• Home
  • Search
  • Challenges in Code Optimization of Parallel Programs
  • Open Access IconOpen Access
  • Cite Icon7
  • https://doi.org/10.1007/978-3-642-00722-4_1Copy DOI Icon

Challenges in Code Optimization of Parallel Programs

  • Jan 1, 2009
  • Vivek Sarkar
Show More
  • Abstract
  • Literature Map
  • Citations
  • Similar Papers
Abstract

Code optimization has a rich history that dates back over half a century, and includes deep innovations that arose in response to changing trends in hardware and programming languages. These innovations have contributed significantly to programmer productivity by reducing the effort that programmers spend on hand-implementing code optimizations and by enabling code to be more portable. Often these innovations were accompanied by paradigm shifts in the foundations of compilers led by the introduction of new ideas such as interprocedural whole program analysis, coloring-based register allocation, static single assignment form, array dependence analysis, pointer alias analysis, loop transformations, adaptive profile-directed optimizations, and dynamic compilation. In this talk, we claim that the current multicore trend in the computer industry is forcing a new paradigm shift in compilers to address the challenge of code optimization of parallel programs , regardless of whether the parallelism is implicit or explicit in the programming model. All computers -- embedded, mainstream, and high-end -- are now being built from multicore processors with little or no increase in clock speed per core. This trend poses multiple challenges for compilers for future systems as the number of cores per socket continues to grow, and the cores become more heterogeneous. In addition, compilers have to keep pace with a proliferation of new parallel languages and libraries. To substantiate our claim, we first highlight some of the anomalies that arise when classical techniques from sequential code optimization are applied to parallel code. We then examine the historical foundations of code optimization including intermediate representations (IR's), abstract execution models, legality and cost analyses of IR transformations and identify paradigm shifts that will be necessary to support optimization of parallel code. We pay special attention to memory consistency models and their impact on code optimization. Finally, we summarize the approach to code optimization of parallel programs being taken in the Habanero Multicore Software Research project at Rice University.

Similar Papers
  • Conference Article
  • Citations3

Code optimization of parallel programs

  • Apr 06, 2008
  • Vivek Sarkar
  • Conference Article
  • Citations9

Towards Constructing the SSA form using Reaching Definitions Over Dominance Frontiers

  • Sep 01, 2019
  • Abu Naser Masud +1
  • Research Article
  • Citations10

More precise construction of static single assignment programs using reaching definitions

  • Apr 06, 2020
  • Journal of Systems and Software
  • Abu Naser Masud +1
  • Conference Article
  • Citations8

A type system equivalent to static single assignment

  • Jul 10, 2006
  • Yutaka Matsuno +1
  • Research Article
  • Citations8

Using static single assignment form in a code optimizer

  • Jun 01, 1992
  • ACM Letters on Programming Languages and Systems
  • Carl Mcconnell +1
  • Conference Article
  • Citations1

Recovering the Memory Behavior of Executable Programs

  • Jul 15, 2010
  • Alain Ketterlin +1
  • Research Article

Recovering memory access patterns of executable programs

  • Sep 06, 2012
  • Science of Computer Programming
  • Alain Ketterlin +1
  • Book Chapter
  • Citations4

Using the SSA-Form in a Code Generator

  • Jan 01, 2014
  • Benoît Dupont De Dinechin
  • Research Article
  • Citations132

Practical improvements to the construction and destruction of static single assignment form

  • Jul 10, 1998
  • Software: Practice and Experience
  • Preston Briggs +3
  • Book Chapter
  • Citations18

Incremental computation of static single assignment form

  • Jan 01, 1996
  • Jong -Deok Choi +2
  • Conference Article

Solving large complex problems using a higher order basis: Parallel out-of-core integral equation solvers involving a million unknowns

  • Mar 01, 2014
  • T K Sarkar +5
  • Research Article
  • Citations44

Solving large complex problems using a higher-order basis: parallel in-core and out-of-core integral-equation solvers

  • Aug 01, 2008
  • IEEE Antennas and Propagation Magazine
  • Yu Zhang +4
  • Research Article
  • Citations3

Structural Encoding of Static Single Assignment Form

  • Nov 25, 2005
  • Electronic Notes in Theoretical Computer Science
  • Andreas Gal +2
  • Research Article
  • Citations10

Acceleration of the Geostatistical Software Library (GSLIB) by code optimization and hybrid parallel programming

  • Sep 26, 2015
  • Computers & Geosciences
  • Oscar Peredo +2
  • Book Chapter

An Equivalence Checking Framework for Array-Intensive Programs

  • Jan 01, 2017
  • Kunal Banerjee +2
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.