TODO:
- MTTKRP CUDA optimization
- Tensor and matrices locate on GPU memory
- Sparse tensor distribution
- Tensor and matrices CANNOT locate on GPU memory
- Finish Element-wise OMP and GPU
- Documentation: Add main introduction
- Matlab: Need to confirm first, then test it.
Future TODO:
- Sequential, OMP, CUDA
- TTV
- single-node tensor contraction
- Tensor Contraction
Qs:
- Haven't check "sptSparseTensorKroneckerMul" and "sptSparseTensorKhatriRaoMul" functions. (TODO)