Implementation of "Sparse GPU Kernels for Deep Learning" paper.
This code is based on Google Sputnik.
In the kernels
folder, run the following commands to build and run the benchmark.
mkdir build && cd build
cmake .. -DCMAKE_BUILD_TYPE=Release
make -j12
The benchmark files will be created in build/sparsekernel
folder.
You can use this Colab link to Build and run the benchmarks.