psqt_vec_t*>(
+          &accumulator.psqtAccumulation[Perspective][j * PsqtTileHeight]);
+        for (std::size_t k = 0; k < NumPsqtRegs; ++k)
+          vec_store_psqt(&accTilePsqt[k], psqt[k]);
+      }
+
+#else
+      std::memcpy(accumulator.accumulation[Perspective], biases,
+          HalfDimensions * sizeof(BiasType));
+
+      for (std::size_t k = 0; k < PSQTBuckets; ++k)
+        accumulator.psqtAccumulation[Perspective][k] = 0;
+
+      for (const auto index : active)
+      {
+        const IndexType offset = HalfDimensions * index;
+
+        for (IndexType j = 0; j < HalfDimensions; ++j)
+          accumulator.accumulation[Perspective][j] += weights[offset + j];
+
+        for (std::size_t k = 0; k < PSQTBuckets; ++k)
+          accumulator.psqtAccumulation[Perspective][k] += psqtWeights[index * PSQTBuckets + k];
+      }
+#endif
+
+  #if defined(USE_MMX)
+      _mm_empty();
+  #endif
+    }
+
+    template<Color Perspective>
+    void hint_common_access_for_perspective(const Position& pos) const {
+
+      // Works like update_accumulator, but performs less work.
+      // Updates ONLY the accumulator for pos.
+
+      // Look for a usable accumulator of an earlier position. We keep track
+      // of the estimated gain in terms of features to be added/subtracted.
+      // Fast early exit.
+      if (pos.state()->accumulator.computed[Perspective])
+        return;
+
+      auto [oldest_st, _] = try_find_computed_accumulator<Perspective>(pos);
+
+      if (oldest_st->accumulator.computed[Perspective])
+      {
+        // Only update current position accumulator to minimize work.
+        StateInfo* states_to_update[2] = { pos.state(), nullptr };
+        update_accumulator_incremental<Perspective, 2>(pos, oldest_st, states_to_update);
+      }
+      else
+      {
+        update_accumulator_refresh<Perspective>(pos);
+      }
+    }
+
+    template<Color Perspective>
+    void update_accumulator(const Position& pos) const {
+
+      auto [oldest_st, next] = try_find_computed_accumulator<Perspective>(pos);
+
+      if (oldest_st->accumulator.computed[Perspective])
+      {
+        if (next == nullptr)
+          return;
+
+        // Now update the accumulators listed in states_to_update[], where the last element is a sentinel.
+        // Currently we update 2 accumulators.
+        //     1. for the current position
+        //     2. the next accumulator after the computed one
+        // The heuristic may change in the future.
+        StateInfo *states_to_update[3] =
+          { next, next == pos.state() ? nullptr : pos.state(), nullptr };
+
+        update_accumulator_incremental<Perspective, 3>(pos, oldest_st, states_to_update);
+      }
+      else
+      {
+        update_accumulator_refresh<Perspective>(pos);
+      }
+    }
+
+    alignas(CacheLineSize) BiasType biases[HalfDimensions];
+    alignas(CacheLineSize) WeightType weights[HalfDimensions * InputDimensions];
+    alignas(CacheLineSize) PSQTWeightType psqtWeights[InputDimensions * PSQTBuckets];
+  };
+
+}  // namespace Stockfish::Eval::NNUE
+
+#endif // #ifndef NNUE_FEATURE_TRANSFORMER_H_INCLUDED

src/Stockfish/src/pawns.cpp

@@ -0,0 +1,305 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include <algorithm>
+#include <cassert>
+
+#include "bitboard.h"
+#include "pawns.h"
+#include "position.h"
+#include "thread.h"
+
+namespace Stockfish {
+
+namespace {
+
+  #define V Value
+  #define S(mg, eg) make_score(mg, eg)
+
+  // Pawn penalties
+  constexpr Score Backward      = S( 6, 19);
+  constexpr Score Doubled       = S(11, 51);
+  constexpr Score DoubledEarly  = S(17,  7);
+  constexpr Score Isolated      = S( 1, 20);
+  constexpr Score WeakLever     = S( 2, 57);
+  constexpr Score WeakUnopposed = S(15, 18);
+
+  // Bonus for blocked pawns at 5th or 6th rank
+  constexpr Score BlockedPawn[2] = { S(-19, -8), S(-7, 3) };
+
+  constexpr Score BlockedStorm[RANK_NB] = {
+    S(0, 0), S(0, 0), S(64, 75), S(-3, 14), S(-12, 19), S(-7, 4), S(-10, 5)
+  };
+
+  // Connected pawn bonus
+  constexpr int Connected[RANK_NB] = { 0, 3, 7, 7, 15, 54, 86 };
+
+  // Strength of pawn shelter for our king by [distance from edge][rank].
+  // RANK_1 = 0 is used for files where we have no pawn, or pawn is behind our king.
+  constexpr Value ShelterStrength[int(FILE_NB) / 2][RANK_NB] = {
+    { V(-2), V(85), V(95), V(53), V(39), V(23), V(25) },
+    { V(-55), V(64), V(32), V(-55), V(-30), V(-11), V(-61) },
+    { V(-11), V(75), V(19), V(-6), V(26), V(9), V(-47) },
+    { V(-41), V(-11), V(-27), V(-58), V(-42), V(-66), V(-163) }
+  };
+
+  // Danger of enemy pawns moving toward our king by [distance from edge][rank].
+  // RANK_1 = 0 is used for files where the enemy has no pawn, or their pawn
+  // is behind our king. Note that UnblockedStorm[0][1-2] accommodate opponent pawn
+  // on edge, likely blocked by our king.
+  constexpr Value UnblockedStorm[int(FILE_NB) / 2][RANK_NB] = {
+    { V(94), V(-280), V(-170), V(90), V(59), V(47), V(53) },
+    { V(43), V(-17), V(128), V(39), V(26), V(-17), V(15) },
+    { V(-9), V(62), V(170), V(34), V(-5), V(-20), V(-11) },
+    { V(-27), V(-19), V(106), V(10), V(2), V(-13), V(-24) }
+  };
+
+
+  // KingOnFile[semi-open Us][semi-open Them] contains bonuses/penalties
+  // for king when the king is on a semi-open or open file.
+  constexpr Score KingOnFile[2][2] = {{ S(-18,11), S(-6,-3)  },
+                                     {  S(  0, 0), S( 5,-4) }};
+
+  #undef S
+  #undef V
+
+
+  /// evaluate() calculates a score for the static pawn structure of the given position.
+  /// We cannot use the location of pieces or king in this function, as the evaluation
+  /// of the pawn structure will be stored in a small cache for speed reasons, and will
+  /// be re-used even when the pieces have moved.
+
+  template<Color Us>
+  Score evaluate(const Position& pos, Pawns::Entry* e) {
+
+    constexpr Color     Them = ~Us;
+    constexpr Direction Up   = pawn_push(Us);
+    constexpr Direction Down = -Up;
+
+    Bitboard neighbours, stoppers, support, phalanx, opposed;
+    Bitboard lever, leverPush, blocked;
+    Square s;
+    bool backward, passed, doubled;
+    Score score = SCORE_ZERO;
+    Bitboard b = pos.pieces(Us, PAWN);
+
+    Bitboard ourPawns   = pos.pieces(  Us, PAWN);
+    Bitboard theirPawns = pos.pieces(Them, PAWN);
+
+    Bitboard doubleAttackThem = pawn_double_attacks_bb<Them>(theirPawns);
+
+    e->passedPawns[Us] = 0;
+    e->kingSquares[Us] = SQ_NONE;
+    e->pawnAttacks[Us] = e->pawnAttacksSpan[Us] = pawn_attacks_bb<Us>(ourPawns);
+    e->blockedCount += popcount(shift<Up>(ourPawns) & (theirPawns | doubleAttackThem));
+
+    // Loop through all pawns of the current color and score each pawn
+    while (b)
+    {
+        s = pop_lsb(b);
+
+        assert(pos.piece_on(s) == make_piece(Us, PAWN));
+
+        Rank r = relative_rank(Us, s);
+
+        // Flag the pawn
+        opposed    = theirPawns & forward_file_bb(Us, s);
+        blocked    = theirPawns & (s + Up);
+        stoppers   = theirPawns & passed_pawn_span(Us, s);
+        lever      = theirPawns & pawn_attacks_bb(Us, s);
+        leverPush  = theirPawns & pawn_attacks_bb(Us, s + Up);
+        doubled    = ourPawns   & (s - Up);
+        neighbours = ourPawns   & adjacent_files_bb(s);
+        phalanx    = neighbours & rank_bb(s);
+        support    = neighbours & rank_bb(s - Up);
+
+        if (doubled)
+        {
+            // Additional doubled penalty if none of their pawns is fixed
+            if (!(ourPawns & shift<Down>(theirPawns | pawn_attacks_bb<Them>(theirPawns))))
+                score -= DoubledEarly;
+        }
+
+        // A pawn is backward when it is behind all pawns of the same color on
+        // the adjacent files and cannot safely advance.
+        backward =  !(neighbours & forward_ranks_bb(Them, s + Up))
+                  && (leverPush | blocked);
+
+        // Compute additional span if pawn is not backward nor blocked
+        if (!backward && !blocked)
+            e->pawnAttacksSpan[Us] |= pawn_attack_span(Us, s);
+
+        // A pawn is passed if one of the three following conditions is true:
+        // (a) there is no stoppers except some levers
+        // (b) the only stoppers are the leverPush, but we outnumber them
+        // (c) there is only one front stopper which can be levered.
+        //     (Refined in Evaluation::passed)
+        passed =   !(stoppers ^ lever)
+                || (   !(stoppers ^ leverPush)
+                    && popcount(phalanx) >= popcount(leverPush))
+                || (   stoppers == blocked && r >= RANK_5
+                    && (shift<Up>(support) & ~(theirPawns | doubleAttackThem)));
+
+        passed &= !(forward_file_bb(Us, s) & ourPawns);
+
+        // Passed pawns will be properly scored later in evaluation when we have
+        // full attack info.
+        if (passed)
+            e->passedPawns[Us] |= s;
+
+        // Score this pawn
+        if (support | phalanx)
+        {
+            int v =  Connected[r] * (2 + bool(phalanx) - bool(opposed))
+                   + 22 * popcount(support);
+
+            score += make_score(v, v * (r - 2) / 4);
+        }
+
+        else if (!neighbours)
+        {
+            if (     opposed
+                &&  (ourPawns & forward_file_bb(Them, s))
+                && !(theirPawns & adjacent_files_bb(s)))
+                score -= Doubled;
+            else
+                score -=  Isolated
+                        + WeakUnopposed * !opposed;
+        }
+
+        else if (backward)
+            score -=  Backward
+                    + WeakUnopposed * !opposed * bool(~(FileABB | FileHBB) & s);
+
+        if (!support)
+            score -=  Doubled * doubled
+                    + WeakLever * more_than_one(lever);
+
+        if (blocked && r >= RANK_5)
+            score += BlockedPawn[r - RANK_5];
+    }
+
+    return score;
+  }
+
+} // namespace
+
+namespace Pawns {
+
+
+/// Pawns::probe() looks up the current position's pawns configuration in
+/// the pawns hash table. It returns a pointer to the Entry if the position
+/// is found. Otherwise a new Entry is computed and stored there, so we don't
+/// have to recompute all when the same pawns configuration occurs again.
+
+Entry* probe(const Position& pos) {
+
+  Key key = pos.pawn_key();
+  Entry* e = pos.this_thread()->pawnsTable[key];
+
+  if (e->key == key)
+      return e;
+
+  e->key = key;
+  e->blockedCount = 0;
+  e->scores[WHITE] = evaluate<WHITE>(pos, e);
+  e->scores[BLACK] = evaluate<BLACK>(pos, e);
+
+  return e;
+}
+
+
+/// Entry::evaluate_shelter() calculates the shelter bonus and the storm
+/// penalty for a king, looking at the king file and the two closest files.
+
+template<Color Us>
+Score Entry::evaluate_shelter(const Position& pos, Square ksq) const {
+
+  constexpr Color Them = ~Us;
+
+  Bitboard b = pos.pieces(PAWN) & ~forward_ranks_bb(Them, ksq);
+  Bitboard ourPawns = b & pos.pieces(Us) & ~pawnAttacks[Them];
+  Bitboard theirPawns = b & pos.pieces(Them);
+
+  Score bonus = make_score(5, 5);
+
+  File center = std::clamp(file_of(ksq), FILE_B, FILE_G);
+  for (File f = File(center - 1); f <= File(center + 1); ++f)
+  {
+      b = ourPawns & file_bb(f);
+      int ourRank = b ? relative_rank(Us, frontmost_sq(Them, b)) : 0;
+
+      b = theirPawns & file_bb(f);
+      int theirRank = b ? relative_rank(Us, frontmost_sq(Them, b)) : 0;
+
+      int d = edge_distance(f);
+      bonus += make_score(ShelterStrength[d][ourRank], 0);
+
+      if (ourRank && (ourRank == theirRank - 1))
+          bonus -= BlockedStorm[theirRank];
+      else
+          bonus -= make_score(UnblockedStorm[d][theirRank], 0);
+  }
+
+  // King On File
+  bonus -= KingOnFile[pos.is_on_semiopen_file(Us, ksq)][pos.is_on_semiopen_file(Them, ksq)];
+
+  return bonus;
+}
+
+
+/// Entry::do_king_safety() calculates a bonus for king safety. It is called only
+/// when king square changes, which is about 20% of total king_safety() calls.
+
+template<Color Us>
+Score Entry::do_king_safety(const Position& pos) {
+
+  Square ksq = pos.square<KING>(Us);
+  kingSquares[Us] = ksq;
+  castlingRights[Us] = pos.castling_rights(Us);
+  auto compare = [](Score a, Score b) { return mg_value(a) < mg_value(b); };
+
+  Score shelter = evaluate_shelter<Us>(pos, ksq);
+
+  // If we can castle use the bonus after castling if it is bigger
+
+  if (pos.can_castle(Us & KING_SIDE))
+      shelter = std::max(shelter, evaluate_shelter<Us>(pos, relative_square(Us, SQ_G1)), compare);
+
+  if (pos.can_castle(Us & QUEEN_SIDE))
+      shelter = std::max(shelter, evaluate_shelter<Us>(pos, relative_square(Us, SQ_C1)), compare);
+
+  // In endgame we like to bring our king near our closest pawn
+  Bitboard pawns = pos.pieces(Us, PAWN);
+  int minPawnDist = 6;
+
+  if (pawns & attacks_bb<KING>(ksq))
+      minPawnDist = 1;
+  else while (pawns)
+      minPawnDist = std::min(minPawnDist, distance(ksq, pop_lsb(pawns)));
+
+  return shelter - make_score(0, 16 * minPawnDist);
+}
+
+// Explicit template instantiation
+template Score Entry::do_king_safety<WHITE>(const Position& pos);
+template Score Entry::do_king_safety<BLACK>(const Position& pos);
+
+} // namespace Pawns
+
+} // namespace Stockfish

src/Stockfish/src/pawns.h

@@ -0,0 +1,70 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef PAWNS_H_INCLUDED
+#define PAWNS_H_INCLUDED
+
+#include "misc.h"
+#include "position.h"
+#include "types.h"
+
+namespace Stockfish::Pawns {
+
+/// Pawns::Entry contains various information about a pawn structure. A lookup
+/// to the pawn hash table (performed by calling the probe function) returns a
+/// pointer to an Entry object.
+
+struct Entry {
+
+  Score pawn_score(Color c) const { return scores[c]; }
+  Bitboard pawn_attacks(Color c) const { return pawnAttacks[c]; }
+  Bitboard passed_pawns(Color c) const { return passedPawns[c]; }
+  Bitboard pawn_attacks_span(Color c) const { return pawnAttacksSpan[c]; }
+  int passed_count() const { return popcount(passedPawns[WHITE] | passedPawns[BLACK]); }
+  int blocked_count() const { return blockedCount; }
+
+  template<Color Us>
+  Score king_safety(const Position& pos) {
+    return  kingSquares[Us] == pos.square<KING>(Us) && castlingRights[Us] == pos.castling_rights(Us)
+          ? kingSafety[Us] : (kingSafety[Us] = do_king_safety<Us>(pos));
+  }
+
+  template<Color Us>
+  Score do_king_safety(const Position& pos);
+
+  template<Color Us>
+  Score evaluate_shelter(const Position& pos, Square ksq) const;
+
+  Key key;
+  Score scores[COLOR_NB];
+  Bitboard passedPawns[COLOR_NB];
+  Bitboard pawnAttacks[COLOR_NB];
+  Bitboard pawnAttacksSpan[COLOR_NB];
+  Square kingSquares[COLOR_NB];
+  Score kingSafety[COLOR_NB];
+  int castlingRights[COLOR_NB];
+  int blockedCount;
+};
+
+using Table = HashTable<Entry, 131072>;
+
+Entry* probe(const Position& pos);
+
+} // namespace Stockfish::Pawns
+
+#endif // #ifndef PAWNS_H_INCLUDED

