Efficient sparse-dense matrix-matrix multiplication on GPUs using the customized sparse storage format
git clone https://github.com/hclhkbu/gcoospdm.git
cd gcoospdm
git submodule update --init --recursive
cd src
make
After sucessfully compiled, one can try the following example:
./main -sparsity=0.900000 -nRowsA=800 -nRowsB=800 -nColsA=800 -nColsB=800 -algo=gcoospdm