Optimize merge algorithm for data sizes equal or greater then 4M items with SLM cache usage
#31
This file contains 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: oneDPL CI Testing | |
| on: | |
| push: | |
| branches: [main] | |
| pull_request: | |
| branches: | |
| - release_oneDPL | |
| - main | |
| - 'release/**' | |
| paths: | |
| - '.github/**/*' | |
| - 'cmake/**/*' | |
| - 'include/**/*' | |
| - 'test/**/*' | |
| - 'CMakeLists.txt' | |
| - '.clang-format' | |
| permissions: read-all | |
| env: | |
| BUILD_CONCURRENCY: 4 | |
| MACOS_BUILD_CONCURRENCY: 3 | |
| TEST_TIMEOUT: 360 | |
| WINDOWS_TBB_DOWNLOAD_LINK: https://registrationcenter-download.intel.com/akdlm/IRC_NAS/8222de4d-2c5e-41bc-bdd5-f557a9edda28/w_tbb_oneapi_p_2021.13.0.627_offline.exe | |
| WINDOWS_ICPX_DOWNLOAD_LINK: https://registrationcenter-download.intel.com/akdlm/IRC_NAS/7991e201-ca0f-4689-bdb6-1ed73a8246fd/w_dpcpp-cpp-compiler_p_2024.2.0.491_offline.exe | |
| WINDOWS_ONEAPI_PATH: C:\Program Files (x86)\Intel\oneAPI | |
| LINUX_ONEAPI_PATH: /opt/intel/oneapi | |
| concurrency: | |
| group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }} | |
| cancel-in-progress: ${{ !contains(github.ref, 'refs/heads/main') }} | |
| jobs: | |
| clang-format: | |
| if: github.event_name == 'pull_request' | |
| runs-on: ubuntu-latest | |
| steps: | |
| - uses: actions/checkout@v4 | |
| with: | |
| fetch-depth: 0 | |
| ref: ${{ github.event.pull_request.head.sha }} | |
| - uses: actions/setup-python@v5 | |
| with: | |
| python-version: '3.x' | |
| - name: Get clang-format | |
| run: sudo apt-get install -yqq clang-format | |
| - name: Applying clang-format for changed files | |
| run: | | |
| MERGE_BASE=$(git merge-base ${{ github.event.pull_request.head.sha }} ${{ github.event.pull_request.base.sha }}) | |
| FILES=$(git diff --diff-filter=d --name-only $MERGE_BASE | grep ^include | grep -v nanorange\.hpp\$ || true) | |
| CLANG_FORMAT_DIFF_PATH=$(which clang-format-diff) | |
| echo $FILES | xargs -n1 -t -r git diff -U0 --no-color --relative $MERGE_BASE | python3 $CLANG_FORMAT_DIFF_PATH -i -p1 -style file | |
| if: github.event_name == 'pull_request' || github.event_name == 'push' | |
| - name: Creating diff | |
| run: git diff > clang-format.diff | |
| - name: Checking if diff is empty | |
| run: if [ -s clang-format.diff ]; then cat clang-format.diff; exit 1; fi | |
| - if: failure() | |
| name: Save artifacts | |
| uses: actions/upload-artifact@v4 | |
| with: | |
| name: clang-format-diff | |
| path: clang-format.diff | |
| retention-days: 3 | |
| linux-testing: | |
| name: ${{ matrix.device_type }},bknd=${{ matrix.backend }},cmplr=${{ matrix.cxx_compiler }},${{ matrix.os }},std=с++${{ matrix.std }},cfg=${{ matrix.build_type }} | |
| runs-on: ['${{ matrix.os }}'] | |
| timeout-minutes: 60 | |
| strategy: | |
| fail-fast: false | |
| matrix: | |
| include: | |
| - os: ubuntu-latest | |
| cxx_compiler: icpx | |
| std: 17 | |
| build_type: release | |
| backend: dpcpp | |
| device_type: CPU | |
| - os: ubuntu-latest | |
| cxx_compiler: icpx | |
