A Combined Data and Program Partitioning Algorithm for Distributed Memory Multiprocessors

IPSJ Web Site 12 References

Search this article

Description

In this paper we propose an algorithm to perform data partitioning and program partitioning simultaneously on an intermediate representation for parallelizing compilers which we have proposed the Data Partitioning Graph. Conventional and therefore conservative parallelizing compilers usually activate program partitioning prior to data partitioning. However on distributed memory multiprocessors it is quite difficult to partition a program effectively with consideration of data partitioning since communication costs change depending on a data partitioning and distribution decision. The proposed algorithm resolves this conflict by handling these inseparable partitioning problems simultaneously with and A* algorithm.

In this paper we propose an algorithm to perform data partitioning and program partitioning simultaneously on an intermediate representation for parallelizing compilers which we have proposed, the Data Partitioning Graph. Conventional and, therefore, conservative parallelizing compilers usually activate program partitioning prior to data partitioning. However, on distributed memory multiprocessors it is quite difficult to partition a program effectively with consideration of data partitioning since communication costs change depending on a data partitioning and distribution decision. The proposed algorithm resolves this conflict by handling these inseparable partitioning problems simultaneously with and A* algorithm.

Journal

References(12)*help

See more

Keywords

Details 詳細情報について

Report a problem

Back to top