Compiler and Runtime Support for Deep Memories
- funded by: ANL
- funding level: $28.346
- duration: 1/22/2021 - 7/21/2021
This project develops and experimentally assesses compiler
transformations and a runtime system to support dynamic memory
allocations for deep hierarchies with volatile and non-volatile
byte-addressable memories.
Publications:
-
PEARS: A Performance-Aware Static and Dynamic Framework for Heterogeneous Memory
by Onkar Patil, Latchesar Ionkov, Jason Lee, Frank Mueller, Michael Lang,
in International Symposium on Memory Systems (MEMSYS), Oct 2021, pages.
-
NVM-based energy and cost efficient HPC clusters
by Onkar Patil, Latchesar Ionkov, Jason Lee, Frank Mueller, Michael Lang,
in International Symposium on Memory Systems (MEMSYS), Oct 2021, pages.
Theses: