COMPILER GENERATED SYNCHRONIZATION FOR DO LOOPS.

Samuel P. Midkiff, David A Padua

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Methods are presented for the compile-time generation of synchronization instructions for parallel loops running on a multiprocessor. A synchronization instruction set and architecture are defined, and using these it is shown how to synchronize Fortran source loops so that dependences on the parallel loop are enforced. Construction of an applicable dependence graph is covered. A program transformation is given that increases the parallelism that can be utilized in a parallel loop containing nested loops. Finally, a generalizable technique is presented for reducing the number of dependences needing synchronizing in a parallel loop.

Original languageEnglish (US)
Title of host publicationProceedings of the International Conference on Parallel Processing
EditorsKai Hwang, Steven M. Jacobs, Earl E. Swartzlander
PublisherIEEE
Pages544-551
Number of pages8
ISBN (Print)0818607246
StatePublished - 1986

Publication series

NameProceedings of the International Conference on Parallel Processing
ISSN (Print)0190-3918

ASJC Scopus subject areas

  • Hardware and Architecture

Fingerprint

Dive into the research topics of 'COMPILER GENERATED SYNCHRONIZATION FOR DO LOOPS.'. Together they form a unique fingerprint.

Cite this