By Topic

Awareness of MPI Virtual Process Topologies on the Single-Chip Cloud Computer

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
Christgau, S. ; Inst. of Comput. Sci., Univ. of Potsdam, Potsdam, Germany ; Schnor, B.

This paper presents an approach to improve the communication bandwidth between MPI tasks by evaluating the application's communication graph and making optimal use of communication buffers. It was implemented on the Intel Single-Chip Cloud Computer (SCC). The SCC is a processor with 48 cores created by Intel Labs. Special to the SCC is the SRAM-based Message Passing Buffer which can be used as fast communication memory. First, we evaluate RCKMPI, the MPI implementation for the SCC which makes use of this MPB. Then we extend RCKMPI to support virtual process topologies like defined in the MPI specification. The presented approach shows a performance improvement up to 44 % for a communication-intensive application.

Published in:

Parallel and Distributed Processing Symposium Workshops & PhD Forum (IPDPSW), 2012 IEEE 26th International

Date of Conference:

21-25 May 2012