DPNP - NumPy*的数据并行扩展

API覆盖摘要

完整文档

DPNP C++后端文档

从源代码构建

确保已安装以下依赖项

• cython
• cmake >=3.21
• dpcpp_linux-64 或 dpcpp_win-64 (取决于您的操作系统)
• dpctl
• mkl-devel-dpcpp
• onedpl-devel
• ninja
• numpy >=1.19,<1.25a0
• python
• scikit-build
• setuptools
• sysroot_linux-64 >=2.28 (仅在Linux操作系统上)
• tbb-devel

完成这些步骤后，可以使用以下方式在调试模式下构建dpnp

git clone https://github.com/IntelPython/dpnp
cd dpnp
python scripts/build_locally.py

通过pip安装Wheel包

安装DPNP

python -m pip install --index-url https://pypi.anaconda.org/intel/simple dpnp

如果使用venv或系统Python，请设置性能库路径

export LD_LIBRARY_PATH=<path_to_your_env>/lib

还需要设置以下环境变量

export OCL_ICD_FILENAMES_RESET=1
export OCL_ICD_FILENAMES=libintelocl.so

运行测试

pytest
# or
pytest tests/test_matmul.py -s -v
# or
python -m unittest tests/test_mixins.py

运行NumPy外部测试

. ./0.env.sh
python -m tests.third_party.numpy_ext
# or
python -m tests.third_party.numpy_ext core/tests/test_umath.py
# or
python -m tests.third_party.numpy_ext core/tests/test_umath.py::TestHypot::test_simple

构建文档

Prerequisites:
$ conda install sphinx sphinx_rtd_theme
Building:
1. Install dpnp into your python environment
2. $ cd doc && make html
3. The documentation will be in doc/_build/html

打包

. ./0.env.sh
conda-build conda-recipe/

运行基准测试

cd benchmarks/

asv run --python=python --bench <filename without .py>
# example:
asv run --python=python --bench bench_elementwise

# or

asv run --python=python --bench <class>.<bench>
# example:
asv run --python=python --bench Elementwise.time_square

# add --quick option to run every case once but looks like first execution has additional overheads and takes a lot of time (need to be investigated)

测试矩阵