src/Stockfish/src/position.h

@@ -0,0 +1,450 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef POSITION_H_INCLUDED
+#define POSITION_H_INCLUDED
+
+#include <cassert>
+#include <deque>
+#include <memory> // For std::unique_ptr
+#include <string>
+
+#include "bitboard.h"
+#include "evaluate.h"
+#include "psqt.h"
+#include "types.h"
+
+#include "nnue/nnue_accumulator.h"
+
+namespace Stockfish {
+
+/// StateInfo struct stores information needed to restore a Position object to
+/// its previous state when we retract a move. Whenever a move is made on the
+/// board (by calling Position::do_move), a StateInfo object must be passed.
+
+struct StateInfo {
+
+  // Copied when making a move
+  Key    pawnKey;
+  Key    materialKey;
+  Value  nonPawnMaterial[COLOR_NB];
+  int    castlingRights;
+  int    rule50;
+  int    pliesFromNull;
+  Square epSquare;
+
+  // Not copied when making a move (will be recomputed anyhow)
+  Key        key;
+  Bitboard   checkersBB;
+  StateInfo* previous;
+  Bitboard   blockersForKing[COLOR_NB];
+  Bitboard   pinners[COLOR_NB];
+  Bitboard   checkSquares[PIECE_TYPE_NB];
+  Piece      capturedPiece;
+  int        repetition;
+
+  // Used by NNUE
+  Eval::NNUE::Accumulator accumulator;
+  DirtyPiece dirtyPiece;
+};
+
+
+/// A list to keep track of the position states along the setup moves (from the
+/// start position to the position just before the search starts). Needed by
+/// 'draw by repetition' detection. Use a std::deque because pointers to
+/// elements are not invalidated upon list resizing.
+using StateListPtr = std::unique_ptr<std::deque<StateInfo>>;
+
+
+/// Position class stores information regarding the board representation as
+/// pieces, side to move, hash keys, castling info, etc. Important methods are
+/// do_move() and undo_move(), used by the search to update node info when
+/// traversing the search tree.
+class Thread;
+
+class Position {
+public:
+  static void init();
+
+  Position() = default;
+  Position(const Position&) = delete;
+  Position& operator=(const Position&) = delete;
+
+  // FEN string input/output
+  Position& set(const std::string& fenStr, bool isChess960, StateInfo* si, Thread* th);
+  Position& set(const std::string& code, Color c, StateInfo* si);
+  std::string fen() const;
+
+  // Position representation
+  Bitboard pieces(PieceType pt) const;
+  template<typename ...PieceTypes> Bitboard pieces(PieceType pt, PieceTypes... pts) const;
+  Bitboard pieces(Color c) const;
+  template<typename ...PieceTypes> Bitboard pieces(Color c, PieceTypes... pts) const;
+  Piece piece_on(Square s) const;
+  Square ep_square() const;
+  bool empty(Square s) const;
+  template<PieceType Pt> int count(Color c) const;
+  template<PieceType Pt> int count() const;
+  template<PieceType Pt> Square square(Color c) const;
+  bool is_on_semiopen_file(Color c, Square s) const;
+
+  // Castling
+  CastlingRights castling_rights(Color c) const;
+  bool can_castle(CastlingRights cr) const;
+  bool castling_impeded(CastlingRights cr) const;
+  Square castling_rook_square(CastlingRights cr) const;
+
+  // Checking
+  Bitboard checkers() const;
+  Bitboard blockers_for_king(Color c) const;
+  Bitboard check_squares(PieceType pt) const;
+  Bitboard pinners(Color c) const;
+
+  // Attacks to/from a given square
+  Bitboard attackers_to(Square s) const;
+  Bitboard attackers_to(Square s, Bitboard occupied) const;
+  Bitboard slider_blockers(Bitboard sliders, Square s, Bitboard& pinners) const;
+  template<PieceType Pt> Bitboard attacks_by(Color c) const;
+
+  // Properties of moves
+  bool legal(Move m) const;
+  bool pseudo_legal(const Move m) const;
+  bool capture(Move m) const;
+  bool capture_stage(Move m) const;
+  bool gives_check(Move m) const;
+  Piece moved_piece(Move m) const;
+  Piece captured_piece() const;
+
+  // Piece specific
+  bool pawn_passed(Color c, Square s) const;
+  bool opposite_bishops() const;
+  int  pawns_on_same_color_squares(Color c, Square s) const;
+
+  // Doing and undoing moves
+  void do_move(Move m, StateInfo& newSt);
+  void do_move(Move m, StateInfo& newSt, bool givesCheck);
+  void undo_move(Move m);
+  void do_null_move(StateInfo& newSt);
+  void undo_null_move();
+
+  // Static Exchange Evaluation
+  bool see_ge(Move m, Value threshold = VALUE_ZERO) const;
+  bool see_ge(Move m, Bitboard& occupied, Value threshold = VALUE_ZERO) const;
+
+  // Accessing hash keys
+  Key key() const;
+  Key key_after(Move m) const;
+  Key material_key() const;
+  Key pawn_key() const;
+
+  // Other properties of the position
+  Color side_to_move() const;
+  int game_ply() const;
+  bool is_chess960() const;
+  Thread* this_thread() const;
+  bool is_draw(int ply) const;
+  bool has_game_cycle(int ply) const;
+  bool has_repeated() const;
+  int rule50_count() const;
+  Score psq_score() const;
+  Value psq_eg_stm() const;
+  Value non_pawn_material(Color c) const;
+  Value non_pawn_material() const;
+
+  // Position consistency check, for debugging
+  bool pos_is_ok() const;
+  void flip();
+
+  // Used by NNUE
+  StateInfo* state() const;
+
+  void put_piece(Piece pc, Square s);
+  void remove_piece(Square s);
+
+private:
+  // Initialization helpers (used while setting up a position)
+  void set_castling_right(Color c, Square rfrom);
+  void set_state() const;
+  void set_check_info() const;
+
+  // Other helpers
+  void move_piece(Square from, Square to);
+  template<bool Do>
+  void do_castling(Color us, Square from, Square& to, Square& rfrom, Square& rto);
+  template<bool AfterMove>
+  Key adjust_key50(Key k) const;
+
+  // Data members
+  Piece board[SQUARE_NB];
+  Bitboard byTypeBB[PIECE_TYPE_NB];
+  Bitboard byColorBB[COLOR_NB];
+  int pieceCount[PIECE_NB];
+  int castlingRightsMask[SQUARE_NB];
+  Square castlingRookSquare[CASTLING_RIGHT_NB];
+  Bitboard castlingPath[CASTLING_RIGHT_NB];
+  Thread* thisThread;
+  StateInfo* st;
+  int gamePly;
+  Color sideToMove;
+  Score psq;
+  bool chess960;
+};
+
+std::ostream& operator<<(std::ostream& os, const Position& pos);
+
+inline Color Position::side_to_move() const {
+  return sideToMove;
+}
+
+inline Piece Position::piece_on(Square s) const {
+  assert(is_ok(s));
+  return board[s];
+}
+
+inline bool Position::empty(Square s) const {
+  return piece_on(s) == NO_PIECE;
+}
+
+inline Piece Position::moved_piece(Move m) const {
+  return piece_on(from_sq(m));
+}
+
+inline Bitboard Position::pieces(PieceType pt = ALL_PIECES) const {
+  return byTypeBB[pt];
+}
+
+template<typename ...PieceTypes>
+inline Bitboard Position::pieces(PieceType pt, PieceTypes... pts) const {
+  return pieces(pt) | pieces(pts...);
+}
+
+inline Bitboard Position::pieces(Color c) const {
+  return byColorBB[c];
+}
+
+template<typename ...PieceTypes>
+inline Bitboard Position::pieces(Color c, PieceTypes... pts) const {
+  return pieces(c) & pieces(pts...);
+}
+
+template<PieceType Pt> inline int Position::count(Color c) const {
+  return pieceCount[make_piece(c, Pt)];
+}
+
+template<PieceType Pt> inline int Position::count() const {
+  return count<Pt>(WHITE) + count<Pt>(BLACK);
+}
+
+template<PieceType Pt> inline Square Position::square(Color c) const {
+  assert(count<Pt>(c) == 1);
+  return lsb(pieces(c, Pt));
+}
+
+inline Square Position::ep_square() const {
+  return st->epSquare;
+}
+
+inline bool Position::is_on_semiopen_file(Color c, Square s) const {
+  return !(pieces(c, PAWN) & file_bb(s));
+}
+
+inline bool Position::can_castle(CastlingRights cr) const {
+  return st->castlingRights & cr;
+}
+
+inline CastlingRights Position::castling_rights(Color c) const {
+  return c & CastlingRights(st->castlingRights);
+}
+
+inline bool Position::castling_impeded(CastlingRights cr) const {
+  assert(cr == WHITE_OO || cr == WHITE_OOO || cr == BLACK_OO || cr == BLACK_OOO);
+
+  return pieces() & castlingPath[cr];
+}
+
+inline Square Position::castling_rook_square(CastlingRights cr) const {
+  assert(cr == WHITE_OO || cr == WHITE_OOO || cr == BLACK_OO || cr == BLACK_OOO);
+
+  return castlingRookSquare[cr];
+}
+
+inline Bitboard Position::attackers_to(Square s) const {
+  return attackers_to(s, pieces());
+}
+
+template<PieceType Pt>
+inline Bitboard Position::attacks_by(Color c) const {
+
+  if constexpr (Pt == PAWN)
+      return c == WHITE ? pawn_attacks_bb<WHITE>(pieces(WHITE, PAWN))
+                        : pawn_attacks_bb<BLACK>(pieces(BLACK, PAWN));
+  else
+  {
+      Bitboard threats = 0;
+      Bitboard attackers = pieces(c, Pt);
+      while (attackers)
+          threats |= attacks_bb<Pt>(pop_lsb(attackers), pieces());
+      return threats;
+  }
+}
+
+inline Bitboard Position::checkers() const {
+  return st->checkersBB;
+}
+
+inline Bitboard Position::blockers_for_king(Color c) const {
+  return st->blockersForKing[c];
+}
+
+inline Bitboard Position::pinners(Color c) const {
+  return st->pinners[c];
+}
+
+inline Bitboard Position::check_squares(PieceType pt) const {
+  return st->checkSquares[pt];
+}
+
+inline bool Position::pawn_passed(Color c, Square s) const {
+  return !(pieces(~c, PAWN) & passed_pawn_span(c, s));
+}
+
+inline int Position::pawns_on_same_color_squares(Color c, Square s) const {
+  return popcount(pieces(c, PAWN) & ((DarkSquares & s) ? DarkSquares : ~DarkSquares));
+}
+
+inline Key Position::key() const {
+  return adjust_key50<false>(st->key);
+}
+
+template<bool AfterMove>
+inline Key Position::adjust_key50(Key k) const
+{
+  return st->rule50 < 14 - AfterMove
+      ? k : k ^ make_key((st->rule50 - (14 - AfterMove)) / 8);
+}
+
+inline Key Position::pawn_key() const {
+  return st->pawnKey;
+}
+
+inline Key Position::material_key() const {
+  return st->materialKey;
+}
+
+inline Score Position::psq_score() const {
+  return psq;
+}
+
+inline Value Position::psq_eg_stm() const {
+  return (sideToMove == WHITE ? 1 : -1) * eg_value(psq);
+}
+
+inline Value Position::non_pawn_material(Color c) const {
+  return st->nonPawnMaterial[c];
+}
+
+inline Value Position::non_pawn_material() const {
+  return non_pawn_material(WHITE) + non_pawn_material(BLACK);
+}
+
+inline int Position::game_ply() const {
+  return gamePly;
+}
+
+inline int Position::rule50_count() const {
+  return st->rule50;
+}
+
+inline bool Position::opposite_bishops() const {
+  return   count<BISHOP>(WHITE) == 1
+        && count<BISHOP>(BLACK) == 1
+        && opposite_colors(square<BISHOP>(WHITE), square<BISHOP>(BLACK));
+}
+
+inline bool Position::is_chess960() const {
+  return chess960;
+}
+
+inline bool Position::capture(Move m) const {
+  assert(is_ok(m));
+  return     (!empty(to_sq(m)) && type_of(m) != CASTLING)
+          ||  type_of(m) == EN_PASSANT;
+}
+
+// returns true if a move is generated from the capture stage
+// having also queen promotions covered, i.e. consistency with the capture stage move generation
+// is needed to avoid the generation of duplicate moves.
+inline bool Position::capture_stage(Move m) const {
+  assert(is_ok(m));
+  return  capture(m) || promotion_type(m) == QUEEN;
+}
+
+inline Piece Position::captured_piece() const {
+  return st->capturedPiece;
+}
+
+inline Thread* Position::this_thread() const {
+  return thisThread;
+}
+
+inline void Position::put_piece(Piece pc, Square s) {
+
+  board[s] = pc;
+  byTypeBB[ALL_PIECES] |= byTypeBB[type_of(pc)] |= s;
+  byColorBB[color_of(pc)] |= s;
+  pieceCount[pc]++;
+  pieceCount[make_piece(color_of(pc), ALL_PIECES)]++;
+  psq += PSQT::psq[pc][s];
+}
+
+inline void Position::remove_piece(Square s) {
+
+  Piece pc = board[s];
+  byTypeBB[ALL_PIECES] ^= s;
+  byTypeBB[type_of(pc)] ^= s;
+  byColorBB[color_of(pc)] ^= s;
+  board[s] = NO_PIECE;
+  pieceCount[pc]--;
+  pieceCount[make_piece(color_of(pc), ALL_PIECES)]--;
+  psq -= PSQT::psq[pc][s];
+}
+
+inline void Position::move_piece(Square from, Square to) {
+
+  Piece pc = board[from];
+  Bitboard fromTo = from | to;
+  byTypeBB[ALL_PIECES] ^= fromTo;
+  byTypeBB[type_of(pc)] ^= fromTo;
+  byColorBB[color_of(pc)] ^= fromTo;
+  board[from] = NO_PIECE;
+  board[to] = pc;
+  psq += PSQT::psq[pc][to] - PSQT::psq[pc][from];
+}
+
+inline void Position::do_move(Move m, StateInfo& newSt) {
+  do_move(m, newSt, gives_check(m));
+}
+
+inline StateInfo* Position::state() const {
+
+  return st;
+}
+
+} // namespace Stockfish
+
+#endif // #ifndef POSITION_H_INCLUDED

src/Stockfish/src/psqt.cpp

@@ -0,0 +1,131 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+
+#include "psqt.h"
+
+#include <algorithm>
+
+#include "bitboard.h"
+#include "types.h"
+
+namespace Stockfish {
+
+namespace
+{
+
+auto constexpr S = make_score;
+
+// 'Bonus' contains Piece-Square parameters.
+// Scores are explicit for files A to D, implicitly mirrored for E to H.
+constexpr Score Bonus[][RANK_NB][int(FILE_NB) / 2] = {
+  { },
+  { },
+  { // Knight
+   { S(-175, -96), S(-92,-65), S(-74,-49), S(-73,-21) },
+   { S( -77, -67), S(-41,-54), S(-27,-18), S(-15,  8) },
+   { S( -61, -40), S(-17,-27), S(  6, -8), S( 12, 29) },
+   { S( -35, -35), S(  8, -2), S( 40, 13), S( 49, 28) },
+   { S( -34, -45), S( 13,-16), S( 44,  9), S( 51, 39) },
+   { S(  -9, -51), S( 22,-44), S( 58,-16), S( 53, 17) },
+   { S( -67, -69), S(-27,-50), S(  4,-51), S( 37, 12) },
+   { S(-201,-100), S(-83,-88), S(-56,-56), S(-26,-17) }
+  },
+  { // Bishop
+   { S(-37,-40), S(-4 ,-21), S( -6,-26), S(-16, -8) },
+   { S(-11,-26), S(  6, -9), S( 13,-12), S(  3,  1) },
+   { S(-5 ,-11), S( 15, -1), S( -4, -1), S( 12,  7) },
+   { S(-4 ,-14), S(  8, -4), S( 18,  0), S( 27, 12) },
+   { S(-8 ,-12), S( 20, -1), S( 15,-10), S( 22, 11) },
+   { S(-11,-21), S(  4,  4), S(  1,  3), S(  8,  4) },
+   { S(-12,-22), S(-10,-14), S(  4, -1), S(  0,  1) },
+   { S(-34,-32), S(  1,-29), S(-10,-26), S(-16,-17) }
+  },
+  { // Rook
+   { S(-31, -9), S(-20,-13), S(-14,-10), S(-5, -9) },
+   { S(-21,-12), S(-13, -9), S( -8, -1), S( 6, -2) },
+   { S(-25,  6), S(-11, -8), S( -1, -2), S( 3, -6) },
+   { S(-13, -6), S( -5,  1), S( -4, -9), S(-6,  7) },
+   { S(-27, -5), S(-15,  8), S( -4,  7), S( 3, -6) },
+   { S(-22,  6), S( -2,  1), S(  6, -7), S(12, 10) },
+   { S( -2,  4), S( 12,  5), S( 16, 20), S(18, -5) },
+   { S(-17, 18), S(-19,  0), S( -1, 19), S( 9, 13) }
+  },
+  { // Queen
+   { S( 3,-69), S(-5,-57), S(-5,-47), S( 4,-26) },
+   { S(-3,-54), S( 5,-31), S( 8,-22), S(12, -4) },
+   { S(-3,-39), S( 6,-18), S(13, -9), S( 7,  3) },
+   { S( 4,-23), S( 5, -3), S( 9, 13), S( 8, 24) },
+   { S( 0,-29), S(14, -6), S(12,  9), S( 5, 21) },
+   { S(-4,-38), S(10,-18), S( 6,-11), S( 8,  1) },
+   { S(-5,-50), S( 6,-27), S(10,-24), S( 8, -8) },
+   { S(-2,-74), S(-2,-52), S( 1,-43), S(-2,-34) }
+  },
+  { // King
+   { S(271,  1), S(327, 45), S(271, 85), S(198, 76) },
+   { S(278, 53), S(303,100), S(234,133), S(179,135) },
+   { S(195, 88), S(258,130), S(169,169), S(120,175) },
+   { S(164,103), S(190,156), S(138,172), S( 98,172) },
+   { S(154, 96), S(179,166), S(105,199), S( 70,199) },
+   { S(123, 92), S(145,172), S( 81,184), S( 31,191) },
+   { S( 88, 47), S(120,121), S( 65,116), S( 33,131) },
+   { S( 59, 11), S( 89, 59), S( 45, 73), S( -1, 78) }
+  }
+};
+
+constexpr Score PBonus[RANK_NB][FILE_NB] =
+  { // Pawn (asymmetric distribution)
+   { },
+   { S(  2, -8), S(  4, -6), S( 11,  9), S( 18,  5), S( 16, 16), S( 21,  6), S(  9, -6), S( -3,-18) },
+   { S( -9, -9), S(-15, -7), S( 11,-10), S( 15,  5), S( 31,  2), S( 23,  3), S(  6, -8), S(-20, -5) },
+   { S( -3,  7), S(-20,  1), S(  8, -8), S( 19, -2), S( 39,-14), S( 17,-13), S(  2,-11), S( -5, -6) },
+   { S( 11, 12), S( -4,  6), S(-11,  2), S(  2, -6), S( 11, -5), S(  0, -4), S(-12, 14), S(  5,  9) },
+   { S(  3, 27), S(-11, 18), S( -6, 19), S( 22, 29), S( -8, 30), S( -5,  9), S(-14,  8), S(-11, 14) },
+   { S( -7, -1), S(  6,-14), S( -2, 13), S(-11, 22), S(  4, 24), S(-14, 17), S( 10,  7), S( -9,  7) }
+  };
+
+} // namespace
+
+
+namespace PSQT
+{
+
+Score psq[PIECE_NB][SQUARE_NB];
+
+// PSQT::init() initializes piece-square tables: the white halves of the tables are
+// copied from Bonus[] and PBonus[], adding the piece value, then the black halves of
+// the tables are initialized by flipping and changing the sign of the white scores.
+void init() {
+
+  for (Piece pc : {W_PAWN, W_KNIGHT, W_BISHOP, W_ROOK, W_QUEEN, W_KING})
+  {
+    Score score = make_score(PieceValue[MG][pc], PieceValue[EG][pc]);
+
+    for (Square s = SQ_A1; s <= SQ_H8; ++s)
+    {
+      File f = File(edge_distance(file_of(s)));
+      psq[ pc][s] = score + (type_of(pc) == PAWN ? PBonus[rank_of(s)][file_of(s)]
+                                                 : Bonus[pc][rank_of(s)][f]);
+      psq[~pc][flip_rank(s)] = -psq[pc][s];
+    }
+  }
+}
+
+} // namespace PSQT
+
+} // namespace Stockfish

src/Stockfish/src/psqt.h

@@ -0,0 +1,38 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+
+#ifndef PSQT_H_INCLUDED
+#define PSQT_H_INCLUDED
+
+
+#include "types.h"
+
+
+namespace Stockfish::PSQT
+{
+
+extern Score psq[PIECE_NB][SQUARE_NB];
+
+// Fill psqt array from a set of internally linked parameters
+void init();
+
+} // namespace Stockfish::PSQT
+
+
+#endif // PSQT_H_INCLUDED

src/Stockfish/src/search.h

@@ -0,0 +1,116 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef SEARCH_H_INCLUDED
+#define SEARCH_H_INCLUDED
+
+#include <vector>
+
+#include "misc.h"
+#include "movepick.h"
+#include "types.h"
+
+namespace Stockfish {
+
+class Position;
+
+namespace Search {
+
+
+/// Stack struct keeps track of the information we need to remember from nodes
+/// shallower and deeper in the tree during the search. Each search thread has
+/// its own array of Stack objects, indexed by the current ply.
+
+struct Stack {
+  Move* pv;
+  PieceToHistory* continuationHistory;
+  int ply;
+  Move currentMove;
+  Move excludedMove;
+  Move killers[2];
+  Value staticEval;
+  int statScore;
+  int moveCount;
+  bool inCheck;
+  bool ttPv;
+  bool ttHit;
+  int doubleExtensions;
+  int cutoffCnt;
+};
+
+
+/// RootMove struct is used for moves at the root of the tree. For each root move
+/// we store a score and a PV (really a refutation in the case of moves which
+/// fail low). Score is normally set at -VALUE_INFINITE for all non-pv moves.
+
+struct RootMove {
+
+  explicit RootMove(Move m) : pv(1, m) {}
+  bool extract_ponder_from_tt(Position& pos);
+  bool operator==(const Move& m) const { return pv[0] == m; }
+  bool operator<(const RootMove& m) const { // Sort in descending order
+    return m.score != score ? m.score < score
+                            : m.previousScore < previousScore;
+  }
+
+  Value score = -VALUE_INFINITE;
+  Value previousScore = -VALUE_INFINITE;
+  Value averageScore = -VALUE_INFINITE;
+  Value uciScore = -VALUE_INFINITE;
+  bool scoreLowerbound = false;
+  bool scoreUpperbound = false;
+  int selDepth = 0;
+  int tbRank = 0;
+  Value tbScore;
+  std::vector<Move> pv;
+};
+
+using RootMoves = std::vector<RootMove>;
+
+
+/// LimitsType struct stores information sent by GUI about available time to
+/// search the current move, maximum depth/time, or if we are in analysis mode.
+
+struct LimitsType {
+
+  LimitsType() { // Init explicitly due to broken value-initialization of non POD in MSVC
+    time[WHITE] = time[BLACK] = inc[WHITE] = inc[BLACK] = npmsec = movetime = TimePoint(0);
+    movestogo = depth = mate = perft = infinite = 0;
+    nodes = 0;
+  }
+
+  bool use_time_management() const {
+    return time[WHITE] || time[BLACK];
+  }
+
+  std::vector<Move> searchmoves;
+  TimePoint time[COLOR_NB], inc[COLOR_NB], npmsec, movetime, startTime;
+  int movestogo, depth, mate, perft, infinite;
+  int64_t nodes;
+};
+
+extern LimitsType Limits;
+
+void init();
+void clear();
+
+} // namespace Search
+
+} // namespace Stockfish
+
+#endif // #ifndef SEARCH_H_INCLUDED

src/Stockfish/src/syzygy/tbprobe.h

@@ -0,0 +1,76 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef TBPROBE_H
+#define TBPROBE_H
+
+#include <ostream>
+
+#include "../search.h"
+
+namespace Stockfish::Tablebases {
+
+enum WDLScore {
+    WDLLoss        = -2, // Loss
+    WDLBlessedLoss = -1, // Loss, but draw under 50-move rule
+    WDLDraw        =  0, // Draw
+    WDLCursedWin   =  1, // Win, but draw under 50-move rule
+    WDLWin         =  2, // Win
+};
+
+// Possible states after a probing operation
+enum ProbeState {
+    FAIL              =  0, // Probe failed (missing file table)
+    OK                =  1, // Probe successful
+    CHANGE_STM        = -1, // DTZ should check the other side
+    ZEROING_BEST_MOVE =  2  // Best move zeroes DTZ (capture or pawn move)
+};
+
+extern int MaxCardinality;
+
+void init(const std::string& paths);
+WDLScore probe_wdl(Position& pos, ProbeState* result);
+int probe_dtz(Position& pos, ProbeState* result);
+bool root_probe(Position& pos, Search::RootMoves& rootMoves);
+bool root_probe_wdl(Position& pos, Search::RootMoves& rootMoves);
+void rank_root_moves(Position& pos, Search::RootMoves& rootMoves);
+
+inline std::ostream& operator<<(std::ostream& os, const WDLScore v) {
+
+    os << (v == WDLLoss        ? "Loss" :
+           v == WDLBlessedLoss ? "Blessed loss" :
+           v == WDLDraw        ? "Draw" :
+           v == WDLCursedWin   ? "Cursed win" :
+           v == WDLWin         ? "Win" : "None");
+
+    return os;
+}
+
+inline std::ostream& operator<<(std::ostream& os, const ProbeState v) {
+
+    os << (v == FAIL              ? "Failed" :
+           v == OK                ? "Success" :
+           v == CHANGE_STM        ? "Probed opponent side" :
+           v == ZEROING_BEST_MOVE ? "Best move zeroes DTZ" : "None");
+
+    return os;
+}
+
+} // namespace Stockfish::Tablebases
+
+#endif

src/Stockfish/src/thread.cpp

@@ -0,0 +1,268 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include <cassert>
+
+#include <algorithm> // For std::count
+#include "movegen.h"
+#include "search.h"
+#include "thread.h"
+#include "uci.h"
+#include "syzygy/tbprobe.h"
+#include "tt.h"
+
+namespace Stockfish {
+
+ThreadPool Threads; // Global object
+
+
+/// Thread constructor launches the thread and waits until it goes to sleep
+/// in idle_loop(). Note that 'searching' and 'exit' should be already set.
+
+Thread::Thread(size_t n) : idx(n), stdThread(&Thread::idle_loop, this) {
+
+  wait_for_search_finished();
+}
+
+
+/// Thread destructor wakes up the thread in idle_loop() and waits
+/// for its termination. Thread should be already waiting.
+
+Thread::~Thread() {
+
+  assert(!searching);
+
+  exit = true;
+  start_searching();
+  stdThread.join();
+}
+
+
+/// Thread::clear() reset histories, usually before a new game
+
+void Thread::clear() {
+
+  counterMoves.fill(MOVE_NONE);
+  mainHistory.fill(0);
+  captureHistory.fill(0);
+
+  for (bool inCheck : { false, true })
+      for (StatsType c : { NoCaptures, Captures })
+          for (auto& to : continuationHistory[inCheck][c])
+              for (auto& h : to)
+                  h->fill(-71);
+}
+
+
+/// Thread::start_searching() wakes up the thread that will start the search
+
+void Thread::start_searching() {
+  mutex.lock();
+  searching = true;
+  mutex.unlock(); // Unlock before notifying saves a few CPU-cycles
+  cv.notify_one(); // Wake up the thread in idle_loop()
+}
+
+
+/// Thread::wait_for_search_finished() blocks on the condition variable
+/// until the thread has finished searching.
+
+void Thread::wait_for_search_finished() {
+
+  std::unique_lock<std::mutex> lk(mutex);
+  cv.wait(lk, [&]{ return !searching; });
+}
+
+
+/// Thread::idle_loop() is where the thread is parked, blocked on the
+/// condition variable, when it has no work to do.
+
+void Thread::idle_loop() {
+
+  // If OS already scheduled us on a different group than 0 then don't overwrite
+  // the choice, eventually we are one of many one-threaded processes running on
+  // some Windows NUMA hardware, for instance in fishtest. To make it simple,
+  // just check if running threads are below a threshold, in this case all this
+  // NUMA machinery is not needed.
+  if (Options["Threads"] > 8)
+      WinProcGroup::bindThisThread(idx);
+
+  while (true)
+  {
+      std::unique_lock<std::mutex> lk(mutex);
+      searching = false;
+      cv.notify_one(); // Wake up anyone waiting for search finished
+      cv.wait(lk, [&]{ return searching; });
+
+      if (exit)
+          return;
+
+      lk.unlock();
+
+      search();
+  }
+}
+
+/// ThreadPool::set() creates/destroys threads to match the requested number.
+/// Created and launched threads will immediately go to sleep in idle_loop.
+/// Upon resizing, threads are recreated to allow for binding if necessary.
+
+void ThreadPool::set(size_t requested) {
+
+  if (threads.size() > 0)   // destroy any existing thread(s)
+  {
+      main()->wait_for_search_finished();
+
+      while (threads.size() > 0)
+          delete threads.back(), threads.pop_back();
+  }
+
+  if (requested > 0)   // create new thread(s)
+  {
+      threads.push_back(new MainThread(0));
+
+      while (threads.size() < requested)
+          threads.push_back(new Thread(threads.size()));
+      clear();
+
+      // Reallocate the hash with the new threadpool size
+      TT.resize(size_t(Options["Hash"]));
+
+      // Init thread number dependent search params.
+      Search::init();
+  }
+}
+
+
+/// ThreadPool::clear() sets threadPool data to initial values
+
+void ThreadPool::clear() {
+
+  for (Thread* th : threads)
+      th->clear();
+
+  main()->callsCnt = 0;
+  main()->bestPreviousScore = VALUE_INFINITE;
+  main()->bestPreviousAverageScore = VALUE_INFINITE;
+  main()->previousTimeReduction = 1.0;
+}
+
+
+/// ThreadPool::start_thinking() wakes up main thread waiting in idle_loop() and
+/// returns immediately. Main thread will wake up other threads and start the search.
+
+void ThreadPool::start_thinking(Position& pos, StateListPtr& states,
+                                const Search::LimitsType& limits, bool ponderMode) {
+
+  main()->wait_for_search_finished();
+
+  main()->stopOnPonderhit = stop = false;
+  increaseDepth = true;
+  main()->ponder = ponderMode;
+  Search::Limits = limits;
+  Search::RootMoves rootMoves;
+
+  for (const auto& m : MoveList<LEGAL>(pos))
+      if (   limits.searchmoves.empty()
+          || std::count(limits.searchmoves.begin(), limits.searchmoves.end(), m))
+          rootMoves.emplace_back(m);
+
+  if (!rootMoves.empty())
+      Tablebases::rank_root_moves(pos, rootMoves);
+
+  // After ownership transfer 'states' becomes empty, so if we stop the search
+  // and call 'go' again without setting a new position states.get() == nullptr.
+  assert(states.get() || setupStates.get());
+
+  if (states.get())
+      setupStates = std::move(states); // Ownership transfer, states is now empty
+
+  // We use Position::set() to set root position across threads. But there are
+  // some StateInfo fields (previous, pliesFromNull, capturedPiece) that cannot
+  // be deduced from a fen string, so set() clears them and they are set from
+  // setupStates->back() later. The rootState is per thread, earlier states are shared
+  // since they are read-only.
+  for (Thread* th : threads)
+  {
+      th->nodes = th->tbHits = th->nmpMinPly = th->bestMoveChanges = 0;
+      th->rootDepth = th->completedDepth = 0;
+      th->rootMoves = rootMoves;
+      th->rootPos.set(pos.fen(), pos.is_chess960(), &th->rootState, th);
+      th->rootState = setupStates->back();
+  }
+
+  main()->start_searching();
+}
+
+Thread* ThreadPool::get_best_thread() const {
+
+    Thread* bestThread = threads.front();
+    std::map<Move, int64_t> votes;
+    Value minScore = VALUE_NONE;
+
+    // Find minimum score of all threads
+    for (Thread* th: threads)
+        minScore = std::min(minScore, th->rootMoves[0].score);
+
+    // Vote according to score and depth, and select the best thread
+    auto thread_value = [minScore](Thread* th) {
+            return (th->rootMoves[0].score - minScore + 14) * int(th->completedDepth);
+        };
+
+    for (Thread* th : threads)
+        votes[th->rootMoves[0].pv[0]] += thread_value(th);
+
+    for (Thread* th : threads)
+        if (abs(bestThread->rootMoves[0].score) >= VALUE_TB_WIN_IN_MAX_PLY)
+        {
+            // Make sure we pick the shortest mate / TB conversion or stave off mate the longest
+            if (th->rootMoves[0].score > bestThread->rootMoves[0].score)
+                bestThread = th;
+        }
+        else if (   th->rootMoves[0].score >= VALUE_TB_WIN_IN_MAX_PLY
+                 || (   th->rootMoves[0].score > VALUE_TB_LOSS_IN_MAX_PLY
+                     && (   votes[th->rootMoves[0].pv[0]] > votes[bestThread->rootMoves[0].pv[0]]
+                         || (   votes[th->rootMoves[0].pv[0]] == votes[bestThread->rootMoves[0].pv[0]]
+                             &&   thread_value(th) * int(th->rootMoves[0].pv.size() > 2)
+                                > thread_value(bestThread) * int(bestThread->rootMoves[0].pv.size() > 2)))))
+            bestThread = th;
+
+    return bestThread;
+}
+
+
+/// Start non-main threads
+
+void ThreadPool::start_searching() {
+
+    for (Thread* th : threads)
+        if (th != threads.front())
+            th->start_searching();
+}
+
+
+/// Wait for non-main threads
+
+void ThreadPool::wait_for_search_finished() const {
+
+    for (Thread* th : threads)
+        if (th != threads.front())
+            th->wait_for_search_finished();
+}
+
+} // namespace Stockfish

