A Hybrid Evolutionary Algorithm for Solving the Register Allocation Problem

Betul Demiroz, Haluk Topcuoglu, Mahmut Kandemir

Research output: Chapter in Book/Report/Conference proceedingChapter

2 Scopus citations

Abstract

One of the strong impacts on runtime performance of a given code is the register allocation phase of compilation. It is crucial to provide aggressive and sophisticated register allocators for the embedded devices, where the excessive compilation time is tolerated because of its off-line nature. In this paper, we present a hybrid evolutionary algorithm for register allocation problem that combines genetic algorithms with a local search technique. Our algorithm exploits a novel, highly-specialized crossover operator that considers domain-specific information. Computational experiments performed on various synthetic benchmarks prove that our method significantly outperform the state-of-the-art algorithms with respect to all given comparison metrics on solution quality.

Original languageEnglish (US)
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
EditorsJens Gottlieb, Gunther R. Raidl
PublisherSpringer Verlag
Pages62-71
Number of pages10
ISBN (Print)3540213678, 9783540213673
DOIs
StatePublished - Jan 1 2004

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume3004
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint

Dive into the research topics of 'A Hybrid Evolutionary Algorithm for Solving the Register Allocation Problem'. Together they form a unique fingerprint.

Cite this