IBM is creating a Deep Computing business unit, intended to link together the company's hardware, software and services offerings for intensive computing projects.
The Deep Computing category encompasses tasks requiring extensive computational power, such as drug testing, medical simulations, film animation, data mining, fraud detection and weather modelling.
IBM has long attracted business from customers requiring exceptionally powerful systems, such as government research labs or petroleum exploration firms.
Leading the group will be Dave Turek, an IBM veteran who, most recently, focused on grid computing and Linux clusters. He reports to IBM vice president of eServer systems Mark Shearer
Turek has informally co-ordinated IBM's high-performance computing initiatives for a number of years , said IBM spokesman Charles Zinkowski. His position at this new organisation formalizes that role.
IBM already has several notable customers purchasing its most powerful systems, such as the Lawrence Livermore National Laboratory, the development partner and intended recipient of Blue Gene/L, a massive supercomputer scheduled for completion in 2005 that will be used to simulate events such as fires or the ageing of materials.
But as companies' computing needs grow more complex, the opportunities increase to spur sales outside the traditional life sciences market for high-performance computing, Turek said.
Operational analysis and digital imaging are two areas in which companies in an array of industries are beginning to need systems built for intensive processing, he noted.