• Home
  • Search
  • A unified framework for optimizing communication in data-parallel programs
  • Cite Icon76
  • https://doi.org/10.1109/71.508249Copy DOI Icon

A unified framework for optimizing communication in data-parallel programs

  • Abstract
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

This paper presents a framework, based on global array data-flow analysis, to reduce communication costs in a program being compiled for a distributed memory machine. We introduce available section descriptor, a novel representation of communication involving array sections. This representation allows us to apply techniques for partial redundancy elimination to obtain powerful communication optimizations. With a single framework, we are able to capture optimizations like (1) vectorizing communication, (2) eliminating communication that is redundant on any control flow path, (3) reducing the amount of data being communicated, (4) reducing the number of processors to which data must be communicated, and (5) moving communication earlier to hide latency, and to subsume previous communication. We show that the bidirectional problem of eliminating partial redundancies can be decomposed into simpler unidirectional problems even in the context of an array section representation, which makes the analysis procedure more efficient. We present results from a preliminary implementation of this framework, which are extremely encouraging, and demonstrate the effectiveness of this analysis in improving the performance of programs.

Similar Papers
  • Book Chapter
  • Citations29

A unified data-flow framework for optimizing communication

  • Jan 01, 1995
  • Manish Gupta +2
  • Research Article
  • Citations2

Predicated Partial Redundancy Elimination using a Cost Analysis

  • Dec 01, 2003
  • Parallel Processing Letters
  • Bernhard Scholz +2
  • Book Chapter

$$\pi $$ Abstraction: Parallelism-Aware Array Data Flow Analysis for OpenMP

  • Jan 01, 2015
  • Fahed Jubair +3
  • Research Article
  • Citations5

Interprocedural partial redundancy elimination with application to distributed memory compilation

  • Jul 01, 1998
  • IEEE Transactions on Parallel and Distributed Systems
  • G Agrawal
  • Research Article
  • Citations9

Interprocedural partial redundancy elimination and its application to distributed memory compilation

  • Jun 01, 1995
  • ACM SIGPLAN Notices
  • Gagan Agrawal +2
  • Conference Article
  • Citations49

Interprocedural partial redundancy elimination and its application to distributed memory compilation

  • Jun 01, 1995
  • Gagan Agrawal +2
  • Research Article
  • Citations4

Complete removal of redundant expressions

  • Apr 01, 2004
  • ACM SIGPLAN Notices
  • Rastislav Bodík +2
  • Research Article
  • Citations22

Complete removal of redundant expressions

  • May 01, 1998
  • ACM SIGPLAN Notices
  • Rastislav Bodík +2
  • Conference Article
  • Citations136

Complete removal of redundant expressions

  • May 01, 1998
  • Rastislav Bodík +2
  • Conference Article
  • Citations37

Optimal and efficient speculation-based partial redundancy elimination

  • Mar 23, 2003
  • Qiong Cai +1
  • Conference Article
  • Citations21

Optimal and efficient speculation-based partial redundancy elimination

  • Dec 22, 2003
  • Qiong Cai +1
  • Book Chapter
  • Citations6

Partial Value Number Redundancy Elimination

  • Jan 01, 2005
  • Rei Odaira +1
  • Research Article
  • Citations4

COMPILE-TIME ANALYSIS AND OPTIMIZATION OF EXPLICITLY PARALLEL PROGRAMS*

  • Jan 01, 1997
  • Parallel Algorithms and Applications
  • Jeanne Ferrante +2
  • Conference Article
  • Citations69

Global communication analysis and optimization

  • May 01, 1996
  • Soumen Chakrabarti +2
  • Research Article
  • Citations6

Global communication analysis and optimization

  • May 01, 1996
  • ACM SIGPLAN Notices
  • Soumen Chakrabarti +2
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.