src/Stockfish/src/thread.h

@@ -0,0 +1,141 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef THREAD_H_INCLUDED
+#define THREAD_H_INCLUDED
+
+#include <atomic>
+#include <condition_variable>
+#include <mutex>
+#include <thread>
+#include <vector>
+
+#include "material.h"
+#include "movepick.h"
+#include "pawns.h"
+#include "position.h"
+#include "search.h"
+#include "thread_win32_osx.h"
+
+namespace Stockfish {
+
+/// Thread class keeps together all the thread-related stuff. We use
+/// per-thread pawn and material hash tables so that once we get a
+/// pointer to an entry its life time is unlimited and we don't have
+/// to care about someone changing the entry under our feet.
+
+class Thread {
+
+  std::mutex mutex;
+  std::condition_variable cv;
+  size_t idx;
+  bool exit = false, searching = true; // Set before starting std::thread
+  NativeThread stdThread;
+
+public:
+  explicit Thread(size_t);
+  virtual ~Thread();
+  virtual void search();
+  void clear();
+  void idle_loop();
+  void start_searching();
+  void wait_for_search_finished();
+  size_t id() const { return idx; }
+
+  Pawns::Table pawnsTable;
+  Material::Table materialTable;
+  size_t pvIdx, pvLast;
+  std::atomic<uint64_t> nodes, tbHits, bestMoveChanges;
+  int selDepth, nmpMinPly;
+  Value bestValue, optimism[COLOR_NB];
+
+  Position rootPos;
+  StateInfo rootState;
+  Search::RootMoves rootMoves;
+  Depth rootDepth, completedDepth;
+  Value rootDelta;
+  CounterMoveHistory counterMoves;
+  ButterflyHistory mainHistory;
+  CapturePieceToHistory captureHistory;
+  ContinuationHistory continuationHistory[2][2];
+};
+
+
+/// MainThread is a derived class specific for main thread
+
+struct MainThread : public Thread {
+
+  using Thread::Thread;
+
+  void search() override;
+  void check_time();
+
+  double previousTimeReduction;
+  Value bestPreviousScore;
+  Value bestPreviousAverageScore;
+  Value iterValue[4];
+  int callsCnt;
+  bool stopOnPonderhit;
+  std::atomic_bool ponder;
+};
+
+
+/// ThreadPool struct handles all the threads-related stuff like init, starting,
+/// parking and, most importantly, launching a thread. All the access to threads
+/// is done through this class.
+
+struct ThreadPool {
+
+  void start_thinking(Position&, StateListPtr&, const Search::LimitsType&, bool = false);
+  void clear();
+  void set(size_t);
+
+  MainThread* main()        const { return static_cast<MainThread*>(threads.front()); }
+  uint64_t nodes_searched() const { return accumulate(&Thread::nodes); }
+  uint64_t tb_hits()        const { return accumulate(&Thread::tbHits); }
+  Thread* get_best_thread() const;
+  void start_searching();
+  void wait_for_search_finished() const;
+
+  std::atomic_bool stop, increaseDepth;
+
+  auto cbegin() const noexcept { return threads.cbegin(); }
+  auto begin() noexcept { return threads.begin(); }
+  auto end() noexcept { return threads.end(); }
+  auto cend() const noexcept { return threads.cend(); }
+  auto size() const noexcept { return threads.size(); }
+  auto empty() const noexcept { return threads.empty(); }
+
+private:
+  StateListPtr setupStates;
+  std::vector<Thread*> threads;
+
+  uint64_t accumulate(std::atomic<uint64_t> Thread::* member) const {
+
+    uint64_t sum = 0;
+    for (Thread* th : threads)
+        sum += (th->*member).load(std::memory_order_relaxed);
+    return sum;
+  }
+};
+
+extern ThreadPool Threads;
+
+} // namespace Stockfish
+
+#endif // #ifndef THREAD_H_INCLUDED

src/Stockfish/src/thread_win32_osx.h

@@ -0,0 +1,74 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef THREAD_WIN32_OSX_H_INCLUDED
+#define THREAD_WIN32_OSX_H_INCLUDED
+
+#include <thread>
+
+/// On OSX threads other than the main thread are created with a reduced stack
+/// size of 512KB by default, this is too low for deep searches, which require
+/// somewhat more than 1MB stack, so adjust it to TH_STACK_SIZE.
+/// The implementation calls pthread_create() with the stack size parameter
+/// equal to the linux 8MB default, on platforms that support it.
+
+#if defined(__APPLE__) || defined(__MINGW32__) || defined(__MINGW64__) || defined(USE_PTHREADS)
+
+#include <pthread.h>
+
+namespace Stockfish {
+
+static const size_t TH_STACK_SIZE = 8 * 1024 * 1024;
+
+template <class T, class P = std::pair<T*, void(T::*)()>>
+void* start_routine(void* ptr)
+{
+   P* p = reinterpret_cast<P*>(ptr);
+   (p->first->*(p->second))(); // Call member function pointer
+   delete p;
+   return nullptr;
+}
+
+class NativeThread {
+
+   pthread_t thread;
+
+public:
+  template<class T, class P = std::pair<T*, void(T::*)()>>
+  explicit NativeThread(void(T::*fun)(), T* obj) {
+    pthread_attr_t attr_storage, *attr = &attr_storage;
+    pthread_attr_init(attr);
+    pthread_attr_setstacksize(attr, TH_STACK_SIZE);
+    pthread_create(&thread, attr, start_routine<T>, new P(obj, fun));
+  }
+  void join() { pthread_join(thread, nullptr); }
+};
+
+} // namespace Stockfish
+
+#else // Default case: use STL classes
+
+namespace Stockfish {
+
+using NativeThread = std::thread;
+
+} // namespace Stockfish
+
+#endif
+
+#endif // #ifndef THREAD_WIN32_OSX_H_INCLUDED

src/Stockfish/src/timeman.cpp

@@ -0,0 +1,109 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include <algorithm>
+#include <cfloat>
+#include <cmath>
+
+#include "search.h"
+#include "timeman.h"
+#include "uci.h"
+
+namespace Stockfish {
+
+TimeManagement Time; // Our global time management object
+
+
+/// TimeManagement::init() is called at the beginning of the search and calculates
+/// the bounds of time allowed for the current game ply. We currently support:
+//      1) x basetime (+ z increment)
+//      2) x moves in y seconds (+ z increment)
+
+void TimeManagement::init(Search::LimitsType& limits, Color us, int ply) {
+
+  // if we have no time, no need to initialize TM, except for the start time,
+  // which is used by movetime.
+  startTime = limits.startTime;
+  if (limits.time[us] == 0)
+      return;
+
+  TimePoint moveOverhead    = TimePoint(Options["Move Overhead"]);
+  TimePoint slowMover       = TimePoint(Options["Slow Mover"]);
+  TimePoint npmsec          = TimePoint(Options["nodestime"]);
+
+  // optScale is a percentage of available time to use for the current move.
+  // maxScale is a multiplier applied to optimumTime.
+  double optScale, maxScale;
+
+  // If we have to play in 'nodes as time' mode, then convert from time
+  // to nodes, and use resulting values in time management formulas.
+  // WARNING: to avoid time losses, the given npmsec (nodes per millisecond)
+  // must be much lower than the real engine speed.
+  if (npmsec)
+  {
+      if (!availableNodes) // Only once at game start
+          availableNodes = npmsec * limits.time[us]; // Time is in msec
+
+      // Convert from milliseconds to nodes
+      limits.time[us] = TimePoint(availableNodes);
+      limits.inc[us] *= npmsec;
+      limits.npmsec = npmsec;
+  }
+
+  // Maximum move horizon of 50 moves
+  int mtg = limits.movestogo ? std::min(limits.movestogo, 50) : 50;
+
+  // Make sure timeLeft is > 0 since we may use it as a divisor
+  TimePoint timeLeft =  std::max(TimePoint(1),
+      limits.time[us] + limits.inc[us] * (mtg - 1) - moveOverhead * (2 + mtg));
+
+  // Use extra time with larger increments
+  double optExtra = std::clamp(1.0 + 12.0 * limits.inc[us] / limits.time[us], 1.0, 1.12);
+
+  // A user may scale time usage by setting UCI option "Slow Mover"
+  // Default is 100 and changing this value will probably lose elo.
+  timeLeft = slowMover * timeLeft / 100;
+
+  // x basetime (+ z increment)
+  // If there is a healthy increment, timeLeft can exceed actual available
+  // game time for the current move, so also cap to 20% of available game time.
+  if (limits.movestogo == 0)
+  {
+      optScale = std::min(0.0120 + std::pow(ply + 3.0, 0.45) * 0.0039,
+                           0.2 * limits.time[us] / double(timeLeft))
+                 * optExtra;
+      maxScale = std::min(7.0, 4.0 + ply / 12.0);
+  }
+
+  // x moves in y seconds (+ z increment)
+  else
+  {
+      optScale = std::min((0.88 + ply / 116.4) / mtg,
+                            0.88 * limits.time[us] / double(timeLeft));
+      maxScale = std::min(6.3, 1.5 + 0.11 * mtg);
+  }
+
+  // Never use more than 80% of the available time for this move
+  optimumTime = TimePoint(optScale * timeLeft);
+  maximumTime = TimePoint(std::min(0.8 * limits.time[us] - moveOverhead, maxScale * optimumTime));
+
+  if (Options["Ponder"])
+      optimumTime += optimumTime / 4;
+}
+
+} // namespace Stockfish

src/Stockfish/src/timeman.h

@@ -0,0 +1,51 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef TIMEMAN_H_INCLUDED
+#define TIMEMAN_H_INCLUDED
+
+#include "misc.h"
+#include "search.h"
+#include "thread.h"
+
+namespace Stockfish {
+
+/// The TimeManagement class computes the optimal time to think depending on
+/// the maximum available time, the game move number and other parameters.
+
+class TimeManagement {
+public:
+  void init(Search::LimitsType& limits, Color us, int ply);
+  TimePoint optimum() const { return optimumTime; }
+  TimePoint maximum() const { return maximumTime; }
+  TimePoint elapsed() const { return Search::Limits.npmsec ?
+                                     TimePoint(Threads.nodes_searched()) : now() - startTime; }
+
+  int64_t availableNodes; // When in 'nodes as time' mode
+
+private:
+  TimePoint startTime;
+  TimePoint optimumTime;
+  TimePoint maximumTime;
+};
+
+extern TimeManagement Time;
+
+} // namespace Stockfish
+
+#endif // #ifndef TIMEMAN_H_INCLUDED

src/Stockfish/src/tt.cpp

@@ -0,0 +1,162 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include <cstring>   // For std::memset
+#include <iostream>
+#include <thread>
+
+#include "bitboard.h"
+#include "misc.h"
+#include "thread.h"
+#include "tt.h"
+#include "uci.h"
+
+namespace Stockfish {
+
+TranspositionTable TT; // Our global transposition table
+
+/// TTEntry::save() populates the TTEntry with a new node's data, possibly
+/// overwriting an old position. Update is not atomic and can be racy.
+
+void TTEntry::save(Key k, Value v, bool pv, Bound b, Depth d, Move m, Value ev) {
+
+  // Preserve any existing move for the same position
+  if (m || (uint16_t)k != key16)
+      move16 = (uint16_t)m;
+
+  // Overwrite less valuable entries (cheapest checks first)
+  if (   b == BOUND_EXACT
+      || (uint16_t)k != key16
+      || d - DEPTH_OFFSET + 2 * pv > depth8 - 4)
+  {
+      assert(d > DEPTH_OFFSET);
+      assert(d < 256 + DEPTH_OFFSET);
+
+      key16     = (uint16_t)k;
+      depth8    = (uint8_t)(d - DEPTH_OFFSET);
+      genBound8 = (uint8_t)(TT.generation8 | uint8_t(pv) << 2 | b);
+      value16   = (int16_t)v;
+      eval16    = (int16_t)ev;
+  }
+}
+
+
+/// TranspositionTable::resize() sets the size of the transposition table,
+/// measured in megabytes. Transposition table consists of a power of 2 number
+/// of clusters and each cluster consists of ClusterSize number of TTEntry.
+
+void TranspositionTable::resize(size_t mbSize) {
+
+  Threads.main()->wait_for_search_finished();
+
+  aligned_large_pages_free(table);
+
+  clusterCount = mbSize * 1024 * 1024 / sizeof(Cluster);
+
+  table = static_cast<Cluster*>(aligned_large_pages_alloc(clusterCount * sizeof(Cluster)));
+  if (!table)
+  {
+      std::cerr << "Failed to allocate " << mbSize
+                << "MB for transposition table." << fakeendl;
+      exit(EXIT_FAILURE);
+  }
+
+  clear();
+}
+
+
+/// TranspositionTable::clear() initializes the entire transposition table to zero,
+//  in a multi-threaded way.
+
+void TranspositionTable::clear() {
+
+  std::vector<std::thread> threads;
+
+  for (size_t idx = 0; idx < size_t(Options["Threads"]); ++idx)
+  {
+      threads.emplace_back([this, idx]() {
+
+          // Thread binding gives faster search on systems with a first-touch policy
+          if (Options["Threads"] > 8)
+              WinProcGroup::bindThisThread(idx);
+
+          // Each thread will zero its part of the hash table
+          const size_t stride = size_t(clusterCount / Options["Threads"]),
+                       start  = size_t(stride * idx),
+                       len    = idx != size_t(Options["Threads"]) - 1 ?
+                                stride : clusterCount - start;
+
+          std::memset(&table[start], 0, len * sizeof(Cluster));
+      });
+  }
+
+  for (std::thread& th : threads)
+      th.join();
+}
+
+
+/// TranspositionTable::probe() looks up the current position in the transposition
+/// table. It returns true and a pointer to the TTEntry if the position is found.
+/// Otherwise, it returns false and a pointer to an empty or least valuable TTEntry
+/// to be replaced later. The replace value of an entry is calculated as its depth
+/// minus 8 times its relative age. TTEntry t1 is considered more valuable than
+/// TTEntry t2 if its replace value is greater than that of t2.
+
+TTEntry* TranspositionTable::probe(const Key key, bool& found) const {
+
+  TTEntry* const tte = first_entry(key);
+  const uint16_t key16 = (uint16_t)key;  // Use the low 16 bits as key inside the cluster
+
+  for (int i = 0; i < ClusterSize; ++i)
+      if (tte[i].key16 == key16 || !tte[i].depth8)
+      {
+          tte[i].genBound8 = uint8_t(generation8 | (tte[i].genBound8 & (GENERATION_DELTA - 1))); // Refresh
+
+          return found = (bool)tte[i].depth8, &tte[i];
+      }
+
+  // Find an entry to be replaced according to the replacement strategy
+  TTEntry* replace = tte;
+  for (int i = 1; i < ClusterSize; ++i)
+      // Due to our packed storage format for generation and its cyclic
+      // nature we add GENERATION_CYCLE (256 is the modulus, plus what
+      // is needed to keep the unrelated lowest n bits from affecting
+      // the result) to calculate the entry age correctly even after
+      // generation8 overflows into the next cycle.
+      if (  replace->depth8 - ((GENERATION_CYCLE + generation8 - replace->genBound8) & GENERATION_MASK)
+          >   tte[i].depth8 - ((GENERATION_CYCLE + generation8 -   tte[i].genBound8) & GENERATION_MASK))
+          replace = &tte[i];
+
+  return found = false, replace;
+}
+
+
+/// TranspositionTable::hashfull() returns an approximation of the hashtable
+/// occupation during a search. The hash is x permill full, as per UCI protocol.
+
+int TranspositionTable::hashfull() const {
+
+  int cnt = 0;
+  for (int i = 0; i < 1000; ++i)
+      for (int j = 0; j < ClusterSize; ++j)
+          cnt += table[i].entry[j].depth8 && (table[i].entry[j].genBound8 & GENERATION_MASK) == generation8;
+
+  return cnt / ClusterSize;
+}
+
+} // namespace Stockfish

