Four assignments of High Performance Computing Labs course, Tsinghua University, Fall 2020, including:
- DGEMM (single thread + OpenMP): traditional + Strassen algorithm
- Stencil computing (MPI): traditional + time-skewed tiling
- SpMV (CUDA): naive + wrap based + sort based + merge based
- SpTRSV (CUDA): elementwise + Capellini + adaptive
All programs except DGEMM require input / answer data files to run, which are not provided due to huge size.
The code is not thoroughly tested and not licensed. Use at you own risk.