Modeling Performance and Energy for Applications Offloaded to Intel Xeon Phi

Document Type

Conference Paper

Publication Date

2015

DOI

10.1145/2834899.2834903

Publication Title

Proceedings of the 2nd International Workshop on Hardware-Software Co-Design for High Performance Computing

Pages

7 (1-8)

Conference Name

2nd International Workshop on Hardware-Software Co-Design for High Performance Computing, 15 November 2015, Austin, Texas

Abstract

Accelerators are adopted to increase performance, reduce time-to-solution, and minimize energy-to-solution. However, employing them efficiently, given system and application characteristics, is often a daunting task. A goal of this work is to propose a general model that predicts performance and power requirements for an application, computational portions of which are offloaded to an accelerator. Intel Xeon Phi is the only accelerator type investigated here, and only in offload execution mode. This mode is also employed by other accelerator types, such as GPU; thus the proposed model is applicable directly. The predictive capabilities of the model are demonstrated by determining the best hardware-software configuration instances with respect to the minimum energy consumption for the CoMD proxy application executed on single or multiple nodes. For the CoMD problem sizes investigated here, the best modeled configuration was relatively close to the best measured configuration with relative error under 5% of the energy consumed for most configurations. Initial model validation also confirmed the model accuracy for a variety of model parameters, such as host computation time and power consumption on the host and accelerator. The model also provides estimates of the total data movement and computational throughput as well as of some key metrics, such as FLOPs-per-joule and bytes-per-joule, which are commonly used to study the energy-performance trade-offs.

Rights

© 2015 Association for Computing Machinery

"ACM treats links as citations (references to objects) rather than as incorporations (embedding of objects). Permission is not needed to create links to citations in The ACM Digital Library or Online Guide to Computing Literature. ACM encourages the widespread distribution of links to the definitive Version of Records of its copyrighted works in the ACM Digital Library and does not require that authors obtain prior permission to include such links in their new works.

However, someone who creates a work or a service whose pattern of links substantially duplicates an ACM-copyrighted volume or issue should get prior permission from ACM. One example: the creator of "A Table of Contents for the Current Issue of TODS" -- consisting of citations and active links to author-versions of the works in the latest issue of TODS -- needs ACM permission because that creator is reproducing an ACM-copyrighted work. If all the links in the "Table of Contents" pointed to the ACM-held definitive Version of Records, ACM would normally give permission because then the new work advertises an ACM work. To avoid misunderstandings, consult with ACM before duplicating an ACM work via links.

If an author wishes to embed a copyrighted object---rather than a link---in a new work, that author needs to obtain the copyright holder's permission."

Original Publication Citation

Lawson, G., Sundriyal, V., Sosonkina, M., & Shen, Y. (2015). Modeling performance and energy for applications offloaded to Intel Xeon Phi. In Proceedings of the 2nd International Workshop on Hardware-Software Co-Design for High Performance Computing (7). Association for Computing Machinery. https://doi.org/10.1145/2834899.2834903

Link to Full Text

Share

COinS