Eliminating Stale Data References through Array Data-Flow Analysis

In this paper, we develop a compiler algorithm for detecting references to stale data in shared-memory multiprocessors. The algorithm consists of two key analysis techniques, stale reference detection and locality preserving analysis. While the stale reference detection finds the memory reference patterns that may violate cache coherence, the locality preserving analysis minimizes the number of such stale references by analyzing both temporal and spatial reuses. By computing the regions referenced by arrays inside loops, we extend the previous scalar algorithms [7, 9] for more precise analysis. We have implemented the algorithm on the Polaris parallelizing compiler [19], and using execution-driven simulations on Perfect Perfect benchmarks we demonstrate how unnecessary cache misses can be eliminated by the automatic stale reference detection.