Coeus: Page clustering mechanism for machine learning-based hybrid memory management

Run Coeus

python run_across_apps.py

Run k-means

python run_cluster.py

Run Coeus + Kleio

python run_cluster_lstm.py

This repo contains the following code:

• sim/ includes the hybrid memory simulation created in Cori and validated against an Intel Optane PMEM platform.
• traces/ includes memory access traces used in the experimental evaluation of Cori and Coeus.
• kleio/ includes code that replicates the system design of Kleio as reported in the paper. It is not the original code developped while at AMD.

Paper references

• Coeus: Clustering (A)like Patterns for PracticalMachine Intelligent Hybrid Memory Management.
  Thaleia Dimitra Doudali, Ada Gavrilovska.
  In Proceedings of the 22nd IEEE/ACM International Symposium on Cluster, Cloud and Internet Computing (CCGrid 2022).

• Cori: Dancing to the Right Beat of Periodic Data Movements over Hybrid Memory Systems.
  Thaleia Dimitra Doudali, Daniel Zahka, Ada Gavrilovska.
  In Proceedings of the 35th IEEE International Parallel and Distributed Processing Symposium (IPDPS 2021).

• Kleio: a Hybrid Memory Page Scheduler with Machine Intelligence.
  Thaleia Dimitra Doudali, Sergey Blagodurov, Abhinav Vishnu, Sudhanva Gurumurthi, Ada Gavrilovska.
  In Proceedings of the 28th International Symposium on High-Performance Parallel and Distributed Computing (HPDC 2019).