Abstract
One of the most severe performance limitations of parallel applications today stems from the performance of I/O operations. Numerous projects have shown, that parallel I/O in combination with parallel file systems can significantly improve the performance of I/O operations. However, as of today there is no support for parallel I/O operations for applications using shared-memory programming models such as OpenMP. This paper introduces parallel I/O interfaces for OpenMP. We discuss the rationale of our design decisions, present the interface specification and a prototype implementation with the OpenUH compiler. We evaluate the performance of our implementation for various benchmarks on different file systems and demonstrate the benefits of the new interfaces. © 2012 Springer-Verlag.
Cite
CITATION STYLE
Mehta, K., Gabriel, E., & Chapman, B. (2012). Specification and performance evaluation of parallel I/O interfaces for OpenMP. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 7312 LNCS, pp. 1–14). https://doi.org/10.1007/978-3-642-30961-8_1
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
