Integrating data distribution and loop transformations
Conference
·
OSTI ID:125600
- Louisiana State Univ., Baton Rouge, LA (United States)
This paper presents a technique for finding good distributions of arrays and suitable loop restructuring transformations so that communication is minimized in the execution of nested loops on message passing machines. For each possible distribution (by one or more dimensions), we derive the best unimodular loop transformation that results in block transfers of data. Unlike other work which focus on either data layout or on program transformations, this paper combines both array distributions and loop transformations resulting in good performance. The techniques described here are suitable for dense linear algebra codes.
- OSTI ID:
- 125600
- Report Number(s):
- CONF-950212-; CNN: Grant CCR-9210422; Contract LEQSF (1991-94)-RD-A-09; TRN: 95:005768-0145
- Resource Relation:
- Conference: 7. Society for Industrial and Applied Mathematics (SIAM) conference on parallel processing for scientific computing, San Francisco, CA (United States), 15-17 Feb 1995; Other Information: PBD: 1995; Related Information: Is Part Of Proceedings of the seventh SIAM conference on parallel processing for scientific computing; Bailey, D.H.; Bjorstad, P.E.; Gilbert, J.R. [eds.] [and others]; PB: 894 p.
- Country of Publication:
- United States
- Language:
- English
Similar Records
Loop partitioning for distributed-memory multiprocessors as unimodular transformations
Enabling unimodular transformations
Generalized unimodular loop transformations for distributed-memory multiprocessors
Technical Report
·
Sat Dec 01 00:00:00 EST 1990
·
OSTI ID:125600
+1 more
Enabling unimodular transformations
Conference
·
Sat Dec 31 00:00:00 EST 1994
·
OSTI ID:125600
Generalized unimodular loop transformations for distributed-memory multiprocessors
Technical Report
·
Tue Jan 01 00:00:00 EST 1991
·
OSTI ID:125600