HPC Clusters in 2010: Moving Forward with Ethernet

It has been well over ten years since the first Linux clusters arrived in the HPC market. Initially, the goal for many early pioneers was “just get it to work.” Today’s production HPC clusters, like the IBM iDataplex, are delivered as ready to run solutions. The iDataplex leverages ‘best-of-breed’ Ethernet connectivity from IBM alliance partner Force10 Networks, which offers ultra-high-density Ethernet technology designed to augment the HPC environment.

The modern HPC cluster offers unprecedented control over issues such as price, performance, power and cooling, space, ease-of-use, and job throughput. Join IBM and Force10 Networks as they explain the path forward for Ethernet based HPC clusters and learn how to optimize the various aspects mentioned above for your specific needs.

Presenters

Ed Suffern, Senior Engineer, Cluster Solutions Group, IBM

Rick Mathews, Sr. System Engineer, HPCC Solutions, Force10

Douglas Eadline, PhD, Senior HPC Editor, linuxdlsazine (Moderator)