Stockfish 16

Official release version of Stockfish 16 Bench: 2593605 --- Stockfish 16 A new major release of Stockfish is now available at https://stockfishchess.org/download/ *Quality of chess play* Stockfish continues to demonstrate its ability to discover superior moves with remarkable speed. In self-play against Stockfish 15, this new release gains up to 50 Elo[1] and wins up to 12 times more game pairs[2] than it loses. In major chess engine tournaments, Stockfish reliably tops the rankings[3] winning the TCEC season 24 Superfinal, Swiss, Fischer Random, and Double Random Chess tournaments and the CCC 19 Bullet, 20 Blitz, and 20 Rapid competitions. Leela Chess Zero[4] was the challenger in most finals, putting top-engine chess now firmly in the hands of teams embracing free and open-source software. *Progress made* This updated version of Stockfish introduces several enhancements, including an upgraded neural net architecture (SFNNv6)[5], improved implementation, and refined parameterization. The ongoing utilization of Leela’s data combined with a novel inference approach exploiting sparsity[6], and network compression[7] ensure a speedy evaluation and modest binary sizes while allowing for more weights and higher accuracy. The search has undergone more optimization, leading to improved performance, particularly in longer analyses[8]. Additionally, the Fishtest framework has been improved and is now able to run the tests needed to validate new ideas with 10000s of CPU cores. *Usability improvements* Stockfish now comes with documentation, found in the wiki folder when downloading it or on GitHub[9]. Additionally, Stockfish now includes a clear and consistent forced tablebase win score, displaying a value of 200 minus the number of plies required to reach a tablebase win[10]. Furthermore, the UCI_Elo option, to reduce its strength, has been calibrated[11]. It is worth noting that the evaluation system remains consistent with Stockfish 15.1[12], maintaining the choice that 100cp means a 50% chance of winning the game against an equal opponent[13]. Finally, binaries of our latest development version are now provided continuously as pre-releases on GitHub making it easier for enthusiasts to download the latest and strongest version of the program[14], we thank Roman Korba for having provided a similar service for a long time. *Thank you* The success of the Stockfish project relies on the vibrant community of passionate enthusiasts (we appreciate each and every one of you!) who generously contribute their knowledge, time, and resources. Together, this dedicated community works towards the common goal of developing a powerful, freely accessible, and open-source chess engine. We invite all chess enthusiasts to join the Fishtest testing framework and contribute to the project[15]. The Stockfish team [1] https://tests.stockfishchess.org/tests/view/649409f0dc7002ce609c99cc [2] https://tests.stockfishchess.org/tests/view/649409d7dc7002ce609c99c6 [3] https://en.wikipedia.org/wiki/Stockfish_(chess)#Competition_results [4] https://lczero.org/ [5] https://github.com/official-stockfish/Stockfish/commit/c1fff71 [6] https://github.com/official-stockfish/Stockfish/commit/38e6166 [7] https://github.com/official-stockfish/Stockfish/commit/a46087e [8] https://github.com/official-stockfish/Stockfish/commit/472e726 [9] https://github.com/official-stockfish/Stockfish/wiki/ [10] https://github.com/official-stockfish/Stockfish/commit/def2966 [11] https://github.com/official-stockfish/Stockfish/commit/a08b8d4 [12] https://github.com/official-stockfish/Stockfish/commit/52e84e4 [13] https://github.com/official-stockfish/Stockfish/wiki/Stockfish-FAQ#interpretation-of-the-stockfish-evaluation [14] https://github.com/official-stockfish/Stockfish/releases?q=prerelease%3Atrue [15] https://stockfishchess.org/get-involved/
Update default net to nn-5af11540bbfe.nnue
2026-05-20 06:17:49 +00:00 · 2023-06-29 08:00:10 +02:00 · 2023-06-22 10:33:19 +02:00 · 2023-06-22 10:26:17 +02:00 · 2023-06-22 10:17:44 +02:00 · 2023-06-22 10:15:51 +02:00
69 changed files with 3442 additions and 2159 deletions
@@ -0,0 +1,65 @@
+name: Report issue
+description: Create a report to help us fix issues with the engine
+body:
+- type: textarea
+  attributes:
+    label: Describe the issue
+    description: A clear and concise description of what you're experiencing.
+  validations:
+    required: true
+
+- type: textarea
+  attributes:
+    label: Expected behavior
+    description: A clear and concise description of what you expected to happen.
+  validations:
+    required: true
+
+- type: textarea
+  attributes:
+    label: Steps to reproduce
+    description: |
+      Steps to reproduce the behavior.
+      You can also use this section to paste the command line output.
+    placeholder: |
+      ```
+      position startpos moves g2g4 e7e5 f2f3
+      go mate 1
+      info string NNUE evaluation using nn-6877cd24400e.nnue enabled
+      info depth 1 seldepth 1 multipv 1 score mate 1 nodes 33 nps 11000 tbhits 0 time 3 pv d8h4
+      bestmove d8h4
+      ```
+  validations:
+    required: true
+
+- type: textarea
+  attributes:
+    label: Anything else?
+    description: |
+      Anything that will give us more context about the issue you are encountering.
+      You can also use this section to propose ideas on how to solve the issue. 
+  validations:
+    required: false
+
+- type: dropdown
+  attributes:
+    label: Operating system
+    options:
+      - All
+      - Windows
+      - Linux
+      - MacOS
+      - Android
+      - Other or N/A
+  validations:
+    required: true
+
+- type: input
+  attributes:
+    label: Stockfish version
+    description: |
+      This can be found by running the engine.
+      You can also use the commit ID.
+    placeholder: Stockfish 15 / e6e324e
+  validations:
+    required: true
@@ -0,0 +1,8 @@
+blank_issues_enabled: false
+contact_links:
+  - name: Discord server
+    url: https://discord.gg/GWDRS3kU6R
+    about: Feel free to ask for support or have a chat with us in our Discord server!
+  - name: Discussions, Q&A, ideas, show us something...
+    url: https://github.com/official-stockfish/Stockfish/discussions/new
+    about: Do you have an idea for Stockfish? Do you want to show something that you made? Please open a discussion about it!
@@ -1,333 +1,51 @@
 name: Stockfish
 on:
  push:
+    tags:        
+      - '*' 
    branches:
      - master
      - tools
      - github_ci
-      - github_ci_armv7
  pull_request:
    branches:
      - master
      - tools
 jobs:
-  Stockfish:
-    name: ${{ matrix.config.name }}
-    runs-on: ${{ matrix.config.os }}
-    env:
-      COMPILER: ${{ matrix.config.compiler }}
-      COMP: ${{ matrix.config.comp }}
-      CXXFLAGS: "-Werror"
-    strategy:
-      matrix:
-        config:
-          # set the variable for the required tests:
-          # run_expensive_tests: true
-          # run_32bit_tests: true
-          # run_64bit_tests: true
-          # run_armv8_tests: true
-          # run_armv7_tests: true
-          - {
-              name: "Ubuntu 20.04 GCC",
-              os: ubuntu-20.04,
-              compiler: g++,
-              comp: gcc,
-              run_expensive_tests: true,
-              run_32bit_tests: true,
-              run_64bit_tests: true,
-              shell: 'bash {0}'
-            }
-          - {
-              name: "Ubuntu 20.04 Clang",
-              os: ubuntu-20.04,
-              compiler: clang++,
-              comp: clang,
-              run_32bit_tests: true,
-              run_64bit_tests: true,
-              shell: 'bash {0}'
-            }
-          - {
-              name: "Ubuntu 20.04 NDK armv8",
-              os: ubuntu-20.04,
-              compiler: aarch64-linux-android21-clang++,
-              comp: ndk,
-              run_armv8_tests: true,
-              shell: 'bash {0}'
-            }
-          - {
-              name: "Ubuntu 20.04 NDK armv7",
-              os: ubuntu-20.04,
-              compiler: armv7a-linux-androideabi21-clang++,
-              comp: ndk,
-              run_armv7_tests: true,
-              shell: 'bash {0}'
-            }
-          - {
-              name: "MacOS 10.15 Apple Clang",
-              os: macos-10.15,
-              compiler: clang++,
-              comp: clang,
-              run_64bit_tests: true,
-              shell: 'bash {0}'
-            }
-          - {
-              name: "MacOS 10.15 GCC 10",
-              os: macos-10.15,
-              compiler: g++-10,
-              comp: gcc,
-              run_64bit_tests: true,
-              shell: 'bash {0}'
-            }
-          - {
-              name: "Windows 2022 Mingw-w64 GCC x86_64",
-              os: windows-2022,
-              compiler: g++,
-              comp: mingw,
-              run_64bit_tests: true,
-              msys_sys: 'mingw64',
-              msys_env: 'x86_64-gcc',
-              shell: 'msys2 {0}'
-            }
-          - {
-              name: "Windows 2022 Mingw-w64 GCC i686",
-              os: windows-2022,
-              compiler: g++,
-              comp: mingw,
-              run_32bit_tests: true,
-              msys_sys: 'mingw32',
-              msys_env: 'i686-gcc',
-              shell: 'msys2 {0}'
-            }
-          - {
-              name: "Windows 2022 Mingw-w64 Clang x86_64",
-              os: windows-2022,
-              compiler: clang++,
-              comp: clang,
-              run_64bit_tests: true,
-              msys_sys: 'clang64',
-              msys_env: 'clang-x86_64-clang',
-              shell: 'msys2 {0}'
-            }
-
-    defaults:
-      run:
-        working-directory: src
-        shell: ${{ matrix.config.shell }}
+  Prerelease:
+    if: github.ref == 'refs/heads/master'
+    runs-on: ubuntu-latest
    steps:
-      - uses: actions/checkout@v2
+      - uses: actions/checkout@v3
        with:
          fetch-depth: 0

-      - name: Download required linux packages
-        if: runner.os == 'Linux'
+      # returns null if no pre-release exists
+      - name: Get Commit SHA of Latest Pre-release
        run: |
-          sudo apt update
-          sudo apt install expect valgrind g++-multilib qemu-user
+          # Install required packages
+          sudo apt-get update
+          sudo apt-get install -y curl jq

-      - name: Setup msys and install required packages
-        if: runner.os == 'Windows'
-        uses: msys2/setup-msys2@v2
+          echo "COMMIT_SHA=$(jq -r 'map(select(.prerelease)) | first | .tag_name' <<< $(curl -s https://api.github.com/repos/${{ github.repository_owner }}/Stockfish/releases))" >> $GITHUB_ENV
+
+      # delete old previous pre-release and tag
+      - uses: dev-drprasad/delete-tag-and-release@v0.2.1
+        if: env.COMMIT_SHA != 'null'
        with:
-          msystem: ${{matrix.config.msys_sys}}
-          install: mingw-w64-${{matrix.config.msys_env}} make git expect
+          tag_name: ${{ env.COMMIT_SHA }}
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

