thek0tyara/llama-cpp-turboquant
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 4
llama-cpp-turboquant/examples/python_naive_matmul
History
Corentin 7f4ec27235 Fix second implementation, add benchmark script 
* Third implementation is broken (WIP)
2021-06-25 02:49:28 +09:00
..
benchmark.py Fix second implementation, add benchmark script 2021-06-25 02:49:28 +09:00
first_example.py Naive matmul benchmark 2021-06-22 18:21:32 +09:00
imp1_naive.py Fix second implementation, add benchmark script 2021-06-25 02:49:28 +09:00
imp2_tiled.py Fix second implementation, add benchmark script 2021-06-25 02:49:28 +09:00
imp2_tiled_debug.py Fix second implementation, add benchmark script 2021-06-25 02:49:28 +09:00
imp3_better_tiling.py Fix second implementation, add benchmark script 2021-06-25 02:49:28 +09:00
matmul_plot.py Fix second implementation, add benchmark script 2021-06-25 02:49:28 +09:00
README.md Automated work group parameters and tiling matmul 2021-06-24 18:36:10 +09:00

README.md

Naive Matmul Implementation

This demonstrate a basic matmul implementation using Python and vulkan-kompute. Many thanks for the very helpful SGEMM in WebGL2-compute article on the public library ibiblio.org.

To test the implementation simply run the matmul.py script :

python matmul.py