Publication:
A Hybrid Evolutionary Algorithm for Solving the Register Allocation Problem

dc.contributor.authorBOZ, BETÜL
dc.contributor.authorTOPCUOĞLU, HALUK RAHMİ
dc.contributor.authorsDemiroz B., Topcuoglu H., Kandemir M.
dc.date.accessioned2022-03-15T01:54:51Z
dc.date.accessioned2026-01-11T08:13:52Z
dc.date.available2022-03-15T01:54:51Z
dc.date.issued2004
dc.description.abstractOne 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. © Springer-Verlag 2004.
dc.identifier.doi10.1007/978-3-540-24652-7_7
dc.identifier.isbn3540213678; 9783540213673
dc.identifier.issn3029743
dc.identifier.urihttps://hdl.handle.net/11424/246624
dc.language.isoeng
dc.publisherSpringer Verlag
dc.relation.ispartofLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
dc.rightsinfo:eu-repo/semantics/closedAccess
dc.titleA Hybrid Evolutionary Algorithm for Solving the Register Allocation Problem
dc.typearticle
dspace.entity.typePublication
oaire.citation.endPage71
oaire.citation.startPage62
oaire.citation.titleLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
oaire.citation.volume3004

Files