-      - name: Download the used network from the fishtest framework
-        run: |
-          make net
-
-      - name: Extract the bench number from the commit history
-        run: |
-          git log HEAD | grep "\b[Bb]ench[ :]\+[0-9]\{7\}" | head -n 1 | sed "s/[^0-9]*\([0-9]*\).*/\1/g" > git_sig
-          [ -s git_sig ] && echo "benchref=$(cat git_sig)" >> $GITHUB_ENV && echo "Reference bench:" $(cat git_sig) || echo "No bench found"
-
-      - name: Check compiler
-        run: |
-          export PATH=$PATH:$ANDROID_NDK_HOME/toolchains/llvm/prebuilt/linux-x86_64/bin
-          $COMPILER -v
-
-      - name: Test help target
-        run: |
-          make help
-
-      # x86-32 tests
-
-      - name: Test debug x86-32 build
-        if: ${{ matrix.config.run_32bit_tests }}
-        run: |
-          export CXXFLAGS="-Werror -D_GLIBCXX_DEBUG"
-          make clean
-          make -j2 ARCH=x86-32 optimize=no debug=yes build
-          ../tests/signature.sh $benchref
-
-      - name: Test x86-32 build
-        if: ${{ matrix.config.run_32bit_tests }}
-        run: |
-          make clean
-          make -j2 ARCH=x86-32 build
-          ../tests/signature.sh $benchref
-
-      - name: Test x86-32-sse41-popcnt build
-        if: ${{ matrix.config.run_32bit_tests }}
-        run: |
-          make clean
-          make -j2 ARCH=x86-32-sse41-popcnt build
-          ../tests/signature.sh $benchref
-
-      - name: Test x86-32-sse2 build
-        if: ${{ matrix.config.run_32bit_tests }}
-        run: |
-          make clean
-          make -j2 ARCH=x86-32-sse2 build
-          ../tests/signature.sh $benchref
-
-      - name: Test general-32 build
-        if: ${{ matrix.config.run_32bit_tests }}
-        run: |
-          make clean
-          make -j2 ARCH=general-32 build
-          ../tests/signature.sh $benchref
-
-      # x86-64 tests
-
-      - name: Test debug x86-64-modern build
-        if: ${{ matrix.config.run_64bit_tests }}
-        run: |
-          export CXXFLAGS="-Werror -D_GLIBCXX_DEBUG"
-          make clean
-          make -j2 ARCH=x86-64-modern optimize=no debug=yes build
-          ../tests/signature.sh $benchref
-
-      - name: Test x86-64-modern build
-        if: ${{ matrix.config.run_64bit_tests }}
-        run: |
-          make clean
-          make -j2 ARCH=x86-64-modern build
-          ../tests/signature.sh $benchref
-
-      - name: Test x86-64-ssse3 build
-        if: ${{ matrix.config.run_64bit_tests }}
-        run: |
-          make clean
-          make -j2 ARCH=x86-64-ssse3 build
-          ../tests/signature.sh $benchref
-
-      - name: Test x86-64-sse3-popcnt build
-        if: ${{ matrix.config.run_64bit_tests }}
-        run: |
-          make clean
-          make -j2 ARCH=x86-64-sse3-popcnt build
-          ../tests/signature.sh $benchref
-
-      - name: Test x86-64 build
-        if: ${{ matrix.config.run_64bit_tests }}
-        run: |
-          make clean
-          make -j2 ARCH=x86-64 build
-          ../tests/signature.sh $benchref
-
-      - name: Test general-64 build
-        if: matrix.config.run_64bit_tests
-        run: |
-          make clean
-          make -j2 ARCH=general-64 build
-          ../tests/signature.sh $benchref
-
-      # x86-64 with newer extensions tests
-
-      - name: Compile x86-64-avx2 build
-        if: ${{ matrix.config.run_64bit_tests }}
-        run: |
-          make clean
-          make -j2 ARCH=x86-64-avx2 build
-
-      - name: Compile x86-64-bmi2 build
-        if: ${{ matrix.config.run_64bit_tests }}
-        run: |
-          make clean
-          make -j2 ARCH=x86-64-bmi2 build
-
-      - name: Compile x86-64-avx512 build
-        if: ${{ matrix.config.run_64bit_tests }}
-        run: |
-          make clean
-          make -j2 ARCH=x86-64-avx512 build
-
-      - name: Compile x86-64-vnni512 build
-        if: ${{ matrix.config.run_64bit_tests }}
-        run: |
-          make clean
-          make -j2 ARCH=x86-64-vnni512 build
-
-      - name: Compile x86-64-vnni256 build
-        if: ${{ matrix.config.run_64bit_tests }}
-        run: |
-          make clean
-          make -j2 ARCH=x86-64-vnni256 build
-
-      # armv8 tests
-
-      - name: Test armv8 build
-        if: ${{ matrix.config.run_armv8_tests }}
-        run: |
-          export PATH=$ANDROID_NDK_HOME/toolchains/llvm/prebuilt/linux-x86_64/bin:$PATH
-          export LDFLAGS="-static -Wno-unused-command-line-argument"
-          make clean
-          make -j2 ARCH=armv8 build
-          ../tests/signature.sh $benchref
-
-      # armv7 tests
-
-      - name: Test armv7 build
-        if: ${{ matrix.config.run_armv7_tests }}
-        run: |
-          export PATH=$ANDROID_NDK_HOME/toolchains/llvm/prebuilt/linux-x86_64/bin:$PATH
-          export LDFLAGS="-static -Wno-unused-command-line-argument"
-          make clean
-          make -j2 ARCH=armv7 build
-          ../tests/signature.sh $benchref
-
-      - name: Test armv7-neon build
-        if: ${{ matrix.config.run_armv7_tests }}
-        run: |
-          export PATH=$ANDROID_NDK_HOME/toolchains/llvm/prebuilt/linux-x86_64/bin:$PATH
-          export LDFLAGS="-static -Wno-unused-command-line-argument"
-          make clean
-          make -j2 ARCH=armv7-neon build
-          ../tests/signature.sh $benchref
-
-      # Other tests
-
-      - name: Check perft and search reproducibility
-        if: ${{ matrix.config.run_64bit_tests }}
-        run: |
-          make clean
-          make -j2 ARCH=x86-64-modern build
-          ../tests/perft.sh
-          ../tests/reprosearch.sh
-
-      # Sanitizers
-
-      - name: Run under valgrind
-        if: ${{ matrix.config.run_expensive_tests }}
-        run: |
-          export CXXFLAGS="-O1 -fno-inline"
-          make clean
-          make -j2 ARCH=x86-64-modern debug=yes optimize=no build > /dev/null
-          ../tests/instrumented.sh --valgrind
-          ../tests/instrumented.sh --valgrind-thread
-
-      - name: Run with UB sanitizer
-        if: ${{ matrix.config.run_expensive_tests }}
-        run: |
-          export CXXFLAGS="-O1 -fno-inline"
-          make clean
-          make -j2 ARCH=x86-64-modern sanitize=undefined optimize=no debug=yes build > /dev/null
-          ../tests/instrumented.sh --sanitizer-undefined
-
-      - name: Run with thread sanitizer
-        if: ${{ matrix.config.run_expensive_tests }}
-        run: |
-          export CXXFLAGS="-O1 -fno-inline"
-          make clean
-          make -j2 ARCH=x86-64-modern sanitize=thread optimize=no debug=yes build > /dev/null
-          ../tests/instrumented.sh --sanitizer-thread
+  Sanitizers:
+    uses: ./.github/workflows/stockfish_sanitizers.yml
+  Tests:
+    uses: ./.github/workflows/stockfish_test.yml
+  Compiles:
+    uses: ./.github/workflows/stockfish_compile_test.yml
+  Binaries:
+    if: github.ref == 'refs/heads/master' || (startsWith(github.ref_name, 'sf_') && github.ref_type == 'tag')
+    uses: ./.github/workflows/stockfish_binaries.yml
+  ARM_Binaries:
+    if: github.ref == 'refs/heads/master' || (startsWith(github.ref_name, 'sf_') && github.ref_type == 'tag')
+    uses: ./.github/workflows/stockfish_arm_binaries.yml
@@ -0,0 +1,158 @@
+name: Stockfish
+on:
+  workflow_call:
+jobs:
+  Stockfish:
+    name: ${{ matrix.config.name }} ${{ matrix.binaries }}
+    runs-on: ${{ matrix.config.os }}
+    env:
+      COMPILER: ${{ matrix.config.compiler }}
+      COMP: ${{ matrix.config.comp }}
+      EMU: ${{ matrix.config.emu }}
+      EXT: ${{ matrix.config.ext }}
+      OS: ${{ matrix.config.os }}
+      BINARY: ${{ matrix.binaries }}
+    strategy:
+      matrix:
+        config:
+          - name: Android NDK aarch64
+            os: ubuntu-22.04
+            compiler: aarch64-linux-android21-clang++
+            emu: qemu-aarch64
+            comp: ndk
+            shell: bash {0}
+          - name: Android NDK arm
+            os: ubuntu-22.04
+            compiler: armv7a-linux-androideabi21-clang++
+            emu: qemu-arm
+            comp: ndk
+            shell: bash {0}
+        binaries:
+          - armv8
+          - armv7
+          - armv7-neon
+        exclude:
+          - binaries: armv8
+            config: {compiler: armv7a-linux-androideabi21-clang++}
+          - binaries: armv7
+            config: {compiler: aarch64-linux-android21-clang++}
+          - binaries: armv7-neon
+            config: {compiler: aarch64-linux-android21-clang++}
+    defaults:
+      run:
+        working-directory: src
+        shell: ${{ matrix.config.shell }}
+    steps:
+      - uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+
+      - name: Download required linux packages
+        if: runner.os == 'Linux'
+        run: |
+          sudo apt update
+          sudo apt install qemu-user
+
+      - name: Install NDK
+        if: runner.os == 'Linux'
+        run: |
+          if [ $COMP == ndk ]; then
+            NDKV="21.4.7075529"
+            ANDROID_ROOT=/usr/local/lib/android
+            ANDROID_SDK_ROOT=$ANDROID_ROOT/sdk
+            SDKMANAGER=$ANDROID_SDK_ROOT/cmdline-tools/latest/bin/sdkmanager
+            echo "y" | $SDKMANAGER "ndk;$NDKV"
+            ANDROID_NDK_ROOT=$ANDROID_SDK_ROOT/ndk/$NDKV
+            ANDROID_NDK_BIN=$ANDROID_NDK_ROOT/toolchains/llvm/prebuilt/linux-x86_64/bin
+            echo "ANDROID_NDK_BIN=$ANDROID_NDK_BIN" >> $GITHUB_ENV
+          fi
+
+      - name: Download the used network from the fishtest framework
+        run: make net
+
+      - name: Check compiler
+        run: |
+          if [ $COMP == ndk ]; then
+            export PATH=${{ env.ANDROID_NDK_BIN }}:$PATH
+          fi
+          $COMPILER -v
+
+      - name: Test help target
+        run: make help
+
+      - name: Check git
+        run: git --version
+
+      # Compile profile guided builds
+
+      - name: Compile ${{ matrix.binaries }} build
+        run: |
+          if [ $COMP == ndk ]; then
+            export PATH=${{ env.ANDROID_NDK_BIN }}:$PATH
+            export LDFLAGS="-static -Wno-unused-command-line-argument"
+          fi
+          make clean
+          make -j2 profile-build ARCH=$BINARY COMP=$COMP WINE_PATH=$EMU
+          make strip ARCH=$BINARY COMP=$COMP
+          mv ./stockfish$EXT ../stockfish-android-$BINARY$EXT
+
+      - name: Remove non src files
+        run: rm -f *.o .depend *.nnue
+
+      - name: Download wiki
+        run: |
+          git clone https://github.com/official-stockfish/Stockfish.wiki.git ../wiki
+          cd ../wiki
+          rm -rf .git
+
+      - name: Create tar archive.
+        run: |
+          cd ..
+          mkdir stockfish
+          cp -r wiki stockfish/
+          cp -r src stockfish/
+          cp stockfish-android-$BINARY$EXT stockfish/
+          cp "Top CPU Contributors.txt" stockfish/
+          cp Copying.txt stockfish/
+          cp AUTHORS stockfish/
+          cp CITATION.cff stockfish/
+          cp README.md stockfish/
+          tar -cvf stockfish-android-$BINARY.tar stockfish
+
+      - name: Upload binaries
+        uses: actions/upload-artifact@v3
+        with:
+          name: stockfish-android-${{ matrix.binaries }}
+          path: stockfish-android-${{ matrix.binaries }}.tar
+
+      - name: Release
+        if: startsWith(github.ref_name, 'sf_') && github.ref_type == 'tag'
+        uses: softprops/action-gh-release@v1
+        with:
+          files: stockfish-android-${{ matrix.binaries }}.tar
+
+      - name: Get last commit sha
+        id: last_commit
+        run: echo "COMMIT_SHA=$(git rev-parse HEAD | cut -c 1-8)" >> $GITHUB_ENV
+
+      - name: Get commit date
+        id: commit_date
+        run: echo "COMMIT_DATE=$(git show -s --date=format:'%Y%m%d' --format=%cd HEAD)" >> $GITHUB_ENV
+
+      # Make sure that an old ci which still runs on master doesn't recreate a prerelease
+      - name: Check Pullable Commits
+        id: check_commits
+        run: |
+          git fetch
+          CHANGES=$(git rev-list HEAD..origin/master --count)
+          echo "CHANGES=$CHANGES" >> $GITHUB_ENV
+
+      - name: Prerelease
+        if: github.ref_name == 'master' && env.CHANGES == '0'
+        continue-on-error: true
+        uses: softprops/action-gh-release@v1
+        with:
+          name: Stockfish dev-${{ env.COMMIT_DATE }}-${{ env.COMMIT_SHA }}
+          tag_name: stockfish-dev-${{ env.COMMIT_DATE }}-${{ env.COMMIT_SHA }}
+          prerelease: true
+          files: stockfish-android-${{ matrix.binaries }}.tar
@@ -0,0 +1,168 @@
+name: Stockfish
+on:
+  workflow_call:
+jobs:
+  Stockfish:
+    name: ${{ matrix.config.name }} ${{ matrix.binaries }}
+    runs-on: ${{ matrix.config.os }}
+    env:
+      COMPILER: ${{ matrix.config.compiler }}
+      COMP: ${{ matrix.config.comp }}
+      EXT: ${{ matrix.config.ext }}
+      NAME: ${{ matrix.config.simple_name }}
+      BINARY: ${{ matrix.binaries }}
+    strategy:
+      matrix:
+        config:
+          - name: Ubuntu 20.04 GCC
+            os: ubuntu-20.04
+            simple_name: ubuntu
+            compiler: g++
+            comp: gcc
+            shell: bash {0}
+            archive_ext: tar
+          - name: MacOS 12 Apple Clang
+            os: macos-12
+            simple_name: macos
+            compiler: clang++
+            comp: clang
+            shell: bash {0}
+            archive_ext: tar
+          - name: Windows 2022 Mingw-w64 GCC x86_64
+            os: windows-2022
+            simple_name: windows
+            compiler: g++
+            comp: mingw
+            msys_sys: mingw64
+            msys_env: x86_64-gcc
+            shell: msys2 {0}
+            ext: .exe
+            archive_ext: zip
+        binaries:
+          - x86-64
+          - x86-64-modern
+          - x86-64-avx2
+        exclude:
+          - binaries: x86-64-avx2
+            config: {os: macos-12}
+    defaults:
+      run:
+        working-directory: src
+        shell: ${{ matrix.config.shell }}
+    steps:
+      - uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+
+      - name: Download required linux packages
+        if: runner.os == 'Linux'
+        run: sudo apt update
+
+      - name: Setup msys and install required packages
+        if: runner.os == 'Windows'
+        uses: msys2/setup-msys2@v2
+        with:
+          msystem: ${{ matrix.config.msys_sys }}
+          install: mingw-w64-${{ matrix.config.msys_env }} make git zip
+
+      - name: Download the used network from the fishtest framework
+        run: make net
+
+      - name: Check compiler
+        run: $COMPILER -v
+
+      - name: Test help target
+        run: make help
+
+      - name: Check git
+        run: git --version
+
+      # Compile profile guided builds
+
+      - name: Compile ${{ matrix.binaries }} build
+        run: |
+          make -j2 profile-build ARCH=$BINARY COMP=$COMP
+          make strip ARCH=$BINARY COMP=$COMP
+          mv ./stockfish$EXT ../stockfish-$NAME-$BINARY$EXT
+
+      - name: Remove non src files
+        run: git clean -fx
+
+      - name: Download wiki
+        run: |
+          git clone https://github.com/official-stockfish/Stockfish.wiki.git ../wiki
+          rm -rf ../wiki/.git
+
+      - name: Create directory.
+        run: |
+          cd ..
+          mkdir stockfish
+          cp -r wiki stockfish/
+          cp -r src stockfish/
+          cp stockfish-$NAME-$BINARY$EXT stockfish/
+          cp "Top CPU Contributors.txt" stockfish/
+          cp Copying.txt stockfish/
+          cp AUTHORS stockfish/
+          cp CITATION.cff stockfish/
+          cp README.md stockfish/
+
+      - name: Create tar
+        if: runner.os != 'Windows'
+        run: |
+          cd ..
+          tar -cvf stockfish-$NAME-$BINARY.tar stockfish
+
+      - name: Create zip
+        if: runner.os == 'Windows'
+        run: |
+          cd ..
+          zip -r stockfish-$NAME-$BINARY.zip stockfish
+
+      - name: Upload binaries
+        if: runner.os != 'Windows'
+        uses: actions/upload-artifact@v3
+        with:
+          name: stockfish-${{ matrix.config.os }}-${{ matrix.binaries }}
+          path: stockfish-${{ matrix.config.simple_name }}-${{ matrix.binaries }}.tar
+
+      # Artifacts automatically get zipped
+      # to avoid double zipping, we use the unzipped directory
+      - name: Upload binaries
+        if: runner.os == 'Windows'
+        uses: actions/upload-artifact@v3
+        with:
+          name: stockfish-${{ matrix.config.os }}-${{ matrix.binaries }}
+          path: stockfish
+
+      - name: Release
+        if: startsWith(github.ref_name, 'sf_') && github.ref_type == 'tag'
+        uses: softprops/action-gh-release@v1
+        with:
+          files: stockfish-${{ matrix.config.simple_name }}-${{ matrix.binaries }}.${{ matrix.config.archive_ext }}
+
+      - name: Get last commit sha
+        id: last_commit
+        run: echo "COMMIT_SHA=$(git rev-parse HEAD | cut -c 1-8)" >> $GITHUB_ENV
+
+      - name: Get commit date
+        id: commit_date
+        run: echo "COMMIT_DATE=$(git show -s --date=format:'%Y%m%d' --format=%cd HEAD)" >> $GITHUB_ENV
+
+      # Make sure that an old ci which still runs on master doesn't recreate a prerelease
+      - name: Check Pullable Commits
+        id: check_commits
+        run: |
+          git fetch
+          CHANGES=$(git rev-list HEAD..origin/master --count)
+          echo "CHANGES=$CHANGES" >> $GITHUB_ENV
+
+      - name: Prerelease
+        if: github.ref_name == 'master' && env.CHANGES == '0'
+        continue-on-error: true
+        uses: softprops/action-gh-release@v1
+        with:
+          name: Stockfish dev-${{ env.COMMIT_DATE }}-${{ env.COMMIT_SHA }}
+          tag_name: stockfish-dev-${{ env.COMMIT_DATE }}-${{ env.COMMIT_SHA }}
+          prerelease: true
+          files: stockfish-${{ matrix.config.simple_name }}-${{ matrix.binaries }}.${{ matrix.config.archive_ext }}
+
@@ -0,0 +1,102 @@
+name: Stockfish
+on:
+  workflow_call:
+jobs:
+  Stockfish:
+    name: ${{ matrix.config.name }}
+    runs-on: ${{ matrix.config.os }}
+    env:
+      COMPILER: ${{ matrix.config.compiler }}
+      COMP: ${{ matrix.config.comp }}
+    strategy:
+      matrix:
+        config:
+          - name: Ubuntu 20.04 GCC
+            os: ubuntu-20.04
+            compiler: g++
+            comp: gcc
+            shell: bash {0}
+          - name: Ubuntu 20.04 Clang
+            os: ubuntu-20.04
+            compiler: clang++
+            comp: clang
+            shell: bash {0}
+          - name: MacOS 12 Apple Clang
+            os: macos-12
+            compiler: clang++
+            comp: clang
+            shell: bash {0}
+          - name: MacOS 12 GCC 11
+            os: macos-12
+            compiler: g++-11
+            comp: gcc
+            shell: bash {0}
+          - name: Windows 2022 Mingw-w64 GCC x86_64
+            os: windows-2022
+            compiler: g++
+            comp: mingw
+            msys_sys: mingw64
+            msys_env: x86_64-gcc
+            shell: msys2 {0}
+          - name: Windows 2022 Mingw-w64 Clang x86_64
+            os: windows-2022
+            compiler: clang++
+            comp: clang
+            msys_sys: clang64
+            msys_env: clang-x86_64-clang
+            shell: msys2 {0}
+
+    defaults:
+      run:
+        working-directory: src
+        shell: ${{ matrix.config.shell }}
+    steps:
+      - uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+
+      - name: Setup msys and install required packages
+        if: runner.os == 'Windows'
+        uses: msys2/setup-msys2@v2
+        with:
+          msystem: ${{matrix.config.msys_sys}}
+          install: mingw-w64-${{matrix.config.msys_env}} make git
+
+      - name: Download the used network from the fishtest framework
+        run: make net
+
+      - name: Check compiler
+        run: $COMPILER -v
+
+      - name: Test help target
+        run: make help
+
+      - name: Check git
+        run: git --version
+
+      # x86-64 with newer extensions tests
+
+      - name: Compile x86-64-avx2 build
+        run: |
+          make clean
+          make -j2 ARCH=x86-64-avx2 build
+
+      - name: Compile x86-64-bmi2 build
+        run: |
+          make clean
+          make -j2 ARCH=x86-64-bmi2 build
+
+      - name: Compile x86-64-avx512 build
+        run: |
+          make clean
+          make -j2 ARCH=x86-64-avx512 build
+
+      - name: Compile x86-64-vnni512 build
+        run: |
+          make clean
+          make -j2 ARCH=x86-64-vnni512 build
+
+      - name: Compile x86-64-vnni256 build
+        run: |
+          make clean
+          make -j2 ARCH=x86-64-vnni256 build
@@ -0,0 +1,66 @@
+name: Stockfish
+on:
+  workflow_call:
+jobs:
+  Stockfish:
+    name: ${{ matrix.sanitizers.name }}
+    runs-on: ${{ matrix.config.os }}
+    env:
+      COMPILER: ${{ matrix.config.compiler }}
+      COMP: ${{ matrix.config.comp }}
+      CXXFLAGS: "-Werror"
+    strategy:
+      matrix:
+        config:
+          - name: Ubuntu 20.04 GCC
+            os: ubuntu-20.04
+            compiler: g++
+            comp: gcc
+            shell: bash {0}
+        sanitizers:
+          - name: Run with thread sanitizer
+            make_option: sanitize=thread
+            instrumented_option: sanitizer-thread
+          - name: Run with UB sanitizer
+            make_option: sanitize=undefined
+            instrumented_option: sanitizer-undefined
+          - name: Run under valgrind
+            make_option: ""
+            instrumented_option: valgrind
+          - name: Run under valgrind-thread
+            make_option: ""
+            instrumented_option: valgrind-thread
+    defaults:
+      run:
+        working-directory: src
+        shell: ${{ matrix.config.shell }}
+    steps:
+      - uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+
+      - name: Download required linux packages
+        run: |
+          sudo apt update
+          sudo apt install expect valgrind g++-multilib
+
+      - name: Download the used network from the fishtest framework
+        run: make net
+
+      - name: Check compiler
+        run: $COMPILER -v
+
+      - name: Test help target
+        run: make help
+
+      - name: Check git
+        run: git --version
+
+      # Sanitizers
+
+      - name: ${{ matrix.sanitizers.name }}
+        run: |
+          export CXXFLAGS="-O1 -fno-inline"
+          make clean
+          make -j2 ARCH=x86-64-modern ${{ matrix.sanitizers.make_option }} debug=yes optimize=no build > /dev/null
+          ../tests/instrumented.sh --${{ matrix.sanitizers.instrumented_option }}
@@ -0,0 +1,256 @@
+name: Stockfish
+on:
+  workflow_call:
+jobs:
+  Stockfish:
+    name: ${{ matrix.config.name }}
+    runs-on: ${{ matrix.config.os }}
+    env:
+      COMPILER: ${{ matrix.config.compiler }}
+      COMP: ${{ matrix.config.comp }}
+      CXXFLAGS: "-Werror"
+    strategy:
+      matrix:
+        config:
+          - name: Ubuntu 20.04 GCC
+            os: ubuntu-20.04
+            compiler: g++
+            comp: gcc
+            run_32bit_tests: true
+            run_64bit_tests: true
+            shell: bash {0}
+          - name: Ubuntu 20.04 Clang
+            os: ubuntu-20.04
+            compiler: clang++
+            comp: clang
+            run_32bit_tests: true
+            run_64bit_tests: true
+            shell: bash {0}
+          - name: Android NDK aarch64
+            os: ubuntu-22.04
+            compiler: aarch64-linux-android21-clang++
+            comp: ndk
+            run_armv8_tests: true
+            shell: bash {0}
+          - name: Android NDK arm
+            os: ubuntu-22.04
+            compiler: armv7a-linux-androideabi21-clang++
+            comp: ndk
+            run_armv7_tests: true
+            shell: bash {0}
+          - name: MacOS 12 Apple Clang
+            os: macos-12
+            compiler: clang++
+            comp: clang
+            run_64bit_tests: true
+            shell: bash {0}
+          - name: MacOS 12 GCC 11
+            os: macos-12
+            compiler: g++-11
+            comp: gcc
+            run_64bit_tests: true
+            shell: bash {0}
+          - name: Windows 2022 Mingw-w64 GCC x86_64
+            os: windows-2022
+            compiler: g++
+            comp: mingw
+            run_64bit_tests: true
+            msys_sys: mingw64
+            msys_env: x86_64-gcc
+            shell: msys2 {0}
+          - name: Windows 2022 Mingw-w64 GCC i686
+            os: windows-2022
+            compiler: g++
+            comp: mingw
+            run_32bit_tests: true
+            msys_sys: mingw32
+            msys_env: i686-gcc
+            shell: msys2 {0}
+          - name: Windows 2022 Mingw-w64 Clang x86_64
+            os: windows-2022
+            compiler: clang++
+            comp: clang
+            run_64bit_tests: true
+            msys_sys: clang64
+            msys_env: clang-x86_64-clang
+            shell: msys2 {0}
+    defaults:
+      run:
+        working-directory: src
+        shell: ${{ matrix.config.shell }}
+    steps:
+      - uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+
+      - name: Download required linux packages
+        if: runner.os == 'Linux'
+        run: |
+          sudo apt update
+          sudo apt install expect valgrind g++-multilib qemu-user
+
+      - name: Install NDK
+        if: runner.os == 'Linux'
+        run: |
+          if [ $COMP == ndk ]; then
+            NDKV="21.4.7075529"
+            ANDROID_ROOT=/usr/local/lib/android
+            ANDROID_SDK_ROOT=$ANDROID_ROOT/sdk
+            SDKMANAGER=$ANDROID_SDK_ROOT/cmdline-tools/latest/bin/sdkmanager
+            echo "y" | $SDKMANAGER "ndk;$NDKV"
+            ANDROID_NDK_ROOT=$ANDROID_SDK_ROOT/ndk/$NDKV
+            ANDROID_NDK_BIN=$ANDROID_NDK_ROOT/toolchains/llvm/prebuilt/linux-x86_64/bin
+            echo "ANDROID_NDK_BIN=$ANDROID_NDK_BIN" >> $GITHUB_ENV
+          fi
+
+      - name: Setup msys and install required packages
+        if: runner.os == 'Windows'
+        uses: msys2/setup-msys2@v2
+        with:
+          msystem: ${{ matrix.config.msys_sys }}
+          install: mingw-w64-${{ matrix.config.msys_env }} make git expect
+
+      - name: Download the used network from the fishtest framework
+        run: make net
+
+      - name: Extract the bench number from the commit history
+        run: |
+          git log HEAD | grep -o "\b[Bb]ench[ :]\+[1-9][0-9]\{5,9\}\b" | head -n 1 | sed "s/[^0-9]//g" > git_sig
+          [ -s git_sig ] && echo "benchref=$(cat git_sig)" >> $GITHUB_ENV && echo "Reference bench:" $(cat git_sig) || echo "No bench found"
+
+      - name: Check compiler
+        run: |
+          if [ $COMP == ndk ]; then
+            export PATH=${{ env.ANDROID_NDK_BIN }}:$PATH
+          fi
+          $COMPILER -v
+
+      - name: Test help target
+        run: make help
+
+      - name: Check git
+        run: git --version
+
+      # x86-32 tests
+
+      - name: Test debug x86-32 build
+        if: ${{ matrix.config.run_32bit_tests }}
+        run: |
+          export CXXFLAGS="-Werror -D_GLIBCXX_DEBUG"
+          make clean
+          make -j2 ARCH=x86-32 optimize=no debug=yes build
+          ../tests/signature.sh $benchref
+
+      - name: Test x86-32 build
+        if: ${{ matrix.config.run_32bit_tests }}
+        run: |
+          make clean
+          make -j2 ARCH=x86-32 build
+          ../tests/signature.sh $benchref
+
+      - name: Test x86-32-sse41-popcnt build
+        if: ${{ matrix.config.run_32bit_tests }}
+        run: |
+          make clean
+          make -j2 ARCH=x86-32-sse41-popcnt build
+          ../tests/signature.sh $benchref
+
+      - name: Test x86-32-sse2 build
+        if: ${{ matrix.config.run_32bit_tests }}
+        run: |
+          make clean
+          make -j2 ARCH=x86-32-sse2 build
+          ../tests/signature.sh $benchref
+
+      - name: Test general-32 build
+        if: ${{ matrix.config.run_32bit_tests }}
+        run: |
+          make clean
+          make -j2 ARCH=general-32 build
+          ../tests/signature.sh $benchref
+
+      # x86-64 tests
+
+      - name: Test debug x86-64-modern build
+        if: ${{ matrix.config.run_64bit_tests }}
+        run: |
+          export CXXFLAGS="-Werror -D_GLIBCXX_DEBUG"
+          make clean
+          make -j2 ARCH=x86-64-modern optimize=no debug=yes build
+          ../tests/signature.sh $benchref
+
+      - name: Test x86-64-modern build
+        if: ${{ matrix.config.run_64bit_tests }}
+        run: |
+          make clean
+          make -j2 ARCH=x86-64-modern build
+          ../tests/signature.sh $benchref
+
+      - name: Test x86-64-ssse3 build
+        if: ${{ matrix.config.run_64bit_tests }}
+        run: |
+          make clean
+          make -j2 ARCH=x86-64-ssse3 build
+          ../tests/signature.sh $benchref
+
+      - name: Test x86-64-sse3-popcnt build
+        if: ${{ matrix.config.run_64bit_tests }}
+        run: |
+          make clean
+          make -j2 ARCH=x86-64-sse3-popcnt build
+          ../tests/signature.sh $benchref
+
+      - name: Test x86-64 build
+        if: ${{ matrix.config.run_64bit_tests }}
+        run: |
+          make clean
+          make -j2 ARCH=x86-64 build
+          ../tests/signature.sh $benchref
+
+      - name: Test general-64 build
+        if: matrix.config.run_64bit_tests
+        run: |
+          make clean
+          make -j2 ARCH=general-64 build
+          ../tests/signature.sh $benchref
+
+      # armv8 tests
+
+      - name: Test armv8 build
+        if: ${{ matrix.config.run_armv8_tests }}
+        run: |
+          export PATH=${{ env.ANDROID_NDK_BIN }}:$PATH
+          export LDFLAGS="-static -Wno-unused-command-line-argument"
+          make clean
+          make -j2 ARCH=armv8 build
+          ../tests/signature.sh $benchref
+
+      # armv7 tests
+
+      - name: Test armv7 build
+        if: ${{ matrix.config.run_armv7_tests }}
+        run: |
+          export PATH=${{ env.ANDROID_NDK_BIN }}:$PATH
+          export LDFLAGS="-static -Wno-unused-command-line-argument"
+          make clean
+          make -j2 ARCH=armv7 build
+          ../tests/signature.sh $benchref
+
+      - name: Test armv7-neon build
+        if: ${{ matrix.config.run_armv7_tests }}
+        run: |
+          export PATH=${{ env.ANDROID_NDK_BIN }}:$PATH
+          export LDFLAGS="-static -Wno-unused-command-line-argument"
+          make clean
+          make -j2 ARCH=armv7-neon build
+          ../tests/signature.sh $benchref
+
+      # Other tests
+
+      - name: Check perft and search reproducibility
+        if: ${{ matrix.config.run_64bit_tests }}
+        run: |
+          make clean
+          make -j2 ARCH=x86-64-modern build
+          ../tests/perft.sh
+          ../tests/reprosearch.sh
@@ -1,17 +1,15 @@
-# List of authors for Stockfish
-
-# Founders of the Stockfish project and fishtest infrastructure
+# 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, NNUE port
+# Authors and inventors of NNUE, training, and NNUE port
 Yu Nasu (ynasu87)
 Motohiro Isozaki (yaneurao)
 Hisayori Noda (nodchip)

-# all other authors of the code in alphabetical order
+# All other authors of Stockfish code (in alphabetical order)
 Aditya (absimaldata)
 Adrian Petrescu (apetresc)
 Ajith Chandy Jose (ajithcj)
@@ -21,6 +19,7 @@ Alexander Kure
 Alexander Pagel (Lolligerhans)
 Alfredo Menezes (lonfom169)
 Ali AlZhrani (Cooffe)
+Andreas Matthies (Matthies)
 Andrei Vetrov (proukornew)
 Andrew Grant (AndyGrant)
 Andrey Neporada (nepal)
@@ -35,18 +34,23 @@ 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)
-Dan Schmidt (dfannius)
 Daniel Axtens (daxtens)
 Daniel Dugovic (ddugovic)
+Dan Schmidt (dfannius)
 Dariusz Orzechowski (dorzechowski)
+David (dav1312)
 David Zar
 Daylen Yang (daylen)
 Deshawn Mohan-Smith (GoldenRare)
@@ -55,12 +59,12 @@ DiscanX
 Dominik Schlösser (domschl)
 double-beep
 Douglas Matos Gomes (dsmsgms)
+Dubslow
 Eduardo Cáceres (eduherminio)
 Eelco de Groot (KingDefender)
 Elvin Liu (solarlight2)
 erbsenzaehler
 Ernesto Gatti
-Linmiao Xu (linrock)
 Fabian Beuke (madnight)
 Fabian Fichter (ianfab)
 Fanael Linithien (Fanael)
@@ -73,32 +77,35 @@ 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
-Günther Demetz (pb00067, pb00068)
 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)
+Jared Kish (Kurtbusch, kurt22i)
 Jarrod Torriero (DU-jdto)
-Jean Gauthier (OuaisBla)
 Jean-Francois Romang (jromang)
+Jean Gauthier (OuaisBla)
 Jekaa
 Jerry Donald Watson (jerrydonaldwatson)
 jjoshua2
-Jonathan Calovski (Mysseno)
 Jonathan Buladas Dumale (SFisGOD)
+Jonathan Calovski (Mysseno)
+Jonathan McDermid (jonathanmcdermid)
 Joost VandeVondele (vondele)
-Jörg Oster (joergoster)
 Joseph Ellis (jhellis3)
 Joseph R. Prostko
+Jörg Oster (joergoster)
 Julian Willemer (NightlyKing)
 jundery
 Justin Blanchard (UncombedCoconut)
@@ -112,6 +119,7 @@ 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)
@@ -125,6 +133,8 @@ 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)
@@ -139,35 +149,39 @@ Mira
 Miroslav Fontán (Hexik)
 Moez Jellouli (MJZ1977)
 Mohammed Li (tthsqe12)
+Muzhen J (XInTheDark)
 Nathan Rugg (nmrugg)
-Nick Pelling (nickpelling)
+Nguyen Pham (nguyenpham)
 Nicklas Persson (NicklasPersson)
+Nick Pelling (nickpelling)
 Niklas Fiekas (niklasf)
 Nikolay Kostov (NikolayIT)
-Nguyen Pham (nguyenpham)
 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)
-pellanda
 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
-Richard Lloyd
+Reuven Peleg (R-Peleg)
+Richard Lloyd (Richard-Lloyd)
 Rodrigo Exterckötter Tjäder
-Ron Britvich (Britvich)
+Rodrigo Roim (roim)
 Ronald de Man (syzygy1, syzygy)
+Ron Britvich (Britvich)
 rqs
 Rui Coelho (ruicoelhopedro)
 Ryan Schmitt
@@ -178,20 +192,22 @@ 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)
-Prokop Randáček (ProkopRandacek)
+Syine Mineta (MinetaS)
 Thanar2
 thaspel
 theo77186
+Tomasz Sobczyk (Sopel97)
 Tom Truscott
 Tom Vijlbrief (tomtor)
-Tomasz Sobczyk (Sopel97)
 Torsten Franz (torfranz, tfranzer)
 Torsten Hellwig (Torom)
 Tracey Emery (basepr1me)
@@ -199,11 +215,12 @@ 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 the development of Stockfish!
+# an amazing and essential framework for Stockfish development!
 #
 # https://github.com/glinscott/fishtest/blob/master/AUTHORS
@@ -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
@@ -1,337 +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

-[![Build Status](https://github.com/official-stockfish/Stockfish/actions/workflows/stockfish.yml/badge.svg)](https://github.com/official-stockfish/Stockfish/actions)
-[![Build Status](https://ci.appveyor.com/api/projects/status/github/official-stockfish/Stockfish?branch=master&svg=true)](https://ci.appveyor.com/project/mcostalba/stockfish/branch/master)
+[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](https://stockfishchess.org) is a free, powerful UCI chess engine
-derived from Glaurung 2.1. Stockfish is not a complete chess program and requires a
-UCI-compatible graphical user interface (GUI) (e.g. XBoard with PolyGlot, Scid,
-Cute Chess, eboard, Arena, Sigma Chess, Shredder, Chess Partner or Fritz) in order
-to be used comfortably. Read the documentation for your GUI of choice for information
-about how to use Stockfish with it.
+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.

-The Stockfish engine features two evaluation functions for chess. The efficiently 
-updatable neural network (NNUE) based evaluation is the default and by far the strongest.
-The classical evaluation based on handcrafted terms remains available. The strongest 
-network is integrated in the binary and downloaded automatically during the build process.
-The NNUE evaluation benefits from the vector intrinsics available on most CPUs (sse2, 
-avx2, neon, or similar).
+See also the Stockfish [documentation][wiki-usage-link] for further usage help.

 ## Files

 This distribution of Stockfish consists of the following files:

-  * [Readme.md](https://github.com/official-stockfish/Stockfish/blob/master/README.md),
-    the file you are currently reading.
+  * [README.md][readme-link], the file you are currently reading.

-  * [Copying.txt](https://github.com/official-stockfish/Stockfish/blob/master/Copying.txt),
-    a text file containing the GNU General Public License version 3.
+  * [Copying.txt][license-link], a text file containing the GNU General Public
+    License version 3.

-  * [AUTHORS](https://github.com/official-stockfish/Stockfish/blob/master/AUTHORS), 
-    a text file with the list of authors for the project
+  * [AUTHORS][authors-link], a text file with the list of authors for the project.

-  * [src](https://github.com/official-stockfish/Stockfish/tree/master/src),
-    a subdirectory containing the full source code, including a Makefile
-    that can be used to compile Stockfish on Unix-like systems.
+  * [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 and available options
+## The UCI protocol

-The Universal Chess Interface (UCI) is a standard 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 as described
-in [the UCI protocol](https://www.shredderchess.com/download/div/uci.zip).
+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 UCI options available in Stockfish by typing
-`./stockfish uci` in a terminal, but the majority of users will typically see them and
-change them via a chess GUI. This is a list of available UCI options in Stockfish:
+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.

