CUDA seeded graph matching (SGM) implementation.
Currently, only tested on undirected, unweighted graphs.
git clone --recursive https://github.com/owensgroup/csgm
See run.sh
for usage.
- [=] Profiling
- [=] Test on connectome graphs
- Performance testing on kasios graphs (esp. compared to fused implementation)
- [=] Re-implement auction algorithm in CUB