Saturn

Small Footprint,
Big Performance
Incredibly easy to Manage
Quiet, Cool and Efficient
View All Saturn Bundles
View Datasheet - (PDF 304K)
View Saturn Revolution Datasheet - (PDF 233K)
View Cluster Toolbox - Cluster Setup & Basic Management Kit
How to Buy?
Contact us via our website - Complete our On-line Form
Phone our Sales Team -
Christy Kennedy, 804.521.3052, christy@jrti.com, or
Joseph Rittenberry, 804.521.3050, joseph@jrti.com
Find a Rep - View our sales coverage map
.
|
Dimensions: 14in. x 12.5in. x 22.5in.
Easily build larger clusters from multiple systems
Four nodes, each with two dual- or single-core AMD Opteron HE processors with Direct Connect Architecture
1MB 16-way set associative L2 cache/128KB L1 cache per processor core
Up to 16GB Registered ECC DDR 400MHz SDRAM per node
8Gbps total low-latency (21µs TCP/IP) gigabit ethernet bandwidth
Included Houston and Ganglia cluster control and monitoring software tools
|
Saturn systems are as easy to manage as a PC. Rocketcalc Houston cluster management software and included Ganglia monitoring software provide a simple, intuitive interface for cluster monitoring and control. Turnkey Saturn UniclusterT systems arrive with Novell SUSE Linux and a wide selection of available, optimized and ready to run HPC software applications, including up to 3 Terabytes of cluster-wide PVFS2 storage. Saturn systems are easy to integrate with existing GNU/Linux, Apple and Windows workstations.
Saturn systems feature the power-efficient, high performance 64-bit AMD Opteron HE processors. Powered from a standard wall outlet, the Saturn chassis provides significant airflow without generating excessive noise.
Saturn clusters deliver best-in class performance/price ratios. The benchmarks below were run on a Saturn 270HE system with 16 2GHz AMD Opteron processor cores, 16GB DDR 400MHz RAM, and a single 400GB SATA hard disk with SuSE 9.3 and gcc 3.3.3 and LAM/MPI 7.1. The math kernel library for Linpack was the libgoto_opteron 64p-r1.00 library by Kazushige Goto (available from http://www.tacc.utexas.edu
/resources/software ).
|