| std: 17 | |
| build_type: release | |
| backend: dpcpp | |
| device_type: FPGA_EMU | |
| - os: ubuntu-latest | |
| cxx_compiler: icpx | |
| std: 17 | |
| build_type: release | |
| backend: tbb | |
| device_type: HOST | |
| - os: ubuntu-latest | |
| cxx_compiler: icpx | |
| std: 17 | |
| build_type: release | |
| backend: omp | |
| device_type: HOST | |
| - os: ubuntu-latest | |
| cxx_compiler: g++ | |
| std: 17 | |
| build_type: release | |
| backend: omp | |
| device_type: HOST | |
| - os: ubuntu-latest | |
| cxx_compiler: g++ | |
| std: 17 | |
| build_type: release | |
| backend: tbb | |
| device_type: HOST | |
| - os: ubuntu-latest | |
| cxx_compiler: g++ | |
| std: 17 | |
| build_type: release | |
| backend: tbb | |
| device_type: HOST | |
| - os: ubuntu-latest | |
| cxx_compiler: clang++ | |
| std: 17 | |
| build_type: release | |
| backend: tbb | |
| device_type: HOST | |
| - os: ubuntu-latest | |
| cxx_compiler: clang++ | |
| std: 17 | |
| build_type: release | |
| backend: omp | |
| device_type: HOST | |
| - os: ubuntu-latest | |
| cxx_compiler: clang++ | |
| std: 17 | |
| build_type: release | |
| backend: tbb | |
| device_type: HOST | |
| - os: ubuntu-latest | |
| cxx_compiler: icpx | |
| std: 17 | |
| build_type: release | |
| backend: tbb | |
| device_type: HOST | |
| - os: ubuntu-latest | |
| cxx_compiler: g++ | |
| std: 17 | |
| build_type: release | |
| backend: serial | |
| device_type: HOST | |
| - os: ubuntu-latest | |
| cxx_compiler: icpx | |
| std: 20 | |
| build_type: release | |
| backend: dpcpp | |
| device_type: CPU | |
| - os: ubuntu-latest | |
| cxx_compiler: icpx | |
| std: 20 | |
| build_type: release | |
| backend: tbb | |
| device_type: HOST | |
| - os: ubuntu-latest | |
| cxx_compiler: icpx | |
| std: 20 | |
| build_type: release | |
| backend: omp | |
| device_type: HOST | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - name: Set up Intel APT repository | |
| if: (matrix.backend == 'tbb' || matrix.backend == 'dpcpp' || matrix.cxx_compiler == 'icpx' || matrix.cxx_compiler == 'icx' || matrix.cxx_compiler == 'icx-cl' || matrix.cxx_compiler == 'dpcpp' || matrix.cxx_compiler == 'dpcpp-cl') | |
| run: | | |
| # https://www.intel.com/content/www/us/en/docs/oneapi/installation-guide-linux/2024-0/apt.html | |
| wget -O- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB \ | |
| | gpg --dearmor | sudo tee /usr/share/keyrings/oneapi-archive-keyring.gpg > /dev/null | |
| echo "deb [signed-by=/usr/share/keyrings/oneapi-archive-keyring.gpg] https://apt.repos.intel.com/oneapi all main" | sudo tee /etc/apt/sources.list.d/oneAPI.list | |
| sudo apt-get update -y | |
| - name: Install Intel® oneAPI Threading Building Blocks | |
| if: (matrix.backend == 'tbb' || matrix.backend == 'dpcpp') | |
| run: | | |
| sudo apt-get install intel-oneapi-tbb-devel -y | |
| - name: Install Intel® oneAPI DPC++/C++ Compiler | |
| if: (matrix.cxx_compiler == 'icpx' || matrix.cxx_compiler == 'icx' || matrix.cxx_compiler == 'icx-cl' || matrix.cxx_compiler == 'dpcpp' || matrix.cxx_compiler == 'dpcpp-cl') | |
