Zobrazit v EDS

Multi-Threaded Parallel I/O for OpenMP Applications.

Uloženo v:
Podrobná bibliografie
Název: Multi-Threaded Parallel I/O for OpenMP Applications.
Autoři: Mehta, Kshitij1 kmehta@cs.uh.edu, Gabriel, Edgar1 gabriel@cs.uh.edu
Zdroj: International Journal of Parallel Programming. Apr2015, Vol. 43 Issue 2, p286-309. 24p.
Témata: *PARALLEL processing, *MATHEMATICAL optimization, *USER interfaces, THREADS (Computer programs), GRAPHICS processing units
Abstrakt: Processing large quantities of data is a common scenario for parallel applications. While distributed memory applications are able to improve the performance of their I/O operations by using parallel I/O libraries, there is no support for parallel I/O operations for applications using shared-memory programming models such as OpenMP available as of today. This paper presents parallel I/O interfaces for OpenMP. We discuss the rationale of our design decisions, present the interface specification, an implementation within the OpenUH compiler and discuss a number of optimizations performed. We demonstrate the benefits of this approach on different file systems for multiple benchmarks and application scenarios. In most cases, we observe significant improvements in I/O performance as compared to the sequential version. Furthermore, we perform a comparison of the OpenMP I/O functions introduced in this paper to message passing interface I/O, and demonstrate the benefits of the new interfaces. [ABSTRACT FROM AUTHOR]
Copyright of International Journal of Parallel Programming is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Databáze: Business Source Index
Popis
Abstrakt:Processing large quantities of data is a common scenario for parallel applications. While distributed memory applications are able to improve the performance of their I/O operations by using parallel I/O libraries, there is no support for parallel I/O operations for applications using shared-memory programming models such as OpenMP available as of today. This paper presents parallel I/O interfaces for OpenMP. We discuss the rationale of our design decisions, present the interface specification, an implementation within the OpenUH compiler and discuss a number of optimizations performed. We demonstrate the benefits of this approach on different file systems for multiple benchmarks and application scenarios. In most cases, we observe significant improvements in I/O performance as compared to the sequential version. Furthermore, we perform a comparison of the OpenMP I/O functions introduced in this paper to message passing interface I/O, and demonstrate the benefits of the new interfaces. [ABSTRACT FROM AUTHOR]
ISSN:08857458
DOI:10.1007/s10766-014-0306-9