Can wear-aware memory allocation be intelligent?

C. Hakert, K.-H. Chen, J.-J. Chen

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

1 Citation (Scopus)


Many non-volatile memories (NVM) suffer from a severe reducedcell endurance and therefore require wear-leveling. Heap memory,as one segment, which potentially is mapped to a NVM, faces astrong application dependent characteristic regarding the amountof memory accesses and allocations. A simple deterministic strategyfor wear leveling of the heap may suffer when the available actionspace becomes too large. Therefore, we investigate the employmentof a reinforcement learning agent as a substitute for such a strategyin this paper. The agent's objective is to learn a strategy, which isoptimal with respect to the total memory wear out. We concludethis work with an evaluation, where we compare the deterministicstrategy with the proposed agent. We report that our proposedagent outperforms the simple deterministic strategy in several cases.However, we also report further optimization potential in the agentdesign and deployment.
Original languageEnglish
Title of host publicationMLCAD 2020 - Proceedings of the 2020 ACM/IEEE Workshop on Machine Learning for CAD
Publication statusPublished - 2020
Externally publishedYes
Event2nd ACM/IEEE Workshop on Machine Learning for CAD, MLCAD 2020 - Virtual Event
Duration: 16 Nov 202020 Nov 2020
Conference number: 2


Conference2nd ACM/IEEE Workshop on Machine Learning for CAD, MLCAD 2020
Abbreviated titleMLCAD 2020
CityVirtual Event


Dive into the research topics of 'Can wear-aware memory allocation be intelligent?'. Together they form a unique fingerprint.

Cite this