src/Stockfish/src/tt.h

@@ -0,0 +1,107 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef TT_H_INCLUDED
+#define TT_H_INCLUDED
+
+#include "misc.h"
+#include "types.h"
+
+namespace Stockfish {
+
+/// TTEntry struct is the 10 bytes transposition table entry, defined as below:
+///
+/// key        16 bit
+/// depth       8 bit
+/// generation  5 bit
+/// pv node     1 bit
+/// bound type  2 bit
+/// move       16 bit
+/// value      16 bit
+/// eval value 16 bit
+
+struct TTEntry {
+
+  Move  move()  const { return (Move )move16; }
+  Value value() const { return (Value)value16; }
+  Value eval()  const { return (Value)eval16; }
+  Depth depth() const { return (Depth)depth8 + DEPTH_OFFSET; }
+  bool is_pv()  const { return (bool)(genBound8 & 0x4); }
+  Bound bound() const { return (Bound)(genBound8 & 0x3); }
+  void save(Key k, Value v, bool pv, Bound b, Depth d, Move m, Value ev);
+
+private:
+  friend class TranspositionTable;
+
+  uint16_t key16;
+  uint8_t  depth8;
+  uint8_t  genBound8;
+  uint16_t move16;
+  int16_t  value16;
+  int16_t  eval16;
+};
+
+
+/// A TranspositionTable is an array of Cluster, of size clusterCount. Each
+/// cluster consists of ClusterSize number of TTEntry. Each non-empty TTEntry
+/// contains information on exactly one position. The size of a Cluster should
+/// divide the size of a cache line for best performance, as the cacheline is
+/// prefetched when possible.
+
+class TranspositionTable {
+
+  static constexpr int ClusterSize = 3;
+
+  struct Cluster {
+    TTEntry entry[ClusterSize];
+    char padding[2]; // Pad to 32 bytes
+  };
+
+  static_assert(sizeof(Cluster) == 32, "Unexpected Cluster size");
+
+  // Constants used to refresh the hash table periodically
+  static constexpr unsigned GENERATION_BITS  = 3;                                // nb of bits reserved for other things
+  static constexpr int      GENERATION_DELTA = (1 << GENERATION_BITS);           // increment for generation field
+  static constexpr int      GENERATION_CYCLE = 255 + (1 << GENERATION_BITS);     // cycle length
+  static constexpr int      GENERATION_MASK  = (0xFF << GENERATION_BITS) & 0xFF; // mask to pull out generation number
+
+public:
+ ~TranspositionTable() { aligned_large_pages_free(table); }
+  void new_search() { generation8 += GENERATION_DELTA; } // Lower bits are used for other things
+  TTEntry* probe(const Key key, bool& found) const;
+  int hashfull() const;
+  void resize(size_t mbSize);
+  void clear();
+
+  TTEntry* first_entry(const Key key) const {
+    return &table[mul_hi64(key, clusterCount)].entry[0];
+  }
+
+private:
+  friend struct TTEntry;
+
+  size_t clusterCount;
+  Cluster* table;
+  uint8_t generation8; // Size must be not bigger than TTEntry::genBound8
+};
+
+extern TranspositionTable TT;
+
+} // namespace Stockfish
+
+#endif // #ifndef TT_H_INCLUDED

src/Stockfish/src/tune.cpp

@@ -0,0 +1,133 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include <algorithm>
+#include <iostream>
+#include <sstream>
+
+#include "types.h"
+#include "misc.h"
+#include "uci.h"
+
+using std::string;
+
+namespace Stockfish {
+
+bool Tune::update_on_last;
+const UCI::Option* LastOption = nullptr;
+static std::map<std::string, int> TuneResults;
+
+string Tune::next(string& names, bool pop) {
+
+  string name;
+
+  do {
+      string token = names.substr(0, names.find(','));
+
+      if (pop)
+          names.erase(0, token.size() + 1);
+
+      std::stringstream ws(token);
+      name += (ws >> token, token); // Remove trailing whitespace
+
+  } while (  std::count(name.begin(), name.end(), '(')
+           - std::count(name.begin(), name.end(), ')'));
+
+  return name;
+}
+
+static void on_tune(const UCI::Option& o) {
+
+  if (!Tune::update_on_last || LastOption == &o)
+      Tune::read_options();
+}
+
+static void make_option(const string& n, int v, const SetRange& r) {
+
+  // Do not generate option when there is nothing to tune (ie. min = max)
+  if (r(v).first == r(v).second)
+      return;
+
+  if (TuneResults.count(n))
+      v = TuneResults[n];
+
+  Options[n] << UCI::Option(v, r(v).first, r(v).second, on_tune);
+  LastOption = &Options[n];
+
+  // Print formatted parameters, ready to be copy-pasted in Fishtest
+  fakeout << n << ","
+            << v << ","
+            << r(v).first << "," << r(v).second << ","
+            << (r(v).second - r(v).first) / 20.0 << ","
+            << "0.0020"
+            << fakeendl;
+}
+
+template<> void Tune::Entry<int>::init_option() { make_option(name, value, range); }
+
+template<> void Tune::Entry<int>::read_option() {
+  if (Options.count(name))
+      value = int(Options[name]);
+}
+
+template<> void Tune::Entry<Value>::init_option() { make_option(name, value, range); }
+
+template<> void Tune::Entry<Value>::read_option() {
+  if (Options.count(name))
+      value = Value(int(Options[name]));
+}
+
+template<> void Tune::Entry<Score>::init_option() {
+  make_option("m" + name, mg_value(value), range);
+  make_option("e" + name, eg_value(value), range);
+}
+
+template<> void Tune::Entry<Score>::read_option() {
+  if (Options.count("m" + name))
+      value = make_score(int(Options["m" + name]), eg_value(value));
+
+  if (Options.count("e" + name))
+      value = make_score(mg_value(value), int(Options["e" + name]));
+}
+
+// Instead of a variable here we have a PostUpdate function: just call it
+template<> void Tune::Entry<Tune::PostUpdate>::init_option() {}
+template<> void Tune::Entry<Tune::PostUpdate>::read_option() { value(); }
+
+} // namespace Stockfish
+
+
+// Init options with tuning session results instead of default values. Useful to
+// get correct bench signature after a tuning session or to test tuned values.
+// Just copy fishtest tuning results in a result.txt file and extract the
+// values with:
+//
+// cat results.txt | sed 's/^param: \([^,]*\), best: \([^,]*\).*/  TuneResults["\1"] = int(round(\2));/'
+//
+// Then paste the output below, as the function body
+
+#include <cmath>
+
+namespace Stockfish {
+
+void Tune::read_results() {
+
+  /* ...insert your values here... */
+}
+
+} // namespace Stockfish

src/Stockfish/src/tune.h

@@ -0,0 +1,163 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef TUNE_H_INCLUDED
+#define TUNE_H_INCLUDED
+
+#include <memory>
+#include <string>
+#include <type_traits>
+#include <vector>
+
+namespace Stockfish {
+
+using Range = std::pair<int, int>; // Option's min-max values
+using RangeFun = Range (int);
+
+// Default Range function, to calculate Option's min-max values
+inline Range default_range(int v) {
+  return v > 0 ? Range(0, 2 * v) : Range(2 * v, 0);
+}
+
+struct SetRange {
+  explicit SetRange(RangeFun f) : fun(f) {}
+  SetRange(int min, int max) : fun(nullptr), range(min, max) {}
+  Range operator()(int v) const { return fun ? fun(v) : range; }
+
+  RangeFun* fun;
+  Range range;
+};
+
+#define SetDefaultRange SetRange(default_range)
+
+
+/// Tune class implements the 'magic' code that makes the setup of a fishtest
+/// tuning session as easy as it can be. Mainly you have just to remove const
+/// qualifiers from the variables you want to tune and flag them for tuning, so
+/// if you have:
+///
+///   const Score myScore = S(10, 15);
+///   const Value myValue[][2] = { { V(100), V(20) }, { V(7), V(78) } };
+///
+/// If you have a my_post_update() function to run after values have been updated,
+/// and a my_range() function to set custom Option's min-max values, then you just
+/// remove the 'const' qualifiers and write somewhere below in the file:
+///
+///   TUNE(SetRange(my_range), myScore, myValue, my_post_update);
+///
+/// You can also set the range directly, and restore the default at the end
+///
+///   TUNE(SetRange(-100, 100), myScore, SetDefaultRange);
+///
+/// In case update function is slow and you have many parameters, you can add:
+///
+///   UPDATE_ON_LAST();
+///
+/// And the values update, including post update function call, will be done only
+/// once, after the engine receives the last UCI option, that is the one defined
+/// and created as the last one, so the GUI should send the options in the same
+/// order in which have been defined.
+
+class Tune {
+
+  using PostUpdate = void (); // Post-update function
+
+  Tune() { read_results(); }
+  Tune(const Tune&) = delete;
+  void operator=(const Tune&) = delete;
+  void read_results();
+
+  static Tune& instance() { static Tune t; return t; } // Singleton
+
+  // Use polymorphism to accommodate Entry of different types in the same vector
+  struct EntryBase {
+    virtual ~EntryBase() = default;
+    virtual void init_option() = 0;
+    virtual void read_option() = 0;
+  };
+
+  template<typename T>
+  struct Entry : public EntryBase {
+
+    static_assert(!std::is_const<T>::value, "Parameter cannot be const!");
+
+    static_assert(   std::is_same<T,   int>::value
+                  || std::is_same<T, Value>::value
+                  || std::is_same<T, Score>::value
+                  || std::is_same<T, PostUpdate>::value, "Parameter type not supported!");
+
+    Entry(const std::string& n, T& v, const SetRange& r) : name(n), value(v), range(r) {}
+    void operator=(const Entry&) = delete; // Because 'value' is a reference
+    void init_option() override;
+    void read_option() override;
+
+    std::string name;
+    T& value;
+    SetRange range;
+  };
+
+  // Our facility to fill the container, each Entry corresponds to a parameter
+  // to tune. We use variadic templates to deal with an unspecified number of
+  // entries, each one of a possible different type.
+  static std::string next(std::string& names, bool pop = true);
+
+  int add(const SetRange&, std::string&&) { return 0; }
+
+  template<typename T, typename... Args>
+  int add(const SetRange& range, std::string&& names, T& value, Args&&... args) {
+    list.push_back(std::unique_ptr<EntryBase>(new Entry<T>(next(names), value, range)));
+    return add(range, std::move(names), args...);
+  }
+
+  // Template specialization for arrays: recursively handle multi-dimensional arrays
+  template<typename T, size_t N, typename... Args>
+  int add(const SetRange& range, std::string&& names, T (&value)[N], Args&&... args) {
+    for (size_t i = 0; i < N; i++)
+        add(range, next(names, i == N - 1) + "[" + std::to_string(i) + "]", value[i]);
+    return add(range, std::move(names), args...);
+  }
+
+  // Template specialization for SetRange
+  template<typename... Args>
+  int add(const SetRange&, std::string&& names, SetRange& value, Args&&... args) {
+    return add(value, (next(names), std::move(names)), args...);
+  }
+
+  std::vector<std::unique_ptr<EntryBase>> list;
+
+public:
+  template<typename... Args>
+  static int add(const std::string& names, Args&&... args) {
+    return instance().add(SetDefaultRange, names.substr(1, names.size() - 2), args...); // Remove trailing parenthesis
+  }
+  static void init() { for (auto& e : instance().list) e->init_option(); read_options(); } // Deferred, due to UCI::Options access
+  static void read_options() { for (auto& e : instance().list) e->read_option(); }
+  static bool update_on_last;
+};
+
+// Some macro magic :-) we define a dummy int variable that compiler initializes calling Tune::add()
+#define STRINGIFY(x) #x
+#define UNIQUE2(x, y) x ## y
+#define UNIQUE(x, y) UNIQUE2(x, y) // Two indirection levels to expand __LINE__
+#define TUNE(...) int UNIQUE(p, __LINE__) = Tune::add(STRINGIFY((__VA_ARGS__)), __VA_ARGS__)
+
+#define UPDATE_ON_LAST() bool UNIQUE(p, __LINE__) = Tune::update_on_last = true
+
+} // namespace Stockfish
+
+#endif // #ifndef TUNE_H_INCLUDED

