Efficient Detection of Determinacy Races in Cilk Programs

Abstract

A parallel multithreaded program that is ostensibly deterministic may nevertheless behave nondeterministically due to bugs in the code. These bugs are called determinacy races, and they result when one thread updates a location in shared memory while another thread is concurrently accessing the location. We have implemented a provably efficient determinacy-race detector for Cilk, an algorithmic multithreaded programming language. If a Cilk program run on a given input data set has a determinacy race, our debugging tool, which we call the "Nondeterminator," guarantees to detect and localize the race.The core of the Nondeterminator is an asymptotically efficient serial algorithm (inspired by Tarjan's nearly linear-time leastcommon-ancestors algorithm) for detecting determinacy races in series-parallel directed acyclic graphs. For a Cilk program that runs in T time on one processor and uses v shared-memory locations, the Nondeterminator runs in O (T α(v; v)) time, where α is Tarjan's functional inverse of Ackermann's function, a very slowly growing function which, for all practical purposes, is bounded above by 4. The Nondeterminator uses at most a constant factor more space than does the original program. On a variety of Cilk program benchmarks, the Nondeterminator exhibits a slowdown of less than 12 compared with the serial execution time of the original optimized code, which we contend is an acceptable slowdown for debugging purposes.

Memory-mapping support for reducer hyperobjects

“…Thus in this example, a cilk_sync statement is inserted at line 10 to ensure that the function does not return until the walk of the left and right children are done. This parallelization is incorrect, however, since it contains a determinacy race [11] (also referred a as general race [28]) on the list l, because the logically parallel subcomputations -the walks of the left and right subtrees -may potentially access the list in parallel.…”

“…Reducer hyperobjects (or reducers for short) [12] have been shown to be a useful linguistic mechanism to avoid determinacy races [11] (also referred as general races [28]) in dynamic multithreaded programs. Reducers allow different logical branches of a parallel computation to maintain coordinated local views of the same nonlocal variable.…”

SingleTrack: A Dynamic Determinism Checker for Multithreaded Programs

“…1 This special case of entirely deterministic applications was also addressed by the Cilk Nondeterminator [7], whereas this paper addresses the problem in a more general setting.…”

“…The Cilk project investigated verifying determinism of entire multithreaded applications, first addressing a more restricted fork-join concurrency structure [7] and later extending that approach to more general locking idioms [2]. While successful for deterministic Cilk applications, this approach does not support applications (like hedc, queue-jg, and queue-mm) that are non-deterministic but contain deterministic subcomputations.…”

CILK: A Multi-Threaded Programming System for Meta-Computers

“…Cilk-5.0 could use operating system threads as well as processes to implement the individual Cilk "workers" that schedule Cilk threads. The Cilk-5.2 release included a debugging tool called the Nondeterminator [12,9], which can help Cilk programmers to localize data-race bugs in their code. (The Nondeterminator is not included in the present Cilk-5.3 release.)…”