A software approach to accelerating memory translation for virtualized clouds

Zhang, Jiyuan

Permalink

https://hdl.handle.net/2142/124263 Copy

Description

Title

A software approach to accelerating memory translation for virtualized clouds

Author(s)

Zhang, Jiyuan

Issue Date

2024-04-25

Director of Research (if dissertation) or Advisor (if thesis)

Xu, Tianyin

Department of Study

Computer Science

Discipline

Computer Science

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

M.S.

Degree Level

Thesis

Keyword(s)

• Cloud Computing
• Virtualization
• Virtual Memory
• Address Translation

Abstract

Expensive page table walks triggered by frequent translation lookaside buffer (TLB) misses have incurred major performance bottlenecks for data-intensive workloads that are dominated by memory accesses with weak locality. Since it is hard to reduce TLB misses for such workloads, reducing page table walk overhead (i.e., the overhead of each TLB miss) is an increasingly important direction for improving application performance. The direction of reducing page table walk overhead is more compelling for workloads running in virtual machines. In virtualized environments, each TLB miss triggers a two-dimensional page table walk, which has a significantly higher overhead than that on native systems. However, a major caveat of research in this area is that most designs require changes in computer hardware. Yet, for practical applications, this requirement is often untenable. To this end, research on methods to reduce page walk overhead without hardware changes becomes a valuable topic. Taking this path, our study proves that even for a hardware-defined page walk flow, it is still possible to improve address translation performance by purely software means. This thesis presents HugeGPT, a software approach to reducing two-dimensional page table walk overhead in virtualized environments. HugeGPT ensures that page tables used in guest systems are physically held in the huge pages formed in the host system. This brings two-fold benefits: 1) the number of steps walking down the host page table is reduced; 2) the misses of page walk caches incurred by accessing the leaf nodes on host page tables can be eliminated. Extensive evaluation based on the prototype implementation and diverse real-world applications shows that HugeGPT can efficiently reduce address translation overhead and improve application performance in virtualized clouds, resulting in up to 50% application performance improvement compared to vanilla Linux/KVM.

Graduation Semester

2024-05

Type of Resource

Thesis

Handle URL

https://hdl.handle.net/2142/124263

Copyright and License Information

In reference to IEEE copyrighted material which is used with permission in this thesis, the IEEE does not endorse any of University of Illinois Urbana-Champaign’s products or services. Internal or personal use of this material is permitted. If interested in reprinting/republishing IEEE copyrighted material for advertising or promotional purposes or for creating new collective works for resale or redistribution, please go to http://www.ieee.org/publications_standards/publications/rights/rights_link.html to learn how to obtain a License from RightsLink. If applicable, University Microfilms and/or ProQuest Library, or the Archives of Canada may supply single copies of the dissertation.

Owning Collections