Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Evaluation of the heuristic approach for program partitioning and allocation using an expert system

Thesis/Dissertation ·
OSTI ID:5213886

Over the last few years, a number of novel computer architectures offering new parallel organizations for computation have been proposed. One class of the resulting computer architectures is designed using the data-flow model of execution. In these architectures the availability of operands triggers the execution of operations. The motivation for these architectures is the desire to utilize fine-grained concurrency to increase computer performance, and to exploit VLSI in the design of computers. The programming methodology required to exploit the performance potential of such architectures is a critical aspect. Program partitioning and allocation are the two main steps in programming a data flow computer. Program partitioning refers to the division of a program into several sections or tasks which can be executed on different processors. Program allocation is the process of assigning these task to individual processors. To study the program partitioning and allocation problem, a general data flow architecture model is defined. A class of data flow-computation graphs is considered. The expert system conducts an heuristic search to find the best partitioning and allocation according to a performance measurement.

Research Organization:
Toledo Univ., OH (USA)
OSTI ID:
5213886
Country of Publication:
United States
Language:
English

Similar Records

Partitioning and allocation of functional programs for data flow processors
Book · Tue Dec 31 23:00:00 EST 1985 · OSTI ID:5045824

On partitioning of algorithms for parallel execution on VLSI circuit architectures
Thesis/Dissertation · Thu Dec 31 23:00:00 EST 1987 · OSTI ID:6930573

Program partitioning for NUMA multiprocessor computer systems. [Nonuniform memory access]
Journal Article · Sun Oct 31 23:00:00 EST 1993 · Journal of Parallel and Distributed Computing; (United States) · OSTI ID:5703692