Added ability to accumulate in FP16. Convert BF16 to FP32. For FP16 and BF16 GEMM in RISC-V (BF16 now works for pre-RVA23) #3850
Workflow file for this run
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| name: mips64 qemu test | |
| on: [push, pull_request] | |
| concurrency: | |
| group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }} | |
| cancel-in-progress: true | |
| permissions: | |
| contents: read # to fetch code (actions/checkout) | |
| jobs: | |
| TEST: | |
| if: "github.repository == 'OpenMathLib/OpenBLAS'" | |
| runs-on: ubuntu-latest | |
| strategy: | |
| fail-fast: false | |
| matrix: | |
| include: | |
| - target: MIPS64_GENERIC | |
| triple: mips64el-linux-gnuabi64 | |
| opts: NO_SHARED=1 TARGET=MIPS64_GENERIC | |
| - target: SICORTEX | |
| triple: mips64el-linux-gnuabi64 | |
| opts: NO_SHARED=1 TARGET=SICORTEX | |
| - target: I6400 | |
| triple: mipsisa64r6el-linux-gnuabi64 | |
| opts: NO_SHARED=1 TARGET=I6400 | |
| - target: P6600 | |
| triple: mipsisa64r6el-linux-gnuabi64 | |
| opts: NO_SHARED=1 TARGET=P6600 | |
| - target: I6500 | |
| triple: mipsisa64r6el-linux-gnuabi64 | |
| opts: NO_SHARED=1 TARGET=I6500 | |
| steps: | |
| - name: Checkout repository | |
| uses: actions/checkout@v3 | |
| - name: install build deps | |
| run: | | |
| sudo apt-get update | |
| sudo apt-get install autoconf automake autotools-dev ninja-build make ccache \ | |
| gcc-${{ matrix.triple }} gfortran-${{ matrix.triple }} libgomp1-mips64el-cross libglib2.0-dev | |
| - name: checkout qemu | |
| uses: actions/checkout@v3 | |
| with: | |
| repository: qemu/qemu | |
| path: qemu | |
| ref: ae35f033b874c627d81d51070187fbf55f0bf1a7 | |
| - name: build qemu | |
| run: | | |
| cd qemu | |
| ./configure --prefix=$GITHUB_WORKSPACE/qemu-install --target-list=mips64el-linux-user --disable-system | |
| make -j$(nproc) | |
| make install | |
| - name: Compilation cache | |
| uses: actions/cache@v3 | |
| with: | |
| path: ~/.ccache | |
| key: ccache-${{ runner.os }}-${{ matrix.target }}-${{ github.ref }}-${{ github.sha }} | |
| restore-keys: | | |
| ccache-${{ runner.os }}-${{ matrix.target }}-${{ github.ref }} | |
| ccache-${{ runner.os }}-${{ matrix.target }} | |
| - name: Configure ccache | |
| run: | | |
| test -d ~/.ccache || mkdir -p ~/.ccache | |
| echo "max_size = 300M" > ~/.ccache/ccache.conf | |
| echo "compression = true" >> ~/.ccache/ccache.conf | |
| ccache -s | |
| - name: build OpenBLAS | |
| run: make CC='ccache ${{ matrix.triple }}-gcc -static' FC='ccache ${{ matrix.triple }}-gfortran -static' ${{ matrix.opts }} HOSTCC='ccache gcc' -j$(nproc) | |
| - name: test | |
| run: | | |
| export PATH=$GITHUB_WORKSPACE/qemu-install/bin/:$PATH | |
| qemu-mips64el ./utest/openblas_utest | |
| qemu-mips64el ./utest/openblas_utest_ext | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./ctest/xscblat1 | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./ctest/xdcblat1 | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./ctest/xccblat1 | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./ctest/xzcblat1 | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./ctest/xscblat2 < ./ctest/sin2 | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./ctest/xdcblat2 < ./ctest/din2 | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./ctest/xccblat2 < ./ctest/cin2 | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./ctest/xzcblat2 < ./ctest/zin2 | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./ctest/xscblat3 < ./ctest/sin3 | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./ctest/xdcblat3 < ./ctest/din3 | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./ctest/xccblat3 < ./ctest/cin3 | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./ctest/xzcblat3 < ./ctest/zin3 | |
| OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-mips64el ./test/sblat1 | |
| OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-mips64el ./test/dblat1 | |
| OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-mips64el ./test/cblat1 | |
| OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-mips64el ./test/zblat1 | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./test/sblat1 | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./test/dblat1 | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./test/cblat1 | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./test/zblat1 | |
| rm -f ./test/?BLAT2.SUMM | |
| OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-mips64el ./test/sblat2 < ./test/sblat2.dat | |
| OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-mips64el ./test/dblat2 < ./test/dblat2.dat | |
| OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-mips64el ./test/cblat2 < ./test/cblat2.dat | |
| OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-mips64el ./test/zblat2 < ./test/zblat2.dat | |
| rm -f ./test/?BLAT2.SUMM | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./test/sblat2 < ./test/sblat2.dat | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./test/dblat2 < ./test/dblat2.dat | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./test/cblat2 < ./test/cblat2.dat | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./test/zblat2 < ./test/zblat2.dat | |
| rm -f ./test/?BLAT3.SUMM | |
| OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-mips64el ./test/sblat3 < ./test/sblat3.dat | |
| OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-mips64el ./test/dblat3 < ./test/dblat3.dat | |
| OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-mips64el ./test/cblat3 < ./test/cblat3.dat | |
| OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-mips64el ./test/zblat3 < ./test/zblat3.dat | |
| rm -f ./test/?BLAT3.SUMM | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./test/sblat3 < ./test/sblat3.dat | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./test/dblat3 < ./test/dblat3.dat | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./test/cblat3 < ./test/cblat3.dat | |
| OPENBLAS_NUM_THREADS=2 qemu-mips64el ./test/zblat3 < ./test/zblat3.dat |