| run: | | |
| sudo apt-get install intel-oneapi-compiler-dpcpp-cpp -y | |
| # Avoid accidental use of a released version, keeping libpstloffload.so | |
| sudo rm -rf ${LINUX_ONEAPI_PATH}/dpl/latest/include | |
| - name: Install Intel® oneAPI DPC++/C++ Compiler SYCL* FPGA Emulator Runtime | |
| if: (matrix.device_type == 'FPGA_EMU') | |
| run: | | |
| sudo apt-get install intel-oneapi-compiler-fpga -y | |
| - name: Install OpenMP dependencies | |
| if: (matrix.os == 'ubuntu-latest' && matrix.cxx_compiler == 'clang++' && matrix.backend == 'omp') | |
| run: | | |
| sudo apt-get install libomp5 libomp-dev -y | |
| - name: Run testing | |
| shell: bash | |
| run: | | |
| set -x | |
| if [[ -f "${LINUX_ONEAPI_PATH}/setvars.sh" ]]; then | |
| source ${LINUX_ONEAPI_PATH}/setvars.sh | |
| fi | |
| echo "::warning::CMake: $(cmake --version)" | |
| echo "::warning::Compiler: $(${{ matrix.cxx_compiler }} --version)" | |
| if [[ "${{ matrix.backend }}" == "dpcpp" ]]; then | |
| make_targets="build-onedpl-sycl_iterator-tests build-onedpl-ranges-tests" | |
| ctest_flags="-R (sycl_iterator_.*)|(std_ranges_.*)\.pass" | |
| echo "::warning::dpcpp backend is set. Compile and run only sycl_iterator tests" | |
| else | |
| make_targets="build-onedpl-tests" | |
| fi | |
| mkdir build && cd build | |
| lscpu | |
| cmake -DCMAKE_CXX_STANDARD=${{ matrix.std }} -DCMAKE_BUILD_TYPE=${{ matrix.build_type }} \ | |
| -DCMAKE_CXX_COMPILER=${{ matrix.cxx_compiler }} -DONEDPL_BACKEND=${{ matrix.backend }} -DONEDPL_DEVICE_TYPE=${{ matrix.device_type }} .. | |
| make VERBOSE=1 -j${BUILD_CONCURRENCY} ${make_targets} | |
| ctest --timeout ${TEST_TIMEOUT} --output-on-failure ${ctest_flags} | |
| windows-testing: | |
| name: ${{ matrix.device_type }},bknd=${{ matrix.backend }},cmplr=${{ matrix.cxx_compiler }},${{ matrix.os }},std=c++${{ matrix.std }},cfg=${{ matrix.build_type }} | |
| runs-on: ['${{ matrix.os }}'] | |
| timeout-minutes: 60 | |
| strategy: | |
| fail-fast: false | |
| matrix: | |
| include: | |
| - os: windows-latest | |
| cxx_compiler: icx-cl | |
| std: 17 | |
| build_type: release | |
| backend: tbb | |
| device_type: HOST | |
| - os: windows-latest | |
| cxx_compiler: cl | |
| std: 17 | |
| build_type: release | |
| backend: tbb | |
| device_type: HOST | |
| - os: windows-latest | |
| cxx_compiler: icx | |
| std: 17 | |
| build_type: release | |
| backend: dpcpp | |
| device_type: CPU | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - name: Install Intel® oneAPI Threading Building Blocks | |
| if: (matrix.backend == 'tbb' || matrix.backend == 'dpcpp') | |
| shell: cmd | |
| run: | | |
| curl %WINDOWS_TBB_DOWNLOAD_LINK% --output tbb_install.exe | |
| tbb_install.exe -s -a --silent --eula accept -p=NEED_VS2019_INTEGRATION=0 -p=NEED_VS2022_INTEGRATION=0 | |
| del tbb_install.exe | |
| - name: Install Intel® oneAPI DPC++/C++ Compiler | |
| shell: cmd | |
| run: | | |
| curl %WINDOWS_ICPX_DOWNLOAD_LINK% --output icpx_install.exe | |
| icpx_install.exe -s -a --silent --eula accept -p=NEED_VS2019_INTEGRATION=0 -p=NEED_VS2022_INTEGRATION=0 | |
