This Technical White Paper is
based on a High
Performance Computing (HPC) worshop sponsored
by Appro and its technology and business partners
to included Raytheon, AMD, Mellanox, and DataDirect.
As you know, commodity clusters have become the primary
choice for High Performance Computing (HPC) applications.
Vendors are beginning to form alliances that allow
them to offer complete turnkey solutions to users.
It all starts with the design of the system; it must
be viewed as a single entity to offer a balanced,
scalable, and reliable platform to end-users. Its
components must be well selected and well integrated.
This document provides examples and technical details
on how to build a state-of-the Art supercomputer class
system that is ready-to-use with enterprisewide scalability
and reliability.