| Title: | DETECTION OF CERTAIN FALSE DATA RACES FROM RUNTIME TRACES |
| DOI No: | 10.1142/9781860948534_0002 |
| Source: | INNOVATIVE APPLICATIONS OF INFORMATION TECHNOLOGY FOR THE DEVELOPING WORLD (pp 8-31)
|
| Author(s): | KOUSHIK SINHA
Address for Correspondence: Koushik Sinha, Honeywell Technology Solutions Lab, India 560076.
Honeywell Technology Solutions Lab, Bangalore, India 560076, India
RAJIV GUPTA
Department of Computer Science, University of Arizona, Tucson, AZ 85719, USA
|
| Abstract: | We present a family of algorithms for the on-the-fly detection of certain false data race conditions from runtime generated trace in a multithreaded environment. Our algorithms do not require instrumentation of the source code for catching data race conditions and can be implemented as low cost hardware solution. By filtering out false and benign race conditions from the reported races, the user is presented with a smaller set of potentially harmful data race conditions, hence speeding up the debugging process, especially for large multithreaded programs. The proposed algorithms are characterized by low overhead in terms of hardware support and effect on program execution time. |
| Full Text: | View full text in PDF format (961KB) |
| TOC: | Back to Table of Contents |
|
|