src/Stockfish/src/types.h

@@ -0,0 +1,491 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef TYPES_H_INCLUDED
+#define TYPES_H_INCLUDED
+
+/// When compiling with provided Makefile (e.g. for Linux and OSX), configuration
+/// is done automatically. To get started type 'make help'.
+///
+/// When Makefile is not used (e.g. with Microsoft Visual Studio) some switches
+/// need to be set manually:
+///
+/// -DNDEBUG      | Disable debugging mode. Always use this for release.
+///
+/// -DNO_PREFETCH | Disable use of prefetch asm-instruction. You may need this to
+///               | run on some very old machines.
+///
+/// -DUSE_POPCNT  | Add runtime support for use of popcnt asm-instruction. Works
+///               | only in 64-bit mode and requires hardware with popcnt support.
+///
+/// -DUSE_PEXT    | Add runtime support for use of pext asm-instruction. Works
+///               | only in 64-bit mode and requires hardware with pext support.
+
+#include <cassert>
+#include <cctype>
+#include <cstdint>
+#include <cstdlib>
+#include <algorithm>
+
+#if defined(_MSC_VER)
+// Disable some silly and noisy warning from MSVC compiler
+#pragma warning(disable: 4127) // Conditional expression is constant
+#pragma warning(disable: 4146) // Unary minus operator applied to unsigned type
+#pragma warning(disable: 4800) // Forcing value to bool 'true' or 'false'
+#endif
+
+/// Predefined macros hell:
+///
+/// __GNUC__           Compiler is gcc, Clang or Intel on Linux
+/// __INTEL_COMPILER   Compiler is Intel
+/// _MSC_VER           Compiler is MSVC or Intel on Windows
+/// _WIN32             Building on Windows (any)
+/// _WIN64             Building on Windows 64 bit
+
+#if defined(__GNUC__ ) && (__GNUC__ < 9 || (__GNUC__ == 9 && __GNUC_MINOR__ <= 2)) && defined(_WIN32) && !defined(__clang__)
+#define ALIGNAS_ON_STACK_VARIABLES_BROKEN
+#endif
+
+#define ASSERT_ALIGNED(ptr, alignment) assert(reinterpret_cast<uintptr_t>(ptr) % alignment == 0)
+
+#if defined(_WIN64) && defined(_MSC_VER) // No Makefile used
+#  include <intrin.h> // Microsoft header for _BitScanForward64()
+#  define IS_64BIT
+#endif
+
+#if defined(USE_POPCNT) && (defined(__INTEL_COMPILER) || defined(_MSC_VER))
+#  include <nmmintrin.h> // Intel and Microsoft header for _mm_popcnt_u64()
+#endif
+
+#if !defined(NO_PREFETCH) && (defined(__INTEL_COMPILER) || defined(_MSC_VER))
+#  include <xmmintrin.h> // Intel and Microsoft header for _mm_prefetch()
+#endif
+
+#if defined(USE_PEXT)
+#  include <immintrin.h> // Header for _pext_u64() intrinsic
+#  define pext(b, m) _pext_u64(b, m)
+#else
+#  define pext(b, m) 0
+#endif
+
+namespace Stockfish {
+
+#ifdef USE_POPCNT
+constexpr bool HasPopCnt = true;
+#else
+constexpr bool HasPopCnt = false;
+#endif
+
+#ifdef USE_PEXT
+constexpr bool HasPext = true;
+#else
+constexpr bool HasPext = false;
+#endif
+
+#ifdef IS_64BIT
+constexpr bool Is64Bit = true;
+#else
+constexpr bool Is64Bit = false;
+#endif
+
+using Key = uint64_t;
+using Bitboard = uint64_t;
+
+constexpr int MAX_MOVES = 256;
+constexpr int MAX_PLY   = 246;
+
+/// A move needs 16 bits to be stored
+///
+/// bit  0- 5: destination square (from 0 to 63)
+/// bit  6-11: origin square (from 0 to 63)
+/// bit 12-13: promotion piece type - 2 (from KNIGHT-2 to QUEEN-2)
+/// bit 14-15: special move flag: promotion (1), en passant (2), castling (3)
+/// NOTE: en passant bit is set only when a pawn can be captured
+///
+/// Special cases are MOVE_NONE and MOVE_NULL. We can sneak these in because in
+/// any normal move destination square is always different from origin square
+/// while MOVE_NONE and MOVE_NULL have the same origin and destination square.
+
+enum Move : int {
+  MOVE_NONE,
+  MOVE_NULL = 65
+};
+
+enum MoveType {
+  NORMAL,
+  PROMOTION = 1 << 14,
+  EN_PASSANT = 2 << 14,
+  CASTLING  = 3 << 14
+};
+
+enum Color {
+  WHITE, BLACK, COLOR_NB = 2
+};
+
+enum CastlingRights {
+  NO_CASTLING,
+  WHITE_OO,
+  WHITE_OOO = WHITE_OO << 1,
+  BLACK_OO  = WHITE_OO << 2,
+  BLACK_OOO = WHITE_OO << 3,
+
+  KING_SIDE      = WHITE_OO  | BLACK_OO,
+  QUEEN_SIDE     = WHITE_OOO | BLACK_OOO,
+  WHITE_CASTLING = WHITE_OO  | WHITE_OOO,
+  BLACK_CASTLING = BLACK_OO  | BLACK_OOO,
+  ANY_CASTLING   = WHITE_CASTLING | BLACK_CASTLING,
+
+  CASTLING_RIGHT_NB = 16
+};
+
+enum Phase {
+  PHASE_ENDGAME,
+  PHASE_MIDGAME = 128,
+  MG = 0, EG = 1, PHASE_NB = 2
+};
+
+enum ScaleFactor {
+  SCALE_FACTOR_DRAW    = 0,
+  SCALE_FACTOR_NORMAL  = 64,
+  SCALE_FACTOR_MAX     = 128,
+  SCALE_FACTOR_NONE    = 255
+};
+
+enum Bound {
+  BOUND_NONE,
+  BOUND_UPPER,
+  BOUND_LOWER,
+  BOUND_EXACT = BOUND_UPPER | BOUND_LOWER
+};
+
+enum Value : int {
+  VALUE_ZERO      = 0,
+  VALUE_DRAW      = 0,
+  VALUE_KNOWN_WIN = 10000,
+  VALUE_MATE      = 32000,
+  VALUE_INFINITE  = 32001,
+  VALUE_NONE      = 32002,
+
+  VALUE_TB_WIN_IN_MAX_PLY  =  VALUE_MATE - 2 * MAX_PLY,
+  VALUE_TB_LOSS_IN_MAX_PLY = -VALUE_TB_WIN_IN_MAX_PLY,
+  VALUE_MATE_IN_MAX_PLY  =  VALUE_MATE - MAX_PLY,
+  VALUE_MATED_IN_MAX_PLY = -VALUE_MATE_IN_MAX_PLY,
+
+  // In the code, we make the assumption that these values
+  // are such that non_pawn_material() can be used to uniquely
+  // identify the material on the board.
+  PawnValueMg   = 126,   PawnValueEg   = 208,
+  KnightValueMg = 781,   KnightValueEg = 854,
+  BishopValueMg = 825,   BishopValueEg = 915,
+  RookValueMg   = 1276,  RookValueEg   = 1380,
+  QueenValueMg  = 2538,  QueenValueEg  = 2682,
+
+  MidgameLimit  = 15258, EndgameLimit  = 3915
+};
+
+enum PieceType {
+  NO_PIECE_TYPE, PAWN, KNIGHT, BISHOP, ROOK, QUEEN, KING,
+  ALL_PIECES = 0,
+  PIECE_TYPE_NB = 8
+};
+
+enum Piece {
+  NO_PIECE,
+  W_PAWN = PAWN,     W_KNIGHT, W_BISHOP, W_ROOK, W_QUEEN, W_KING,
+  B_PAWN = PAWN + 8, B_KNIGHT, B_BISHOP, B_ROOK, B_QUEEN, B_KING,
+  PIECE_NB = 16
+};
+
+constexpr Value PieceValue[PHASE_NB][PIECE_NB] = {
+  { VALUE_ZERO, PawnValueMg, KnightValueMg, BishopValueMg, RookValueMg, QueenValueMg, VALUE_ZERO, VALUE_ZERO,
+    VALUE_ZERO, PawnValueMg, KnightValueMg, BishopValueMg, RookValueMg, QueenValueMg, VALUE_ZERO, VALUE_ZERO },
+  { VALUE_ZERO, PawnValueEg, KnightValueEg, BishopValueEg, RookValueEg, QueenValueEg, VALUE_ZERO, VALUE_ZERO,
+    VALUE_ZERO, PawnValueEg, KnightValueEg, BishopValueEg, RookValueEg, QueenValueEg, VALUE_ZERO, VALUE_ZERO }
+};
+
+using Depth = int;
+
+enum : int {
+  DEPTH_QS_CHECKS     =  0,
+  DEPTH_QS_NO_CHECKS  = -1,
+  DEPTH_QS_RECAPTURES = -5,
+
+  DEPTH_NONE   = -6,
+
+  DEPTH_OFFSET = -7 // value used only for TT entry occupancy check
+};
+
+enum Square : int {
+  SQ_A1, SQ_B1, SQ_C1, SQ_D1, SQ_E1, SQ_F1, SQ_G1, SQ_H1,
+  SQ_A2, SQ_B2, SQ_C2, SQ_D2, SQ_E2, SQ_F2, SQ_G2, SQ_H2,
+  SQ_A3, SQ_B3, SQ_C3, SQ_D3, SQ_E3, SQ_F3, SQ_G3, SQ_H3,
+  SQ_A4, SQ_B4, SQ_C4, SQ_D4, SQ_E4, SQ_F4, SQ_G4, SQ_H4,
+  SQ_A5, SQ_B5, SQ_C5, SQ_D5, SQ_E5, SQ_F5, SQ_G5, SQ_H5,
+  SQ_A6, SQ_B6, SQ_C6, SQ_D6, SQ_E6, SQ_F6, SQ_G6, SQ_H6,
+  SQ_A7, SQ_B7, SQ_C7, SQ_D7, SQ_E7, SQ_F7, SQ_G7, SQ_H7,
+  SQ_A8, SQ_B8, SQ_C8, SQ_D8, SQ_E8, SQ_F8, SQ_G8, SQ_H8,
+  SQ_NONE,
+
+  SQUARE_ZERO = 0,
+  SQUARE_NB   = 64
+};
+
+enum Direction : int {
+  NORTH =  8,
+  EAST  =  1,
+  SOUTH = -NORTH,
+  WEST  = -EAST,
+
+  NORTH_EAST = NORTH + EAST,
+  SOUTH_EAST = SOUTH + EAST,
+  SOUTH_WEST = SOUTH + WEST,
+  NORTH_WEST = NORTH + WEST
+};
+
+enum File : int {
+  FILE_A, FILE_B, FILE_C, FILE_D, FILE_E, FILE_F, FILE_G, FILE_H, FILE_NB
+};
+
+enum Rank : int {
+  RANK_1, RANK_2, RANK_3, RANK_4, RANK_5, RANK_6, RANK_7, RANK_8, RANK_NB
+};
+
+// Keep track of what a move changes on the board (used by NNUE)
+struct DirtyPiece {
+
+  // Number of changed pieces
+  int dirty_num;
+
+  // Max 3 pieces can change in one move. A promotion with capture moves
+  // both the pawn and the captured piece to SQ_NONE and the piece promoted
+  // to from SQ_NONE to the capture square.
+  Piece piece[3];
+
+  // From and to squares, which may be SQ_NONE
+  Square from[3];
+  Square to[3];
+};
+
+/// Score enum stores a middlegame and an endgame value in a single integer (enum).
+/// The least significant 16 bits are used to store the middlegame value and the
+/// upper 16 bits are used to store the endgame value. We have to take care to
+/// avoid left-shifting a signed int to avoid undefined behavior.
+enum Score : int { SCORE_ZERO };
+
+constexpr Score make_score(int mg, int eg) {
+  return Score((int)((unsigned int)eg << 16) + mg);
+}
+
+/// Extracting the signed lower and upper 16 bits is not so trivial because
+/// according to the standard a simple cast to short is implementation defined
+/// and so is a right shift of a signed integer.
+inline Value eg_value(Score s) {
+  union { uint16_t u; int16_t s; } eg = { uint16_t(unsigned(s + 0x8000) >> 16) };
+  return Value(eg.s);
+}
+
+inline Value mg_value(Score s) {
+  union { uint16_t u; int16_t s; } mg = { uint16_t(unsigned(s)) };
+  return Value(mg.s);
+}
+
+#define ENABLE_BASE_OPERATORS_ON(T)                                \
+constexpr T operator+(T d1, int d2) { return T(int(d1) + d2); }    \
+constexpr T operator-(T d1, int d2) { return T(int(d1) - d2); }    \
+constexpr T operator-(T d) { return T(-int(d)); }                  \
+inline T& operator+=(T& d1, int d2) { return d1 = d1 + d2; }       \
+inline T& operator-=(T& d1, int d2) { return d1 = d1 - d2; }
+
+#define ENABLE_INCR_OPERATORS_ON(T)                                \
+inline T& operator++(T& d) { return d = T(int(d) + 1); }           \
+inline T& operator--(T& d) { return d = T(int(d) - 1); }
+
+#define ENABLE_FULL_OPERATORS_ON(T)                                \
+ENABLE_BASE_OPERATORS_ON(T)                                        \
+constexpr T operator*(int i, T d) { return T(i * int(d)); }        \
+constexpr T operator*(T d, int i) { return T(int(d) * i); }        \
+constexpr T operator/(T d, int i) { return T(int(d) / i); }        \
+constexpr int operator/(T d1, T d2) { return int(d1) / int(d2); }  \
+inline T& operator*=(T& d, int i) { return d = T(int(d) * i); }    \
+inline T& operator/=(T& d, int i) { return d = T(int(d) / i); }
+
+ENABLE_FULL_OPERATORS_ON(Value)
+ENABLE_FULL_OPERATORS_ON(Direction)
+
+ENABLE_INCR_OPERATORS_ON(Piece)
+ENABLE_INCR_OPERATORS_ON(PieceType)
+ENABLE_INCR_OPERATORS_ON(Square)
+ENABLE_INCR_OPERATORS_ON(File)
+ENABLE_INCR_OPERATORS_ON(Rank)
+
+ENABLE_BASE_OPERATORS_ON(Score)
+
+#undef ENABLE_FULL_OPERATORS_ON
+#undef ENABLE_INCR_OPERATORS_ON
+#undef ENABLE_BASE_OPERATORS_ON
+
+/// Additional operators to add a Direction to a Square
+constexpr Square operator+(Square s, Direction d) { return Square(int(s) + int(d)); }
+constexpr Square operator-(Square s, Direction d) { return Square(int(s) - int(d)); }
+inline Square& operator+=(Square& s, Direction d) { return s = s + d; }
+inline Square& operator-=(Square& s, Direction d) { return s = s - d; }
+
+/// Only declared but not defined. We don't want to multiply two scores due to
+/// a very high risk of overflow. So user should explicitly convert to integer.
+Score operator*(Score, Score) = delete;
+
+/// Division of a Score must be handled separately for each term
+inline Score operator/(Score s, int i) {
+  return make_score(mg_value(s) / i, eg_value(s) / i);
+}
+
+/// Multiplication of a Score by an integer. We check for overflow in debug mode.
+inline Score operator*(Score s, int i) {
+
+  Score result = Score(int(s) * i);
+
+  assert(eg_value(result) == (i * eg_value(s)));
+  assert(mg_value(result) == (i * mg_value(s)));
+  assert((i == 0) || (result / i) == s);
+
+  return result;
+}
+
+/// Multiplication of a Score by a boolean
+inline Score operator*(Score s, bool b) {
+  return b ? s : SCORE_ZERO;
+}
+
+constexpr Color operator~(Color c) {
+  return Color(c ^ BLACK); // Toggle color
+}
+
+constexpr Square flip_rank(Square s) { // Swap A1 <-> A8
+  return Square(s ^ SQ_A8);
+}
+
+constexpr Square flip_file(Square s) { // Swap A1 <-> H1
+  return Square(s ^ SQ_H1);
+}
+
+constexpr Piece operator~(Piece pc) {
+  return Piece(pc ^ 8); // Swap color of piece B_KNIGHT <-> W_KNIGHT
+}
+
+constexpr CastlingRights operator&(Color c, CastlingRights cr) {
+  return CastlingRights((c == WHITE ? WHITE_CASTLING : BLACK_CASTLING) & cr);
+}
+
+constexpr Value mate_in(int ply) {
+  return VALUE_MATE - ply;
+}
+
+constexpr Value mated_in(int ply) {
+  return -VALUE_MATE + ply;
+}
+
+constexpr Square make_square(File f, Rank r) {
+  return Square((r << 3) + f);
+}
+
+constexpr Piece make_piece(Color c, PieceType pt) {
+  return Piece((c << 3) + pt);
+}
+
+constexpr PieceType type_of(Piece pc) {
+  return PieceType(pc & 7);
+}
+
+inline Color color_of(Piece pc) {
+  assert(pc != NO_PIECE);
+  return Color(pc >> 3);
+}
+
+constexpr bool is_ok(Move m) {
+  return m != MOVE_NONE && m != MOVE_NULL;
+}
+
+constexpr bool is_ok(Square s) {
+  return s >= SQ_A1 && s <= SQ_H8;
+}
+
+constexpr File file_of(Square s) {
+  return File(s & 7);
+}
+
+constexpr Rank rank_of(Square s) {
+  return Rank(s >> 3);
+}
+
+constexpr Square relative_square(Color c, Square s) {
+  return Square(s ^ (c * 56));
+}
+
+constexpr Rank relative_rank(Color c, Rank r) {
+  return Rank(r ^ (c * 7));
+}
+
+constexpr Rank relative_rank(Color c, Square s) {
+  return relative_rank(c, rank_of(s));
+}
+
+constexpr Direction pawn_push(Color c) {
+  return c == WHITE ? NORTH : SOUTH;
+}
+
+constexpr Square from_sq(Move m) {
+  assert(is_ok(m));
+  return Square((m >> 6) & 0x3F);
+}
+
+constexpr Square to_sq(Move m) {
+  assert(is_ok(m));
+  return Square(m & 0x3F);
+}
+
+constexpr int from_to(Move m) {
+  return m & 0xFFF;
+}
+
+constexpr MoveType type_of(Move m) {
+  return MoveType(m & (3 << 14));
+}
+
+constexpr PieceType promotion_type(Move m) {
+  return PieceType(((m >> 12) & 3) + KNIGHT);
+}
+
+constexpr Move make_move(Square from, Square to) {
+  return Move((from << 6) + to);
+}
+
+template<MoveType T>
+constexpr Move make(Square from, Square to, PieceType pt = KNIGHT) {
+  return Move(T + ((pt - KNIGHT) << 12) + (from << 6) + to);
+}
+
+/// Based on a congruential pseudo random number generator
+constexpr Key make_key(uint64_t seed) {
+  return seed * 6364136223846793005ULL + 1442695040888963407ULL;
+}
+
+} // namespace Stockfish
+
+#endif // #ifndef TYPES_H_INCLUDED
+
+#include "tune.h" // Global visibility to tuning setup

