System and method for controlling execution of nested loops in parallel

Abstract

A system and method for controlling execution of nested loops in parallel in a computer including multiple processors, and a compiler for generating code therefor. The code enables the computer to operate in the following manner. Each processor processes an iteration of an outer loop in a set of nested loops. If the outer loop contains more iterations than processors in the system, the processors are initially assigned early iterations, and the later iterations are assigned to the processors as they finish their earlier iterations, until the processors have processed all of the iterations. Each processor, during processing of an outer loop iteration runs the iterations comprising the inner loop serially. In order to enforce dependencies between the loops, each processor reports its progress in its iterations of the inner loop to the processor executing the succeeding outer loop iteration. In addition, each processor, before processing the inner loop iterations of an assigned outer loop iteration, determines whether the processor computing the preceding outer loop is ahead or behind in processing its inner loop iteration by an amount which guarantees that dependencies will be enforced.