Accession Number:

ADA281194

Title:

A Framework for Exploiting Data and Functional Parallelism on Distributed Memory Multicomputers

Descriptive Note:

Technical rept.

Corporate Author:

ILLINOIS UNIV AT URBANA CENTER FOR RELIABLE AND HIGH-PERFORMANCE COMPUTING

Personal Author(s):

• Ramaswamy, Shankar
• Sapatnekar, Sachin
• Banerjee, Prithviraj

Report Date:

1994-06-21

Pagination or Media Count:

40.0

Abstract:

Recent research efforts have shown the benefits of integrating functional and data parallelism over using either pure data parallelism or pure functional parallelism. The work in this paper presents a theoretical framework for deciding on a good execution strategy for a given program based on the available functional and data parallelism in the program. The framework is based on assumptions about the form of computation and communication cost functions for multicomputer systems. We present mathematical functions for these costs and show that these functions are realistic. The framework also requires specification of the available functional and data parallelism for a given problem. For this purpose, we have developed a graphical programming tool. Currently, we have tested our approach using three benchmark programs on the Thinking Machines CM-5 and Intel Paragon. Results presented show that the approach is very effective and can provide a two- to three-fold increase in speedups over approaches using only data parallelism. Functional and data parallelism, Macro dataflow graphs, Graphical programming, Processor allocation, Scheduling, DMM

Descriptors:

• *DISTRIBUTED DATA PROCESSING
• *COMPUTER PROGRAMMING
• *PARALLEL PROCESSING
• ALGORITHMS
• COMPUTATIONS
• SCHEDULING
• ALLOCATIONS
• FIELDS(COMPUTER PROGRAMS)
• EXECUTIVE ROUTINES
• BENEFITS
• FUNCTIONAL ANALYSIS
• COSTS
• SPECIFICATIONS
• COMPUTER COMMUNICATIONS
• STRATEGY

Subject Categories:

• Computer Programming and Software

Distribution Statement:

APPROVED FOR PUBLIC RELEASE