IBM to commercialise Blue Gene supercomputer
- 11 November, 2004 14:15
Fresh from setting a record for performance among supercomputers just a few days ago, IBM has announced it is making a commercial version of its Blue Gene system available to be aimed at businesses and scientific researchers.
Called the IBM eServer Blue Gene, the Power-based system has employed a mix of cooling and clustering technologies to achieve a performance of 5.7 teraflops. The system has a footprint of less than one square metre. Price of the system, which is available now, starts at US$1.5 million. For the first time, users can also rent the system from one of IBM's Deep Computing On Demand centres located in the United States or Europe, a spokesman said.
"We think this system will allow us to introduce a new class of high-performance computing capability to industry-specific businesses and for out clients to deliver better optimised functionality," said Colin Parris, vice-president with IBM's eServer Product Management group.
Parris said his company was actively working with a number of industry business partners to put together an optimal delivery of Blue Gene's capabilities aimed at "key marketplaces".
IBM has been consistently working with partners over the past few years to make Blue Gene more applicable to handle a variety of different computing workloads. IBM and a large number of national testing labs and universities are working on a list of high-performance computing applications in the areas of life sciences, financial modelling, hydrodynamics, quantum chemistry, molecular dynamics, astronomy, and climate modelling.
The company is also looking at more commercial applications involving grid computing, business intelligence, manufacturing processes, and risk and compliance, company officials said.
The new system will be available in configurations ranging anywhere from 1 to 64 racks. Partially populated racks with less than 1024 nodes are also available, a company spokesman said.
Blue Gene was born as a result of a five-year project that cost US$100 million in research alone. It started life as a research vehicle for protein folding.