Skip to content

Commit

Permalink
Improve CUDA speed using O3 compile argument
Browse files Browse the repository at this point in the history
Co-authored-by: James Zhao <[email protected]>
  • Loading branch information
0cc4m and Noir-Lime committed May 18, 2023
1 parent 8eb27ed commit 0236f5a
Showing 1 changed file with 3 additions and 0 deletions.
3 changes: 3 additions & 0 deletions setup.py
Original file line number Diff line number Diff line change
Expand Up @@ -13,12 +13,15 @@
ext_modules=[
cpp_extension.CUDAExtension(
"quant_cuda_v1", ["quant_cuda_v1/quant_cuda.cpp", "quant_cuda_v1/quant_cuda_kernel.cu"],
extra_compile_args={'nvcc': ['-O3']},
),
cpp_extension.CUDAExtension(
"quant_cuda_v2", ["quant_cuda_v2/quant_cuda.cpp", "quant_cuda_v2/quant_cuda_kernel.cu"],
extra_compile_args={'nvcc': ['-O3']},
),
cpp_extension.CUDAExtension(
"quant_cuda_v3", ["quant_cuda_v3/quant_cuda.cpp", "quant_cuda_v3/quant_cuda_kernel.cu"],
extra_compile_args={'nvcc': ['-O3']},
),
],
cmdclass={"build_ext": cpp_extension.BuildExtension},
Expand Down

0 comments on commit 0236f5a

Please sign in to comment.