To search, Click below search items. |
| All Published Papers Search Service |
|
|
|
Title |
A Lightweight Static Data Race Checker for OpenMP Software |
|
Author |
Hend Alshede |
| Citation |
Vol. 26 No. 1 pp. 102-106 |
|
Abstract |
OpenMP has become one of the most widely adopted interfaces for parallel programming, particularly for managing shared-memory parallelism on multi-core architectures. With the increasing prevalence of multi-core systems, a large body of sequential software has been parallelized using OpenMP. However, writing correct parallel programs remains challenging, and concurrency errors such as data races and deadlocks are common. This paper presents the design and conceptual evaluation of a lightweight static data race checker for OpenMP programs. The proposed approach relies on lexical, syntactic, and semantic analysis performed at compile time, without requiring runtime instrumentation. By focusing on simplicity and early error detection, the checker aims to provide programmers with fast and useful feedback about potential data races. Compared to dynamic approaches, the proposed method can identify likely race conditions earlier in the development process with lower analysis overhead. |
|
Keywords |
OpenMP; data race detection; Parallel programming; High-Performance Computing; Static analysis. |
|
URL |