-  * #### Threads
-    The number of CPU threads used for searching a position. For best performance, set
-    this equal to the number of CPU cores available.
+For more information on UCI or debug commands, see our [documentation][wiki-commands-link].

-  * #### Hash
-    The size of the hash table in MB. It is recommended to set Hash after setting Threads.
+## Compiling Stockfish

-  * #### Clear Hash
-    Clear the hash table.
+Stockfish has support for 32 or 64-bit CPUs, certain hardware instructions,
+big-endian machines such as Power PC, and other platforms.

-  * #### Ponder
-    Let Stockfish ponder its next move while the opponent is thinking.
-
-  * #### MultiPV
-    Output the N best lines (principal variations, PVs) when searching.
-    Leave at 1 for best performance.
-
-  * #### Use NNUE
-    Toggle between the NNUE and classical evaluation functions. If set to "true",
-    the network parameters must be available to load from file (see also EvalFile),
-    if they are not embedded in the binary.
-
-  * #### EvalFile
-    The name of the file of the NNUE evaluation parameters. Depending on the GUI the
-    filename might have to include the full path to the folder/directory that contains
-    the file. Other locations, such as the directory that contains the binary and the
-    working directory, are also searched.
-
-  * #### UCI_AnalyseMode
-    An option handled by your GUI.
-
-  * #### UCI_Chess960
-    An option handled by your GUI. If true, Stockfish will play Chess960.
-
-  * #### UCI_ShowWDL
-    If enabled, show approximate WDL statistics as part of the engine output.
-    These WDL numbers model expected game outcomes for a given evaluation and
-    game ply for engine self-play at fishtest LTC conditions (60+0.6s per game).
-
-  * #### UCI_LimitStrength
-    Enable weaker play aiming for an Elo rating as set by UCI_Elo. This option overrides Skill Level.
-
-  * #### UCI_Elo
-    If enabled by UCI_LimitStrength, aim for an engine strength of the given Elo.
-    This Elo rating has been calibrated at a time control of 60s+0.6s and anchored to CCRL 40/4.
-
-  * #### Skill Level
-    Lower the Skill Level in order to make Stockfish play weaker (see also UCI_LimitStrength).
-    Internally, MultiPV is enabled, and with a certain probability depending on the Skill Level a
-    weaker move will be played.
-
-  * #### SyzygyPath
-    Path to the folders/directories storing the Syzygy tablebase files. Multiple
-    directories are to be separated by ";" on Windows and by ":" on Unix-based
-    operating systems. Do not use spaces around the ";" or ":".
-
-    Example: `C:\tablebases\wdl345;C:\tablebases\wdl6;D:\tablebases\dtz345;D:\tablebases\dtz6`
-
-    It is recommended to store .rtbw files on an SSD. There is no loss in storing
-    the .rtbz files on a regular HDD. It is recommended to verify all md5 checksums
-    of the downloaded tablebase files (`md5sum -c checksum.md5`) as corruption will
-    lead to engine crashes.
-
-  * #### SyzygyProbeDepth
-    Minimum remaining search depth for which a position is probed. Set this option
-    to a higher value to probe less aggressively if you experience too much slowdown
-    (in terms of nps) due to tablebase probing.
-
-  * #### Syzygy50MoveRule
-    Disable to let fifty-move rule draws detected by Syzygy tablebase probes count
-    as wins or losses. This is useful for ICCF correspondence games.
-
-  * #### SyzygyProbeLimit
-    Limit Syzygy tablebase probing to positions with at most this many pieces left
-    (including kings and pawns).
-
-  * #### Move Overhead
-    Assume a time delay of x ms due to network and GUI overheads. This is useful to
-    avoid losses on time in those cases.
-
-  * #### Slow Mover
-    Lower values will make Stockfish take less time in games, higher values will
-    make it think longer.
-
-  * #### nodestime
-    Tells the engine to use nodes searched instead of wall time to account for
-    elapsed time. Useful for engine testing.
-
-  * #### Debug Log File
-    Write all communication to and from the engine into a text file.
-
-For developers the following non-standard commands might be of interest, mainly useful for debugging:
-
-  * #### bench *ttSize threads limit fenFile limitType evalType*
-    Performs a standard benchmark using various options. The signature of a version 
-    (standard node count) is obtained using all defaults. `bench` is currently 
-    `bench 16 1 13 default depth mixed`.
-
-  * #### compiler
-    Give information about the compiler and environment used for building a binary.
-
-  * #### d
-    Display the current position, with ascii art and fen.
-
-  * #### eval
-    Return the evaluation of the current position.
-
-  * #### export_net [filename]
-    Exports the currently loaded network to a file.
-    If the currently loaded network is the embedded network and the filename
-    is not specified then the network is saved to the file matching the name
-    of the embedded network, as defined in evaluate.h.
-    If the currently loaded network is not the embedded network (some net set
-    through the UCI setoption) then the filename parameter is required and the
-    network is saved into that file.
-
-  * #### flip
-    Flips the side to move.
-
-
-## A note on classical evaluation versus NNUE evaluation
-
-Both approaches assign a value to a position that is used in alpha-beta (PVS) search
-to find the best move. The classical evaluation computes this value as a function
-of various chess concepts, handcrafted by experts, tested and tuned using fishtest.
-The NNUE evaluation computes this value with a neural network based on basic
-inputs (e.g. piece positions only). The network is optimized and trained
-on the evaluations of millions of positions at moderate search depth.
-
-The NNUE evaluation was first introduced in shogi, and ported to Stockfish afterward.
-It can be evaluated efficiently on CPUs, and exploits the fact that only parts
-of the neural network need to be updated after a typical chess move.
-[The nodchip repository](https://github.com/nodchip/Stockfish) provided the first 
-version of the needed tools to train and develop the NNUE networks. Today, more 
-advanced training tools are available in 
-[the nnue-pytorch repository](https://github.com/glinscott/nnue-pytorch/), 
-while data generation tools are available in 
-[a dedicated branch](https://github.com/official-stockfish/Stockfish/tree/tools).
-
-On CPUs supporting modern vector instructions (avx2 and similar), the NNUE evaluation
-results in much stronger playing strength, even if the nodes per second computed by
-the engine is somewhat lower (roughly 80% of nps is typical).
-
-Notes:
-
-1) the NNUE evaluation depends on the Stockfish binary and the network parameter file 
-(see the EvalFile UCI option). Not every parameter file is compatible with a given
-Stockfish binary, but the default value of the EvalFile UCI option is the name of a 
-network that is guaranteed to be compatible with that binary.
-
-2) to use the NNUE evaluation, the additional data file with neural network parameters
-needs to be available. Normally, this file is already embedded in the binary or it can
-be downloaded. The filename for the default (recommended) net can be found as the default
-value of the `EvalFile` UCI option, with the format `nn-[SHA256 first 12 digits].nnue`
-(for instance, `nn-c157e0a5755b.nnue`). This file can be downloaded from
-```
-https://tests.stockfishchess.org/api/nn/[filename]
-```
-replacing `[filename]` as needed.
-
-## What to expect from the Syzygy tablebases?
-
-If the engine is searching a position that is not in the tablebases (e.g.
-a position with 8 pieces), it will access the tablebases during the search.
-If the engine reports a very large score (typically 153.xx), this means
-it has found a winning line into a tablebase position.
-
-If the engine is given a position to search that is in the tablebases, it
-will use the tablebases at the beginning of the search to preselect all
-good moves, i.e. all moves that preserve the win or preserve the draw while
-taking into account the 50-move rule.
-It will then perform a search only on those moves. **The engine will not move
-immediately**, unless there is only a single good move. **The engine likely
-will not report a mate score, even if the position is known to be won.**
-
-It is therefore clear that this behaviour is not identical to what one might
-be used to with Nalimov tablebases. There are technical reasons for this
-difference, the main technical reason being that Nalimov tablebases use the
-DTM metric (distance-to-mate), while the Syzygy tablebases use a variation of the
-DTZ metric (distance-to-zero, zero meaning any move that resets the 50-move
-counter). This special metric is one of the reasons that the Syzygy tablebases are
-more compact than Nalimov tablebases, while still storing all information
-needed for optimal play and in addition being able to take into account
-the 50-move rule.
-
-## Large Pages
-
-Stockfish supports large pages on Linux and Windows. Large pages make
-the hash access more efficient, improving the engine speed, especially
-on large hash sizes. Typical increases are 5..10% in terms of nodes per
-second, but speed increases up to 30% have been measured. The support is
-automatic. Stockfish attempts to use large pages when available and
-will fall back to regular memory allocation when this is not the case.
-
-### Support on Linux
-
-Large page support on Linux is obtained by the Linux kernel
-transparent huge pages functionality. Typically, transparent huge pages
-are already enabled, and no configuration is needed.
-
-### Support on Windows
-
-The use of large pages requires "Lock Pages in Memory" privilege. See
-[Enable the Lock Pages in Memory Option (Windows)](https://docs.microsoft.com/en-us/sql/database-engine/configure-windows/enable-the-lock-pages-in-memory-option-windows)
-on how to enable this privilege, then run [RAMMap](https://docs.microsoft.com/en-us/sysinternals/downloads/rammap)
-to double-check that large pages are used. We suggest that you reboot
-your computer after you have enabled large pages, because long Windows
-sessions suffer from memory fragmentation, which may prevent Stockfish
-from getting large pages: a fresh session is better in this regard.
-
-## Compiling Stockfish yourself from the sources
-
-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.
+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 help
-    make net
-    make build ARCH=x86-64-modern
+cd src
+make -j build ARCH=x86-64-modern
 ```

-When not using the Makefile to compile (for instance, with Microsoft MSVC) you
-need to manually set/unset some switches in the compiler command line; see
-file *types.h* for a quick reference.
+Detailed compilation instructions for all platforms can be found in our
+[documentation][wiki-compile-link].

-When reporting an issue or a bug, please tell us which Stockfish version
-and which compiler you used to create your executable. This information
-can be found by typing the following command in a console:
-
-```
-    ./stockfish compiler
-```
-
-## Understanding the code base and participating in the project
-
-Stockfish's improvement over the last decade has been a great community
-effort. There are a few ways to help contribute to its growth.
+## Contributing

 ### Donating hardware

-Improving Stockfish requires a massive amount of testing. You can donate
-your hardware resources by installing the [Fishtest Worker](https://github.com/glinscott/fishtest/wiki/Running-the-worker:-overview)
-and view the current tests on [Fishtest](https://tests.stockfishchess.org/tests).
+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

-If you want to help improve the code, there are several valuable resources:
-
-* [In this wiki,](https://www.chessprogramming.org) many techniques used in
+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 section on Stockfish](https://www.chessprogramming.org/Stockfish)
-describes many features and techniques used by Stockfish. However, it is
-generic rather than being focused on Stockfish's precise implementation.
-Nevertheless, a helpful resource.
-
-* The latest source can always be found on [GitHub](https://github.com/official-stockfish/Stockfish).
-Discussions about Stockfish take place these days mainly in the [FishCooking](https://groups.google.com/forum/#!forum/fishcooking)
-group and on the [Stockfish Discord channel](https://discord.gg/nv8gDtt).
-The engine testing is done on [Fishtest](https://tests.stockfishchess.org/tests).
-If you want to help improve Stockfish, please read this [guideline](https://github.com/glinscott/fishtest/wiki/Creating-my-first-test)
+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**
-(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.
+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 the GPL v3.
+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.

-For full details, read the copy of the GPL v3 found in the file named
-[*Copying.txt*](https://github.com/official-stockfish/Stockfish/blob/master/Copying.txt).
+
+[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
@@ -1,235 +1,267 @@
-Contributors to Fishtest with >10,000 CPU hours, as of 2022-04-14.
+Contributors to Fishtest with >10,000 CPU hours, as of 2023-06-20.
 Thank you!

 Username                                CPU Hours     Games played
 ------------------------------------------------------------------
-noobpwnftw                               31714850       2267266129
-mlang                                     2954099        198421098
-technologov                               2324150        102449398
-dew                                       1670874         99276012
-grandphish2                               1134273         68070459
-okrout                                     901194         77738874
-TueRens                                    821388         50207666
-tvijlbrief                                 795993         51894442
-pemo                                       744463         32486677
-JojoM                                      724378         43660674
+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
-linrock                                    626939         17408017
-gvreuls                                    534079         34352532
-cw                                         507221         34006775
-fastgm                                     489749         29344518
+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
-CSU_Dynasty                                424643         28525220
-ctoks                                      415771         27364603
-oz                                         369200         27017658
-bcross                                     342642         23671289
+maximmasiutin                              424795         26577722
+bcross                                     415722         29060963
+olafm                                      395922         32268020
+rpngn                                      348378         24560289
+velislav                                   342567         22138992
 Fisherman                                  327231         21829379
-velislav                                   325670         20911076
-leszek                                     321295         19874113
-Dantist                                    274747         16910258
-mgrabiak                                   237604         15418700
-robal                                      217959         13840386
-glinscott                                  217799         13780820
-nordlandia                                 211692         13484886
-drabel                                     201967         13798360
+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
-mhoram                                     194862         12261809
 Thanar                                     179852         12365359
 vdv                                        175544          9904472
 spams                                      157128         10319326
-rpngn                                      154081          9652139
-marrco                                     150300          9402229
 sqrt2                                      147963          9724586
-vdbergh                                    137430          8955097
+DesolatedDodo                              146350          9536172
+Calis007                                   143165          9478764
+vdbergh                                    138650          9064413
 CoffeeOne                                  137100          5024116
+armo9494                                   136191          9460264
 malala                                     136182          8002293
 xoto                                       133759          9159372
-davar                                      125240          8117121
+davar                                      129023          8376525
+DMBK                                       122960          8980062
 dsmith                                     122059          7570238
-amicic                                     119659          7937885
+amicic                                     119661          7938029
 Data                                       113305          8220352
 BrunoBanani                                112960          7436849
-CypressChess                               108321          7759588
-DesolatedDodo                              106811          6776980
+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
-skiminki                                    98123          6478402
+cuistot                                     98853          6069816
+bigpen0r                                    94809          6529203
 brabos                                      92118          6186135
-cuistot                                     90358          5351004
+Wolfgang                                    91939          6105872
 psk                                         89957          5984901
-racerschmacer                               85712          6119648
+sschnee                                     88235          5268000
+racerschmacer                               85805          6122790
+Fifis                                       85722          5709729
+Dubslow                                     84986          6042456
 Vizvezdenec                                 83761          5344740
-zeryl                                       83680          5250995
-sschnee                                     83003          4840890
 0x3C33                                      82614          5271253
 BRAVONE                                     81239          5054681
 nssy                                        76497          5259388
+jromang                                     76106          5236025
 teddybaer                                   75125          5407666
-jromang                                     74796          5175825
+tolkki963                                   74762          5149662
+megaman7de                                  74351          4940352
+Wencey                                      74181          4711488
 Pking_cda                                   73776          5293873
-Calis007                                    72477          4088576
+yurikvelo                                   73150          5004382
+markkulix                                   72607          5304642
+Bobo1239                                    70579          4794999
 solarlight                                  70517          5028306
 dv8silencer                                 70287          3883992
-Bobo1239                                    68515          4652287
 manap                                       66273          4121774
-yurikvelo                                   65716          4457300
 tinker                                      64333          4268790
-Wolfgang                                    62644          3817410
 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
-megaman7de                                  52434          3243016
-MaxKlaxxMiner                               51977          3153032
+javran                                      53785          4627608
 finfish                                     51360          3370515
 eva42                                       51272          3599691
-eastorwest                                  51058          3451555
+eastorwest                                  51117          3454811
 rap                                         49985          3219146
-pb00067                                     49727          3298270
-Spprtr                                      48920          3161711
-bigpen0r                                    47667          3336927
+pb00067                                     49733          3298934
+OuaisBla                                    48626          3445134
 ronaldjerum                                 47654          3240695
 biffhero                                    46564          3111352
-Fifis                                       45843          3088497
 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
-armo9494                                    39064          2832326
-oryx                                        38867          2976992
+Garf                                        37741          2999686
 SC                                          37299          2731694
-Garf                                        37213          2986270
-tolkki963                                   37059          2154330
 csnodgrass                                  36207          2688994
-jmdana                                      36157          2210661
 strelock                                    34716          2074055
-DMBK                                        34010          2482916
+szupaw                                      34102          2880346
 EthanOConnor                                33370          2090311
 slakovv                                     32915          2021889
-gopeto                                      30993          2028106
+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
-ncfish1                                     29105          1704011
-belzedar94                                  27935          1789106
-OuaisBla                                    27636          1578800
+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
-rodneyc                                     24275          1410450
+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                                     23339          1602636
-Ente                                        23270          1651432
-cisco2015                                   22897          1762669
-MarcusTullius                               22688          1274821
+Norabor                                     23371          1603244
+cisco2015                                   22920          1763301
+jsys14                                      22824          1591906
 Zirie                                       22542          1472937
 team-oh                                     22272          1636708
+Roady                                       22220          1465606
 MazeOfGalious                               21978          1629593
-sg4032                                      21947          1643265
+sg4032                                      21947          1643353
 ianh2105                                    21725          1632562
 xor12                                       21628          1680365
 dex                                         21612          1467203
 nesoneg                                     21494          1463031
-Roady                                       21323          1433822
+user213718                                  21454          1404128
 sphinx                                      21211          1384728
-user213718                                  21196          1397710
-spcc                                        21065          1311338
+AndreasKrug                                 21097          1634811
 jjoshua2                                    21001          1423089
+Zake9298                                    20938          1565848
 horst.prack                                 20878          1465656
 0xB00B1ES                                   20590          1208666
 j3corre                                     20405           941444
-kdave                                       20364          1389254
 Adrian.Schmidt123                           20316          1281436
-Ulysses                                     20217          1351500
-markkulix                                   19976          1115258
 wei                                         19973          1745989
+notchris                                    19958          1800128
+Serpensin                                   19840          1697528
+Gaster319                                   19712          1677310
+fishtester                                  19617          1257388
 rstoesser                                   19569          1293588
 eudhan                                      19274          1283717
-fishtester                                  18995          1238686
+votoanthuan                                 19108          1609992
 vulcan                                      18871          1729392
+Karpovbot                                   18766          1053178
+qoo_charly_cai                              18543          1284937
 jundery                                     18445          1115855
-iisiraider                                  18247          1101015
