DMCA
Collective optimization (2009)
| Venue: | In Proceedings of the International Conference on High Performance Embedded Architectures & Compilers (HiPEAC |
| Citations: | 25 - 12 self |
BibTeX
@INPROCEEDINGS{Fursin09collectiveoptimization,
author = {Grigori Fursin and Olivier Temam and Hipeac Members},
title = {Collective optimization},
booktitle = {In Proceedings of the International Conference on High Performance Embedded Architectures & Compilers (HiPEAC},
year = {2009}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Abstract. Iterative compilation is an efficient approach to optimize programs on rapidly evolving hardware, but it is still only scarcely used in practice due to a necessity to gather a large number of runs often with the same data set and on the same environment in order to test many different optimizations and to select the most appropriate ones. Naturally, in many cases, users cannot afford a training phase, will run each data set once, develop new programs which are not yet known, and may regularly change the environment the programs are run on. In this article, we propose to overcome that practical obstacle using Collective Optimization, where the task of optimizing a program leverages the experience of many other users, rather than being performed in isolation, and often redundantly, by each user. Collective optimization is an unobtrusive approach, where performance information obtained after each run is sent back to a central database, which is then queried for optimizations suggestions, and the program is then recompiled accordingly. We show that it is possible to learn across data sets, programs and architectures in non-dynamic environments using static function cloning and run-time adaptation without even a reference run to compute speedups over the baseline optimization. We also show that it is possible to simultaneously learn and improve performance, since there are no longer two separate training and test phases, as in most studies. We demonstrate that extensively relying on competition among pairs of optimizations (program reaction to optimizations) provides a robust and efficient method for capturing the impact of optimizations, and for reusing this knowledge across data sets, programs and environments. We implemented our approach in GCC and will publicly disseminate it in the near future. 1
Keyphrases
collective optimization data set many case efficient method static function cloning reference run practical obstacle program reaction appropriate one many user new program non-dynamic environment performance information run-time adaptation efficient approach iterative compilation training phase many different optimization unobtrusive approach separate training optimization suggestion central database near future baseline optimization large number test phase
