CPAR-language extensions to C for irregular and adaptive parallel computations

Author

Dong-Yuan Chen ; Chen, Marina C.

Author_Institution

Dept. of Comput. Sci., Yale Univ., New Haven, CT, USA

fYear

1994

fDate

23-25 May 1994

Firstpage

501

Lastpage

508

Abstract

The CPAR language consists of a set of extensions to the C language, designed to support parallel programming on distributed-memory multiprocessors, in particular, irregular and adaptive computations which are hard to do well in languages such as Fortran-90 or High Performance Fortran (HPF). CPAR provides data accesses over a global name space using concepts of communication functions and object streams. These constructs allow programmers to specify interprocessor data transfer between arrays or pointer-based data structures without getting involved with the split phases of “send” and “receive” of a communication. Our experience of using CPAR to program an N-body simulation problem is briefly described

Keywords

C language; FORTRAN; N-body problems; data structures; distributed memory systems; parallel languages; parallel programming; C; CPAR language extensions; Fortran-90; High Performance Fortran; N-body simulation problem; adaptive parallel computations; arrays; communication functions; distributed-memory multiprocessors; global name space; interprocessor data transfer; irregular parallel computations; object streams; parallel programming; pointer-based data structures; split phases; Computational modeling; Computer science; Computer simulation; Concurrent computing; Data structures; Distributed computing; High performance computing; Parallel processing; Phased arrays; Programming profession;

fLanguage

English

Publisher

ieee

Conference_Titel

Scalable High-Performance Computing Conference, 1994., Proceedings of the

Conference_Location

Knoxville, TN

Print_ISBN

0-8186-5680-8

Type

conf

DOI

10.1109/SHPCC.1994.296684

Filename

296684