Model-checking task parallel programs for data-race

Abstract

Data-race detection is the problem of determining if a concurrent program has a data-race in some execution and input; it has been long studied and often solved. The research in this paper reprises the problem in the context of task parallel programs with the intent to prove, via model checking, the absence of data-race on any feasible schedule for a given input. Many of the correctness properties afforded by task parallel programming models such as OpenMP, Cilk, X10, Chapel, Habanero, etc. rely on data-race freedom. Model checking for data-race, presented here, is in contrast to recent work using run-time monitoring, log analysis, or static analysis which are complete or sound but never both. The model checking algorithm builds a happens-before relation from the program execution and uses that relation to detect data-race similar to many solutions that reason over a single observed execution. Unlike those solutions, model checking generates additional program schedules sufficient to prove data-race freedom over all schedules on the given input. The approach is evaluated in a Java implementation of Habanero using the JavaPathfinder model checker. The results, when compared to existing data-race detectors in Java Pathfinder, show a significant reduction in the time required for proving data race freedom.