Rigorous Benchmarking in Reasonable Time

Kalibera, Tomas and Jones, Richard E. (2013) Rigorous Benchmarking in Reasonable Time. In: ACM SIGPLAN International Symposium on Memory Management (ISMM 2013), 20-12 June, 2013, Seattle, Washington, USA. (doi:https://doi.org/10.1145/2464157.2464160) (Full text available)

PDF (Corrections to variance estimators in section 9.2) - Updated Version
Download (248kB) Preview
[img]
Preview

Abstract

Experimental evaluation is key to systems research. Because modern systems are complex and non-deterministic, good experimental methodology demands that researchers account for uncertainty. To obtain valid results, they are expected to run many iterations of benchmarks, invoke virtual machines (VMs) several times, or even rebuild VM or benchmark binaries more than once. All this repetition costs time to complete experiments. Currently, many evaluations give up on sufficient repetition or rigorous statistical methods, or even run benchmarks only in training sizes. The results reported often lack proper variation estimates and, when a small difference between two systems is reported, some are simply unreliable. In contrast, we provide a statistically rigorous methodology for repetition and summarising results that makes efficient use of experimentation time. Time efficiency comes from two key observations. First, a given benchmark on a given platform is typically prone to much less non-determinism than the common worst-case of published corner-case studies. Second, repetition is most needed where most uncertainty arises (whether between builds, between executions or between iterations). We capture experimentation cost with a novel mathematical model, which we use to identify the number of repetitions at each level of an experiment necessary and sufficient to obtain a given level of precision. We present our methodology as a cookbook that guides researchers on the number of repetitions they should run to obtain reliable results. We also show how to present results with an effect size confidence interval. As an example, we show how to use our methodology to conduct throughput experiments with the DaCapo and SPEC CPU benchmarks on three recent platforms. NOTE: this version corrects the ISMM 2013 version

Item Type: Conference or workshop item (Paper)
Projects: [UNSPECIFIED] Garbage Collection for Multicore Systems
Uncontrolled keywords: Benchmarking methodology; statistical methods; DaCapo; SPEC CPU
Subjects: Q Science > QA Mathematics (inc Computing science) > QA 76 Software, computer programming, > QA76.76 Computer software
Divisions: Faculties > Sciences > School of Computing > Programming Languages and Systems Group
Depositing User: Richard Jones
Date Deposited: 15 Apr 2013 15:48 UTC
Last Modified: 17 Jan 2017 10:20 UTC
Resource URI: https://kar.kent.ac.uk/id/eprint/33611 (The current URI for this page, for reference purposes)
  • Depositors only (login required):

Downloads

Downloads per month over past year