 ville                                       17883          1384026
 chris                                       17698          1487385
 purplefishies                               17595          1092533
-dju                                         17353           978595
-Wencey                                      17125           805964
+dju                                         17414           981289
+iisiraider                                  17275          1049015
 DragonLord                                  17014          1162790
-thirdlife                                   16996           447356
+redstone59                                  16842          1461780
+Alb11747                                    16787          1213926
 IgorLeMasson                                16064          1147232
+Karby                                       15982           979610
+scuzzi                                      15757           968735
 ako027ako                                   15671          1173203
-AndreasKrug                                 15550          1194497
 Nikolay.IT                                  15154          1068349
 Andrew Grant                                15114           895539
-scuzzi                                      14928           953313
+Naven94                                     15054           834762
 OssumOpossum                                14857          1007129
-Karby                                       14808           867120
-jsys14                                      14652           855642
+ZacHFX                                      14783          1021842
 enedene                                     14476           905279
-bpfliegel                                   14298           884523
+bpfliegel                                   14233           882523
 mpx86                                       14019           759568
 jpulman                                     13982           870599
+Skiff84                                     13826           721996
 crocogoat                                   13803          1117422
-joster                                      13794           950160
 Nesa92                                      13786          1114691
+joster                                      13710           946160
 mbeier                                      13650          1044928
 Hjax                                        13535           915487
+Nullvalue                                   13468          1140498
 Dark_wizzie                                 13422          1007152
-Jopo12321                                   13367           678852
 Rudolphous                                  13244           883140
+pirt                                        13100          1009897
 Machariel                                   13010           863104
+infinigon                                   12991           943216
 mabichito                                   12903           749391
 thijsk                                      12886           722107
 AdrianSA                                    12860           804972
-infinigon                                   12807           937332
 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
-pirt                                        12008           923149
+dbernier                                    12103           860824
+getraideBFF                                 12072          1024966
 stocky                                      11954           699440
 mschmidt                                    11941           803401
-dbernier                                    11609           818636
-Maxim                                       11543           836024
+MooTheCow                                   11870           773598
+FormazChar                                  11766           885707
+whelanh                                     11557           245188
+3cho                                        11494          1031076
 infinity                                    11470           727027
-aga                                         11409           695071
+aga                                         11412           695127
 torbjo                                      11395           729145
 Thomas A. Anderson                          11372           732094
 savage84                                    11358           670860
-FormazChar                                  11349           850327
 d64                                         11263           789184
-MooTheCow                                   11237           720174
+ali-al-zhrani                               11245           779246
 snicolet                                    11106           869170
-ali-al-zhrani                               11098           768494
-whelanh                                     11067           235676
-Jackfish                                    10978           720078
-deflectooor                                 10886           520116
+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
-Garruk                                      10334           704065
-ols                                         10259           570669
-lbraesch                                    10252           647825
-qoo_charly_cai                              10212           620407
-Naven94                                     10069           503192
@@ -1,5 +1,5 @@
 # Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-# Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+# 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
@@ -69,32 +69,33 @@ VPATH = syzygy:nnue:nnue/features
 ### Section 2. High-level Configuration
 ### ==========================================================================
 #
-# flag                --- Comp switch      --- Description
+# flag                --- Comp switch        --- Description
 # ----------------------------------------------------------------------------
 #
-# debug = yes/no      --- -DNDEBUG         --- Enable/Disable debug mode
+# debug = yes/no      --- -DNDEBUG           --- Enable/Disable debug mode
 # sanitize = none/<sanitizer> ... (-fsanitize )
-#                     --- ( undefined )    --- enable undefined behavior checks
-#                     --- ( thread    )    --- enable threading error checks
-#                     --- ( address   )    --- enable memory access checks
-#                     --- ...etc...        --- see compiler documentation for supported sanitizers
-# optimize = yes/no   --- (-O3/-fast etc.) --- Enable/Disable optimizations
-# arch = (name)       --- (-arch)          --- Target architecture
-# bits = 64/32        --- -DIS_64BIT       --- 64-/32-bit operating system
-# prefetch = yes/no   --- -DUSE_PREFETCH   --- Use prefetch asm-instruction
-# popcnt = yes/no     --- -DUSE_POPCNT     --- Use popcnt asm-instruction
-# pext = yes/no       --- -DUSE_PEXT       --- Use pext x86_64 asm-instruction
-# sse = yes/no        --- -msse            --- Use Intel Streaming SIMD Extensions
-# mmx = yes/no        --- -mmmx            --- Use Intel MMX instructions
-# sse2 = yes/no       --- -msse2           --- Use Intel Streaming SIMD Extensions 2
-# ssse3 = yes/no      --- -mssse3          --- Use Intel Supplemental Streaming SIMD Extensions 3
-# sse41 = yes/no      --- -msse4.1         --- Use Intel Streaming SIMD Extensions 4.1
-# avx2 = yes/no       --- -mavx2           --- Use Intel Advanced Vector Extensions 2
-# avxvnni = yes/no    --- -mavxvnni        --- Use Intel Vector Neural Network Instructions AVX
-# avx512 = yes/no     --- -mavx512bw       --- Use Intel Advanced Vector Extensions 512
-# vnni256 = yes/no    --- -mavx512vnni     --- Use Intel Vector Neural Network Instructions 256
-# vnni512 = yes/no    --- -mavx512vnni     --- Use Intel Vector Neural Network Instructions 512
-# neon = yes/no       --- -DUSE_NEON       --- Use ARM SIMD architecture
+#                     --- ( undefined )      --- enable undefined behavior checks
+#                     --- ( thread    )      --- enable threading error checks
+#                     --- ( address   )      --- enable memory access checks
+#                     --- ...etc...          --- see compiler documentation for supported sanitizers
+# optimize = yes/no   --- (-O3/-fast etc.)   --- Enable/Disable optimizations
+# arch = (name)       --- (-arch)            --- Target architecture
+# bits = 64/32        --- -DIS_64BIT         --- 64-/32-bit operating system
+# prefetch = yes/no   --- -DUSE_PREFETCH     --- Use prefetch asm-instruction
+# popcnt = yes/no     --- -DUSE_POPCNT       --- Use popcnt asm-instruction
+# pext = yes/no       --- -DUSE_PEXT         --- Use pext x86_64 asm-instruction
+# sse = yes/no        --- -msse              --- Use Intel Streaming SIMD Extensions
+# mmx = yes/no        --- -mmmx              --- Use Intel MMX instructions
+# sse2 = yes/no       --- -msse2             --- Use Intel Streaming SIMD Extensions 2
+# ssse3 = yes/no      --- -mssse3            --- Use Intel Supplemental Streaming SIMD Extensions 3
+# sse41 = yes/no      --- -msse4.1           --- Use Intel Streaming SIMD Extensions 4.1
+# avx2 = yes/no       --- -mavx2             --- Use Intel Advanced Vector Extensions 2
+# avxvnni = yes/no    --- -mavxvnni          --- Use Intel Vector Neural Network Instructions AVX
+# avx512 = yes/no     --- -mavx512bw         --- Use Intel Advanced Vector Extensions 512
+# vnni256 = yes/no    --- -mavx256vnni       --- Use Intel Vector Neural Network Instructions 512 with 256bit operands
+# vnni512 = yes/no    --- -mavx512vnni       --- Use Intel Vector Neural Network Instructions 512
+# neon = yes/no       --- -DUSE_NEON         --- Use ARM SIMD architecture
+# dotprod = yes/no    --- -DUSE_NEON_DOTPROD --- Use ARM advanced SIMD Int8 dot product instructions
 #
 # Note that Makefile is space sensitive, so when adding new architectures
 # or modifying existing flags, you have to make sure there are no extra spaces
@@ -116,7 +117,7 @@ ifeq ($(ARCH), $(filter $(ARCH), \
                 x86-64-vnni512 x86-64-vnni256 x86-64-avx512 x86-64-avxvnni x86-64-bmi2 \
                 x86-64-avx2 x86-64-sse41-popcnt x86-64-modern x86-64-ssse3 x86-64-sse3-popcnt \
                 x86-64 x86-32-sse41-popcnt x86-32-sse2 x86-32 ppc-64 ppc-32 e2k \
-                 armv7 armv7-neon armv8 apple-silicon general-64 general-32))
+                 armv7 armv7-neon armv8 armv8-dotprod apple-silicon general-64 general-32 riscv64))
   SUPPORTED_ARCH=true
 else
   SUPPORTED_ARCH=false
@@ -140,6 +141,7 @@ avx512 = no
 vnni256 = no
 vnni512 = no
 neon = no
+dotprod = no
 arm_version = 0
 STRIP = strip

@@ -308,11 +310,21 @@ ifeq ($(ARCH),armv8)
 	arm_version = 8
 endif

+ifeq ($(ARCH),armv8-dotprod)
+	arch = armv8
+	prefetch = yes
+	popcnt = yes
+	neon = yes
+	dotprod = yes
+	arm_version = 8
+endif
+
 ifeq ($(ARCH),apple-silicon)
 	arch = arm64
 	prefetch = yes
 	popcnt = yes
 	neon = yes
+	dotprod = yes
 	arm_version = 8
 endif

@@ -338,16 +350,26 @@ ifeq ($(findstring e2k,$(ARCH)),e2k)
 	popcnt = yes
 endif

+ifeq ($(ARCH),riscv64)
+	arch = riscv64
 endif
+endif
+

 ### ==========================================================================
 ### Section 3. Low-level Configuration
 ### ==========================================================================

 ### 3.1 Selecting compiler (default = gcc)
-CXXFLAGS += -Wall -Wcast-qual -fno-exceptions -std=c++17 $(EXTRACXXFLAGS)
-DEPENDFLAGS += -std=c++17
-LDFLAGS += $(EXTRALDFLAGS)
+ifeq ($(MAKELEVEL),0)
+       export ENV_CXXFLAGS := $(CXXFLAGS)
+       export ENV_DEPENDFLAGS := $(DEPENDFLAGS)
+       export ENV_LDFLAGS := $(LDFLAGS)
+endif
+
+CXXFLAGS = $(ENV_CXXFLAGS) -Wall -Wcast-qual -fno-exceptions -std=c++17 $(EXTRACXXFLAGS)
+DEPENDFLAGS = $(ENV_DEPENDFLAGS) -std=c++17
+LDFLAGS = $(ENV_LDFLAGS) $(EXTRALDFLAGS)

 ifeq ($(COMP),)
 	COMP=gcc
@@ -356,13 +378,16 @@ endif
 ifeq ($(COMP),gcc)
 	comp=gcc
 	CXX=g++
-	CXXFLAGS += -pedantic -Wextra -Wshadow
+	CXXFLAGS += -pedantic -Wextra -Wshadow -Wmissing-declarations

-	ifeq ($(arch),$(filter $(arch),armv7 armv8))
+	ifeq ($(arch),$(filter $(arch),armv7 armv8 riscv64))
 		ifeq ($(OS),Android)
 			CXXFLAGS += -m$(bits)
 			LDFLAGS += -m$(bits)
 		endif
+		ifeq ($(ARCH),riscv64)
+			CXXFLAGS += -latomic
+		endif
 	else
 		CXXFLAGS += -m$(bits)
 		LDFLAGS += -m$(bits)
@@ -397,13 +422,14 @@ ifeq ($(COMP),mingw)
 			CXX=i686-w64-mingw32-c++-posix
 		endif
 	endif
-	CXXFLAGS += -pedantic -Wextra -Wshadow
+	CXXFLAGS += -pedantic -Wextra -Wshadow -Wmissing-declarations
 endif

-ifeq ($(COMP),icc)
-	comp=icc
-	CXX=icpc
-	CXXFLAGS += -diag-disable 1476,10120 -Wcheck -Wabi -Wdeprecated -strict-ansi
+ifeq ($(COMP),icx)
+	comp=icx
+	CXX=icpx
+	CXXFLAGS += --intel -pedantic -Wextra -Wshadow -Wmissing-prototypes \
+		-Wconditional-uninitialized -Wabi -Wdeprecated
 endif

 ifeq ($(COMP),clang)
@@ -413,7 +439,8 @@ ifeq ($(COMP),clang)
 		CXX=x86_64-w64-mingw32-clang++
 	endif

-	CXXFLAGS += -pedantic -Wextra -Wshadow
+	CXXFLAGS += -pedantic -Wextra -Wshadow -Wmissing-prototypes \
+	            -Wconditional-uninitialized

 	ifeq ($(filter $(KERNEL),Darwin OpenBSD FreeBSD),)
 	ifeq ($(target_windows),)
@@ -423,11 +450,14 @@ ifeq ($(COMP),clang)
 	endif
 	endif

-	ifeq ($(arch),$(filter $(arch),armv7 armv8))
+	ifeq ($(arch),$(filter $(arch),armv7 armv8 riscv64))
 		ifeq ($(OS),Android)
 			CXXFLAGS += -m$(bits)
 			LDFLAGS += -m$(bits)
 		endif
+		ifeq ($(ARCH),riscv64)
+			CXXFLAGS += -latomic
+		endif
 	else
 		CXXFLAGS += -m$(bits)
 		LDFLAGS += -m$(bits)
@@ -470,9 +500,9 @@ ifeq ($(COMP),ndk)
 	LDFLAGS += -static-libstdc++ -pie -lm -latomic
 endif

-ifeq ($(comp),icc)
-	profile_make = icc-profile-make
-	profile_use = icc-profile-use
+ifeq ($(comp),icx)
+	profile_make = icx-profile-make
+	profile_use = icx-profile-use
 else ifeq ($(comp),clang)
 	profile_make = clang-profile-make
 	profile_use = clang-profile-use
@@ -496,7 +526,7 @@ endif

 ### Sometimes gcc is really clang
 ifeq ($(COMP),gcc)
-	gccversion = $(shell $(CXX) --version)
+	gccversion = $(shell $(CXX) --version 2>/dev/null)
 	gccisclang = $(findstring clang,$(gccversion))
 	ifneq ($(gccisclang),)
 		profile_make = clang-profile-make
@@ -542,20 +572,23 @@ ifeq ($(optimize),yes)
 		endif
 	endif

-    ifeq ($(KERNEL),Darwin)
-        ifeq ($(comp),$(filter $(comp),clang icc))
-            CXXFLAGS += -mdynamic-no-pic
-        endif
+	ifeq ($(KERNEL),Darwin)
+		ifeq ($(comp),$(filter $(comp),clang icx))
+			CXXFLAGS += -mdynamic-no-pic
+		endif

-        ifeq ($(comp),gcc)
-            ifneq ($(arch),arm64)
-                CXXFLAGS += -mdynamic-no-pic
-            endif
-        endif
-    endif
+		ifeq ($(comp),gcc)
+			ifneq ($(arch),arm64)
+				CXXFLAGS += -mdynamic-no-pic
+			endif
+		endif
+	endif

 	ifeq ($(comp),clang)
-		CXXFLAGS += -fexperimental-new-pass-manager
+		clangmajorversion = $(shell $(CXX) -dumpversion 2>/dev/null | cut -f1 -d.)
+		ifeq ($(shell expr $(clangmajorversion) \< 16),1)
+			CXXFLAGS += -fexperimental-new-pass-manager
+		endif
 	endif
 endif

@@ -576,8 +609,6 @@ endif
 ifeq ($(popcnt),yes)
 	ifeq ($(arch),$(filter $(arch),ppc64 armv7 armv8 arm64))
 		CXXFLAGS += -DUSE_POPCNT
-	else ifeq ($(comp),icc)
-		CXXFLAGS += -msse3 -DUSE_POPCNT
 	else
 		CXXFLAGS += -msse3 -mpopcnt -DUSE_POPCNT
 	endif
@@ -586,63 +617,63 @@ endif
 ### 3.6 SIMD architectures
 ifeq ($(avx2),yes)
 	CXXFLAGS += -DUSE_AVX2
-	ifeq ($(comp),$(filter $(comp),gcc clang mingw))
-		CXXFLAGS += -mavx2
+	ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
+		CXXFLAGS += -mavx2 -mbmi
 	endif
 endif

 ifeq ($(avxvnni),yes)
 	CXXFLAGS += -DUSE_VNNI -DUSE_AVXVNNI
-	ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+	ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
 		CXXFLAGS += -mavxvnni
 	endif
 endif

 ifeq ($(avx512),yes)
 	CXXFLAGS += -DUSE_AVX512
-	ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+	ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
 		CXXFLAGS += -mavx512f -mavx512bw
 	endif
 endif

 ifeq ($(vnni256),yes)
 	CXXFLAGS += -DUSE_VNNI
-	ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+	ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
 		CXXFLAGS += -mavx512f -mavx512bw -mavx512vnni -mavx512dq -mavx512vl -mprefer-vector-width=256
 	endif
 endif

 ifeq ($(vnni512),yes)
 	CXXFLAGS += -DUSE_VNNI
-	ifeq ($(comp),$(filter $(comp),gcc clang mingw))
-		CXXFLAGS += -mavx512vnni -mavx512dq -mavx512vl
+	ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
+		CXXFLAGS += -mavx512f -mavx512bw -mavx512vnni -mavx512dq -mavx512vl -mprefer-vector-width=512
 	endif
 endif

 ifeq ($(sse41),yes)
 	CXXFLAGS += -DUSE_SSE41
-	ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+	ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
 		CXXFLAGS += -msse4.1
 	endif
 endif

 ifeq ($(ssse3),yes)
 	CXXFLAGS += -DUSE_SSSE3
-	ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+	ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
 		CXXFLAGS += -mssse3
 	endif
 endif

 ifeq ($(sse2),yes)
 	CXXFLAGS += -DUSE_SSE2
-	ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+	ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
 		CXXFLAGS += -msse2
 	endif
 endif

 ifeq ($(mmx),yes)
 	CXXFLAGS += -DUSE_MMX
-	ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+	ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
 		CXXFLAGS += -mmmx
 	endif
 endif
@@ -658,21 +689,40 @@ ifeq ($(neon),yes)
 	endif
 endif

+ifeq ($(dotprod),yes)
+	CXXFLAGS += -march=armv8.2-a+dotprod -DUSE_NEON_DOTPROD
+endif
+
 ### 3.7 pext
 ifeq ($(pext),yes)
 	CXXFLAGS += -DUSE_PEXT
-	ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+	ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
 		CXXFLAGS += -mbmi2
 	endif
 endif

+### 3.7.1 Try to include git commit sha for versioning
+GIT_SHA = $(shell git rev-parse HEAD 2>/dev/null | cut -c 1-8)
+ifneq ($(GIT_SHA), )
+	CXXFLAGS += -DGIT_SHA=$(GIT_SHA)
+endif
+
+### 3.7.2 Try to include git commit date for versioning
+GIT_DATE = $(shell git show -s --date=format:'%Y%m%d' --format=%cd HEAD 2>/dev/null)
+ifneq ($(GIT_DATE), )
+	CXXFLAGS += -DGIT_DATE=$(GIT_DATE)
+endif
+
 ### 3.8 Link Time Optimization
 ### This is a mix of compile and link time options because the lto link phase
 ### needs access to the optimization flags.
 ifeq ($(optimize),yes)
 ifeq ($(debug), no)
-	ifeq ($(comp),clang)
-		CXXFLAGS += -flto
+	ifeq ($(comp),$(filter $(comp),clang icx))
+		CXXFLAGS += -flto=full
+		ifeq ($(comp),icx)
+			CXXFLAGS += -fwhole-program-vtables
+                endif
 		ifeq ($(target_windows),yes)
 			CXXFLAGS += -fuse-ld=lld
 		endif
@@ -682,21 +732,19 @@ ifeq ($(debug), no)
 # GCC on some systems.
 	else ifeq ($(comp),gcc)
 	ifeq ($(gccisclang),)
-		CXXFLAGS += -flto
+		CXXFLAGS += -flto -flto-partition=one
 		LDFLAGS += $(CXXFLAGS) -flto=jobserver
 	else
-		CXXFLAGS += -flto
+		CXXFLAGS += -flto=full
 		LDFLAGS += $(CXXFLAGS)
 	endif

 # To use LTO and static linking on Windows,
 # the tool chain requires gcc version 10.1 or later.
 	else ifeq ($(comp),mingw)
-	ifneq ($(arch),i386)
-		CXXFLAGS += -flto
+		CXXFLAGS += -flto -flto-partition=one
 		LDFLAGS += $(CXXFLAGS) -save-temps
 	endif
-	endif
 endif
 endif

@@ -721,19 +769,19 @@ help:
 	@echo "Supported targets:"
 	@echo ""
 	@echo "help                    > Display architecture details"
-	@echo "build                   > Standard build"
+	@echo "profile-build           > standard build with profile-guided optimization"
+	@echo "build                   > skip profile-guided optimization"
 	@echo "net                     > Download the default nnue net"
-	@echo "profile-build           > Faster build (with profile-guided optimization)"
 	@echo "strip                   > Strip executable"
 	@echo "install                 > Install executable"
 	@echo "clean                   > Clean up"
 	@echo ""
 	@echo "Supported archs:"
 	@echo ""
-	@echo "x86-64-vnni512          > x86 64-bit with vnni support 512bit wide"
-	@echo "x86-64-vnni256          > x86 64-bit with vnni support 256bit wide"
+	@echo "x86-64-vnni512          > x86 64-bit with vnni 512bit support"
+	@echo "x86-64-vnni256          > x86 64-bit with vnni 512bit support, limit operands to 256bit wide"
 	@echo "x86-64-avx512           > x86 64-bit with avx512 support"
-	@echo "x86-64-avxvnni          > x86 64-bit with avxvnni support"
+	@echo "x86-64-avxvnni          > x86 64-bit with vnni 256bit support"
 	@echo "x86-64-bmi2             > x86 64-bit with bmi2 support"
 	@echo "x86-64-avx2             > x86 64-bit with avx2 support"
 	@echo "x86-64-sse41-popcnt     > x86 64-bit with sse41 and popcnt support"
@@ -749,27 +797,30 @@ help:
 	@echo "armv7                   > ARMv7 32-bit"
 	@echo "armv7-neon              > ARMv7 32-bit with popcnt and neon"
 	@echo "armv8                   > ARMv8 64-bit with popcnt and neon"
+	@echo "armv8-dotprod           > ARMv8 64-bit with popcnt, neon and dot product support"
 	@echo "e2k                     > Elbrus 2000"
 	@echo "apple-silicon           > Apple silicon ARM64"
 	@echo "general-64              > unspecified 64-bit"
 	@echo "general-32              > unspecified 32-bit"
+	@echo "riscv64                 > RISC-V 64-bit"
 	@echo ""
 	@echo "Supported compilers:"
 	@echo ""
 	@echo "gcc                     > Gnu compiler (default)"
 	@echo "mingw                   > Gnu compiler with MinGW under Windows"
 	@echo "clang                   > LLVM Clang compiler"
-	@echo "icc                     > Intel compiler"
+	@echo "icx                     > Intel oneAPI DPC++/C++ Compiler"
 	@echo "ndk                     > Google NDK to cross-compile for Android"
 	@echo ""
-	@echo "Simple examples. If you don't know what to do, you likely want to run: "
+	@echo "Simple examples. If you don't know what to do, you likely want to run one of: "
 	@echo ""
-	@echo "make -j build ARCH=x86-64  (A portable, slow compile for 64-bit systems)"
-	@echo "make -j build ARCH=x86-32  (A portable, slow compile for 32-bit systems)"
+	@echo "make -j profile-build ARCH=x86-64-avx2    # typically a fast compile for common systems "
+	@echo "make -j profile-build ARCH=x86-64-modern  # A more portable compile for 64-bit systems "
+	@echo "make -j profile-build ARCH=x86-64         # A portable compile for 64-bit systems "
 	@echo ""
-	@echo "Advanced examples, for experienced users looking for performance: "
+	@echo "Advanced examples, for experienced users: "
 	@echo ""
-	@echo "make    help  ARCH=x86-64-bmi2"
+	@echo "make -j profile-build ARCH=x86-64-bmi2"
 	@echo "make -j profile-build ARCH=x86-64-bmi2 COMP=gcc COMPCXX=g++-9.0"
 	@echo "make -j build ARCH=x86-64-ssse3 COMP=clang"
 	@echo ""
@@ -784,8 +835,10 @@ endif


 .PHONY: help build profile-build strip install clean net objclean profileclean \
-        config-sanity icc-profile-use icc-profile-make gcc-profile-use gcc-profile-make \
-        clang-profile-use clang-profile-make
+	config-sanity \
+	icx-profile-use icx-profile-make \
+	gcc-profile-use gcc-profile-make \
+	clang-profile-use clang-profile-make FORCE

 build: net config-sanity
 	$(MAKE) ARCH=$(ARCH) COMP=$(COMP) all
@@ -821,25 +874,37 @@ clean: objclean profileclean
 net:
 	$(eval nnuenet := $(shell grep EvalFileDefaultName evaluate.h | grep define | sed 's/.*\(nn-[a-z0-9]\{12\}.nnue\).*/\1/'))
 	@echo "Default net: $(nnuenet)"
-	$(eval nnuedownloadurl := https://tests.stockfishchess.org/api/nn/$(nnuenet))
+	$(eval nnuedownloadurl1 := https://tests.stockfishchess.org/api/nn/$(nnuenet))
+	$(eval nnuedownloadurl2 := https://github.com/official-stockfish/networks/raw/master/$(nnuenet))
 	$(eval curl_or_wget := $(shell if hash curl 2>/dev/null; then echo "curl -skL"; elif hash wget 2>/dev/null; then echo "wget -qO-"; fi))
-	@if test -f "$(nnuenet)"; then \
-            echo "Already available."; \
-         else \
-            if [ "x$(curl_or_wget)" = "x" ]; then \
-               echo "Automatic download failed: neither curl nor wget is installed. Install one of these tools or download the net manually"; exit 1; \
-            else \
-               echo "Downloading $(nnuedownloadurl)"; $(curl_or_wget) $(nnuedownloadurl) > $(nnuenet);\
-            fi; \
-        fi;
+	@if [ "x$(curl_or_wget)" = "x" ]; then \
+	    echo "Neither curl nor wget is installed. Install one of these tools unless the net has been downloaded manually"; \
+        fi
 	$(eval shasum_command := $(shell if hash shasum 2>/dev/null; then echo "shasum -a 256 "; elif hash sha256sum 2>/dev/null; then echo "sha256sum "; fi))
-	@if [ "x$(shasum_command)" != "x" ]; then \
-	    if [ "$(nnuenet)" != "nn-"`$(shasum_command) $(nnuenet) | cut -c1-12`".nnue" ]; then \
-                echo "Failed download or $(nnuenet) corrupted, please delete!"; exit 1; \
-            fi \
-         else \
+	@if [ "x$(shasum_command)" = "x" ]; then \
            echo "shasum / sha256sum not found, skipping net validation"; \
        fi
+	@for nnuedownloadurl in "$(nnuedownloadurl1)" "$(nnuedownloadurl2)"; do \
+	   if test -f "$(nnuenet)"; then \
+	      echo "$(nnuenet) available."; \
+	   else \
+	      if [ "x$(curl_or_wget)" != "x" ]; then \
+                 echo "Downloading $${nnuedownloadurl}"; $(curl_or_wget) $${nnuedownloadurl} > $(nnuenet);\
+              else \
+                 echo "No net found and download not possible"; exit 1;\
+	      fi; \
+	   fi; \
+	   if [ "x$(shasum_command)" != "x" ]; then \
+	      if [ "$(nnuenet)" != "nn-"`$(shasum_command) $(nnuenet) | cut -c1-12`".nnue" ]; then \
+	         echo "Removing failed download"; rm -f $(nnuenet); \
+	      else \
+	         echo "Network validated"; break; \
+	      fi; \
+	   fi; \
+	done
+	@if ! test -f "$(nnuenet)"; then \
+	    echo "Failed to download $(nnuenet)."; \
+	fi

 # clean binaries and objects
 objclean:
