Maximizing multiprocessor performance with the SUIF compiler

Author

Hall, M.W. ; Anderson, J.M. ; Murphy, B.R. ; Shih-Wei Liao ; Bugnion, E. ; Lam, M.S.

Author_Institution

USC Inf. Scis. Inst., Marina del Rey, CA, USA

Volume

29

Issue

12

fYear

1996

Firstpage

84

Lastpage

89

Abstract

This article describes automatic parallelization techniques in the SUIF (Stanford University Intermediate Format) compiler that result in good multiprocessor performance for array-based numerical programs. Parallelizing compilers for multiprocessors face many hurdles. However, SUIF´s robust analysis and memory optimization techniques enabled speedups on three fourths of the NAS and SPECfp95 benchmark programs.

Keywords

mathematics computing; multiprocessing systems; optimising compilers; parallelising compilers; performance evaluation; NAS; SPECfp95; SUIF compiler; Stanford University Intermediate Format compiler; array-based numerical programs; automatic parallelization techniques; benchmark programs; memory optimization techniques; multiprocessor performance; parallelizing compilers; robust analysis techniques; speedups; Cache memory; Concurrent computing; Information analysis; Multiprocessing systems; Parallel processing; Phased arrays; Privatization; Program processors; Programming profession; Throughput;

fLanguage

English

Journal_Title

Computer

Publisher

ieee

ISSN

0018-9162

Type

jour

DOI

10.1109/2.546613

Filename

546613