By Topic

Network-based processing versus host-based processing: Lessons learned

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
$33 $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)
Gilad Shainer ; Mellanox Technologies Sunnyvale, California ; Amir Ancel

Interconnect technologies play a great role in the efficiency and scalability of high-performance systems. The capability of low latency, high bandwidth, high message rate and low CPU overhead greatly influence application performance and overall system productivity. In the paper we explore two different architectures of high-speed interconnects-offloading interconnects (where the interconnect handles the transport protocol) and onloading interconnects (where the host CPU is responsible for handling the transport protocol. We will demonstrate that offloading solutions are preferred for high-performance systems by presenting several cases of latency, message rate and the ability to support advanced capabilities.

Published in:

Cluster Computing Workshops and Posters (CLUSTER WORKSHOPS), 2010 IEEE International Conference on

Date of Conference:

20-24 Sept. 2010