xsimd/.github/workflows/windows.yml at master · xtensor-stack/xsimd · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
name: Windows build
on: [push, pull_request]
concurrency:
  group: ${{ github.workflow }}-${{ github.job }}-${{ github.ref }}
  cancel-in-progress: true

jobs:
  build-windows-x86:
    name: 'MSVC ${{ matrix.os }}, ${{ matrix.target }} ${{ matrix.sys.set }}'
    defaults:
      run:
        shell: bash {0}
    strategy:
      matrix:
        os:
          - 2022
          - 2025
        target:
          - x86
          - x64
        sys:
          - { set: SSE, flags: "/arch:SSE2" }
          - { set: AVX, flags: "/arch:AVX" }
          - { set: AVX2, flags: "/arch:AVX2" }
          - { set: AVX512, flags: "/arch:AVX512" }
        exclude:
          # AVX on both platforms has a codegen error
          # On 2019 in _mm256_rsqrt_ps, on 2022 in _mm256_blend_p*
          - { sys: { set: AVX } }
          # On both platforms x86 + AVX512 triggers a compiler crash
          - { target: x86, sys: { set: AVX512 } }
          # /arch:SSE2 is not available on x64 platforms (SSE2 is enabled by default)
          - { target: x64, sys: { set: SSE} }
    runs-on: windows-${{ matrix.os }}
    steps:
    - name: Setup compiler
      uses: ilammy/msvc-dev-cmd@v1
      with:
        arch: ${{ matrix.target }}
    - name: Setup Ninja
      run: |
        python3 -m pip install --upgrade pip setuptools wheel
        python3 -m pip install ninja
    - name: Checkout xsimd
      uses: actions/checkout@v6
    - name: Setup
      run: cmake -B _build -DBUILD_TESTS=ON -DDOWNLOAD_DOCTEST=ON -DBUILD_BENCHMARK=ON -DBUILD_EXAMPLES=ON -DCMAKE_BUILD_TYPE=Release -DCMAKE_CXX_FLAGS="${{ matrix.sys.flags }}" -G Ninja
    - name: Build
      run: cmake --build _build
    - name: Testing xsimd
      if: ${{ !startsWith(matrix.sys.set, 'AVX512') }}
      env:
        # Set CPU feature test expectations
        # Assuming the runner always has AVX2 (independent of compilation option)
        XSIMD_TEST_CPU_ASSUME_NEON64: "0"
        XSIMD_TEST_CPU_ASSUME_SSE4_2: "1"
        XSIMD_TEST_CPU_ASSUME_AVX2: "1"
        XSIMD_TEST_CPU_ASSUME_MANUFACTURER: "intel,amd"
      run: ./_build/test/test_xsimd

  build-windows-mingw:
    name: 'MSYS2 ${{ matrix.msystem }}'
    runs-on: windows-2022
    defaults:
      run:
        shell: msys2 {0}
    strategy:
      matrix:
        # Temporarily remove MINGW64 and UCRT64 builds because
        # GCC 12 gives an unexpected overflow warning for __builtin_memmove
        # see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106199
        msystem: [ MINGW32, CLANG64 ]
        #msystem: [ MINGW32, MINGW64, UCRT64, CLANG32, CLANG64 ]
      fail-fast: false
    steps:
      - name: Use MinGW from MSYS2
        uses: msys2/setup-msys2@v2
        with:
          msystem: ${{ matrix.msystem }}
          update: true
          path-type: minimal
          pacboy: >-
            cc:p
            cmake:p
            ninja:p
      - name: Checkout xsimd
        uses: actions/checkout@v6
      - name: Configure
        run: cmake -B _build -DBUILD_TESTS=ON -DBUILD_BENCHMARK=ON -DBUILD_EXAMPLES=ON -DCMAKE_BUILD_TYPE=Release -DDOWNLOAD_DOCTEST=ON -G Ninja
      - name: Build
        run: cmake --build _build
      - name: Test
        run: ./_build/test/test_xsimd

  build-windows-clang-cl:
    name: 'clang-cl x64 ${{ matrix.config.name }}'
    defaults:
      run:
        shell: bash {0}
    strategy:
      matrix:
        config:
          - { name: "AVX2", flags: "/arch:AVX2", benchmark: "ON", examples: "ON" }
          - { name: "/fp:fast", flags: "/fp:fast", benchmark: "OFF", examples: "OFF" }
    runs-on: windows-2025
    steps:
    - name: Setup compiler
      uses: ilammy/msvc-dev-cmd@v1
      with:
        arch: amd64
    - name: Check clang-cl
      run: |
        command -v clang-cl
        clang-cl --version
    - name: Setup Ninja
      run: |
        python3 -m pip install --upgrade pip setuptools wheel
        python3 -m pip install ninja
    - name: Checkout xsimd
      uses: actions/checkout@v3
    - name: Setup
      run: |
        cmake -B _build \
              -DBUILD_TESTS=ON \
              -DDOWNLOAD_DOCTEST=ON \
              -DBUILD_BENCHMARK=${{ matrix.config.benchmark }} \
              -DBUILD_EXAMPLES=${{ matrix.config.examples }} \
              -DCMAKE_BUILD_TYPE=Release \
              -DCMAKE_C_COMPILER=clang-cl \
              -DCMAKE_CXX_COMPILER=clang-cl \
              -DCMAKE_CXX_FLAGS="${{ matrix.config.flags }} -DXSIMD_REASSOCIATIVE_MATH=1" \
              -G Ninja
    - name: Build
      run: cmake --build _build
    - name: Testing xsimd
      run: ./_build/test/test_xsimd

  build-windows-arm64:
    name: 'MSVC arm64'
    defaults:
      run:
        shell: bash {0}
    runs-on: windows-11-arm
    steps:
    - name: Setup compiler
      uses: ilammy/msvc-dev-cmd@v1
      with:
        arch: amd64
    - name: Setup Ninja
      run: |
        python3 -m pip install --upgrade pip setuptools wheel
        python3 -m pip install ninja
    - name: Checkout xsimd
      uses: actions/checkout@v6
    - name: Setup
      run: cmake -B _build -DBUILD_TESTS=ON -DDOWNLOAD_DOCTEST=ON -DBUILD_BENCHMARK=ON -DBUILD_EXAMPLES=ON -DCMAKE_BUILD_TYPE=Release -G Ninja
    - name: Build
      run: cmake --build _build
    - name: Testing xsimd
      run: ./_build/test/test_xsimd