DMCA
Static Analysis to Reduce Synchronization Costs in Data-Parallel Programs (1996)
| Venue: | IN PROC. ACM CONFERENCE ON PRINCIPLES OF PROGRAMMING LANGUAGES |
| Citations: | 16 - 1 self |
BibTeX
@INPROCEEDINGS{Gupta96staticanalysis,
author = {Manish Gupta and Edith Schonberg},
title = {Static Analysis to Reduce Synchronization Costs in Data-Parallel Programs},
booktitle = {IN PROC. ACM CONFERENCE ON PRINCIPLES OF PROGRAMMING LANGUAGES},
year = {1996},
pages = {322--332},
publisher = {ACM Press}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
For a program with sufficient parallelism, reducing synchronization costs is one of the most important objectives for achieving efficient execution on any parallel machine. This paper presents a novel methodology for reducing synchronization costs of programs compiled for SPMD execution. This methodology combines data flow analysis with communication analysis to determine the ordering between production and consumption of data on different processors, which helps in identifying redundant synchronization. The resulting framework is more powerful than any that have been previously presented, as it provides the first algorithm that can eliminate synchronization messages even from computations that need communication. We show that several commonly occurring computation patterns such as reductions and stencil computations with reciprocal producer-consumer relationship between processors lend themselves well to this optimization, an observation that is confirmed by an examination of some HPF...
Keyphrases
data-parallel program static analysis reduce synchronization synchronization cost synchronization message data flow analysis novel methodology sufficient parallelism important objective parallel machine computation pattern efficient execution first algorithm communication analysis different processor reciprocal producer-consumer relationship spmd execution stencil computation redundant synchronization
