Efficient Filtering Technique for Reducing Time Overhead of Dynamic Data Race Detection in Multithread Programs

Abstract

Data races are the hardest defect to handle in multithread programs because they may lead to unpredictable results of the program caused by nondeterministic interleaving of concurrent threads. The main drawback of dynamic data race detection is the heavy additional overhead to monitor and analyze memory operations and thread operations during an execution of the program. It is important to reduce the additional overheads for debugging the concurrency bug. This paper presents a monitoring filtering technique that rules out repeatedly executing regions of parallel loops from the monitoring targets.