| del icpx_install.exe | |
| :: Avoid accidental use of a released version | |
| rd /s /q "%WINDOWS_ONEAPI_PATH%\dpl" | |
| - name: Run testing | |
| shell: cmd | |
| run: | | |
| if exist "%WINDOWS_ONEAPI_PATH%\setvars.bat" ( | |
| call "%WINDOWS_ONEAPI_PATH%\setvars.bat" | |
| ) | |
| if "${{ matrix.cxx_compiler }}" == "cl" ( | |
| call "C:\Program Files (x86)\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 | |
| ) | |
| powershell $output = cmake --version; Write-Host ::warning::CMake: $output | |
| powershell $output = ${{ matrix.cxx_compiler }} --version; Write-Host ::warning::Compiler: $output | |
| if "${{ matrix.backend }}" == "dpcpp" ( | |
| set ninja_targets="build-onedpl-sycl_iterator-tests" | |
| set ctest_flags=-R sycl_iterator_.*\.pass | |
| echo ::warning::dpcpp backend is set. Compile and run only sycl_iterator tests | |
| ) else ( | |
| set ninja_targets=build-onedpl-tests | |
| ) | |
| mkdir build && cd build | |
| cmake -G "Ninja" -DCMAKE_BUILD_TYPE=${{ matrix.build_type }} -DCMAKE_CXX_STANDARD=${{ matrix.std }} -DCMAKE_CXX_COMPILER=${{ matrix.cxx_compiler }} -DONEDPL_BACKEND=${{ matrix.backend }} -DONEDPL_DEVICE_TYPE=${{ matrix.device_type }} .. || goto :short_circuit_fail | |
| ninja -j 2 -v %ninja_targets% || goto :short_circuit_fail | |
| ctest --timeout %TEST_TIMEOUT% -C ${{ matrix.build_type }} --output-on-failure %ctest_flags% || goto :short_circuit_fail | |
| exit /b 0 | |
| :: modify the default behaviour of shell:cmd, which exits with the status of a last command, in order not to unintentially miss an error | |
| :short_circuit_fail | |
| exit /b %errorlevel% | |
| macos-testing: | |
| name: HOST,bknd=${{ matrix.backend }},cmplr=${{ matrix.cxx_compiler }},${{ matrix.os }},std=c++${{ matrix.std }},cfg=${{ matrix.build_type }} | |
| runs-on: ['${{ matrix.os }}'] | |
| timeout-minutes: 60 | |
| strategy: | |
| fail-fast: false | |
| matrix: | |
| include: | |
| - os: macos-latest | |
| cxx_compiler: clang++ | |
| std: 17 | |
| build_type: release | |
| backend: omp | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - name: Install OpenMP for Clang++ | |
| run: | | |
| brew install libomp | |
| - name: Run testing | |
| shell: bash | |
| run: | | |
| set -x | |
| echo "::warning::CMake: $(cmake --version)" | |
| sysctl -a | grep machdep.cpu | |
| # workaround for CMake not being able to find OpenMP: see https://discourse.cmake.org/t/how-to-find-openmp-with-clang-on-macos/8860 | |
| # -DCMAKE_POLICY_DEFAULT_CMP0074=NEW below is forced to make sure CMake uses <PackageName>_ROOT variables. | |
| export OpenMP_ROOT=$(brew --prefix)/opt/libomp | |
| mkdir build && cd build | |
| cmake -DCMAKE_CXX_STANDARD=${{ matrix.std }} -DCMAKE_BUILD_TYPE=${{ matrix.build_type }} \ | |
| -DCMAKE_CXX_COMPILER=${{ matrix.cxx_compiler }} -DONEDPL_BACKEND=${{ matrix.backend }} -DCMAKE_POLICY_DEFAULT_CMP0074=NEW .. | |
| make VERBOSE=1 build-onedpl-tests -j${MACOS_BUILD_CONCURRENCY} | |
| ctest --timeout ${TEST_TIMEOUT} --output-on-failure -E "${EXCLUDE_FROM_TESTING}" |