This is a repository for an artifact evaluation of "Unified Convolution Framework: A compiler-based approach to support sparse convolutions".
Build TACO-UCF using CMake 3.4.0 or greater:
cd <UCF-directory>
mkdir build
cd build
cmake -DCMAKE_BUILD_TYPE=Release ..
make -j8
export LD_LIBRARY_PATH=$(pwd)/lib/:$LD_LIBRARY_PATH
To build TACO-UCF for NVIDIA CUDA, add -DCUDA=ON to the cmake line above. For example:
cmake -DCMAKE_BUILD_TYPE=Release -DCUDA=ON ..
Please also make sure that you have CUDA installed properly and that the following environment variables are set correctly:
export PATH=/usr/local/cuda/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH
export LIBRARY_PATH=/usr/local/cuda/lib64:$LIBRARY_PATH
The generated CUDA code will require compute capability 6.1 or higher to run.
To run filter-sparse convolution:
cd <taco-directory>/benchmark/filter_sparse_convolution
To run masked(submanifold) sparse convolution:
cd <taco-directory>/benchmark/submanifold_sparse_convolution