src/Stockfish/src/uci.cpp

@@ -0,0 +1,398 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include <cassert>
+#include <cmath>
+#include <iostream>
+#include <sstream>
+#include <string>
+
+#include "benchmark.h"
+#include "evaluate.h"
+#include "movegen.h"
+#include "position.h"
+#include "search.h"
+#include "thread.h"
+#include "timeman.h"
+#include "tt.h"
+#include "uci.h"
+#include "syzygy/tbprobe.h"
+#include "nnue/evaluate_nnue.h"
+
+using namespace std;
+
+namespace Stockfish {
+
+namespace {
+
+  // FEN string for the initial position in standard chess
+  const char* StartFEN = "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1";
+
+
+  // position() is called when the engine receives the "position" UCI command.
+  // It sets up the position that is described in the given FEN string ("fen") or
+  // the initial position ("startpos") and then makes the moves given in the following
+  // move list ("moves").
+
+  void position(Position& pos, istringstream& is, StateListPtr& states) {
+
+    Move m;
+    string token, fen;
+
+    is >> token;
+
+    if (token == "startpos")
+    {
+        fen = StartFEN;
+        is >> token; // Consume the "moves" token, if any
+    }
+    else if (token == "fen")
+        while (is >> token && token != "moves")
+            fen += token + " ";
+    else
+        return;
+
+    states = StateListPtr(new std::deque<StateInfo>(1)); // Drop the old state and create a new one
+    pos.set(fen, Options["UCI_Chess960"], &states->back(), Threads.main());
+
+    // Parse the move list, if any
+    while (is >> token && (m = UCI::to_move(pos, token)) != MOVE_NONE)
+    {
+        states->emplace_back();
+        pos.do_move(m, states->back());
+    }
+  }
+
+  // trace_eval() prints the evaluation of the current position, consistent with
+  // the UCI options set so far.
+
+  void trace_eval(Position& pos) {
+
+    StateListPtr states(new std::deque<StateInfo>(1));
+    Position p;
+    p.set(pos.fen(), Options["UCI_Chess960"], &states->back(), Threads.main());
+
+    Eval::NNUE::verify();
+
+    sync_cout << "\n" << Eval::trace(p) << sync_endl;
+  }
+
+
+  // setoption() is called when the engine receives the "setoption" UCI command.
+  // The function updates the UCI option ("name") to the given value ("value").
+
+  void setoption(istringstream& is) {
+
+    string token, name, value;
+
+    is >> token; // Consume the "name" token
+
+    // Read the option name (can contain spaces)
+    while (is >> token && token != "value")
+        name += (name.empty() ? "" : " ") + token;
+
+    // Read the option value (can contain spaces)
+    while (is >> token)
+        value += (value.empty() ? "" : " ") + token;
+
+    if (Options.count(name))
+        Options[name] = value;
+    else
+        sync_cout << "No such option: " << name << sync_endl;
+  }
+
+
+  // go() is called when the engine receives the "go" UCI command. The function
+  // sets the thinking time and other parameters from the input string, then starts
+  // with a search.
+
+  void go(Position& pos, istringstream& is, StateListPtr& states) {
+
+    Search::LimitsType limits;
+    string token;
+    bool ponderMode = false;
+
+    limits.startTime = now(); // The search starts as early as possible
+
+    while (is >> token)
+        if (token == "searchmoves") // Needs to be the last command on the line
+            while (is >> token)
+                limits.searchmoves.push_back(UCI::to_move(pos, token));
+
+        else if (token == "wtime")     is >> limits.time[WHITE];
+        else if (token == "btime")     is >> limits.time[BLACK];
+        else if (token == "winc")      is >> limits.inc[WHITE];
+        else if (token == "binc")      is >> limits.inc[BLACK];
+        else if (token == "movestogo") is >> limits.movestogo;
+        else if (token == "depth")     is >> limits.depth;
+        else if (token == "nodes")     is >> limits.nodes;
+        else if (token == "movetime")  is >> limits.movetime;
+        else if (token == "mate")      is >> limits.mate;
+        else if (token == "perft")     is >> limits.perft;
+        else if (token == "infinite")  limits.infinite = 1;
+        else if (token == "ponder")    ponderMode = true;
+
+    Threads.start_thinking(pos, states, limits, ponderMode);
+  }
+
+
+  // bench() is called when the engine receives the "bench" command.
+  // Firstly, a list of UCI commands is set up according to the bench
+  // parameters, then it is run one by one, printing a summary at the end.
+
+  void bench(Position& pos, istream& args, StateListPtr& states) {
+
+    string token;
+    uint64_t num, nodes = 0, cnt = 1;
+
+    vector<string> list = setup_bench(pos, args);
+    num = count_if(list.begin(), list.end(), [](const string& s) { return s.find("go ") == 0 || s.find("eval") == 0; });
+
+    TimePoint elapsed = now();
+
+    for (const auto& cmd : list)
+    {
+        istringstream is(cmd);
+        is >> skipws >> token;
+
+        if (token == "go" || token == "eval")
+        {
+            cerr << "\nPosition: " << cnt++ << '/' << num << " (" << pos.fen() << ")" << endl;
+            if (token == "go")
+            {
+               go(pos, is, states);
+               Threads.main()->wait_for_search_finished();
+               nodes += Threads.nodes_searched();
+            }
+            else
+               trace_eval(pos);
+        }
+        else if (token == "setoption")  setoption(is);
+        else if (token == "position")   position(pos, is, states);
+        else if (token == "ucinewgame") { Search::clear(); elapsed = now(); } // Search::clear() may take a while
+    }
+
+    elapsed = now() - elapsed + 1; // Ensure positivity to avoid a 'divide by zero'
+
+    dbg_print();
+
+    cerr << "\n==========================="
+         << "\nTotal time (ms) : " << elapsed
+         << "\nNodes searched  : " << nodes
+         << "\nNodes/second    : " << 1000 * nodes / elapsed << endl;
+  }
+
+  // The win rate model returns the probability of winning (in per mille units) given an
+  // eval and a game ply. It fits the LTC fishtest statistics rather accurately.
+  int win_rate_model(Value v, int ply) {
+
+     // The model only captures up to 240 plies, so limit the input and then rescale
+     double m = std::min(240, ply) / 64.0;
+
+     // The coefficients of a third-order polynomial fit is based on the fishtest data
+     // for two parameters that need to transform eval to the argument of a logistic
+     // function.
+     constexpr double as[] = {   0.38036525,   -2.82015070,   23.17882135,  307.36768407};
+     constexpr double bs[] = {  -2.29434733,   13.27689788,  -14.26828904,   63.45318330 };
+
+     // Enforce that NormalizeToPawnValue corresponds to a 50% win rate at ply 64
+     static_assert(UCI::NormalizeToPawnValue == int(as[0] + as[1] + as[2] + as[3]));
+
+     double a = (((as[0] * m + as[1]) * m + as[2]) * m) + as[3];
+     double b = (((bs[0] * m + bs[1]) * m + bs[2]) * m) + bs[3];
+
+     // Transform the eval to centipawns with limited range
+     double x = std::clamp(double(v), -4000.0, 4000.0);
+
+     // Return the win rate in per mille units rounded to the nearest value
+     return int(0.5 + 1000 / (1 + std::exp((a - x) / b)));
+  }
+
+} // namespace
+
+
+/// UCI::loop() waits for a command from the stdin, parses it and then calls the appropriate
+/// function. It also intercepts an end-of-file (EOF) indication from the stdin to ensure a
+/// graceful exit if the GUI dies unexpectedly. When called with some command-line arguments,
+/// like running 'bench', the function returns immediately after the command is executed.
+/// In addition to the UCI ones, some additional debug commands are also supported.
+
+void UCI::loop(int argc, char* argv[]) {
+
+  Position pos;
+  string token, cmd;
+  StateListPtr states(new std::deque<StateInfo>(1));
+
+  pos.set(StartFEN, false, &states->back(), Threads.main());
+
+  for (int i = 1; i < argc; ++i)
+      cmd += std::string(argv[i]) + " ";
+
+  do {
+      if (argc == 1 && !getline(fakein, cmd)) // Wait for an input or an end-of-file (EOF) indication
+          cmd = "quit";
+
+      istringstream is(cmd);
+
+      token.clear(); // Avoid a stale if getline() returns nothing or a blank line
+      is >> skipws >> token;
+
+      if (    token == "quit"
+          ||  token == "stop")
+          Threads.stop = true;
+
+      // The GUI sends 'ponderhit' to tell that the user has played the expected move.
+      // So, 'ponderhit' is sent if pondering was done on the same move that the user
+      // has played. The search should continue, but should also switch from pondering
+      // to the normal search.
+      else if (token == "ponderhit")
+          Threads.main()->ponder = false; // Switch to the normal search
+
+      else if (token == "uci")
+          sync_cout << "id name " << engine_info(true)
+                    << "\n"       << Options
+                    << "\nuciok"  << sync_endl;
+
+      else if (token == "setoption")  setoption(is);
+      else if (token == "go")         go(pos, is, states);
+      else if (token == "position")   position(pos, is, states);
+      else if (token == "ucinewgame") Search::clear();
+      else if (token == "isready")    sync_cout << "readyok" << sync_endl;
+
+      // Add custom non-UCI commands, mainly for debugging purposes.
+      // These commands must not be used during a search!
+      else if (token == "flip")     pos.flip();
+      else if (token == "bench")    bench(pos, is, states);
+      else if (token == "d")        sync_cout << pos << sync_endl;
+      else if (token == "eval")     trace_eval(pos);
+      else if (token == "compiler") sync_cout << compiler_info() << sync_endl;
+      else if (token == "export_net")
+      {
+          std::optional<std::string> filename;
+          std::string f;
+          if (is >> skipws >> f)
+              filename = f;
+          Eval::NNUE::save_eval(filename);
+      }
+      else if (token == "--help" || token == "help" || token == "--license" || token == "license")
+          sync_cout << "\nStockfish is a powerful chess engine for playing and analyzing."
+                       "\nIt is released as free software licensed under the GNU GPLv3 License."
+                       "\nStockfish is normally used with a graphical user interface (GUI) and implements"
+                       "\nthe Universal Chess Interface (UCI) protocol to communicate with a GUI, an API, etc."
+                       "\nFor any further information, visit https://github.com/official-stockfish/Stockfish#readme"
+                       "\nor read the corresponding README.md and Copying.txt files distributed along with this program.\n" << sync_endl;
+      else if (!token.empty() && token[0] != '#')
+          sync_cout << "Unknown command: '" << cmd << "'. Type help for more information." << sync_endl;
+
+  } while (token != "quit" && argc == 1); // The command-line arguments are one-shot
+}
+
+
+/// UCI::value() converts a Value to a string by adhering to the UCI protocol specification:
+///
+/// cp <x>    The score from the engine's point of view in centipawns.
+/// mate <y>  Mate in 'y' moves (not plies). If the engine is getting mated,
+///           uses negative values for 'y'.
+
+string UCI::value(Value v) {
+
+  assert(-VALUE_INFINITE < v && v < VALUE_INFINITE);
+
+  stringstream ss;
+
+  if (abs(v) < VALUE_TB_WIN_IN_MAX_PLY)
+      ss << "cp " << v * 100 / NormalizeToPawnValue;
+  else if (abs(v) < VALUE_MATE_IN_MAX_PLY)
+  {
+      const int ply = VALUE_MATE_IN_MAX_PLY - 1 - std::abs(v);  // recompute ss->ply
+      ss << "cp " << (v > 0 ? 20000 - ply : -20000 + ply);
+  }
+  else
+      ss << "mate " << (v > 0 ? VALUE_MATE - v + 1 : -VALUE_MATE - v) / 2;
+
+  return ss.str();
+}
+
+
+/// UCI::wdl() reports the win-draw-loss (WDL) statistics given an evaluation
+/// and a game ply based on the data gathered for fishtest LTC games.
+
+string UCI::wdl(Value v, int ply) {
+
+  stringstream ss;
+
+  int wdl_w = win_rate_model( v, ply);
+  int wdl_l = win_rate_model(-v, ply);
+  int wdl_d = 1000 - wdl_w - wdl_l;
+  ss << " wdl " << wdl_w << " " << wdl_d << " " << wdl_l;
+
+  return ss.str();
+}
+
+
+/// UCI::square() converts a Square to a string in algebraic notation (g1, a7, etc.)
+
+std::string UCI::square(Square s) {
+  return std::string{ char('a' + file_of(s)), char('1' + rank_of(s)) };
+}
+
+
+/// UCI::move() converts a Move to a string in coordinate notation (g1f3, a7a8q).
+/// The only special case is castling where the e1g1 notation is printed in
+/// standard chess mode and in e1h1 notation it is printed in Chess960 mode.
+/// Internally, all castling moves are always encoded as 'king captures rook'.
+
+string UCI::move(Move m, bool chess960) {
+
+  if (m == MOVE_NONE)
+      return "(none)";
+
+  if (m == MOVE_NULL)
+      return "0000";
+
+  Square from = from_sq(m);
+  Square to = to_sq(m);
+
+  if (type_of(m) == CASTLING && !chess960)
+      to = make_square(to > from ? FILE_G : FILE_C, rank_of(from));
+
+  string move = UCI::square(from) + UCI::square(to);
+
+  if (type_of(m) == PROMOTION)
+      move += " pnbrqk"[promotion_type(m)];
+
+  return move;
+}
+
+
+/// UCI::to_move() converts a string representing a move in coordinate notation
+/// (g1f3, a7a8q) to the corresponding legal Move, if any.
+
+Move UCI::to_move(const Position& pos, string& str) {
+
+  if (str.length() == 5)
+      str[4] = char(tolower(str[4])); // The promotion piece character must be lowercased
+
+  for (const auto& m : MoveList<LEGAL>(pos))
+      if (str == UCI::move(m, pos.is_chess960()))
+          return m;
+
+  return MOVE_NONE;
+}
+
+} // namespace Stockfish

src/Stockfish/src/uci.h

@@ -0,0 +1,92 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#ifndef UCI_H_INCLUDED
+#define UCI_H_INCLUDED
+
+#include <map>
+#include <string>
+
+#include "types.h"
+
+namespace Stockfish {
+
+class Position;
+
+namespace UCI {
+
+// Normalizes the internal value as reported by evaluate or search
+// to the UCI centipawn result used in output. This value is derived from
+// the win_rate_model() such that Stockfish outputs an advantage of
+// "100 centipawns" for a position if the engine has a 50% probability to win
+// from this position in selfplay at fishtest LTC time control.
+const int NormalizeToPawnValue = 328;
+
+class Option;
+
+/// Define a custom comparator, because the UCI options should be case-insensitive
+struct CaseInsensitiveLess {
+  bool operator() (const std::string&, const std::string&) const;
+};
+
+/// The options container is defined as a std::map
+using OptionsMap = std::map<std::string, Option, CaseInsensitiveLess>;
+
+/// The Option class implements each option as specified by the UCI protocol
+class Option {
+
+  using OnChange = void (*)(const Option&);
+
+public:
+  Option(OnChange = nullptr);
+  Option(bool v, OnChange = nullptr);
+  Option(const char* v, OnChange = nullptr);
+  Option(double v, int minv, int maxv, OnChange = nullptr);
+  Option(const char* v, const char* cur, OnChange = nullptr);
+
+  Option& operator=(const std::string&);
+  void operator<<(const Option&);
+  operator int() const;
+  operator std::string() const;
+  bool operator==(const char*) const;
+
+private:
+  friend std::ostream& operator<<(std::ostream&, const OptionsMap&);
+
+  std::string defaultValue, currentValue, type;
+  int min, max;
+  size_t idx;
+  OnChange on_change;
+};
+
+void init(OptionsMap&);
+void loop(int argc, char* argv[]);
+std::string value(Value v);
+std::string square(Square s);
+std::string move(Move m, bool chess960);
+std::string pv(const Position& pos, Depth depth);
+std::string wdl(Value v, int ply);
+Move to_move(const Position& pos, std::string& str);
+
+} // namespace UCI
+
+extern UCI::OptionsMap Options;
+
+} // namespace Stockfish
+
+#endif // #ifndef UCI_H_INCLUDED

