This is a fork from the official SST Github page with implementations maintained by SPEAR group at IIT.
Visit sst-simulator.org to learn more about SST.
The current master branch has the implementation of Q-adatpive routing in Merlin Dragonfly topology.
Note: if you use this SST implementation, please cite the following papers:
Yao Kang, Xin Wang, and Zhiling Lan. "Q-adaptive: A Multi-Agent Reinforcement Learning Based Routing on Dragonfly Network". In Proceedings of the 30th International Symposium on High-Performance Parallel and Distributed Computing (HPDC ’21).
Clone sst-core and sst-elements from this repository. Use other SST versions may cause compatible issue.
Follow installation guide to install SST.
mpirun -np 4 sst sst-elements/tests/dragonfly_q-adaptive.py