Abstract
This paper discusses a software tool to determine parallelism in sequential programs automatically and to determine the granularity size of parallel tasks to be executed on parallel machines. It includes a discussion of the techniques to detect parallelism in sequential programs and the techniques for determining the best way to partition parallel tasks into modules called grains, which will then be scheduled on a parallel machine to give near optimal parallel execution time.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have