Operating System I/O Speculation:
How two invocations are faster than one

Keir Fraser, University of Cambridge Computer Laboratory¹
Fay Chang, Google Inc.*

Keir.Fraser@cl.cam.ac.uk Fay@google.com

Abstract
Introduction
Speculative execution
Experimental setup
- Benchmark applications
In-kernel speculative execution
- Basic design
- Performance of the basic design
Improved prefetching for swapping applications
Controlling memory overhead
- Reactive cost-benefit analysis
- Evaluation of memory control
Comparison to previous design
Related and future work
Conclusions
Bibliography

Keir Fraser 2003-04-04

This paper was originally published in the Proceedings of the USENIX Annual Technical Conference (General Track), June 9 – 14, 2003, San Antonio, TX, USA
Last changed: 3 Jun 2003 aw

Technical Program

USENIX 2003 Annual Technical Conference Home

USENIX home

Operating System I/O Speculation: How two invocations are faster than one

Operating System I/O Speculation:
How two invocations are faster than one