@@ -888,7 +953,9 @@ config-sanity: net
 	@echo "vnni256: '$(vnni256)'"
 	@echo "vnni512: '$(vnni512)'"
 	@echo "neon: '$(neon)'"
+	@echo "dotprod: '$(dotprod)'"
 	@echo "arm_version: '$(arm_version)'"
+	@echo "target_windows: '$(target_windows)'"
 	@echo ""
 	@echo "Flags:"
 	@echo "CXX: $(CXX)"
@@ -902,7 +969,7 @@ config-sanity: net
 	@test "$(SUPPORTED_ARCH)" = "true"
 	@test "$(arch)" = "any" || test "$(arch)" = "x86_64" || test "$(arch)" = "i386" || \
 	 test "$(arch)" = "ppc64" || test "$(arch)" = "ppc" || test "$(arch)" = "e2k" || \
-	 test "$(arch)" = "armv7" || test "$(arch)" = "armv8" || test "$(arch)" = "arm64"
+	 test "$(arch)" = "armv7" || test "$(arch)" = "armv8" || test "$(arch)" = "arm64" || test "$(arch)" = "riscv64"
 	@test "$(bits)" = "32" || test "$(bits)" = "64"
 	@test "$(prefetch)" = "yes" || test "$(prefetch)" = "no"
 	@test "$(popcnt)" = "yes" || test "$(popcnt)" = "no"
@@ -917,12 +984,16 @@ config-sanity: net
 	@test "$(vnni256)" = "yes" || test "$(vnni256)" = "no"
 	@test "$(vnni512)" = "yes" || test "$(vnni512)" = "no"
 	@test "$(neon)" = "yes" || test "$(neon)" = "no"
-	@test "$(comp)" = "gcc" || test "$(comp)" = "icc" || test "$(comp)" = "mingw" || test "$(comp)" = "clang" \
+	@test "$(comp)" = "gcc" || test "$(comp)" = "icx" || test "$(comp)" = "mingw" || test "$(comp)" = "clang" \
 	|| test "$(comp)" = "armv7a-linux-androideabi16-clang"  || test "$(comp)" = "aarch64-linux-android21-clang"

 $(EXE): $(OBJS)
 	+$(CXX) -o $@ $(OBJS) $(LDFLAGS)

+# Force recompilation to ensure version info is up-to-date
+misc.o: FORCE
+FORCE:
+
 clang-profile-make:
 	$(MAKE) ARCH=$(ARCH) COMP=$(COMP) \
 	EXTRACXXFLAGS='-fprofile-instr-generate ' \
@@ -951,18 +1022,22 @@ gcc-profile-use:
 	EXTRALDFLAGS='-lgcov' \
 	all

-icc-profile-make:
-	@mkdir -p profdir
+icx-profile-make:
 	$(MAKE) ARCH=$(ARCH) COMP=$(COMP) \
-	EXTRACXXFLAGS='-prof-gen=srcpos -prof_dir ./profdir' \
+	EXTRACXXFLAGS='-fprofile-instr-generate ' \
+	EXTRALDFLAGS=' -fprofile-instr-generate' \
 	all

-icc-profile-use:
+icx-profile-use:
+	$(XCRUN) llvm-profdata merge -output=stockfish.profdata *.profraw
 	$(MAKE) ARCH=$(ARCH) COMP=$(COMP) \
-	EXTRACXXFLAGS='-prof_use -prof_dir ./profdir' \
+	EXTRACXXFLAGS='-fprofile-instr-use=stockfish.profdata' \
+	EXTRALDFLAGS='-fprofile-use ' \
 	all

 .depend: $(SRCS)
 	-@$(CXX) $(DEPENDFLAGS) -MM $(SRCS) > $@ 2> /dev/null

+ifeq (, $(filter $(MAKECMDGOALS), help strip install clean net objclean profileclean config-sanity))
 -include .depend
+endif
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -16,6 +16,8 @@
  along with this program.  If not, see <http://www.gnu.org/licenses/>.
 */

+#include "benchmark.h"
+
 #include <fstream>
 #include <iostream>
 #include <istream>
@@ -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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -27,7 +27,6 @@ namespace Stockfish {
 uint8_t PopCnt16[1 << 16];
 uint8_t SquareDistance[SQUARE_NB][SQUARE_NB];

-Bitboard SquareBB[SQUARE_NB];
 Bitboard LineBB[SQUARE_NB][SQUARE_NB];
 Bitboard BetweenBB[SQUARE_NB][SQUARE_NB];
 Bitboard PseudoAttacks[PIECE_TYPE_NB][SQUARE_NB];
@@ -82,9 +81,6 @@ void Bitboards::init() {
  for (unsigned i = 0; i < (1 << 16); ++i)
      PopCnt16[i] = uint8_t(std::bitset<16>(i).count());

-  for (Square s = SQ_A1; s <= SQ_H8; ++s)
-      SquareBB[s] = (1ULL << s);
-
  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));
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -74,7 +74,6 @@ constexpr Bitboard KingFlank[FILE_NB] = {
 extern uint8_t PopCnt16[1 << 16];
 extern uint8_t SquareDistance[SQUARE_NB][SQUARE_NB];

-extern Bitboard SquareBB[SQUARE_NB];
 extern Bitboard BetweenBB[SQUARE_NB][SQUARE_NB];
 extern Bitboard LineBB[SQUARE_NB][SQUARE_NB];
 extern Bitboard PseudoAttacks[PIECE_TYPE_NB][SQUARE_NB];
@@ -108,7 +107,7 @@ extern Magic BishopMagics[SQUARE_NB];

 inline Bitboard square_bb(Square s) {
  assert(is_ok(s));
-  return SquareBB[s];
+  return (1ULL << s);
 }


@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -36,7 +36,7 @@
 #include "timeman.h"
 #include "uci.h"
 #include "incbin/incbin.h"
-
+#include "nnue/evaluate_nnue.h"

 // Macro to embed the default efficiently updatable neural network (NNUE) file
 // data in the engine binary (using incbin.h, by Dale Weiler).
@@ -82,20 +82,18 @@ namespace Eval {
        eval_file = EvalFileDefaultName;

    #if defined(DEFAULT_NNUE_DIRECTORY)
-    #define stringify2(x) #x
-    #define stringify(x) stringify2(x)
    vector<string> dirs = { "<internal>" , "" , CommandLine::binaryDirectory , stringify(DEFAULT_NNUE_DIRECTORY) };
    #else
    vector<string> dirs = { "<internal>" , "" , CommandLine::binaryDirectory };
    #endif

-    for (string directory : dirs)
+    for (const string& directory : dirs)
        if (currentEvalFileName != eval_file)
        {
            if (directory != "<internal>")
            {
                ifstream stream(directory + eval_file, ios::binary);
-                if (load_eval(eval_file, stream))
+                if (NNUE::load_eval(eval_file, stream))
                    currentEvalFileName = eval_file;
            }

@@ -111,7 +109,7 @@ namespace Eval {
                (void) gEmbeddedNNUEEnd; // Silence warning on unused variable

                istream stream(&buffer);
-                if (load_eval(eval_file, stream))
+                if (NNUE::load_eval(eval_file, stream))
                    currentEvalFileName = eval_file;
            }
        }
@@ -159,24 +157,24 @@ namespace Trace {

  Score scores[TERM_NB][COLOR_NB];

-  double to_cp(Value v) { return double(v) / PawnValueEg; }
+  static double to_cp(Value v) { return double(v) / UCI::NormalizeToPawnValue; }

-  void add(int idx, Color c, Score s) {
+  static void add(int idx, Color c, Score s) {
    scores[idx][c] = s;
  }

-  void add(int idx, Score w, Score b = SCORE_ZERO) {
+  static void add(int idx, Score w, Score b = SCORE_ZERO) {
    scores[idx][WHITE] = w;
    scores[idx][BLACK] = b;
  }

-  std::ostream& operator<<(std::ostream& os, Score s) {
+  static std::ostream& operator<<(std::ostream& os, Score s) {
    os << std::setw(5) << to_cp(mg_value(s)) << " "
       << std::setw(5) << to_cp(eg_value(s));
    return os;
  }

-  std::ostream& operator<<(std::ostream& os, Term t) {
+  static std::ostream& operator<<(std::ostream& os, Term t) {

    if (t == MATERIAL || t == IMBALANCE || t == WINNABLE || t == TOTAL)
        os << " ----  ----"    << " | " << " ----  ----";
@@ -193,8 +191,8 @@ using namespace Trace;
 namespace {

  // Threshold for lazy and space evaluation
-  constexpr Value LazyThreshold1    =  Value(3631);
-  constexpr Value LazyThreshold2    =  Value(2084);
+  constexpr Value LazyThreshold1    =  Value(3622);
+  constexpr Value LazyThreshold2    =  Value(1962);
  constexpr Value SpaceThreshold    =  Value(11551);

  // KingAttackWeights[PieceType] contains king attack weights by piece type
@@ -388,10 +386,10 @@ namespace {
  template<Tracing T> template<Color Us, PieceType Pt>
  Score Evaluation<T>::pieces() {

-    constexpr Color     Them = ~Us;
-    constexpr Direction Down = -pawn_push(Us);
-    constexpr Bitboard OutpostRanks = (Us == WHITE ? Rank4BB | Rank5BB | Rank6BB
-                                                   : Rank5BB | Rank4BB | Rank3BB);
+    constexpr Color Them = ~Us;
+    [[maybe_unused]] constexpr Direction Down = -pawn_push(Us);
+    [[maybe_unused]] constexpr Bitboard OutpostRanks = (Us == WHITE ? Rank4BB | Rank5BB | Rank6BB
+                                                                    : Rank5BB | Rank4BB | Rank3BB);
    Bitboard b1 = pos.pieces(Us, Pt);
    Bitboard b, bb;
    Score score = SCORE_ZERO;
@@ -430,7 +428,7 @@ namespace {
        int mob = popcount(b & mobilityArea[Us]);
        mobility[Us] += MobilityBonus[Pt - 2][mob];

-        if (Pt == BISHOP || Pt == KNIGHT)
+        if constexpr (Pt == BISHOP || Pt == KNIGHT)
        {
            // Bonus if the piece is on an outpost square or can reach one
            // Bonus for knights (UncontestedOutpost) if few relevant targets
@@ -981,7 +979,7 @@ namespace {
    // Initialize score by reading the incrementally updated scores included in
    // the position object (material + piece square tables) and the material
    // imbalance. Score is computed internally from the white point of view.
-    Score score = pos.psq_score() + me->imbalance() + pos.this_thread()->trend;
+    Score score = pos.psq_score() + me->imbalance();

    // Probe the pawn hash table
    pe = Pawns::probe(pos);
@@ -1042,38 +1040,6 @@ make_v:
    return v;
  }

-
-  /// Fisher Random Chess: correction for cornered bishops, to fix chess960 play with NNUE
-
-  Value fix_FRC(const Position& pos) {
-
-    constexpr Bitboard Corners =  1ULL << SQ_A1 | 1ULL << SQ_H1 | 1ULL << SQ_A8 | 1ULL << SQ_H8;
-
-    if (!(pos.pieces(BISHOP) & Corners))
-        return VALUE_ZERO;
-
-    int correction = 0;
-
-    if (   pos.piece_on(SQ_A1) == W_BISHOP
-        && pos.piece_on(SQ_B2) == W_PAWN)
-        correction -= CorneredBishop;
-
-    if (   pos.piece_on(SQ_H1) == W_BISHOP
-        && pos.piece_on(SQ_G2) == W_PAWN)
-        correction -= CorneredBishop;
-
-    if (   pos.piece_on(SQ_A8) == B_BISHOP
-        && pos.piece_on(SQ_B7) == B_PAWN)
-        correction += CorneredBishop;
-
-    if (   pos.piece_on(SQ_H8) == B_BISHOP
-        && pos.piece_on(SQ_G7) == B_PAWN)
-        correction += CorneredBishop;
-
-    return pos.side_to_move() == WHITE ?  Value(3 * correction)
-                                       : -Value(3 * correction);
-  }
-
 } // namespace Eval


@@ -1082,38 +1048,35 @@ make_v:

 Value Eval::evaluate(const Position& pos) {

+  assert(!pos.checkers());
+
  Value v;
-  bool useClassical = false;
+  Value psq = pos.psq_eg_stm();

-  // Deciding between classical and NNUE eval (~10 Elo): for high PSQ imbalance we use classical,
-  // but we switch to NNUE during long shuffling or with high material on the board.
-  if (  !useNNUE
-      || ((pos.this_thread()->depth > 9 || pos.count<ALL_PIECES>() > 7) &&
-          abs(eg_value(pos.psq_score())) * 5 > (856 + pos.non_pawn_material() / 64) * (10 + pos.rule50_count())))
+  // We use the much less accurate but faster Classical eval when the NNUE
+  // option is set to false. Otherwise we use the NNUE eval unless the
+  // PSQ advantage is decisive. (~4 Elo at STC, 1 Elo at LTC)
+  bool useClassical = !useNNUE || abs(psq) > 2048;
+
+  if (useClassical)
+      v = Evaluation<NO_TRACE>(pos).value();
+  else
  {
-      v = Evaluation<NO_TRACE>(pos).value();          // classical
-      useClassical = abs(v) >= 297;
-  }
+      int nnueComplexity;
+      int npm = pos.non_pawn_material() / 64;

-  // If result of a classical evaluation is much lower than threshold fall back to NNUE
-  if (useNNUE && !useClassical)
-  {
-       Value nnue     = NNUE::evaluate(pos, true);     // NNUE
-       int scale      = 1036 + 22 * pos.non_pawn_material() / 1024;
-       Color stm      = pos.side_to_move();
-       Value optimism = pos.this_thread()->optimism[stm];
-       Value psq      = (stm == WHITE ? 1 : -1) * eg_value(pos.psq_score());
-       int complexity = 35 * abs(nnue - psq) / 256;
+      Color stm = pos.side_to_move();
+      Value optimism = pos.this_thread()->optimism[stm];

-       optimism = optimism * (44 + complexity) / 31;
-       v = (nnue + optimism) * scale / 1024 - optimism;
+      Value nnue = NNUE::evaluate(pos, true, &nnueComplexity);

-       if (pos.is_chess960())
-           v += fix_FRC(pos);
+      // Blend optimism with nnue complexity and (semi)classical complexity
+      optimism += optimism * (nnueComplexity + abs(psq - nnue)) / 512;
+      v = (nnue * (945 + npm) + optimism * (150 + npm)) / 1024;
  }

  // Damp down the evaluation linearly when shuffling
-  v = v * (195 - pos.rule50_count()) / 211;
+  v = v * (200 - pos.rule50_count()) / 214;

  // Guarantee evaluation does not hit the tablebase range
  v = std::clamp(v, VALUE_TB_LOSS_IN_MAX_PLY + 1, VALUE_TB_WIN_IN_MAX_PLY - 1);
@@ -1139,8 +1102,6 @@ std::string Eval::trace(Position& pos) {
  std::memset(scores, 0, sizeof(scores));

  // Reset any global variable used in eval
-  pos.this_thread()->depth           = 0;
-  pos.this_thread()->trend           = SCORE_ZERO;
  pos.this_thread()->bestValue       = VALUE_ZERO;
  pos.this_thread()->optimism[WHITE] = VALUE_ZERO;
  pos.this_thread()->optimism[BLACK] = VALUE_ZERO;
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -39,20 +39,13 @@ namespace Eval {
  // 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-6877cd24400e.nnue"
+  #define EvalFileDefaultName   "nn-5af11540bbfe.nnue"

  namespace NNUE {

-    std::string trace(Position& pos);
-    Value evaluate(const Position& pos, bool adjusted = false);
-
    void init();
    void verify();

-    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 NNUE

 } // namespace Eval
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -28,7 +28,7 @@ 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 NULL, meaning that the
+/// 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
@@ -62,7 +62,7 @@ struct Entry {
  uint8_t factor[COLOR_NB];
 };

-typedef HashTable<Entry, 8192> Table;
+using Table = HashTable<Entry, 8192>;

 Entry* probe(const Position& pos);

@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -32,21 +32,26 @@
 // 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" {
-typedef bool(*fun1_t)(LOGICAL_PROCESSOR_RELATIONSHIP,
-                      PSYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX, PDWORD);
-typedef bool(*fun2_t)(USHORT, PGROUP_AFFINITY);
-typedef bool(*fun3_t)(HANDLE, CONST GROUP_AFFINITY*, PGROUP_AFFINITY);
-typedef bool(*fun4_t)(USHORT, PGROUP_AFFINITY, USHORT, PUSHORT);
-typedef WORD(*fun5_t)();
+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>
-#include <cstdlib>

 #if defined(__linux__) && !defined(__ANDROID__)
 #include <stdlib.h>
@@ -67,9 +72,8 @@ namespace Stockfish {

 namespace {

-/// Version number. If Version is left empty, then compile date in the format
-/// DD-MM-YY and show in engine_info.
-const string Version = "15";
+/// 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
@@ -138,23 +142,41 @@ public:
 } // namespace


-/// engine_info() returns the full name of the current Stockfish version. This
-/// will be either "Stockfish <Tag> DD-MM-YY" (where DD-MM-YY is the date when
-/// the program was compiled) or "Stockfish <Version>", depending on whether
-/// Version is empty.
+/// 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');

-  const string months("Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec");
-  string month, day, year;
-  stringstream ss, date(__DATE__); // From compiler, format is "Sep 21 2008"
-
-  ss << "Stockfish " << Version << setfill('0');
-
-  if (Version.empty())
+  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 << setw(2) << day << setw(2) << (1 + months.find(month) / 4) << year.substr(2);
+      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 ")
@@ -168,8 +190,6 @@ string engine_info(bool to_uci) {

 std::string compiler_info() {

-  #define stringify2(x) #x
-  #define stringify(x) stringify2(x)
  #define make_version_string(major, minor, patch) stringify(major) "." stringify(minor) "." stringify(patch)

 /// Predefined macros hell:
@@ -281,21 +301,94 @@ std::string compiler_info() {


 /// Debug functions used mainly to collect run-time statistics
-static std::atomic<int64_t> hits[2], means[2];
+constexpr int MaxDebugSlots = 32;

-void dbg_hit_on(bool b) { ++hits[0]; if (b) ++hits[1]; }
-void dbg_hit_on(bool c, bool b) { if (c) dbg_hit_on(b); }
-void dbg_mean_of(int v) { ++means[0]; means[1] += v; }
+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() {

-  if (hits[0])
-      cerr << "Total " << hits[0] << " Hits " << hits[1]
-           << " hit rate (%) " << 100 * hits[1] / hits[0] << endl;
+    int64_t n;
+    auto E   = [&n](int64_t x) { return double(x) / n; };
+    auto sqr = [](double x) { return x * x; };

-  if (means[0])
-      cerr << "Total " << means[0] << " Mean "
-           << (double)means[1] / means[0] << endl;
+    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])
+                      << std::endl;
+
+    for (int i = 0; i < MaxDebugSlots; ++i)
+        if ((n = mean[i][0]))
+        {
+            std::cerr << "Mean #" << i
+                      << ": Total " << n << " Mean " << E(mean[i][1])
+                      << std::endl;
+        }
+
+    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
+                      << std::endl;
+        }
+
+    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
+                      << std::endl;
+        }
 }


@@ -356,8 +449,10 @@ 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)
+#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
@@ -367,8 +462,10 @@ void std_aligned_free(void* ptr) {

 #if defined(POSIXALIGNEDALLOC)
  free(ptr);
-#elif defined(_WIN32)
+#elif defined(_WIN32) && !defined(_M_ARM) && !defined(_M_ARM64)
  _mm_free(ptr);
+#elif defined(_WIN32)
+  _aligned_free(ptr);
 #else
  free(ptr);
 #endif
@@ -378,10 +475,9 @@ void std_aligned_free(void* ptr) {

 #if defined(_WIN32)

-static void* aligned_large_pages_alloc_windows(size_t allocSize) {
+static void* aligned_large_pages_alloc_windows([[maybe_unused]] size_t allocSize) {

  #if !defined(_WIN64)
-    (void)allocSize; // suppress unused-parameter compiler warning
    return nullptr;
  #else

@@ -393,11 +489,30 @@ static void* aligned_large_pages_alloc_windows(size_t allocSize) {
  if (!largePageSize)
      return nullptr;

-  // We need SeLockMemoryPrivilege, so try to enable it for the process
-  if (!OpenProcessToken(GetCurrentProcess(), TOKEN_ADJUST_PRIVILEGES | TOKEN_QUERY, &hProcessToken))
+  // 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;

-  if (LookupPrivilegeValue(NULL, SE_LOCK_MEMORY_NAME, &luid))
+  // 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 { };
@@ -409,17 +524,18 @@ static void* aligned_large_pages_alloc_windows(size_t allocSize) {

      // 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 (AdjustTokenPrivileges(
+      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(
-              NULL, allocSize, MEM_RESERVE | MEM_COMMIT | MEM_LARGE_PAGES, PAGE_READWRITE);
+              nullptr, allocSize, MEM_RESERVE | MEM_COMMIT | MEM_LARGE_PAGES, PAGE_READWRITE);

          // Privilege no longer needed, restore previous state
-          AdjustTokenPrivileges(hProcessToken, FALSE, &prevTp, 0, NULL, NULL);
+          fun8( // AdjustTokenPrivileges ()
+              hProcessToken, FALSE, &prevTp, 0, nullptr, nullptr);
      }
  }

@@ -437,7 +553,7 @@ void* aligned_large_pages_alloc(size_t allocSize) {

  // Fall back to regular, page aligned, allocation if necessary
  if (!mem)
-      mem = VirtualAlloc(NULL, allocSize, MEM_RESERVE | MEM_COMMIT, PAGE_READWRITE);
+      mem = VirtualAlloc(nullptr, allocSize, MEM_RESERVE | MEM_COMMIT, PAGE_READWRITE);

  return mem;
 }
@@ -501,7 +617,7 @@ void bindThisThread(size_t) {}
 /// API and returns the best node id for the thread with index idx. Original
 /// code from Texel by Peter Österlund.

-int best_node(size_t idx) {
+static int best_node(size_t idx) {

  int threads = 0;
  int nodes = 0;
@@ -510,7 +626,7 @@ int best_node(size_t idx) {
  DWORD byteOffset = 0;

  // Early exit if the needed API is not available at runtime
-  HMODULE k32 = GetModuleHandle("Kernel32.dll");
+  HMODULE k32 = GetModuleHandle(TEXT("Kernel32.dll"));
  auto fun1 = (fun1_t)(void(*)())GetProcAddress(k32, "GetLogicalProcessorInformationEx");
  if (!fun1)
      return -1;
@@ -580,7 +696,7 @@ void bindThisThread(size_t idx) {
      return;

  // Early exit if the needed API are not available at runtime
-  HMODULE k32 = GetModuleHandle("Kernel32.dll");
+  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");
@@ -626,8 +742,7 @@ string argv0;            // path+name of the executable binary, as given by argv
 string binaryDirectory;  // path of the executable directory
 string workingDirectory; // path of the working directory

-void init(int argc, char* argv[]) {
-    (void)argc;
+void init([[maybe_unused]] int argc, char* argv[]) {
    string pathSeparator;

    // extract the path+name of the executable binary
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -28,6 +28,9 @@

 #include "types.h"

+#define stringify2(x) #x
+#define stringify(x) stringify2(x)
+
 namespace Stockfish {

 std::string engine_info(bool to_uci = false);
@@ -39,12 +42,13 @@ 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 b);
-void dbg_hit_on(bool c, bool b);
-void dbg_mean_of(int v);
+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();

-typedef std::chrono::milliseconds::rep TimePoint; // A value in milliseconds
+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>
@@ -85,87 +89,21 @@ static inline const union { uint32_t i; char c[4]; } Le = { 0x01020304 };
 static inline const bool IsLittleEndian = (Le.c[0] == 4);


-// RunningAverage : a class to calculate a running average of a series of values.
-// For efficiency, all computations are done with integers.
-class RunningAverage {
-  public:
-
-      // Reset the running average to rational value p / q
-      void set(int64_t p, int64_t q)
-        { average = p * PERIOD * RESOLUTION / q; }
-
-      // Update average with value v
-      void update(int64_t v)
-        { average = RESOLUTION * v + (PERIOD - 1) * average / PERIOD; }
-
-      // Test if average is strictly greater than rational a / b
-      bool is_greater(int64_t a, int64_t b) const
-        { return b * average > a * (PERIOD * RESOLUTION); }
-
-      int64_t value() const
-        { return average / (PERIOD * RESOLUTION); }
-
-  private :
-      static constexpr int64_t PERIOD     = 4096;
-      static constexpr int64_t RESOLUTION = 1024;
-      int64_t average;
-};
-
 template <typename T, std::size_t MaxSize>
 class ValueList {

 public:
  std::size_t size() const { return size_; }
-  void resize(std::size_t newSize) { size_ = newSize; }
  void push_back(const T& value) { values_[size_++] = value; }
-  T& operator[](std::size_t index) { return values_[index]; }
-  T* begin() { return values_; }
-  T* end() { return values_ + size_; }
-  const T& operator[](std::size_t index) const { return values_[index]; }
  const T* begin() const { return values_; }
  const T* end() const { return values_ + size_; }

-  void swap(ValueList& other) {
-    const std::size_t maxSize = std::max(size_, other.size_);
-    for (std::size_t i = 0; i < maxSize; ++i) {
-      std::swap(values_[i], other.values_[i]);
-    }
-    std::swap(size_, other.size_);
-  }
-
 private:
  T values_[MaxSize];
  std::size_t size_ = 0;
 };


-/// sigmoid(t, x0, y0, C, P, Q) implements a sigmoid-like function using only integers,
-/// with the following properties:
-///
-///  -  sigmoid is centered in (x0, y0)
-///  -  sigmoid has amplitude [-P/Q , P/Q] instead of [-1 , +1]
-///  -  limit is (y0 - P/Q) when t tends to -infinity
-///  -  limit is (y0 + P/Q) when t tends to +infinity
-///  -  the slope can be adjusted using C > 0, smaller C giving a steeper sigmoid
-///  -  the slope of the sigmoid when t = x0 is P/(Q*C)
-///  -  sigmoid is increasing with t when P > 0 and Q > 0
-///  -  to get a decreasing sigmoid, change sign of P
-///  -  mean value of the sigmoid is y0
-///
-/// Use <https://www.desmos.com/calculator/jhh83sqq92> to draw the sigmoid
-
-inline int64_t sigmoid(int64_t t, int64_t x0,
-                                  int64_t y0,
-                                  int64_t  C,
-                                  int64_t  P,
-                                  int64_t  Q)
-{
-   assert(C > 0);
-   assert(Q != 0);
-   return y0 + P * (t-x0) / (Q * (std::abs(t-x0) + C)) ;
-}
-
-
 /// xorshift64star Pseudo-Random Number Generator
 /// This class is based on original code written and dedicated
 /// to the public domain by Sebastiano Vigna (2014).
@@ -204,7 +142,7 @@ public:

 inline uint64_t mul_hi64(uint64_t a, uint64_t b) {
 #if defined(__GNUC__) && defined(IS_64BIT)
-    __extension__ typedef unsigned __int128 uint128;
+    __extension__ using uint128 = unsigned __int128;
    return ((uint128)a * (uint128)b) >> 64;
 #else
    uint64_t aL = (uint32_t)a, aH = a >> 32;
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -25,13 +25,21 @@ namespace Stockfish {

 namespace {

-  template<GenType Type, Direction D>
-  ExtMove* make_promotions(ExtMove* moveList, Square to) {
+  template<GenType Type, Direction D, bool Enemy>
+  ExtMove* make_promotions(ExtMove* moveList, [[maybe_unused]] Square to) {

-    if (Type == CAPTURES || Type == EVASIONS || Type == NON_EVASIONS)
+    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 (Type == QUIETS || Type == EVASIONS || Type == NON_EVASIONS)
+    if constexpr ((Type == QUIETS && !Enemy) || Type == EVASIONS || Type == NON_EVASIONS)
    {
        *moveList++ = make<PROMOTION>(to - D, to, ROOK);
        *moveList++ = make<PROMOTION>(to - D, to, BISHOP);
@@ -60,18 +68,18 @@ namespace {
    Bitboard pawnsNotOn7 = pos.pieces(Us, PAWN) & ~TRank7BB;

    // Single and double pawn pushes, no promotions
-    if (Type != CAPTURES)
+    if constexpr (Type != CAPTURES)
    {
        Bitboard b1 = shift<Up>(pawnsNotOn7)   & emptySquares;
        Bitboard b2 = shift<Up>(b1 & TRank3BB) & emptySquares;

-        if (Type == EVASIONS) // Consider only blocking squares
+        if constexpr (Type == EVASIONS) // Consider only blocking squares
        {
            b1 &= target;
            b2 &= target;
        }

-        if (Type == QUIET_CHECKS)
+        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.
@@ -102,21 +110,21 @@ namespace {
        Bitboard b2 = shift<UpLeft >(pawnsOn7) & enemies;
