Parallelism extraction in acyclic code

Author

Luís, João Paulo ; Carvalho, Carlos Gongalves ; Delgado, José Carlos

Author_Institution

Parallel Archit. Group, INESC, Lisbon, Portugal

fYear

1996

Firstpage

437

Lastpage

447

Abstract

We describe a framework for parallelism extraction and partitioning in acyclic code regions. This framework is an extension of M. Girkar´s (1981) work on functional parallelism, using a Petri net model to represent parallel code, and applying modified optimization techniques to minimize the overheads of explicit synchronization. The modifications introduced are directed towards the generation of efficient multi threaded code. We also describe a simple partitioning technique that can be used to artificially increase the granularity of the extracted parallelism up to a desired level

Keywords

Petri nets; optimising compilers; parallel programming; parallelising compilers; Petri net model; acyclic code regions; explicit synchronization; functional parallelism; modified optimization techniques; multi threaded code; parallel code; parallelism extraction; parallelizing compilers; simple partitioning technique; Data mining; Flow graphs; Hardware; Operating systems; Parallel architectures; Parallel programming; Partitioning algorithms; Tree graphs;

fLanguage

English

Publisher

ieee

Conference_Titel

Parallel and Distributed Processing, 1996. PDP '96. Proceedings of the Fourth Euromicro Workshop on

Conference_Location

Braga

Print_ISBN

0-8186-7376-1

Type

conf

DOI

10.1109/EMPDP.1996.500617

Filename

500617