Efficient Implementation of High Performance Fortran via Adaptive Scheduling -- An Overview
Authors:
Edward A. Kornkven, Laxmikant V. Kale
Parallel Programming Laboratory, Department of Computer Science, University
of Illinois at Urbana-Champaign
Proceedings of the International Workshop on Parallel Processing. Dec 1994, Bangalore, India.
We have been developing a compiler for a subset of High Performance Fortran. We have shown that generating messagedriven code provides an opportunity for improved efficiency in the presence of communication. By utilizing a notation called Dagger, we are able to schedule work efficiently without relying on complicated and unreliable compile-time approaches. This paper gives an overview of our approach and reports on the project's status.