https://github.com/tensor-compiler/array-programming-benchmarks
https://github.com/tensor-compiler/array-programming-benchmarks
Last synced: 7 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/tensor-compiler/array-programming-benchmarks
- Owner: tensor-compiler
- License: mit
- Created: 2021-02-12T20:48:34.000Z (about 5 years ago)
- Default Branch: main
- Last Pushed: 2021-06-22T00:18:47.000Z (over 4 years ago)
- Last Synced: 2025-05-30T04:58:14.241Z (9 months ago)
- Language: Shell
- Size: 7.13 MB
- Stars: 3
- Watchers: 3
- Forks: 0
- Open Issues: 7
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Setup
Create a virtualenv (`python3 -m venv venv`) and activate it. Install requirements with `pip install -r requirements.txt`. Add benchmarks as python functions prefixed with `test_` and have the function take in the `benchmark` fixture. An example is in `numpy/windowing.py`.
To run the benchmarks, use `make python-bench` to run all `python` benchmarks. Use the `BENCHES` flag to choose a particular set of benchmarks to run.
To write C++ benchmarks, we use the `google/benchmark` repository. Look at the example of `taco/windowing.cpp` for an example benchmark. Run `make taco-bench` to compile the `taco` benchmarks.