Collection of performance benchmarks used to present optimizations implemented for Intel(R) Distribution for Python*.

To install Python environments from Intel channel along with pip-installed packages

• conda env create -f environments/intel.yaml
• conda activate intel_env

• python numpy/umath/umath_mem_bench.py -v --size 10 --goal-time 0.01 --repeats 1

• To run python benchmarks: python numpy/umath/umath_mem_bench.py
• To compile and run native benchmarks (requires icx): make -C numpy/umath

• To run python benchmarks: python numpy/random/rng.py
• To compile and run native benchmarks (requires icx): make -C numpy/random

"Accelerating Scientific Python with Intel Optimizations" by Oleksandr Pavlyk, Denis Nagorny, Andres Guzman-Ballen, Anton Malakhov, Hai Liu, Ehsan Totoni, Todd A. Anderson, Sergey Maidanov. Proceedings of the 16th Python in Science Conference (SciPy 2017), July 10 - July 16, Austin, Texas