Abstract: "Eliminating partially dead code has proved to be a powerful technique for the runtime optimization of sequential programs. In this article, we show how this technique can be adapted to explicitly parallel programs with shared memory and interleaving semantics on the basis of a recent...

Ausführliche Beschreibung

Bibliographische Detailangaben

1. Verfasser:	Knoop, Jens (VerfasserIn)
Format:	Buch
Sprache:	English
Veröffentlicht:	Passau 1996
Schriftenreihe:	Universität <Passau> / Fakultät für Mathematik und Informatik: MIP 1996,18
Schlagworte:	Code generators Coding theory Parallel processing (Electronic computers) Parallel programming (Computer science) Theoretische Informatik Informatik Mathematik
Zusammenfassung:	Abstract: "Eliminating partially dead code has proved to be a powerful technique for the runtime optimization of sequential programs. In this article, we show how this technique can be adapted to explicitly parallel programs with shared memory and interleaving semantics on the basis of a recently presented framework for efficient and precise bitvector analyses for this program setting. Whereas the framework underlying our approach allows a straightforward adaptation of the required data flow analyses to the parallel case, the transformation part of the optimization requires special care in order to preserve parallelism. This preservation is an absolute must in order to guarantee that the optimization does never impair efficiency. The introduction of an appropiate [sic] natural side condition suffices to lift even the optimality result known from the sequential setting to the parallel setting."
Beschreibung:	26, 5 S.