Short Paper: Exploring the Performance and Mapping of HPC Applications to Platforms in the Cloud

ACM Symposium on High-Performance Parallel and Distributed Computing (HPDC) 2012

Pulication Type: Paper

Download: pdf ps

Abstract

This paper presents a scheme to optimize the mapping of HPC applications to a set of hybrid dedicated and cloud resources. First, we characterize application performance on dedicated clusters and cloud to obtain application sig- natures. Then, we propose an algorithm to match these signatures to resources such that performance is maximized and cost is minimized. Finally, we show simulation results revealing that in a concrete scenario our proposed scheme re- duces the cost by 60% at only 10-15% performance penalty vs. a non optimized configuration. We also find that the ex- ecution overhead in cloud can be minimized to a negligible level using thin hypervisors or OS-level containers.

Research Areas

High Performance Computing in the Cloud

Text Ref

BibTex

@inproceedings{Gupta:2012:EPM:2287076.2287093,
 author = {Gupta, Abhishek and Kal{\'e}, Laxmikant V. and Milojicic, Dejan S. and Faraboschi, Paolo and Kaufmann, Richard and March, Verdi and Gioachin, Filippo and Suen, Chun Hui and Lee, Bu-Sung},
 title = {Exploring the performance and mapping of HPC applications to platforms in the cloud},
 booktitle = {Proceedings of the 21st international symposium on High-Performance Parallel and Distributed Computing},
 series = {HPDC '12},
 year = {2012},
 isbn = {978-1-4503-0805-2},
 location = {Delft, The Netherlands},
 pages = {121--122},
 numpages = {2},
 url = {http://doi.acm.org/10.1145/2287076.2287093},
 doi = {10.1145/2287076.2287093},
 acmid = {2287093},
 publisher = {ACM},
 address = {New York, NY, USA},
 keywords = {clouds, high performance computing, resource scheduling},
}