Name		Name	Last commit message	Last commit date
mobilenet		mobilenet
rn90_3x3		rn90_3x3
scripts		scripts
sparsednn		sparsednn
.gitignore		.gitignore
BC.npy		BC.npy
LICENSE		LICENSE
README.md		README.md
autotune_conv_float.sh		autotune_conv_float.sh
autotune_float.sh		autotune_float.sh
exe		exe
install.sh		install.sh
notchpeak.env.sh		notchpeak.env.sh
ptx_result.npy		ptx_result.npy
ref.npy		ref.npy
requirements.txt		requirements.txt
runtime		runtime
temp_stub732.cu		temp_stub732.cu
temp_stub732.ptx		temp_stub732.ptx
temp_stub997.cu		temp_stub997.cu
temp_stub997.ptx		temp_stub997.ptx
testing.cubin		testing.cubin
testing.ptx		testing.ptx

Repository files navigation

gpu-sparsert

More details will be forthcoming. This will basically work on an Amazon T4 instance with the deep learning AMI.

Dependencies:

Cnpy. Please install it and update the include path in the scripts.
Cuda toolchain, etc.
Pytorch

To try SpMM, type autotune_float.sh 0 To try sparse convolution, type bash autotune_conv_float.sh 512 512 7 filter_bg4.npy

About

No description, website, or topics provided.

18 stars

2 watching

Report repository

Releases

Packages

Contributors

Languages