src/Stockfish/src/ucioption.cpp

@@ -0,0 +1,194 @@
+/*
+  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
+
+  Stockfish is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  Stockfish is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
+
+  You should have received a copy of the GNU General Public License
+  along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include <algorithm>
+#include <cassert>
+#include <ostream>
+#include <sstream>
+
+#include "evaluate.h"
+#include "misc.h"
+#include "search.h"
+#include "thread.h"
+#include "tt.h"
+#include "uci.h"
+#include "syzygy/tbprobe.h"
+
+using std::string;
+
+namespace Stockfish {
+
+UCI::OptionsMap Options; // Global object
+
+namespace UCI {
+
+/// 'On change' actions, triggered by an option's value change
+static void on_clear_hash(const Option&) { Search::clear(); }
+static void on_hash_size(const Option& o) { TT.resize(size_t(o)); }
+static void on_logger(const Option& o) { start_logger(o); }
+static void on_threads(const Option& o) { Threads.set(size_t(o)); }
+static void on_tb_path(const Option& o) { Tablebases::init(o); }
+static void on_use_NNUE(const Option&) { Eval::NNUE::init(); }
+static void on_eval_file(const Option&) { Eval::NNUE::init(); }
+
+/// Our case insensitive less() function as required by UCI protocol
+bool CaseInsensitiveLess::operator() (const string& s1, const string& s2) const {
+
+  return std::lexicographical_compare(s1.begin(), s1.end(), s2.begin(), s2.end(),
+         [](char c1, char c2) { return tolower(c1) < tolower(c2); });
+}
+
+
+/// UCI::init() initializes the UCI options to their hard-coded default values
+
+void init(OptionsMap& o) {
+
+  constexpr int MaxHashMB = Is64Bit ? 33554432 : 2048;
+
+  o["Debug Log File"]        << Option("", on_logger);
+  o["Threads"]               << Option(1, 1, 1024, on_threads);
+  o["Hash"]                  << Option(16, 1, MaxHashMB, on_hash_size);
+  o["Clear Hash"]            << Option(on_clear_hash);
+  o["Ponder"]                << Option(false);
+  o["MultiPV"]               << Option(1, 1, 500);
+  o["Skill Level"]           << Option(20, 0, 20);
+  o["Move Overhead"]         << Option(10, 0, 5000);
+  o["Slow Mover"]            << Option(100, 10, 1000);
+  o["nodestime"]             << Option(0, 0, 10000);
+  o["UCI_Chess960"]          << Option(false);
+  o["UCI_AnalyseMode"]       << Option(false);
+  o["UCI_LimitStrength"]     << Option(false);
+  o["UCI_Elo"]               << Option(1320, 1320, 3190);
+  o["UCI_ShowWDL"]           << Option(false);
+  o["SyzygyPath"]            << Option("<empty>", on_tb_path);
+  o["SyzygyProbeDepth"]      << Option(1, 1, 100);
+  o["Syzygy50MoveRule"]      << Option(true);
+  o["SyzygyProbeLimit"]      << Option(7, 0, 7);
+  o["Use NNUE"]              << Option(true, on_use_NNUE);
+  o["EvalFile"]              << Option(EvalFileDefaultName, on_eval_file);
+}
+
+
+/// operator<<() is used to print all the options default values in chronological
+/// insertion order (the idx field) and in the format defined by the UCI protocol.
+
+std::ostream& operator<<(std::ostream& os, const OptionsMap& om) {
+
+  for (size_t idx = 0; idx < om.size(); ++idx)
+      for (const auto& it : om)
+          if (it.second.idx == idx)
+          {
+              const Option& o = it.second;
+              os << "\noption name " << it.first << " type " << o.type;
+
+              if (o.type == "string" || o.type == "check" || o.type == "combo")
+                  os << " default " << o.defaultValue;
+
+              if (o.type == "spin")
+                  os << " default " << int(stof(o.defaultValue))
+                     << " min "     << o.min
+                     << " max "     << o.max;
+
+              break;
+          }
+
+  return os;
+}
+
+
+/// Option class constructors and conversion operators
+
+Option::Option(const char* v, OnChange f) : type("string"), min(0), max(0), on_change(f)
+{ defaultValue = currentValue = v; }
+
+Option::Option(bool v, OnChange f) : type("check"), min(0), max(0), on_change(f)
+{ defaultValue = currentValue = (v ? "true" : "false"); }
+
+Option::Option(OnChange f) : type("button"), min(0), max(0), on_change(f)
+{}
+
+Option::Option(double v, int minv, int maxv, OnChange f) : type("spin"), min(minv), max(maxv), on_change(f)
+{ defaultValue = currentValue = std::to_string(v); }
+
+Option::Option(const char* v, const char* cur, OnChange f) : type("combo"), min(0), max(0), on_change(f)
+{ defaultValue = v; currentValue = cur; }
+
+Option::operator int() const {
+  assert(type == "check" || type == "spin");
+  return (type == "spin" ? std::stoi(currentValue) : currentValue == "true");
+}
+
+Option::operator std::string() const {
+  assert(type == "string");
+  return currentValue;
+}
+
+bool Option::operator==(const char* s) const {
+  assert(type == "combo");
+  return   !CaseInsensitiveLess()(currentValue, s)
+        && !CaseInsensitiveLess()(s, currentValue);
+}
+
+
+/// operator<<() inits options and assigns idx in the correct printing order
+
+void Option::operator<<(const Option& o) {
+
+  static size_t insert_order = 0;
+
+  *this = o;
+  idx = insert_order++;
+}
+
+
+/// operator=() updates currentValue and triggers on_change() action. It's up to
+/// the GUI to check for option's limits, but we could receive the new value
+/// from the user by console window, so let's check the bounds anyway.
+
+Option& Option::operator=(const string& v) {
+
+  assert(!type.empty());
+
+  if (   (type != "button" && type != "string" && v.empty())
+      || (type == "check" && v != "true" && v != "false")
+      || (type == "spin" && (stof(v) < min || stof(v) > max)))
+      return *this;
+
+  if (type == "combo")
+  {
+      OptionsMap comboMap; // To have case insensitive compare
+      string token;
+      std::istringstream ss(defaultValue);
+      while (ss >> token)
+          comboMap[token] << Option();
+      if (!comboMap.count(v) || v == "var")
+          return *this;
+  }
+
+  if (type != "button")
+      currentValue = v;
+
+  if (on_change)
+      on_change(*this);
+
+  return *this;
+}
+
+} // namespace UCI
+
+} // namespace Stockfish

src/Stockfish/tests/instrumented.sh

@@ -0,0 +1,151 @@
+#!/bin/bash
+# check for errors under valgrind or sanitizers.
+
+error()
+{
+  echo "instrumented testing failed on line $1"
+  exit 1
+}
+trap 'error ${LINENO}' ERR
+
+# define suitable post and prefixes for testing options
+case $1 in
+  --valgrind)
+    echo "valgrind testing started"
+    prefix=''
+    exeprefix='valgrind --error-exitcode=42 --errors-for-leak-kinds=all --leak-check=full'
+    postfix='1>/dev/null'
+    threads="1"
+  ;;
+  --valgrind-thread)
+    echo "valgrind-thread testing started"
+    prefix=''
+    exeprefix='valgrind --fair-sched=try --error-exitcode=42'
+    postfix='1>/dev/null'
+    threads="2"
+  ;;
+  --sanitizer-undefined)
+    echo "sanitizer-undefined testing started"
+    prefix='!'
+    exeprefix=''
+    postfix='2>&1 | grep -A50 "runtime error:"'
+    threads="1"
+  ;;
+  --sanitizer-thread)
+    echo "sanitizer-thread testing started"
+    prefix='!'
+    exeprefix=''
+    postfix='2>&1 | grep -A50 "WARNING: ThreadSanitizer:"'
+    threads="2"
+
+cat << EOF > tsan.supp
+race:Stockfish::TTEntry::move
+race:Stockfish::TTEntry::depth
+race:Stockfish::TTEntry::bound
+race:Stockfish::TTEntry::save
+race:Stockfish::TTEntry::value
+race:Stockfish::TTEntry::eval
+race:Stockfish::TTEntry::is_pv
+
+race:Stockfish::TranspositionTable::probe
+race:Stockfish::TranspositionTable::hashfull
+
+EOF
+
+    export TSAN_OPTIONS="suppressions=./tsan.supp"
+
+  ;;
+  *)
+    echo "unknown testing started"
+    prefix=''
+    exeprefix=''
+    postfix=''
+    threads="1"
+  ;;
+esac
+
+# simple command line testing
+for args in "eval" \
+            "go nodes 1000" \
+            "go depth 10" \
+            "go movetime 1000" \
+            "go wtime 8000 btime 8000 winc 500 binc 500" \
+            "bench 128 $threads 8 default depth" \
+            "export_net verify.nnue"
+do
+
+   echo "$prefix $exeprefix ./stockfish $args $postfix"
+   eval "$prefix $exeprefix ./stockfish $args $postfix"
+
+done
+
+# verify the generated net equals the base net
+network=`./stockfish uci | grep 'option name EvalFile type string default' | awk '{print $NF}'`
+echo "Comparing $network to the written verify.nnue"
+diff $network verify.nnue
+
+# more general testing, following an uci protocol exchange
+cat << EOF > game.exp
+ set timeout 240
+ spawn $exeprefix ./stockfish
+
+ send "uci\n"
+ expect "uciok"
+
+ send "setoption name Threads value $threads\n"
+
+ send "ucinewgame\n"
+ send "position startpos\n"
+ send "go nodes 1000\n"
+ expect "bestmove"
+
+ send "position startpos moves e2e4 e7e6\n"
+ send "go nodes 1000\n"
+ expect "bestmove"
+
+ send "position fen 5rk1/1K4p1/8/8/3B4/8/8/8 b - - 0 1\n"
+ send "go depth 10\n"
+ expect "bestmove"
+
+ send "quit\n"
+ expect eof
+
+ # return error code of the spawned program, useful for valgrind
+ lassign [wait] pid spawnid os_error_flag value
+ exit \$value
+EOF
+
+#download TB as needed
+if [ ! -d ../tests/syzygy ]; then
+   curl -sL https://api.github.com/repos/niklasf/python-chess/tarball/9b9aa13f9f36d08aadfabff872882f4ab1494e95 | tar -xzf -
+   mv niklasf-python-chess-9b9aa13 ../tests/syzygy
+fi
+
+cat << EOF > syzygy.exp
+ set timeout 240
+ spawn $exeprefix ./stockfish
+ send "uci\n"
+ send "setoption name SyzygyPath value ../tests/syzygy/\n"
+ expect "info string Found 35 tablebases" {} timeout {exit 1}
+ send "bench 128 1 8 default depth\n"
+ send "quit\n"
+ expect eof
+
+ # return error code of the spawned program, useful for valgrind
+ lassign [wait] pid spawnid os_error_flag value
+ exit \$value
+EOF
+
+for exp in game.exp syzygy.exp
+do
+
+  echo "$prefix expect $exp $postfix"
+  eval "$prefix expect $exp $postfix"
+
+  rm $exp
+
+done
+
+rm -f tsan.supp
+
+echo "instrumented testing OK"

src/Stockfish/tests/perft.sh

@@ -0,0 +1,32 @@
+#!/bin/bash
+# verify perft numbers (positions from www.chessprogramming.org/Perft_Results)
+
+error()
+{
+  echo "perft testing failed on line $1"
+  exit 1
+}
+trap 'error ${LINENO}' ERR
+
+echo "perft testing started"
+
+cat << EOF > perft.exp
+   set timeout 10
+   lassign \$argv pos depth result
+   spawn ./stockfish
+   send "position \$pos\\ngo perft \$depth\\n"
+   expect "Nodes searched? \$result" {} timeout {exit 1}
+   send "quit\\n"
+   expect eof
+EOF
+
+expect perft.exp startpos 5 4865609 > /dev/null
+expect perft.exp "fen r3k2r/p1ppqpb1/bn2pnp1/3PN3/1p2P3/2N2Q1p/PPPBBPPP/R3K2R w KQkq -" 5 193690690 > /dev/null
+expect perft.exp "fen 8/2p5/3p4/KP5r/1R3p1k/8/4P1P1/8 w - -" 6 11030083 > /dev/null
+expect perft.exp "fen r3k2r/Pppp1ppp/1b3nbN/nP6/BBP1P3/q4N2/Pp1P2PP/R2Q1RK1 w kq - 0 1" 5 15833292 > /dev/null
+expect perft.exp "fen rnbq1k1r/pp1Pbppp/2p5/8/2B5/8/PPP1NnPP/RNBQK2R w KQ - 1 8" 5 89941194 > /dev/null
+expect perft.exp "fen r4rk1/1pp1qppp/p1np1n2/2b1p1B1/2B1P1b1/P1NP1N2/1PP1QPPP/R4RK1 w - - 0 10" 5 164075551 > /dev/null
+
+rm perft.exp
+
+echo "perft testing OK"

src/Stockfish/tests/reprosearch.sh

@@ -0,0 +1,61 @@
+#!/bin/bash
+# verify reproducible search
+
+error()
+{
+  echo "reprosearch testing failed on line $1"
+  exit 1
+}
+trap 'error ${LINENO}' ERR
+
+echo "reprosearch testing started"
+
+# repeat two short games, separated by ucinewgame.
+# with go nodes $nodes they should result in exactly
+# the same node count for each iteration.
+cat << EOF > repeat.exp
+ set timeout 10
+ spawn ./stockfish
+ lassign \$argv nodes
+
+ send "uci\n"
+ expect "uciok"
+
+ send "ucinewgame\n"
+ send "position startpos\n"
+ send "go nodes \$nodes\n"
+ expect "bestmove"
+
+ send "position startpos moves e2e4 e7e6\n"
+ send "go nodes \$nodes\n"
+ expect "bestmove"
+
+ send "ucinewgame\n"
+ send "position startpos\n"
+ send "go nodes \$nodes\n"
+ expect "bestmove"
+
+ send "position startpos moves e2e4 e7e6\n"
+ send "go nodes \$nodes\n"
+ expect "bestmove"
+
+ send "quit\n"
+ expect eof
+EOF
+
+# to increase the likelihood of finding a non-reproducible case,
+# the allowed number of nodes are varied systematically
+for i in `seq 1 20`
+do
+
+  nodes=$((100*3**i/2**i))
+  echo "reprosearch testing with $nodes nodes"
+
+  # each line should appear exactly an even number of times
+  expect repeat.exp $nodes 2>&1 | grep -o "nodes [0-9]*" | sort | uniq -c | awk '{if ($1%2!=0) exit(1)}'
+
+done
+
+rm repeat.exp
+
+echo "reprosearch testing OK"

src/Stockfish/tests/signature.sh

@@ -0,0 +1,31 @@
+#!/bin/bash
+# obtain and optionally verify Bench / signature
+# if no reference is given, the output is deliberately limited to just the signature
+
+error()
+{
+  echo "running bench for signature failed on line $1"
+  exit 1
+}
+trap 'error ${LINENO}' ERR
+
+# obtain
+
+signature=`./stockfish bench 2>&1 | grep "Nodes searched  : " | awk '{print $4}'`
+
+if [ $# -gt 0 ]; then
+   # compare to given reference
+   if [ "$1" != "$signature" ]; then
+      if [ -z "$signature" ]; then
+         echo "No signature obtained from bench. Code crashed or assert triggered ?"
+      else
+         echo "signature mismatch: reference $1 obtained: $signature ."
+      fi
+      exit 1
+   else
+      echo "signature OK: $signature"
+   fi
+else
+   # just report signature
+   echo $signature
+fi

src/create_stockfish.sh

@@ -0,0 +1,21 @@
+curl -s https://api.github.com/repos/official-stockfish/Stockfish/releases/latest \
+| grep "tarball_url" \
+| cut -d : -f 2,3 \
+| tr -d "\", " \
+| wget -qi - -O Stockfish.tar.gz
+rm -rf Stockfish
+mkdir -p Stockfish
+tar -xzf Stockfish.tar.gz --directory Stockfish --strip-components=1
+rm Stockfish.tar.gz
+find ./Stockfish/ -type f -exec sed -i \
+    -e 's/std::cout/fakeout/g' \
+    -e 's/std::cin/fakein/g' \
+    -e 's/std::endl/fakeendl/g' \
+    -e 's/getline(cin, cmd)/getline(fakein, cmd)/g' \
+    {} +
+
+nnue_name=$(grep EvalFileDefaultName Stockfish/src/evaluate.h \
+| grep define \
+| sed 's/.*\(nn-[a-z0-9]\{12\}.nnue\).*/\1/')
+
+sed -i "s/set(NNUE_NAME nn-[^)]*.nnue)/set(NNUE_NAME $nnue_name)/" CMakeLists.txt

src/stream_fix.cpp

@@ -17,7 +17,7 @@ bool FakeStream::is_closed() { return closed; }
 
 std::streambuf* FakeStream::rdbuf() { return nullptr; }
 
-std::streambuf* FakeStream::rdbuf(std::streambuf* __sb) { return nullptr; }
+std::streambuf* FakeStream::rdbuf(std::streambuf* buf) { return nullptr; }
 
 bool std::getline(FakeStream& is, std::string& str) {
     if (is.is_closed()) return false;

windows/.gitignore

@@ -0,0 +1,17 @@
+flutter/
+
+# Visual Studio user-specific files.
+*.suo
+*.user
+*.userosscache
+*.sln.docstates
+
+# Visual Studio build-related files.
+x64/
+x86/
+
+# Visual Studio cache files
+# files ending in .cache can be ignored
+*.[Cc]ache
+# but keep track of directories ending in .cache
+!*.[Cc]ache/