Amortised deep parameter optimisation of GPGPU work group size for OpenCV

Jeongju Sohn, Seongmin Lee, Shin Yoo

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

6 Scopus citations

Abstract

GPGPU (General Purpose computing on Graphics Processing Units) enables massive parallelism by taking advantage of the Single Instruction Multiple Data (SIMD) architecture of the large number of cores found on modern graphics cards. A parameter called local work group size controls how many work items are concurrently executed on a single compute unit. Though critical to the performance, there is no deterministic way to tune it, leaving developers to manual trial and error. This paper applies amortised optimisation to determine the best local work group size for GPGPU implementations of OpenCV template matching feature. The empirical evaluation shows that optimised local work group size can outperform the default value with large effect sizes.

Original languageEnglish
Title of host publicationSearch Based Software Engineering - 8th International Symposium, SSBSE 2016, Proceedings
EditorsFederica Sarro, Kalyanmoy Deb
PublisherSpringer Verlag
Pages211-217
Number of pages7
ISBN (Print)9783319471051
DOIs
StatePublished - 2016
Event8th International Symposium on Search Based Software Engineering, SSBSE 2016 - Raleigh, United States
Duration: 8 Oct 201610 Oct 2016

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9962 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference8th International Symposium on Search Based Software Engineering, SSBSE 2016
Country/TerritoryUnited States
CityRaleigh
Period8/10/1610/10/16

Fingerprint

Dive into the research topics of 'Amortised deep parameter optimisation of GPGPU work group size for OpenCV'. Together they form a unique fingerprint.

Cite this