        Bitboard b3 = shift<Up     >(pawnsOn7) & emptySquares;

-        if (Type == EVASIONS)
+        if constexpr (Type == EVASIONS)
            b3 &= target;

        while (b1)
-            moveList = make_promotions<Type, UpRight>(moveList, pop_lsb(b1));
+            moveList = make_promotions<Type, UpRight, true>(moveList, pop_lsb(b1));

        while (b2)
-            moveList = make_promotions<Type, UpLeft >(moveList, pop_lsb(b2));
+            moveList = make_promotions<Type, UpLeft, true>(moveList, pop_lsb(b2));

        while (b3)
-            moveList = make_promotions<Type, Up     >(moveList, pop_lsb(b3));
+            moveList = make_promotions<Type, Up,    false>(moveList, pop_lsb(b3));
    }

    // Standard and en passant captures
-    if (Type == CAPTURES || Type == EVASIONS || Type == NON_EVASIONS)
+    if constexpr (Type == CAPTURES || Type == EVASIONS || Type == NON_EVASIONS)
    {
        Bitboard b1 = shift<UpRight>(pawnsNotOn7) & enemies;
        Bitboard b2 = shift<UpLeft >(pawnsNotOn7) & enemies;
@@ -264,7 +272,7 @@ ExtMove* generate<LEGAL>(const Position& pos, ExtMove* moveList) {
  moveList = pos.checkers() ? generate<EVASIONS    >(pos, moveList)
                            : generate<NON_EVASIONS>(pos, moveList);
  while (cur != moveList)
-      if (  ((pinned && pinned & from_sq(*cur)) || from_sq(*cur) == ksq || type_of(*cur) == EN_PASSANT)
+      if (  ((pinned & from_sq(*cur)) || from_sq(*cur) == ksq || type_of(*cur) == EN_PASSANT)
          && !pos.legal(*cur))
          *cur = (--moveList)->move;
      else
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -82,82 +82,71 @@ MovePicker::MovePicker(const Position& p, Move ttm, Depth d, const ButterflyHist

  stage = (pos.checkers() ? EVASION_TT : QSEARCH_TT) +
          !(   ttm
-            && (pos.checkers() || depth > DEPTH_QS_RECAPTURES || to_sq(ttm) == recaptureSquare)
            && 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, Depth d, const CapturePieceToHistory* cph)
-           : pos(p), captureHistory(cph), ttMove(ttm), threshold(th), depth(d)
+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(ttm)
+  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 histories.
+/// 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");

-  Bitboard threatened, threatenedByPawn, threatenedByMinor, threatenedByRook;
+  [[maybe_unused]] Bitboard threatenedByPawn, threatenedByMinor, threatenedByRook, threatenedPieces;
  if constexpr (Type == QUIETS)
  {
      Color us = pos.side_to_move();
-      // squares threatened by pawns
+
      threatenedByPawn  = pos.attacks_by<PAWN>(~us);
-      // squares threatened by minors or pawns
      threatenedByMinor = pos.attacks_by<KNIGHT>(~us) | pos.attacks_by<BISHOP>(~us) | threatenedByPawn;
-      // squares threatened by rooks, minors or pawns
      threatenedByRook  = pos.attacks_by<ROOK>(~us) | threatenedByMinor;

-      // pieces threatened by pieces of lesser material value
-      threatened =  (pos.pieces(us, QUEEN) & threatenedByRook)
-                  | (pos.pieces(us, ROOK)  & threatenedByMinor)
-                  | (pos.pieces(us, KNIGHT, BISHOP) & threatenedByPawn);
-  }
-  else
-  {
-      // Silence unused variable warnings
-      (void) threatened;
-      (void) threatenedByPawn;
-      (void) threatenedByMinor;
-      (void) threatenedByRook;
+      // 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 =  6 * 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)))];
+          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 =      (*mainHistory)[pos.side_to_move()][from_to(m)]
+          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)]
-                   +     (threatened & from_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);
-
+                          :                                                                           0)
+                   +     bool(pos.check_squares(type_of(pos.moved_piece(m))) & to_sq(m)) * 16384;
      else // Type == EVASIONS
      {
-          if (pos.capture(m))
+          if (pos.capture_stage(m))
              m.value =  PieceValue[MG][pos.piece_on(to_sq(m))]
-                       - Value(type_of(pos.moved_piece(m)));
+                       - Value(type_of(pos.moved_piece(m)))
+                       + (1 << 28);
          else
-              m.value =      (*mainHistory)[pos.side_to_move()][from_to(m)]
-                       + 2 * (*continuationHistory[0])[pos.moved_piece(m)][to_sq(m)]
-                       - (1 << 28);
+              m.value =  (*mainHistory)[pos.side_to_move()][from_to(m)]
+                       + (*continuationHistory[0])[pos.moved_piece(m)][to_sq(m)];
      }
 }

@@ -168,7 +157,7 @@ Move MovePicker::select(Pred filter) {

  while (cur < endMoves)
  {
-      if (T == Best)
+      if constexpr (T == Best)
          std::swap(*cur, *std::max_element(cur, endMoves));

      if (*cur != ttMove && filter())
@@ -201,13 +190,13 @@ top:
      endMoves = generate<CAPTURES>(pos, cur);

      score<CAPTURES>();
-      partial_insertion_sort(cur, endMoves, -3000 * depth);
+      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(-69 * cur->value / 1024)) ?
+                       return pos.see_ge(*cur, Value(-cur->value)) ?
                              // Move losing capture to endBadCaptures to be tried later
                              true : (*endBadCaptures++ = *cur, false); }))
          return *(cur - 1);
@@ -226,7 +215,7 @@ top:

  case REFUTATION:
      if (select<Next>([&](){ return    *cur != MOVE_NONE
-                                    && !pos.capture(*cur)
+                                    && !pos.capture_stage(*cur)
                                    &&  pos.pseudo_legal(*cur); }))
          return *(cur - 1);
      ++stage;
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -62,14 +62,14 @@ public:
 template <typename T, int D, int Size, int... Sizes>
 struct Stats : public std::array<Stats<T, D, Sizes...>, Size>
 {
-  typedef Stats<T, D, Size, Sizes...> stats;
+  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);

-    typedef StatsEntry<T, D> entry;
+    using entry = StatsEntry<T, D>;
    entry* p = reinterpret_cast<entry*>(this);
    std::fill(p, p + sizeof(*this) / sizeof(entry), v);
  }
@@ -86,22 +86,24 @@ enum StatsType { NoCaptures, Captures };
 /// 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
-typedef Stats<int16_t, 14365, COLOR_NB, int(SQUARE_NB) * int(SQUARE_NB)> ButterflyHistory;
+/// (~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
-typedef Stats<Move, NOT_USED, PIECE_NB, SQUARE_NB> CounterMoveHistory;
+using CounterMoveHistory = Stats<Move, NOT_USED, PIECE_NB, SQUARE_NB>;

 /// CapturePieceToHistory is addressed by a move's [piece][to][captured piece type]
-typedef Stats<int16_t, 10692, PIECE_NB, SQUARE_NB, PIECE_TYPE_NB> CapturePieceToHistory;
+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]
-typedef Stats<int16_t, 29952, PIECE_NB, SQUARE_NB> PieceToHistory;
+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.
-typedef Stats<PieceToHistory, NOT_USED, PIECE_NB, SQUARE_NB> ContinuationHistory;
+/// (~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
@@ -126,7 +128,7 @@ public:
                                           const CapturePieceToHistory*,
                                           const PieceToHistory**,
                                           Square);
-  MovePicker(const Position&, Move, Value, Depth, const CapturePieceToHistory*);
+  MovePicker(const Position&, Move, Value, const CapturePieceToHistory*);
  Move next_move(bool skipQuiets = false);

 private:
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -18,15 +18,15 @@

 // Code for calculating NNUE evaluation function

+#include <fstream>
+#include <iomanip>
 #include <iostream>
 #include <set>
 #include <sstream>
-#include <iomanip>
-#include <fstream>
+#include <string_view>

 #include "../evaluate.h"
 #include "../position.h"
-#include "../misc.h"
 #include "../uci.h"
 #include "../types.h"

@@ -83,7 +83,7 @@ namespace Stockfish::Eval::NNUE {
  }  // namespace Detail

  // Initialize the evaluation function parameters
-  void initialize() {
+  static void initialize() {

    Detail::initialize(featureTransformer);
    for (std::size_t i = 0; i < LayerStacks; ++i)
@@ -91,7 +91,7 @@ namespace Stockfish::Eval::NNUE {
  }

  // Read network header
-  bool read_header(std::istream& stream, std::uint32_t* hashValue, std::string* desc)
+  static bool read_header(std::istream& stream, std::uint32_t* hashValue, std::string* desc)
  {
    std::uint32_t version, size;

@@ -105,7 +105,7 @@ namespace Stockfish::Eval::NNUE {
  }

  // Write network header
-  bool write_header(std::ostream& stream, std::uint32_t hashValue, const std::string& desc)
+  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);
@@ -115,7 +115,7 @@ namespace Stockfish::Eval::NNUE {
  }

  // Read network parameters
-  bool read_parameters(std::istream& stream) {
+  static bool read_parameters(std::istream& stream) {

    std::uint32_t hashValue;
    if (!read_header(stream, &hashValue, &netDescription)) return false;
@@ -127,7 +127,7 @@ namespace Stockfish::Eval::NNUE {
  }

  // Write network parameters
-  bool write_parameters(std::ostream& stream) {
+  static bool write_parameters(std::ostream& stream) {

    if (!write_header(stream, HashValue, netDescription)) return false;
    if (!Detail::write_parameters(stream, *featureTransformer)) return false;
@@ -136,14 +136,19 @@ namespace Stockfish::Eval::NNUE {
    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) {
+  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;
-    int delta = 10 - pos.non_pawn_material() / 1515;
+    constexpr int delta = 24;

 #if defined(ALIGNAS_ON_STACK_VARIABLES_BROKEN)
    TransformedFeatureType transformedFeaturesUnaligned[
@@ -161,9 +166,12 @@ namespace Stockfish::Eval::NNUE {
    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>(((128 - delta) * psqt + (128 + delta) * positional) / 128 / OutputScale);
+        return static_cast<Value>(((1024 - delta) * psqt + (1024 + delta) * positional) / (1024 * OutputScale));
    else
        return static_cast<Value>((psqt + positional) / OutputScale);
  }
@@ -208,7 +216,7 @@ namespace Stockfish::Eval::NNUE {
    return t;
  }

-  static const std::string PieceToChar(" PNBRQK  pnbrqk");
+  constexpr std::string_view PieceToChar(" PNBRQK  pnbrqk");


  // format_cp_compact() converts a Value into (centi)pawns and writes it in a buffer.
@@ -217,7 +225,7 @@ namespace Stockfish::Eval::NNUE {

    buffer[0] = (v < 0 ? '-' : v > 0 ? '+' : ' ');

-    int cp = std::abs(100 * v / PawnValueEg);
+    int cp = std::abs(100 * v / UCI::NormalizeToPawnValue);
    if (cp >= 10000)
    {
        buffer[1] = '0' + cp / 10000; cp %= 10000;
@@ -242,14 +250,15 @@ namespace Stockfish::Eval::NNUE {
  }


-  // format_cp_aligned_dot() converts a Value into (centi)pawns and writes it in a buffer,
-  // always keeping two decimals. The buffer must have capacity for at least 7 chars.
-  static void format_cp_aligned_dot(Value v, char* buffer) {
+  // 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;

-    buffer[0] = (v < 0 ? '-' : v > 0 ? '+' : ' ');
-
-    double cp = 1.0 * std::abs(int(v)) / PawnValueEg;
-    sprintf(&buffer[1], "%6.2f", cp);
+    stream << (v < 0 ? '-' : v > 0 ? '+' : ' ')
+           << std::setiosflags(std::ios::fixed)
+           << std::setw(6)
+           << std::setprecision(2)
+           << cp;
  }


@@ -329,17 +338,10 @@ namespace Stockfish::Eval::NNUE {

    for (std::size_t bucket = 0; bucket < LayerStacks; ++bucket)
    {
-      char buffer[3][8];
-      std::memset(buffer, '\0', sizeof(buffer));
-
-      format_cp_aligned_dot(t.psqt[bucket], buffer[0]);
-      format_cp_aligned_dot(t.positional[bucket], buffer[1]);
-      format_cp_aligned_dot(t.psqt[bucket] + t.positional[bucket], buffer[2]);
-
-      ss <<  "|  " << bucket    << "        "
-         << " |  " << buffer[0] << "  "
-         << " |  " << buffer[1] << "  "
-         << " |  " << buffer[2] << "  "
+      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";
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -31,6 +31,7 @@ namespace Stockfish::Eval::NNUE {
  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 {
@@ -54,6 +55,14 @@ namespace Stockfish::Eval::NNUE {
  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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -24,50 +24,51 @@

 namespace Stockfish::Eval::NNUE::Features {

-  // Orient a square according to perspective (rotates by 180 for black)
-  inline Square HalfKAv2_hm::orient(Color perspective, Square s, Square ksq) {
-    return Square(int(s) ^ (bool(perspective) * SQ_A8) ^ ((file_of(ksq) < FILE_E) * SQ_H1));
-  }
-
  // Index of a feature for a given king position and another piece on some square
-  inline IndexType HalfKAv2_hm::make_index(Color perspective, Square s, Piece pc, Square ksq) {
-    Square o_ksq = orient(perspective, ksq, ksq);
-    return IndexType(orient(perspective, s, ksq) + PieceSquareIndex[perspective][pc] + PS_NB * KingBuckets[o_ksq]);
+  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,
-    Color perspective,
    IndexList& active
  ) {
-    Square ksq = pos.square<KING>(perspective);
+    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));
+      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,
-    Color perspective,
    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));
+        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));
+        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;
  }
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -49,8 +49,8 @@ namespace Stockfish::Eval::NNUE::Features {
      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
+      PS_KING     = 10 * SQUARE_NB,
+      PS_NB       = 11 * SQUARE_NB
    };

    static constexpr IndexType PieceSquareIndex[COLOR_NB][PIECE_NB] = {
@@ -62,11 +62,9 @@ namespace Stockfish::Eval::NNUE::Features {
        PS_NONE, PS_W_PAWN, PS_W_KNIGHT, PS_W_BISHOP, PS_W_ROOK, PS_W_QUEEN, PS_KING, PS_NONE }
    };

-    // Orient a square according to perspective (rotates by 180 for black)
-    static Square orient(Color perspective, Square s, Square ksq);
-
    // Index of a feature for a given king position and another piece on some square
-    static IndexType make_index(Color perspective, Square s, Piece pc, Square ksq);
+    template<Color Perspective>
+    static IndexType make_index(Square s, Piece pc, Square ksq);

   public:
    // Feature name
@@ -79,15 +77,45 @@ namespace Stockfish::Eval::NNUE::Features {
    static constexpr IndexType Dimensions =
        static_cast<IndexType>(SQUARE_NB) * static_cast<IndexType>(PS_NB) / 2;

-    static constexpr int KingBuckets[64] = {
-      -1, -1, -1, -1, 31, 30, 29, 28,
-      -1, -1, -1, -1, 27, 26, 25, 24,
-      -1, -1, -1, -1, 23, 22, 21, 20,
-      -1, -1, -1, -1, 19, 18, 17, 16,
-      -1, -1, -1, -1, 15, 14, 13, 12,
-      -1, -1, -1, -1, 11, 10,  9,  8,
-      -1, -1, -1, -1,  7,  6,  5,  4,
-      -1, -1, -1, -1,  3,  2,  1,  0
+#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.
@@ -95,16 +123,16 @@ namespace Stockfish::Eval::NNUE::Features {
    using IndexList = ValueList<IndexType, MaxActiveDimensions>;

    // Get a list of indices for active features
+    template<Color Perspective>
    static void append_active_indices(
      const Position& pos,
-      Color perspective,
      IndexList& active);

    // Get a list of indices for recently changed features
+    template<Color Perspective>
    static void append_changed_indices(
      Square ksq,
      const DirtyPiece& dp,
-      Color perspective,
      IndexList& removed,
      IndexList& added
    );
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -25,13 +25,13 @@
 #include <algorithm>
 #include <type_traits>
 #include "../nnue_common.h"
-#include "../../simd.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:
+  Approach 1 (a specialization for large inputs):
    - used when the PaddedInputDimensions >= 128
    - uses AVX512 if possible
    - processes inputs in batches of 2*InputSimdWidth
@@ -42,9 +42,8 @@
      depends on the architecture (the amount of registers)
    - accumulate + hadd is used

-  Approach 2:
+  Approach 2 (a specialization for small inputs):
    - used when the PaddedInputDimensions < 128
-    - does not use AVX512
    - 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
@@ -72,6 +71,10 @@ namespace Stockfish::Eval::NNUE::Layers {
    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);
@@ -123,6 +126,14 @@ namespace Stockfish::Eval::NNUE::Layers {
      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]);
@@ -151,9 +162,15 @@ namespace Stockfish::Eval::NNUE::Layers {
  template <IndexType InDims, IndexType OutDims, typename Enabled = void>
  class AffineTransform;

-  // A specialization for large inputs.
+#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) >= 2*64)>> {
+  class AffineTransform<InDims, OutDims, std::enable_if_t<(ceil_to_multiple<IndexType>(InDims, MaxSimdWidth) >= LargeInputSize)>> {
   public:
    // Input/output type
    using InputType = std::uint8_t;
@@ -170,36 +187,39 @@ namespace Stockfish::Eval::NNUE::Layers {

    using OutputBuffer = OutputType[PaddedOutputDimensions];

-    static_assert(PaddedInputDimensions >= 128, "Something went wrong. This specialization should not have been chosen.");
+    static_assert(PaddedInputDimensions >= LargeInputSize, "Something went wrong. This specialization (for large inputs) should not have been chosen.");

 #if defined (USE_AVX512)
-    static constexpr const IndexType InputSimdWidth = 64;
-    static constexpr const IndexType MaxNumOutputRegs = 16;
+    static constexpr IndexType InputSimdWidth = 64;
+    static constexpr IndexType MaxNumOutputRegs = 16;
 #elif defined (USE_AVX2)
-    static constexpr const IndexType InputSimdWidth = 32;
-    static constexpr const IndexType MaxNumOutputRegs = 8;
+    static constexpr IndexType InputSimdWidth = 32;
+    static constexpr IndexType MaxNumOutputRegs = 8;
 #elif defined (USE_SSSE3)
-    static constexpr const IndexType InputSimdWidth = 16;
-    static constexpr const IndexType MaxNumOutputRegs = 8;
+    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 const IndexType InputSimdWidth = 8;
-    static constexpr const IndexType MaxNumOutputRegs = 8;
+    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 const IndexType InputSimdWidth = 1;
-    static constexpr const IndexType MaxNumOutputRegs = 1;
+    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 const IndexType NumOutputRegs = std::min(MaxNumOutputRegs, OutputDimensions);
-    static constexpr const IndexType SmallBlockSize = InputSimdWidth;
-    static constexpr const IndexType BigBlockSize = NumOutputRegs * PaddedInputDimensions;
-    static constexpr const IndexType NumSmallBlocksInBigBlock = BigBlockSize / SmallBlockSize;
-    static constexpr const IndexType NumSmallBlocksPerOutput = PaddedInputDimensions / SmallBlockSize;
-    static constexpr const IndexType NumBigBlocks = OutputDimensions / NumOutputRegs;
+    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);

@@ -235,8 +255,7 @@ namespace Stockfish::Eval::NNUE::Layers {

    // Read network parameters
    bool read_parameters(std::istream& stream) {
-      for (IndexType i = 0; i < OutputDimensions; ++i)
-        biases[i] = read_little_endian<BiasType>(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);
@@ -246,8 +265,7 @@ namespace Stockfish::Eval::NNUE::Layers {

    // Write network parameters
    bool write_parameters(std::ostream& stream) const {
-      for (IndexType i = 0; i < OutputDimensions; ++i)
-          write_little_endian<BiasType>(stream, biases[i]);
+      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)]);
@@ -286,6 +304,15 @@ namespace Stockfish::Eval::NNUE::Layers {
      #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;
@@ -368,8 +395,9 @@ namespace Stockfish::Eval::NNUE::Layers {
    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) < 2*64)>> {
+  class AffineTransform<InDims, OutDims, std::enable_if_t<(ceil_to_multiple<IndexType>(InDims, MaxSimdWidth) < LargeInputSize)>> {
   public:
    // Input/output type
    // Input/output type
@@ -387,12 +415,7 @@ namespace Stockfish::Eval::NNUE::Layers {

    using OutputBuffer = OutputType[PaddedOutputDimensions];

-    static_assert(PaddedInputDimensions < 128, "Something went wrong. This specialization should not have been chosen.");
-
-#if defined (USE_SSSE3)
-    static constexpr const IndexType OutputSimdWidth = SimdWidth / 4;
-    static constexpr const IndexType InputSimdWidth = SimdWidth;
-#endif
+    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) {
@@ -422,8 +445,7 @@ namespace Stockfish::Eval::NNUE::Layers {

    // Read network parameters
    bool read_parameters(std::istream& stream) {
-      for (IndexType i = 0; i < OutputDimensions; ++i)
-        biases[i] = read_little_endian<BiasType>(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);

@@ -432,8 +454,7 @@ namespace Stockfish::Eval::NNUE::Layers {

    // Write network parameters
    bool write_parameters(std::ostream& stream) const {
-      for (IndexType i = 0; i < OutputDimensions; ++i)
-        write_little_endian<BiasType>(stream, biases[i]);
+      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)]);
@@ -444,29 +465,34 @@ namespace Stockfish::Eval::NNUE::Layers {
    const OutputType* propagate(
        const InputType* input, OutputType* output) const {

-#if defined (USE_AVX2)
+#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_add_dpbusd_32x4 Simd::m256_add_dpbusd_epi32x4
      #define vec_hadd Simd::m256_hadd
-      #define vec_haddx4 Simd::m256_haddx4
 #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_add_dpbusd_32x4 Simd::m128_add_dpbusd_epi32x4
      #define vec_hadd Simd::m128_hadd
-      #define vec_haddx4 Simd::m128_haddx4
 #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)
@@ -512,9 +538,7 @@ namespace Stockfish::Eval::NNUE::Layers {
 # undef vec_set_32
 # undef vec_add_dpbusd_32
 # undef vec_add_dpbusd_32x2
-# undef vec_add_dpbusd_32x4
 # undef vec_hadd
-# undef vec_haddx4
 #else
      // Use old implementation for the other architectures.
      affine_transform_non_ssse3<
@@ -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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -153,7 +153,7 @@ namespace Stockfish::Simd {
      asm(
        "vpdpbusd %[b0], %[a0], %[acc]\n\t"
        "vpdpbusd %[b1], %[a1], %[acc]\n\t"
-        : [acc]"+v"(acc)
+        : [acc]"+&v"(acc)
        : [a0]"v"(a0), [b0]"vm"(b0), [a1]"v"(a1), [b1]"vm"(b1)
      );
 #   else
@@ -165,18 +165,19 @@ namespace Stockfish::Simd {
      __m512i tmp0 = _mm512_maddubs_epi16(a0, b0);
      __m512i tmp1 = _mm512_maddubs_epi16(a1, b1);
      asm(
-          "vpaddsw     %[tmp0], %[tmp1], %[tmp0]\n\t"
          "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))
+          : [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_adds_epi16(product0, product1);
      product0 = _mm512_madd_epi16(product0, _mm512_set1_epi16(1));
-      acc = _mm512_add_epi32(acc, product0);
+      product1 = _mm512_madd_epi16(product1, _mm512_set1_epi16(1));
+      acc = _mm512_add_epi32(acc, _mm512_add_epi32(product0, product1));
 #   endif
 # endif
    }
@@ -249,7 +250,7 @@ namespace Stockfish::Simd {
      asm(
        VNNI_PREFIX "vpdpbusd %[b0], %[a0], %[acc]\n\t"
        VNNI_PREFIX "vpdpbusd %[b1], %[a1], %[acc]\n\t"
-        : [acc]"+v"(acc)
+        : [acc]"+&v"(acc)
        : [a0]"v"(a0), [b0]"vm"(b0), [a1]"v"(a1), [b1]"vm"(b1)
      );
 #   else
@@ -261,18 +262,19 @@ namespace Stockfish::Simd {
      __m256i tmp0 = _mm256_maddubs_epi16(a0, b0);
      __m256i tmp1 = _mm256_maddubs_epi16(a1, b1);
      asm(
-          "vpaddsw     %[tmp0], %[tmp1], %[tmp0]\n\t"
          "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))
+          : [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_adds_epi16(product0, product1);
      product0 = _mm256_madd_epi16(product0, _mm256_set1_epi16(1));
-      acc = _mm256_add_epi32(acc, product0);
+      product1 = _mm256_madd_epi16(product1, _mm256_set1_epi16(1));
+      acc = _mm256_add_epi32(acc, _mm256_add_epi32(product0, product1));
 #   endif
 # endif
    }
@@ -326,23 +328,37 @@ namespace Stockfish::Simd {
      __m128i tmp0 = _mm_maddubs_epi16(a0, b0);
      __m128i tmp1 = _mm_maddubs_epi16(a1, b1);
      asm(
-          "paddsw     %[tmp1], %[tmp0]\n\t"
          "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))
+          : [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_adds_epi16(product0, product1);
      product0 = _mm_madd_epi16(product0, _mm_set1_epi16(1));
-      acc = _mm_add_epi32(acc, product0);
+      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) {
@@ -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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -27,8 +27,10 @@

 #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"

@@ -38,7 +40,7 @@ namespace Stockfish::Eval::NNUE {
 using FeatureSet = Features::HalfKAv2_hm;

 // Number of input feature dimensions after conversion
-constexpr IndexType TransformedFeatureDimensions = 1024;
+constexpr IndexType TransformedFeatureDimensions = 1536;
 constexpr IndexType PSQTBuckets = 8;
 constexpr IndexType LayerStacks = 8;

@@ -47,9 +49,10 @@ struct Network
  static constexpr int FC_0_OUTPUTS = 15;
  static constexpr int FC_1_OUTPUTS = 32;

-  Layers::AffineTransform<TransformedFeatureDimensions, FC_0_OUTPUTS + 1> fc_0;
+  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, FC_1_OUTPUTS> fc_1;
+  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;

