Research on improving the page table organization to improve address translation
Title: Research on improving the page table organization to improve address translation
SNIC Project: SNIC 2020/13-37
Project Type: SNIC Small Compute
Principal Investigator: Chang Hyun Park <chang.hyun.park@it.uu.se>
Affiliation: Uppsala universitet
Duration: 2020-05-06 – 2021-06-01
Classification: 10201
Keywords:

Abstract

Computer systems use the virtual memory system to manage memory and allow the operating system to manage memory for the user process. This virtual memory has a tax of requiring an address translation for every Load or Store instruction (which is about 20% of all program instructions). Lately, this address translation has become a bottleneck for performance. In this proposal, we would like to explore the benefits of re-organizing the page table (using different data structures, etc.).