increase your computing power not your computing budget
line
Home About us Clustering Technologies Streamline Cluster Systems Software Division Support News
Press Releases 
Press Releases 
Online Sales Contact us Site map


1PLs Company - Online Payday LoansSmall loans at 1PLs Company

Streamline Computing to ship UK's first Intel Itanium 2-based clusters

Streamline Computing announces the release of its new range of JetStream servers and cluster products based on the Intel® Itanium® 2 processors and server platforms. The newly released Intel® SR870BH2 server platform supports up to two Intel® Itanium® 2 processors with 6MB of L3 cache in a 2U rack-optimized form factor and the currently shipping 4-way platform supportiing Itanium® 2 with 3MB of L3 cache in a compact 4U rack-optimized chassis.

The JetStream-2 and JetStream-4 servers provide outstanding price/performance for High Performance Computing (HPC) applications that require 64-bit addressing to access large memory images. With support for up to 32GB of RAM (128GB with an additional memory expansion kit), the 4-way JetStream-4 is ideal for shared-memory parallel applications such as the finite element analysis codes that are used extensively in the engineering sector. A cluster architecture based on the 2-way JetStream-2 provides the most cost-effective solution for scalable MPI applications - such as academic and ISV codes for Bioinformatics, CEM, CFD, Computational Chemistry or Oil Recovery. General-purpose architectures may be built using clusters of JetStream-2 and JetStream-4 servers, with Myrinet® or Gigabit as the communication backplane.

Streamline will shortly deliver the UK's first Intel® Itanium® 2-based cluster using Myrinet 2000 in Europe. A 24-processor system using JetStream-4 servers and 1.3Ghz Itanium 2 processors has been acquired by the School of Computer Science, The University of Birmingham to support collaborative work with businesses and industry through the High Performance Computing Centre and through CERCIA, the new Centre of Excellence for Research in Computational Intelligence and Applications (www.cercia.ac.uk).

Streamline will also announce a number of JetStream-4 systems that have been acquired by two of the UK's largest engineering companies. These are to be used for CFD, as well as finite element analysis using the industry-leading Abaqus® software suite.

"The 2-way and 4-way Itanium 2 platforms are a welcome addition to our range of server and cluster solutions", said Dr Michael Rudgyard, Managing Director of Streamline Computing. "They provide us with powerful, cost-effective and reliable building-blocks for HPC customers who need 64-bit computing capabilities. The JetStream-2 and JetStream-4 are particular well-suited for running the ISV applications that many of our customers rely on - indeed, the Abaqus® benchmarks that we recently ran for one customer were the fastest seen on any platform, even on pre-production 1.2GHz Intel Itanium 2 processors !".

"With Streamline's own parallel debugger, the Distributed Debugging Tool, already ported to the Itanium® 2 platform and fully compatible with Intel's own tools and compilers, we can also provide developers of 64-bit MPI and OpenMP applications with a complete solution".

For more information on Streamline's new JetStream clusters and the Distributed Debugging Tool for Itanium® 2-based systems visit the company's web site at www.streamline-computing.com.

Streamline Computing Ltd is the UK's leading supplier of HPC clusters and tools. Formed by ex-academics from Warwick and Oxford Universities, the company has unrivalled expertise in scientific computing, parallel computing, and Linux clustering. Their customers include the majority of the UK's research universities, and a number of high-profile commercial customers. Whilst maintaining its focus on growing its turn-key cluster business, Streamline has invested significantly in the development of software for distributed computing. Its first product - the Distributed Debugging Tool - was launched in November 2002, and is set to become the defacto standard for parallel debugging on clusters.

right margin graphic
S||C
Top | Home