@@ -70,22 +73,20 @@ struct Network

  // Read network parameters
  bool read_parameters(std::istream& stream) {
-    if (!fc_0.read_parameters(stream)) return false;
-    if (!ac_0.read_parameters(stream)) return false;
-    if (!fc_1.read_parameters(stream)) return false;
-    if (!ac_1.read_parameters(stream)) return false;
-    if (!fc_2.read_parameters(stream)) return false;
-    return true;
+    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);
  }

-  // Read network parameters
+  // Write network parameters
  bool write_parameters(std::ostream& stream) const {
-    if (!fc_0.write_parameters(stream)) return false;
-    if (!ac_0.write_parameters(stream)) return false;
-    if (!fc_1.write_parameters(stream)) return false;
-    if (!ac_1.write_parameters(stream)) return false;
-    if (!fc_2.write_parameters(stream)) return false;
-    return true;
+    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)
@@ -93,6 +94,7 @@ struct Network
    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;
@@ -114,8 +116,10 @@ struct Network
 #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);
-    fc_1.propagate(buffer.ac_0_out, buffer.fc_1_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);

@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -57,6 +57,9 @@ namespace Stockfish::Eval::NNUE {
  // 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;
@@ -159,6 +162,80 @@ namespace Stockfish::Eval::NNUE {
              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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -25,6 +25,7 @@
 #include "nnue_architecture.h"

 #include <cstring> // std::memset()
+#include <utility> // std::pair

 namespace Stockfish::Eval::NNUE {

@@ -41,8 +42,8 @@ namespace Stockfish::Eval::NNUE {
    "Per feature PSQT values cannot be processed at granularity lower than 8 at a time.");

  #ifdef USE_AVX512
-  typedef __m512i vec_t;
-  typedef __m256i psqt_vec_t;
+  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)
@@ -65,8 +66,8 @@ namespace Stockfish::Eval::NNUE {
  #define MaxChunkSize 64

  #elif USE_AVX2
-  typedef __m256i vec_t;
-  typedef __m256i psqt_vec_t;
+  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)
@@ -89,8 +90,8 @@ namespace Stockfish::Eval::NNUE {
  #define MaxChunkSize 32

  #elif USE_SSE2
-  typedef __m128i vec_t;
-  typedef __m128i psqt_vec_t;
+  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)
@@ -110,8 +111,8 @@ namespace Stockfish::Eval::NNUE {
  #define MaxChunkSize 16

  #elif USE_MMX
-  typedef __m64 vec_t;
-  typedef __m64 psqt_vec_t;
+  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)
@@ -120,12 +121,12 @@ namespace Stockfish::Eval::NNUE {
  #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));
+    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));
+    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))
@@ -138,8 +139,8 @@ namespace Stockfish::Eval::NNUE {
  #define MaxChunkSize 8

  #elif USE_NEON
-  typedef int16x8_t vec_t;
-  typedef int32x4_t psqt_vec_t;
+  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)
@@ -150,10 +151,10 @@ namespace Stockfish::Eval::NNUE {
  #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);
+    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)
@@ -252,9 +253,9 @@ namespace Stockfish::Eval::NNUE {
    // Read network parameters
    bool read_parameters(std::istream& stream) {

-      read_little_endian<BiasType      >(stream, biases     , HalfDimensions                  );
-      read_little_endian<WeightType    >(stream, weights    , HalfDimensions * InputDimensions);
-      read_little_endian<PSQTWeightType>(stream, psqtWeights, PSQTBuckets    * InputDimensions);
+      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();
    }
@@ -262,17 +263,17 @@ namespace Stockfish::Eval::NNUE {
    // Write network parameters
    bool write_parameters(std::ostream& stream) const {

-      write_little_endian<BiasType      >(stream, biases     , HalfDimensions                  );
-      write_little_endian<WeightType    >(stream, weights    , HalfDimensions * InputDimensions);
-      write_little_endian<PSQTWeightType>(stream, psqtWeights, PSQTBuckets    * InputDimensions);
+      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(pos, WHITE);
-      update_accumulator(pos, BLACK);
+      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;
@@ -290,7 +291,7 @@ namespace Stockfish::Eval::NNUE {

 #if defined(VECTOR)

-	  constexpr IndexType OutputChunkSize = MaxChunkSize;
+          constexpr IndexType OutputChunkSize = MaxChunkSize;
          static_assert((HalfDimensions / 2) % OutputChunkSize == 0);
          constexpr IndexType NumOutputChunks = HalfDimensions / 2 / OutputChunkSize;

@@ -332,18 +333,41 @@ namespace Stockfish::Eval::NNUE {
 #endif

      return psqt;
+    } // end of function transform()

-   } // 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:
-    void update_accumulator(const Position& pos, const Color perspective) const {
+    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 };
+    }

-      // 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.
+    // 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
@@ -352,232 +376,294 @@ namespace Stockfish::Eval::NNUE {
      psqt_vec_t psqt[NumPsqtRegs];
  #endif

-      // 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])
+      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];
+
      {
-        // 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;
+        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]);
+        }
      }

-      if (st->accumulator.computed[perspective])
+      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)
      {
-        if (next == nullptr)
-          return;
+        // 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]);

-        // Update incrementally in two steps. First, we update the "next"
-        // accumulator. Then, we update the current accumulator (pos.state()).
-
-        // Gather all features to be updated.
-        const Square ksq = pos.square<KING>(perspective);
-        FeatureSet::IndexList removed[2], added[2];
-        FeatureSet::append_changed_indices(
-          ksq, next->dirtyPiece, perspective, removed[0], added[0]);
-        for (StateInfo *st2 = pos.state(); st2 != next; st2 = st2->previous)
-          FeatureSet::append_changed_indices(
-            ksq, st2->dirtyPiece, perspective, removed[1], added[1]);
-
-        // Mark the accumulators as computed.
-        next->accumulator.computed[perspective] = true;
-        pos.state()->accumulator.computed[perspective] = true;
-
-        // Now update the accumulators listed in states_to_update[], where the last element is a sentinel.
-        StateInfo *states_to_update[3] =
-          { next, next == pos.state() ? nullptr : pos.state(), nullptr };
-  #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];
+            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;
-
-            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
-      }
-      else
-      {
-        // Refresh the accumulator
-        auto& accumulator = pos.state()->accumulator;
-        accumulator.computed[perspective] = true;
-        FeatureSet::IndexList active;
-        FeatureSet::append_active_indices(pos, perspective, 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)
+            for (IndexType 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++)
+          // 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)
+      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)
        {
-          for (std::size_t k = 0; k < NumPsqtRegs; ++k)
-            psqt[k] = vec_zero_psqt();
-
-          for (const auto index : active)
+          // 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]);
          }

-          auto accTilePsqt = reinterpret_cast<psqt_vec_t*>(
-            &accumulator.psqtAccumulation[perspective][j * PsqtTileHeight]);
+          // 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
-        std::memcpy(accumulator.accumulation[perspective], biases,
+#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)
-          accumulator.psqtAccumulation[perspective][k] = 0;
+          states_to_update[i]->accumulator.psqtAccumulation[Perspective][k] = st->accumulator.psqtAccumulation[Perspective][k];

-        for (const auto index : active)
+        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)
-            accumulator.accumulation[perspective][j] += weights[offset + j];
+            st->accumulator.accumulation[Perspective][j] -= weights[offset + j];

          for (std::size_t k = 0; k < PSQTBuckets; ++k)
-            accumulator.psqtAccumulation[perspective][k] += psqtWeights[index * 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
      }
+#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];
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -61,7 +61,7 @@ struct Entry {
  int blockedCount;
 };

-typedef HashTable<Entry, 131072> Table;
+using Table = HashTable<Entry, 131072>;

 Entry* probe(const Position& pos);

@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -22,6 +22,7 @@
 #include <cstring> // For std::memset, std::memcmp
 #include <iomanip>
 #include <sstream>
+#include <string_view>

 #include "bitboard.h"
 #include "misc.h"
@@ -46,7 +47,7 @@ namespace Zobrist {

 namespace {

-const string PieceToChar(" PNBRQK  pnbrqk");
+constexpr std::string_view PieceToChar(" PNBRQK  pnbrqk");

 constexpr Piece Pieces[] = { W_PAWN, W_KNIGHT, W_BISHOP, W_ROOK, W_QUEEN, W_KING,
                             B_PAWN, B_KNIGHT, B_BISHOP, B_ROOK, B_QUEEN, B_KING };
@@ -96,7 +97,7 @@ std::ostream& operator<<(std::ostream& os, const Position& pos) {

 // Marcel van Kervinck's cuckoo algorithm for fast detection of "upcoming repetition"
 // situations. Description of the algorithm in the following paper:
-// https://marcelk.net/2013-04-06/paper/upcoming-rep-v2.pdf
+// http://web.archive.org/web/20201107002606/https://marcelk.net/2013-04-06/paper/upcoming-rep-v2.pdf

 // First and second hash functions for indexing the cuckoo tables
 inline int H1(Key h) { return h & 0x1fff; }
@@ -129,7 +130,7 @@ void Position::init() {
  // Prepare the cuckoo tables
  std::memset(cuckoo, 0, sizeof(cuckoo));
  std::memset(cuckooMove, 0, sizeof(cuckooMove));
-  int count = 0;
+  [[maybe_unused]] int count = 0;
  for (Piece pc : Pieces)
      for (Square s1 = SQ_A1; s1 <= SQ_H8; ++s1)
          for (Square s2 = Square(s1 + 1); s2 <= SQ_H8; ++s2)
@@ -281,7 +282,7 @@ Position& Position::set(const string& fenStr, bool isChess960, StateInfo* si, Th

  chess960 = isChess960;
  thisThread = th;
-  set_state(st);
+  set_state();

  assert(pos_is_ok());

@@ -312,60 +313,59 @@ void Position::set_castling_right(Color c, Square rfrom) {

 /// Position::set_check_info() sets king attacks to detect if a move gives check

-void Position::set_check_info(StateInfo* si) const {
+void Position::set_check_info() const {

-  si->blockersForKing[WHITE] = slider_blockers(pieces(BLACK), square<KING>(WHITE), si->pinners[BLACK]);
-  si->blockersForKing[BLACK] = slider_blockers(pieces(WHITE), square<KING>(BLACK), si->pinners[WHITE]);
+  st->blockersForKing[WHITE] = slider_blockers(pieces(BLACK), square<KING>(WHITE), st->pinners[BLACK]);
+  st->blockersForKing[BLACK] = slider_blockers(pieces(WHITE), square<KING>(BLACK), st->pinners[WHITE]);

  Square ksq = square<KING>(~sideToMove);

-  si->checkSquares[PAWN]   = pawn_attacks_bb(~sideToMove, ksq);
-  si->checkSquares[KNIGHT] = attacks_bb<KNIGHT>(ksq);
-  si->checkSquares[BISHOP] = attacks_bb<BISHOP>(ksq, pieces());
-  si->checkSquares[ROOK]   = attacks_bb<ROOK>(ksq, pieces());
-  si->checkSquares[QUEEN]  = si->checkSquares[BISHOP] | si->checkSquares[ROOK];
-  si->checkSquares[KING]   = 0;
+  st->checkSquares[PAWN]   = pawn_attacks_bb(~sideToMove, ksq);
+  st->checkSquares[KNIGHT] = attacks_bb<KNIGHT>(ksq);
+  st->checkSquares[BISHOP] = attacks_bb<BISHOP>(ksq, pieces());
+  st->checkSquares[ROOK]   = attacks_bb<ROOK>(ksq, pieces());
+  st->checkSquares[QUEEN]  = st->checkSquares[BISHOP] | st->checkSquares[ROOK];
+  st->checkSquares[KING]   = 0;
 }


 /// Position::set_state() computes the hash keys of the position, and other
 /// data that once computed is updated incrementally as moves are made.
-/// The function is only used when a new position is set up, and to verify
-/// the correctness of the StateInfo data when running in debug mode.
+/// The function is only used when a new position is set up

-void Position::set_state(StateInfo* si) const {
+void Position::set_state() const {

-  si->key = si->materialKey = 0;
-  si->pawnKey = Zobrist::noPawns;
-  si->nonPawnMaterial[WHITE] = si->nonPawnMaterial[BLACK] = VALUE_ZERO;
-  si->checkersBB = attackers_to(square<KING>(sideToMove)) & pieces(~sideToMove);
+  st->key = st->materialKey = 0;
+  st->pawnKey = Zobrist::noPawns;
+  st->nonPawnMaterial[WHITE] = st->nonPawnMaterial[BLACK] = VALUE_ZERO;
+  st->checkersBB = attackers_to(square<KING>(sideToMove)) & pieces(~sideToMove);

-  set_check_info(si);
+  set_check_info();

  for (Bitboard b = pieces(); b; )
  {
      Square s = pop_lsb(b);
      Piece pc = piece_on(s);
-      si->key ^= Zobrist::psq[pc][s];
+      st->key ^= Zobrist::psq[pc][s];

      if (type_of(pc) == PAWN)
-          si->pawnKey ^= Zobrist::psq[pc][s];
+          st->pawnKey ^= Zobrist::psq[pc][s];

      else if (type_of(pc) != KING)
-          si->nonPawnMaterial[color_of(pc)] += PieceValue[MG][pc];
+          st->nonPawnMaterial[color_of(pc)] += PieceValue[MG][pc];
  }

-  if (si->epSquare != SQ_NONE)
-      si->key ^= Zobrist::enpassant[file_of(si->epSquare)];
+  if (st->epSquare != SQ_NONE)
+      st->key ^= Zobrist::enpassant[file_of(st->epSquare)];

  if (sideToMove == BLACK)
-      si->key ^= Zobrist::side;
+      st->key ^= Zobrist::side;

-  si->key ^= Zobrist::castling[si->castlingRights];
+  st->key ^= Zobrist::castling[st->castlingRights];

  for (Piece pc : Pieces)
      for (int cnt = 0; cnt < pieceCount[pc]; ++cnt)
-          si->materialKey ^= Zobrist::psq[pc][cnt];
+          st->materialKey ^= Zobrist::psq[pc][cnt];
 }


@@ -568,8 +568,7 @@ bool Position::pseudo_legal(const Move m) const {
                        : MoveList<NON_EVASIONS>(*this).contains(m);

  // Is not a promotion, so promotion piece must be empty
-  if (promotion_type(m) - KNIGHT != NO_PIECE_TYPE)
-      return false;
+  assert(promotion_type(m) - KNIGHT == NO_PIECE_TYPE);

  // If the 'from' square is not occupied by a piece belonging to the side to
  // move, the move is obviously not legal.
@@ -765,9 +764,6 @@ void Position::do_move(Move m, StateInfo& newSt, bool givesCheck) {
      // Update board and piece lists
      remove_piece(capsq);

-      if (type_of(m) == EN_PASSANT)
-          board[capsq] = NO_PIECE;
-
      // Update material hash key and prefetch access to materialTable
      k ^= Zobrist::psq[captured][capsq];
      st->materialKey ^= Zobrist::psq[captured][pieceCount[captured]];
@@ -868,7 +864,7 @@ void Position::do_move(Move m, StateInfo& newSt, bool givesCheck) {
  sideToMove = ~sideToMove;

  // Update king attacks used for fast check detection
-  set_check_info(st);
+  set_check_info();

  // Calculate the repetition info. It is the ply distance from the previous
  // occurrence of the same position, negative in the 3-fold case, or zero
@@ -1020,7 +1016,7 @@ void Position::do_null_move(StateInfo& newSt) {

  sideToMove = ~sideToMove;

-  set_check_info(st);
+  set_check_info();

  st->repetition = 0;

@@ -1054,7 +1050,10 @@ Key Position::key_after(Move m) const {
  if (captured)
      k ^= Zobrist::psq[captured][to];

-  return k ^ Zobrist::psq[pc][to] ^ Zobrist::psq[pc][from];
+  k ^= Zobrist::psq[pc][to] ^ Zobrist::psq[pc][from];
+
+  return (captured || type_of(pc) == PAWN)
+      ? k : adjust_key50<true>(k);
 }


@@ -1062,7 +1061,7 @@ Key Position::key_after(Move m) const {
 /// SEE value of move is greater or equal to the given threshold. We'll use an
 /// algorithm similar to alpha-beta pruning with a null window.

-bool Position::see_ge(Move m, Value threshold) const {
+bool Position::see_ge(Move m, Bitboard& occupied, Value threshold) const {

  assert(is_ok(m));

@@ -1081,7 +1080,7 @@ bool Position::see_ge(Move m, Value threshold) const {
      return true;

  assert(color_of(piece_on(from)) == sideToMove);
-  Bitboard occupied = pieces() ^ from ^ to;
+  occupied = pieces() ^ from ^ to; // xoring to is important for pinned piece logic
  Color stm = sideToMove;
  Bitboard attackers = attackers_to(to, occupied);
  Bitboard stmAttackers, bb;
@@ -1099,10 +1098,12 @@ bool Position::see_ge(Move m, Value threshold) const {
      // Don't allow pinned pieces to attack as long as there are
      // pinners on their original square.
      if (pinners(~stm) & occupied)
+      {
          stmAttackers &= ~blockers_for_king(stm);

-      if (!stmAttackers)
-          break;
+          if (!stmAttackers)
+              break;
+      }

      res ^= 1;

@@ -1110,45 +1111,44 @@ bool Position::see_ge(Move m, Value threshold) const {
      // the bitboard 'attackers' any X-ray attackers behind it.
      if ((bb = stmAttackers & pieces(PAWN)))
      {
+          occupied ^= least_significant_square_bb(bb);
          if ((swap = PawnValueMg - swap) < res)
              break;

-          occupied ^= least_significant_square_bb(bb);
          attackers |= attacks_bb<BISHOP>(to, occupied) & pieces(BISHOP, QUEEN);
      }

      else if ((bb = stmAttackers & pieces(KNIGHT)))
      {
+          occupied ^= least_significant_square_bb(bb);
          if ((swap = KnightValueMg - swap) < res)
              break;
-
-          occupied ^= least_significant_square_bb(bb);
      }

      else if ((bb = stmAttackers & pieces(BISHOP)))
      {
+          occupied ^= least_significant_square_bb(bb);
          if ((swap = BishopValueMg - swap) < res)
              break;

-          occupied ^= least_significant_square_bb(bb);
          attackers |= attacks_bb<BISHOP>(to, occupied) & pieces(BISHOP, QUEEN);
      }

      else if ((bb = stmAttackers & pieces(ROOK)))
      {
+          occupied ^= least_significant_square_bb(bb);
          if ((swap = RookValueMg - swap) < res)
              break;

-          occupied ^= least_significant_square_bb(bb);
          attackers |= attacks_bb<ROOK>(to, occupied) & pieces(ROOK, QUEEN);
      }

      else if ((bb = stmAttackers & pieces(QUEEN)))
      {
+          occupied ^= least_significant_square_bb(bb);
          if ((swap = QueenValueMg - swap) < res)
              break;

-          occupied ^= least_significant_square_bb(bb);
          attackers |=  (attacks_bb<BISHOP>(to, occupied) & pieces(BISHOP, QUEEN))
                      | (attacks_bb<ROOK  >(to, occupied) & pieces(ROOK  , QUEEN));
      }
@@ -1162,6 +1162,11 @@ bool Position::see_ge(Move m, Value threshold) const {
  return bool(res);
 }

+bool Position::see_ge(Move m, Value threshold) const {
+    Bitboard occupied;
+    return see_ge(m, occupied, threshold);
+}
+

 /// Position::is_draw() tests whether the position is drawn by 50-move rule
 /// or by repetition. It does not detect stalemates.
@@ -1318,12 +1323,6 @@ bool Position::pos_is_ok() const {
          if (p1 != p2 && (pieces(p1) & pieces(p2)))
              assert(0 && "pos_is_ok: Bitboards");

-  StateInfo si = *st;
-  ASSERT_ALIGNED(&si, Eval::NNUE::CacheLineSize);
-
-  set_state(&si);
-  if (std::memcmp(&si, st, sizeof(StateInfo)))
-      assert(0 && "pos_is_ok: State");

  for (Piece pc : Pieces)
      if (   pieceCount[pc] != popcount(pieces(color_of(pc), type_of(pc)))
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -68,7 +68,7 @@ struct StateInfo {
 /// 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.
-typedef std::unique_ptr<std::deque<StateInfo>> StateListPtr;
+using StateListPtr = std::unique_ptr<std::deque<StateInfo>>;


 /// Position class stores information regarding the board representation as
@@ -92,10 +92,9 @@ public:

  // Position representation
  Bitboard pieces(PieceType pt) const;
-  Bitboard pieces(PieceType pt1, PieceType pt2) const;
+  template<typename ...PieceTypes> Bitboard pieces(PieceType pt, PieceTypes... pts) const;
  Bitboard pieces(Color c) const;
-  Bitboard pieces(Color c, PieceType pt) const;
-  Bitboard pieces(Color c, PieceType pt1, PieceType pt2) 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;
@@ -126,6 +125,7 @@ public:
  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;
@@ -144,6 +144,7 @@ public:

  // 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;
@@ -161,6 +162,7 @@ public:
  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;

@@ -177,13 +179,15 @@ public:
 private:
  // Initialization helpers (used while setting up a position)
  void set_castling_right(Color c, Square rfrom);
-  void set_state(StateInfo* si) const;
-  void set_check_info(StateInfo* si) const;
+  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];
@@ -201,7 +205,7 @@ private:
  bool chess960;
 };

-extern std::ostream& operator<<(std::ostream& os, const Position& pos);
+std::ostream& operator<<(std::ostream& os, const Position& pos);

 inline Color Position::side_to_move() const {
  return sideToMove;
@@ -224,20 +228,18 @@ inline Bitboard Position::pieces(PieceType pt = ALL_PIECES) const {
  return byTypeBB[pt];
 }

-inline Bitboard Position::pieces(PieceType pt1, PieceType pt2) const {
-  return pieces(pt1) | pieces(pt2);
+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];
 }

-inline Bitboard Position::pieces(Color c, PieceType pt) const {
-  return pieces(c) & pieces(pt);
-}
-
-inline Bitboard Position::pieces(Color c, PieceType pt1, PieceType pt2) const {
-  return pieces(c) & (pieces(pt1) | pieces(pt2));
+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 {
@@ -326,8 +328,14 @@ inline int Position::pawns_on_same_color_squares(Color c, Square s) const {
 }

 inline Key Position::key() const {
-  return st->rule50 < 14 ? st->key
-                         : st->key ^ make_key((st->rule50 - 14) / 8);
+  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 {
@@ -342,6 +350,10 @@ 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];
 }
@@ -370,8 +382,16 @@ inline bool Position::is_chess960() const {

 inline bool Position::capture(Move m) const {
  assert(is_ok(m));
-  // Castling is encoded as "king captures rook"
-  return (!empty(to_sq(m)) && type_of(m) != CASTLING) || type_of(m) == EN_PASSANT;
+  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 {
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -30,7 +30,7 @@ namespace Stockfish::PSQT
 extern Score psq[PIECE_NB][SQUARE_NB];

 // Fill psqt array from a set of internally linked parameters
-extern void init();
+void init();

 } // namespace Stockfish::PSQT

@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -31,9 +31,6 @@ class Position;

 namespace Search {

-/// Threshold used for countermoves based pruning
-constexpr int CounterMovePruneThreshold = 0;
-

 /// 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
@@ -47,13 +44,13 @@ struct Stack {
  Move excludedMove;
  Move killers[2];
  Value staticEval;
-  Depth depth;
  int statScore;
  int moveCount;
  bool inCheck;
  bool ttPv;
  bool ttHit;
  int doubleExtensions;
+  int cutoffCnt;
 };


@@ -74,13 +71,16 @@ struct RootMove {
  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;
 };

-typedef std::vector<RootMove> RootMoves;
+using RootMoves = std::vector<RootMove>;


 /// LimitsType struct stores information sent by GUI about available time to
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -24,9 +24,10 @@
 #include <fstream>
 #include <iostream>
 #include <list>
-#include <sstream>
-#include <type_traits>
 #include <mutex>
+#include <sstream>
+#include <string_view>
+#include <type_traits>

 #include "../bitboard.h"
 #include "../movegen.h"
@@ -59,6 +60,7 @@ namespace Stockfish {
 namespace {

 constexpr int TBPIECES = 7; // Max number of supported pieces
+constexpr int MAX_DTZ = 1 << 18; // Max DTZ supported, large enough to deal with the syzygy TB limit.

 enum { BigEndian, LittleEndian };
 enum TBType { WDL, DTZ }; // Used as template parameter
@@ -69,7 +71,7 @@ enum TBFlag { STM = 1, Mapped = 2, WinPlies = 4, LossPlies = 8, Wide = 16, Singl
 inline WDLScore operator-(WDLScore d) { return WDLScore(-int(d)); }
 inline Square operator^(Square s, int i) { return Square(int(s) ^ i); }

-const std::string PieceToChar = " PNBRQK  pnbrqk";
+constexpr std::string_view PieceToChar = " PNBRQK  pnbrqk";

 int MapPawns[SQUARE_NB];
 int MapB1H1H7[SQUARE_NB];
@@ -140,7 +142,7 @@ struct SparseEntry {

 static_assert(sizeof(SparseEntry) == 6, "SparseEntry must be 6 bytes");

-typedef uint16_t Sym; // Huffman symbol
+using Sym = uint16_t; // Huffman symbol

 struct LR {
    enum Side { Left, Right };
@@ -198,13 +200,10 @@ public:
        }
    }

-    // Memory map the file and check it. File should be already open and will be
-    // closed after mapping.
+    // Memory map the file and check it.
    uint8_t* map(void** baseAddress, uint64_t* mapping, TBType type) {
-
-        assert(is_open());
-
-        close(); // Need to re-open to get native file descriptor
+        if (is_open())
+            close(); // Need to re-open to get native file descriptor

 #ifndef _WIN32
        struct stat statbuf;
@@ -235,7 +234,7 @@ public:
        }
 #else
        // Note FILE_FLAG_RANDOM_ACCESS is only a hint to Windows and as such may get ignored.
-        HANDLE fd = CreateFile(fname.c_str(), GENERIC_READ, FILE_SHARE_READ, nullptr,
+        HANDLE fd = CreateFileA(fname.c_str(), GENERIC_READ, FILE_SHARE_READ, nullptr,
                               OPEN_EXISTING, FILE_FLAG_RANDOM_ACCESS, nullptr);

