Region partitioning method of nested loops with non-uniform dependences

Abstract

Many methods have been proposed in order to parallelize loops with non-uniform dependences, but most of these approaches perform poorly due to irregular and complex dependence constraints. Our approach is based on the convex hull theory which has adequate information to handle non-uniform dependences. By parallelizing the anti-dependence region using variable renaming, we divide the iteration space into two parallel regions and one or less sequential region. This paper proposes an improved region partitioning method (IRP) for minimizing the size of the sequential region. Comparison with other schemes shows more parallelism than the existing techniques.