A Technique for Parallel Loop Execution

Authors:
Volodymyr Beletskyy
Affiliations:
-
Venue:
PARA '02 Proceedings of the 6th International Conference on Applied Parallel Computing Advanced Scientific Computing
Year:
2002

Citing 12
Cited 0

Methods and applications of error-free computation

Methods and applications of error-free computation
A practical algorithm for exact array dependence analysis

Communications of the ACM
Non-unimodular transformations of nested loops

Proceedings of the 1992 ACM/IEEE conference on Supercomputing
Some efficient solutions to the affine scheduling problem: I. One-dimensional time

International Journal of Parallel Programming
The privatizing DOALL test: a run-time technique for DOALL loop identification and array privatization

ICS '94 Proceedings of the 8th international conference on Supercomputing
Compiler transformations for high-performance computing

ACM Computing Surveys (CSUR)
Maximizing parallelism and minimizing synchronization with affine partitions

Parallel Computing - Special issues on languages and compilers for parallel computers
Generation of Efficient Nested Loops from Polyhedra

International Journal of Parallel Programming - Special issue on instruction-level parallelism and parallelizing compilation, part 2
Dependence Analysis for Supercomputing

Dependence Analysis for Supercomputing
High Performance Compilers for Parallel Computing

High Performance Compilers for Parallel Computing
An Efficient Data Dependence Analysis for Parallelizing Compilers

IEEE Transactions on Parallel and Distributed Systems
The Power Test for Data Dependence

IEEE Transactions on Parallel and Distributed Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

A technique and its modification to execute in parallel perfectly and imperfectly nested loops with linear dependencies are considered. The technique obliges that firstly, all the loop iterations are executed in parallel. Then, the ends of pairs of dependent iterations are re-executed serially in lexicographical order. With the modified technique, first, all independent iterations and the iterations that are the beginnings of pairs of dependent ones are executed in parallel. Then, the iterations that are the ends of those are executed serially. Possibilities of implementing the technique and its modification by means of the Omega calculator are presented. Experimental results and future research are considered.