        if (fd == INVALID_HANDLE_VALUE)
@@ -328,7 +327,7 @@ struct PairsData {
 // first access, when the corresponding file is memory mapped.
 template<TBType Type>
 struct TBTable {
-    typedef typename std::conditional<Type == WDL, WDLScore, int>::type Ret;
+    using Ret = typename std::conditional<Type == WDL, WDLScore, int>::type;

    static constexpr int Sides = Type == WDL ? 2 : 1;

@@ -1290,7 +1289,7 @@ void Tablebases::init(const std::string& paths) {
    for (auto s : diagonal)
        MapA1D1D4[s] = code++;

-    // MapKK[] encodes all the 461 possible legal positions of two kings where
+    // MapKK[] encodes all the 462 possible legal positions of two kings where
    // the first is in the a1-d1-d4 triangle. If the first king is on the a1-d4
    // diagonal, the other one shall not to be above the a1-h8 diagonal.
    std::vector<std::pair<int, Square>> bothOnDiagonal;
@@ -1513,7 +1512,7 @@ int Tablebases::probe_dtz(Position& pos, ProbeState* result) {
 // A return value false indicates that not all probes were successful.
 bool Tablebases::root_probe(Position& pos, Search::RootMoves& rootMoves) {

-    ProbeState result;
+    ProbeState result = OK;
    StateInfo st;

    // Obtain 50-move counter for the root position
@@ -1522,7 +1521,7 @@ bool Tablebases::root_probe(Position& pos, Search::RootMoves& rootMoves) {
    // Check whether a position was repeated since the last zeroing move.
    bool rep = pos.has_repeated();

-    int dtz, bound = Options["Syzygy50MoveRule"] ? 900 : 1;
+    int dtz, bound = Options["Syzygy50MoveRule"] ? (MAX_DTZ - 100) : 1;

    // Probe and rank each move
    for (auto& m : rootMoves)
@@ -1565,8 +1564,8 @@ bool Tablebases::root_probe(Position& pos, Search::RootMoves& rootMoves) {

        // Better moves are ranked higher. Certain wins are ranked equally.
        // Losing moves are ranked equally unless a 50-move draw is in sight.
-        int r =  dtz > 0 ? (dtz + cnt50 <= 99 && !rep ? 1000 : 1000 - (dtz + cnt50))
-               : dtz < 0 ? (-dtz * 2 + cnt50 < 100 ? -1000 : -1000 + (-dtz + cnt50))
+        int r =  dtz > 0 ? (dtz + cnt50 <= 99 && !rep ? MAX_DTZ : MAX_DTZ - (dtz + cnt50))
+               : dtz < 0 ? (-dtz * 2 + cnt50 < 100 ? -MAX_DTZ : -MAX_DTZ + (-dtz + cnt50))
               : 0;
        m.tbRank = r;

@@ -1574,9 +1573,9 @@ bool Tablebases::root_probe(Position& pos, Search::RootMoves& rootMoves) {
        // 1 cp to cursed wins and let it grow to 49 cp as the positions gets
        // closer to a real win.
        m.tbScore =  r >= bound ? VALUE_MATE - MAX_PLY - 1
-                   : r >  0     ? Value((std::max( 3, r - 800) * int(PawnValueEg)) / 200)
+                   : r >  0     ? Value((std::max( 3, r - (MAX_DTZ - 200)) * int(PawnValueEg)) / 200)
                   : r == 0     ? VALUE_DRAW
-                   : r > -bound ? Value((std::min(-3, r + 800) * int(PawnValueEg)) / 200)
+                   : r > -bound ? Value((std::min(-3, r + (MAX_DTZ - 200)) * int(PawnValueEg)) / 200)
                   :             -VALUE_MATE + MAX_PLY + 1;
    }

@@ -1590,9 +1589,9 @@ bool Tablebases::root_probe(Position& pos, Search::RootMoves& rootMoves) {
 // A return value false indicates that not all probes were successful.
 bool Tablebases::root_probe_wdl(Position& pos, Search::RootMoves& rootMoves) {

-    static const int WDL_to_rank[] = { -1000, -899, 0, 899, 1000 };
+    static const int WDL_to_rank[] = { -MAX_DTZ, -MAX_DTZ + 101, 0, MAX_DTZ - 101, MAX_DTZ };

-    ProbeState result;
+    ProbeState result = OK;
    StateInfo st;
    WDLScore wdl;

@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -31,8 +31,6 @@ enum WDLScore {
    WDLDraw        =  0, // Draw
    WDLCursedWin   =  1, // Win, but draw under 50-move rule
    WDLWin         =  2, // Win
-
-    WDLScoreNone  = -1000
 };

 // Possible states after a probing operation
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -63,21 +63,18 @@ void Thread::clear() {

  for (bool inCheck : { false, true })
      for (StatsType c : { NoCaptures, Captures })
-      {
          for (auto& to : continuationHistory[inCheck][c])
-                for (auto& h : to)
-                      h->fill(-71);
-          continuationHistory[inCheck][c][NO_PIECE][0]->fill(Search::CounterMovePruneThreshold - 1);
-      }
+              for (auto& h : to)
+                  h->fill(-71);
 }


 /// Thread::start_searching() wakes up the thread that will start the search

 void Thread::start_searching() {
-
-  std::lock_guard<std::mutex> lk(mutex);
+  mutex.lock();
  searching = true;
+  mutex.unlock(); // Unlock before notifying saves a few CPU-cycles
  cv.notify_one(); // Wake up the thread in idle_loop()
 }

@@ -127,20 +124,20 @@ void Thread::idle_loop() {

 void ThreadPool::set(size_t requested) {

-  if (size() > 0)   // destroy any existing thread(s)
+  if (threads.size() > 0)   // destroy any existing thread(s)
  {
      main()->wait_for_search_finished();

-      while (size() > 0)
-          delete back(), pop_back();
+      while (threads.size() > 0)
+          delete threads.back(), threads.pop_back();
  }

  if (requested > 0)   // create new thread(s)
  {
-      push_back(new MainThread(0));
+      threads.push_back(new MainThread(0));

-      while (size() < requested)
-          push_back(new Thread(size()));
+      while (threads.size() < requested)
+          threads.push_back(new Thread(threads.size()));
      clear();

      // Reallocate the hash with the new threadpool size
@@ -156,7 +153,7 @@ void ThreadPool::set(size_t requested) {

 void ThreadPool::clear() {

-  for (Thread* th : *this)
+  for (Thread* th : threads)
      th->clear();

  main()->callsCnt = 0;
@@ -189,7 +186,7 @@ void ThreadPool::start_thinking(Position& pos, StateListPtr& states,
      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() == NULL.
+  // and call 'go' again without setting a new position states.get() == nullptr.
  assert(states.get() || setupStates.get());

  if (states.get())
@@ -200,7 +197,7 @@ void ThreadPool::start_thinking(Position& pos, StateListPtr& states,
  // 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 : *this)
+  for (Thread* th : threads)
  {
      th->nodes = th->tbHits = th->nmpMinPly = th->bestMoveChanges = 0;
      th->rootDepth = th->completedDepth = 0;
@@ -214,20 +211,23 @@ void ThreadPool::start_thinking(Position& pos, StateListPtr& states,

 Thread* ThreadPool::get_best_thread() const {

-    Thread* bestThread = front();
+    Thread* bestThread = threads.front();
    std::map<Move, int64_t> votes;
    Value minScore = VALUE_NONE;

    // Find minimum score of all threads
-    for (Thread* th: *this)
+    for (Thread* th: threads)
        minScore = std::min(minScore, th->rootMoves[0].score);

    // Vote according to score and depth, and select the best thread
-    for (Thread* th : *this)
-    {
-        votes[th->rootMoves[0].pv[0]] +=
-            (th->rootMoves[0].score - minScore + 14) * int(th->completedDepth);
+    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
@@ -236,9 +236,11 @@ Thread* ThreadPool::get_best_thread() const {
        }
        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]]
+                         || (   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;
 }
@@ -248,8 +250,8 @@ Thread* ThreadPool::get_best_thread() const {

 void ThreadPool::start_searching() {

-    for (Thread* th : *this)
-        if (th != front())
+    for (Thread* th : threads)
+        if (th != threads.front())
            th->start_searching();
 }

@@ -258,8 +260,8 @@ void ThreadPool::start_searching() {

 void ThreadPool::wait_for_search_finished() const {

-    for (Thread* th : *this)
-        if (th != front())
+    for (Thread* th : threads)
+        if (th != threads.front())
            th->wait_for_search_finished();
 }

@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -60,22 +60,19 @@ public:
  Pawns::Table pawnsTable;
  Material::Table materialTable;
  size_t pvIdx, pvLast;
-  RunningAverage complexityAverage;
  std::atomic<uint64_t> nodes, tbHits, bestMoveChanges;
  int selDepth, nmpMinPly;
-  Color nmpColor;
  Value bestValue, optimism[COLOR_NB];

  Position rootPos;
  StateInfo rootState;
  Search::RootMoves rootMoves;
-  Depth rootDepth, completedDepth, depth;
+  Depth rootDepth, completedDepth;
  Value rootDelta;
  CounterMoveHistory counterMoves;
  ButterflyHistory mainHistory;
  CapturePieceToHistory captureHistory;
  ContinuationHistory continuationHistory[2][2];
-  Score trend;
 };


@@ -102,13 +99,13 @@ struct MainThread : public Thread {
 /// parking and, most importantly, launching a thread. All the access to threads
 /// is done through this class.

-struct ThreadPool : public std::vector<Thread*> {
+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*>(front()); }
+  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;
@@ -117,13 +114,21 @@ struct ThreadPool : public std::vector<Thread*> {

  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 : *this)
+    for (Thread* th : threads)
        sum += (th->*member).load(std::memory_order_relaxed);
    return sum;
  }
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -41,7 +41,7 @@ void* start_routine(void* ptr)
   P* p = reinterpret_cast<P*>(ptr);
   (p->first->*(p->second))(); // Call member function pointer
   delete p;
-   return NULL;
+   return nullptr;
 }

 class NativeThread {
@@ -56,7 +56,7 @@ public:
    pthread_attr_setstacksize(attr, TH_STACK_SIZE);
    pthread_create(&thread, attr, start_routine<T>, new P(obj, fun));
  }
-  void join() { pthread_join(thread, NULL); }
+  void join() { pthread_join(thread, nullptr); }
 };

 } // namespace Stockfish
@@ -65,7 +65,7 @@ public:

 namespace Stockfish {

-typedef std::thread NativeThread;
+using NativeThread = std::thread;

 } // namespace Stockfish

@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -36,6 +36,12 @@ TimeManagement Time; // Our global time management object

 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"]);
@@ -59,8 +65,6 @@ void TimeManagement::init(Search::LimitsType& limits, Color us, int ply) {
      limits.npmsec = npmsec;
  }

-  startTime = limits.startTime;
-
  // Maximum move horizon of 50 moves
  int mtg = limits.movestogo ? std::min(limits.movestogo, 50) : 50;

@@ -80,7 +84,7 @@ void TimeManagement::init(Search::LimitsType& limits, Color us, int ply) {
  // game time for the current move, so also cap to 20% of available game time.
  if (limits.movestogo == 0)
  {
-      optScale = std::min(0.0084 + std::pow(ply + 3.0, 0.5) * 0.0042,
+      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);
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -87,7 +87,7 @@ void TranspositionTable::clear() {

  std::vector<std::thread> threads;

-  for (size_t idx = 0; idx < Options["Threads"]; ++idx)
+  for (size_t idx = 0; idx < size_t(Options["Threads"]); ++idx)
  {
      threads.emplace_back([this, idx]() {

@@ -98,7 +98,7 @@ void TranspositionTable::clear() {
          // 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 != Options["Threads"] - 1 ?
+                       len    = idx != size_t(Options["Threads"]) - 1 ?
                                stride : clusterCount - start;

          std::memset(&table[start], 0, len * sizeof(Cluster));
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -26,8 +26,8 @@

 namespace Stockfish {

-typedef std::pair<int, int> Range; // Option's min-max values
-typedef Range (RangeFun) (int);
+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) {
@@ -75,7 +75,7 @@ struct SetRange {

 class Tune {

-  typedef void (PostUpdate) (); // Post-update function
+  using PostUpdate = void (); // Post-update function

  Tune() { read_results(); }
  Tune(const Tune&) = delete;
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -103,8 +103,8 @@ constexpr bool Is64Bit = true;
 constexpr bool Is64Bit = false;
 #endif

-typedef uint64_t Key;
-typedef uint64_t Bitboard;
+using Key = uint64_t;
+using Bitboard = uint64_t;

 constexpr int MAX_MOVES = 256;
 constexpr int MAX_PLY   = 246;
@@ -186,6 +186,9 @@ enum Value : int {
  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,
@@ -215,7 +218,7 @@ constexpr Value PieceValue[PHASE_NB][PIECE_NB] = {
    VALUE_ZERO, PawnValueEg, KnightValueEg, BishopValueEg, RookValueEg, QueenValueEg, VALUE_ZERO, VALUE_ZERO }
 };

-typedef int Depth;
+using Depth = int;

 enum : int {
  DEPTH_QS_CHECKS     =  0,
@@ -413,6 +416,10 @@ inline Color color_of(Piece pc) {
  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;
 }
@@ -442,15 +449,17 @@ constexpr Direction pawn_push(Color c) {
 }

 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;
+  return m & 0xFFF;
 }

 constexpr MoveType type_of(Move m) {
@@ -470,10 +479,6 @@ constexpr Move make(Square from, Square to, PieceType pt = KNIGHT) {
  return Move(T + ((pt - KNIGHT) << 12) + (from << 6) + to);
 }

-constexpr bool is_ok(Move m) {
-  return from_sq(m) != to_sq(m); // Catch MOVE_NULL and MOVE_NONE
-}
-
 /// Based on a congruential pseudo random number generator
 constexpr Key make_key(uint64_t seed) {
  return seed * 6364136223846793005ULL + 1442695040888963407ULL;
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -22,6 +22,7 @@
 #include <sstream>
 #include <string>

+#include "benchmark.h"
 #include "evaluate.h"
 #include "movegen.h"
 #include "position.h"
@@ -31,23 +32,22 @@
 #include "tt.h"
 #include "uci.h"
 #include "syzygy/tbprobe.h"
+#include "nnue/evaluate_nnue.h"

 using namespace std;

 namespace Stockfish {

-extern vector<string> setup_bench(const Position&, istream&);
-
 namespace {

-  // FEN string of the initial position, normal chess
+  // 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 engine receives the "position" UCI command.
-  // The function sets up the position described in the given FEN string ("fen")
-  // or the starting position ("startpos") and then makes the moves given in the
-  // following move list ("moves").
+  // 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) {

@@ -59,7 +59,7 @@ namespace {
    if (token == "startpos")
    {
        fen = StartFEN;
-        is >> token; // Consume "moves" token if any
+        is >> token; // Consume the "moves" token, if any
    }
    else if (token == "fen")
        while (is >> token && token != "moves")
@@ -67,10 +67,10 @@ namespace {
    else
        return;

-    states = StateListPtr(new std::deque<StateInfo>(1)); // Drop old and create a new one
+    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 move list (if any)
+    // Parse the move list, if any
    while (is >> token && (m = UCI::to_move(pos, token)) != MOVE_NONE)
    {
        states->emplace_back();
@@ -78,8 +78,8 @@ namespace {
    }
  }

-  // trace_eval() prints the evaluation for the current position, consistent with the UCI
-  // options set so far.
+  // trace_eval() prints the evaluation of the current position, consistent with
+  // the UCI options set so far.

  void trace_eval(Position& pos) {

@@ -93,20 +93,20 @@ namespace {
  }


-  // setoption() is called when engine receives the "setoption" UCI command. The
-  // function updates the UCI option ("name") to the given value ("value").
+  // 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 "name" token
+    is >> token; // Consume the "name" token

-    // Read option name (can contain spaces)
+    // Read the option name (can contain spaces)
    while (is >> token && token != "value")
        name += (name.empty() ? "" : " ") + token;

-    // Read option value (can contain spaces)
+    // Read the option value (can contain spaces)
    while (is >> token)
        value += (value.empty() ? "" : " ") + token;

@@ -117,9 +117,9 @@ namespace {
  }


-  // go() is called when engine receives the "go" UCI command. The function sets
-  // the thinking time and other parameters from the input string, then starts
-  // the search.
+  // 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) {

@@ -127,7 +127,7 @@ namespace {
    string token;
    bool ponderMode = false;

-    limits.startTime = now(); // As early as possible!
+    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
@@ -151,9 +151,9 @@ namespace {
  }


-  // bench() is called when engine receives the "bench" command. Firstly
-  // a list of UCI commands is setup according to bench parameters, then
-  // it is run one by one printing a summary at the end.
+  // 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) {

@@ -161,7 +161,7 @@ namespace {
    uint64_t num, nodes = 0, cnt = 1;

    vector<string> list = setup_bench(pos, args);
-    num = count_if(list.begin(), list.end(), [](string s) { return s.find("go ") == 0 || s.find("eval") == 0; });
+    num = count_if(list.begin(), list.end(), [](const string& s) { return s.find("go ") == 0 || s.find("eval") == 0; });

    TimePoint elapsed = now();

@@ -184,12 +184,12 @@ namespace {
        }
        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 some while
+        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(); // Just before exiting
+    dbg_print();

    cerr << "\n==========================="
         << "\nTotal time (ms) : " << elapsed
@@ -197,36 +197,40 @@ namespace {
         << "\nNodes/second    : " << 1000 * nodes / elapsed << endl;
  }

-  // The win rate model returns the probability (per mille) of winning given an eval
-  // and a game-ply. The model fits rather accurately the LTC fishtest statistics.
+  // 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 captures only up to 240 plies, so limit input (and rescale)
+     // The model only captures up to 240 plies, so limit the input and then rescale
     double m = std::min(240, ply) / 64.0;

-     // Coefficients of a 3rd order polynomial fit based on fishtest data
-     // for two parameters needed to transform eval to the argument of a
-     // logistic function.
-     double as[] = {-1.17202460e-01, 5.94729104e-01, 1.12065546e+01, 1.22606222e+02};
-     double bs[] = {-1.79066759,  11.30759193, -17.43677612,  36.47147479};
+     // 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 eval to centipawns with limited range
-     double x = std::clamp(double(100 * v) / PawnValueEg, -2000.0, 2000.0);
+     // Transform the eval to centipawns with limited range
+     double x = std::clamp(double(v), -4000.0, 4000.0);

-     // Return win rate in per mille (rounded to nearest)
+     // 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 stdin, parses it and calls the appropriate
-/// function. Also intercepts EOF from stdin to ensure gracefully exiting if the
-/// GUI dies unexpectedly. When called with some command line arguments, e.g. to
-/// run 'bench', once the command is executed the function returns immediately.
-/// In addition to the UCI ones, also some additional debug commands are supported.
+/// 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[]) {

@@ -240,24 +244,24 @@ void UCI::loop(int argc, char* argv[]) {
      cmd += std::string(argv[i]) + " ";

  do {
-      if (argc == 1 && !getline(cin, cmd)) // Block here waiting for input or EOF
+      if (argc == 1 && !getline(cin, 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 empty or blank line
+      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 us the user has played the expected move.
-      // So 'ponderhit' will be sent if we were told to ponder on the same move the
-      // user has played. We should continue searching but switch from pondering to
-      // normal search.
+      // 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 normal search
+          Threads.main()->ponder = false; // Switch to the normal search

      else if (token == "uci")
          sync_cout << "id name " << engine_info(true)
@@ -270,8 +274,8 @@ void UCI::loop(int argc, char* argv[]) {
      else if (token == "ucinewgame") Search::clear();
      else if (token == "isready")    sync_cout << "readyok" << sync_endl;

-      // Additional custom non-UCI commands, mainly for debugging.
-      // Do not use these commands during a search!
+      // 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;
@@ -285,19 +289,25 @@ void UCI::loop(int argc, char* argv[]) {
              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 << sync_endl;
+          sync_cout << "Unknown command: '" << cmd << "'. Type help for more information." << sync_endl;

-  } while (token != "quit" && argc == 1); // Command line args are one-shot
+  } while (token != "quit" && argc == 1); // The command-line arguments are one-shot
 }


-/// UCI::value() converts a Value to a string suitable for use with the UCI
-/// protocol specification:
+/// 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
-///           use negative values for y.
+/// mate <y>  Mate in 'y' moves (not plies). If the engine is getting mated,
+///           uses negative values for 'y'.

 string UCI::value(Value v) {

@@ -305,8 +315,13 @@ string UCI::value(Value v) {

  stringstream ss;

-  if (abs(v) < VALUE_MATE_IN_MAX_PLY)
-      ss << "cp " << v * 100 / PawnValueEg;
+  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;

@@ -314,8 +329,8 @@ string UCI::value(Value v) {
 }


-/// UCI::wdl() report WDL statistics given an evaluation and a game ply, based on
-/// data gathered for fishtest LTC games.
+/// 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) {

@@ -338,21 +353,21 @@ std::string UCI::square(Square s) {


 /// UCI::move() converts a Move to a string in coordinate notation (g1f3, a7a8q).
-/// The only special case is castling, where we print in the e1g1 notation in
-/// normal chess mode, and in e1h1 notation in chess960 mode. Internally all
-/// castling moves are always encoded as 'king captures rook'.
+/// 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) {

-  Square from = from_sq(m);
-  Square to = to_sq(m);
-
  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));

@@ -370,8 +385,8 @@ string UCI::move(Move m, bool chess960) {

 Move UCI::to_move(const Position& pos, string& str) {

-  if (str.length() == 5) // Junior could send promotion piece in uppercase
-      str[4] = char(tolower(str[4]));
+  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()))
@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -30,20 +30,27 @@ 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;

-/// Custom comparator because UCI options should be case insensitive
+/// Define a custom comparator, because the UCI options should be case-insensitive
 struct CaseInsensitiveLess {
  bool operator() (const std::string&, const std::string&) const;
 };

-/// Our options container is actually a std::map
-typedef std::map<std::string, Option, CaseInsensitiveLess> OptionsMap;
+/// The options container is defined as a std::map
+using OptionsMap = std::map<std::string, Option, CaseInsensitiveLess>;

-/// Option class implements an option as defined by UCI protocol
+/// The Option class implements each option as specified by the UCI protocol
 class Option {

-  typedef void (*OnChange)(const Option&);
+  using OnChange = void (*)(const Option&);

 public:
  Option(OnChange = nullptr);
@@ -54,7 +61,7 @@ public:

  Option& operator=(const std::string&);
  void operator<<(const Option&);
-  operator double() const;
+  operator int() const;
  operator std::string() const;
  bool operator==(const char*) const;

@@ -72,7 +79,7 @@ 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, Value alpha, Value beta);
+std::string pv(const Position& pos, Depth depth);
 std::string wdl(Value v, int ply);
 Move to_move(const Position& pos, std::string& str);

@@ -1,6 +1,6 @@
 /*
  Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  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
@@ -38,13 +38,13 @@ UCI::OptionsMap Options; // Global object
 namespace UCI {

 /// 'On change' actions, triggered by an option's value change
-void on_clear_hash(const Option&) { Search::clear(); }
-void on_hash_size(const Option& o) { TT.resize(size_t(o)); }
-void on_logger(const Option& o) { start_logger(o); }
-void on_threads(const Option& o) { Threads.set(size_t(o)); }
-void on_tb_path(const Option& o) { Tablebases::init(o); }
-void on_use_NNUE(const Option& ) { Eval::NNUE::init(); }
-void on_eval_file(const Option& ) { Eval::NNUE::init(); }
+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 {
@@ -61,7 +61,7 @@ void init(OptionsMap& o) {
  constexpr int MaxHashMB = Is64Bit ? 33554432 : 2048;

  o["Debug Log File"]        << Option("", on_logger);
-  o["Threads"]               << Option(1, 1, 512, on_threads);
+  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);
@@ -73,7 +73,7 @@ void init(OptionsMap& o) {
  o["UCI_Chess960"]          << Option(false);
  o["UCI_AnalyseMode"]       << Option(false);
  o["UCI_LimitStrength"]     << Option(false);
-  o["UCI_Elo"]               << Option(1350, 1350, 2850);
+  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);
@@ -128,9 +128,9 @@ Option::Option(double v, int minv, int maxv, OnChange f) : type("spin"), min(min
 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 double() const {
+Option::operator int() const {
  assert(type == "check" || type == "spin");
-  return (type == "spin" ? stof(currentValue) : currentValue == "true");
+  return (type == "spin" ? std::stoi(currentValue) : currentValue == "true");
 }

 Option::operator std::string() const {
@@ -70,7 +70,8 @@ for args in "eval" \
            "go depth 10" \
            "go movetime 1000" \
            "go wtime 8000 btime 8000 winc 500 binc 500" \
-            "bench 128 $threads 8 default depth"
+            "bench 128 $threads 8 default depth" \
+            "export_net verify.nnue"
 do

   echo "$prefix $exeprefix ./stockfish $args $postfix"
@@ -78,6 +79,11 @@ do

 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