University of Illinois Urbana-Champaign

Page migration and placement in hybrid memory systems using machine learning algorithms

Vemulapati, Vibhakar

This item is only available for download by members of the University of Illinois community. Students, faculty, and staff at the U of I may log in with your NetID and password to view the item. If you are trying to access an Illinois-restricted dissertation or thesis, you can request a copy through your library's Inter-Library Loan office or purchase a copy directly from ProQuest.

Permalink

https://hdl.handle.net/2142/97887

Description

Title
Page migration and placement in hybrid memory systems using machine learning algorithms
Author(s)
Vemulapati, Vibhakar
Contributor(s)
Kim, Nam Sung
Issue Date
2017-12
Keyword(s)
  • hybrid memory system
  • phase change memory
  • prefetching
  • machine learning
Abstract
As we reach the end of DRAM technology scaling, the prevalence of new memory technology in computers is inevitable. Phase-change memory (PCM) is an emerging non-volatile memory technology which can be denser than existing DRAM cells, but is slower. A possible solution is a hybrid PCM/DRAM memory system where we have a large capacity PCM and a DRAM used as a buffer between the processor and PCM. We will investigate various page migration algorithms on the hybrid system to maximize utilization of faster DRAM to mitigate the performance slowdown associated with using PCM as main memory. We primarily investigate the effects of speculative pre-fetching of pages in memory from PCM to DRAM using machine learning algorithms. The page rank program that we tested had irregular memory access patterns that made it difficult to predict the pages that had to be pre-fetched, causing performance slowdown when compared to running the program without modification. The overhead of predicting the prefetch page far outweighed the performance increase achieved.
Type of Resource
text
Language
en
Permalink
http://hdl.handle.net/2142/97887

Owning Collections

Senior Theses - Electrical and Computer Engineering PRIMARY
The best of ECE undergraduate research