Identifying similarities in TMBL programs with alignment to quicken their compilation for GPUs: computational intelligence on consumer games and graphics hardware

Authors:
Tony E. Lewis;George D. Magoulas
Affiliations:
University of London, London, United Kingdom;University of London, London, United Kingdom
Venue:
Proceedings of the 13th annual conference companion on Genetic and evolutionary computation
Year:
2011

Citing 4
Cited 0

Fast Genetic Programming and Artificial Developmental Systems on GPUs

HPCS '07 Proceedings of the 21st International Symposium on High Performance Computing Systems and Applications
A data parallel approach to genetic programming using programmable graphics hardware

Proceedings of the 9th annual conference on Genetic and evolutionary computation
A SIMD interpreter for genetic programming on GPU graphics cards

EuroGP'08 Proceedings of the 11th European conference on Genetic programming
Population parallel GP on the G80 GPU

EuroGP'08 Proceedings of the 11th European conference on Genetic programming

Quantified Score

Hi-index	0.00

Visualization

Abstract

The most impressive accelerations of Genetic Programming (GP) using the Graphics Processing Unit (GPU) have been achieved by dynamically compiling new GPU code for each batch of individuals to be evaluated. This approach suffers an overhead in compilation time. We aim to reduce this penalty by pre-processing the individuals to identify and draw out their similarities, hence reducing duplication in compilation work. We use this approach with Tweaking Mutation Behaviour Learning (TMBL), a form focused on long term fitness growth. For individuals of 300 instructions, the technique is found to reduce compilation time 4.817 times whilst only reducing evaluation speed by 3.656%.