Supercomputer chosen for scalable shared memory.
March 16, 2010 | SGI’s recent sale of its Altix UV system to a major European cancer research center is evidence that the former supercomputing player is still a force to be reckoned with after emerging from bankruptcy.
In 2009, Rackable (see “Out of the Gate,” Bio•IT World, Sept 2009) bought the assets of SGI and changed the name of the company to SGI. “It’s really a merger and we’re now equally focused on HPC as well as the Internet and other markets that Rackable focused on,” said Geoffrey Noer, SGI senior director of product marketing, adding that the integration went extremely smoothly.
The new SGI’s chief products include the Octane III, a desk-side device that combines the power of an HPC cluster with the portability of a workstation that can address a variety of technical and graphical computing applications.
Launched at SC09, the Altix UV has an open systems architecture. Noer says the Altix is in any respect “the world’s fastest supercomputer,” which can scale up to 2,048 cores and 16 Terabytes of global shared memory. The hardware platform is has modular blades for ‘plug and solve’ flexibility. “The response has been phenomenal,” Noer told Bio•IT World. “Many of the customers purchasing Altix UV systems are doing so to complement scale-out clusters.” Among the early customers are the University of Tennessee, Oak Ridge National Laboratory, the North German Supercomputing Alliance, and other centers in France and Japan.
The most intriguing life sciences customer announced thus far is the Institute of Cancer Research (ICR) in London, which was a longtime customer of SGI as well as Rackable servers.
ICR selected the Altix UV to take advantage of its massively scalable shared memory system to process the explosion of data in fields including MRI imaging, mass spectrometry, and deep-sequencing information, across thousands of CPUs. ICR staff hope the Altix UV will allow large, diverse data sets to be processed quickly and facilitate the correlation of medical and biological data. Rune Linding, ICR team leader in cellular and molecular logic, said: “Eventually, this will lead to network-based cancer models that will be used to streamline the process of drug development.”
“Systems biology demands massive integration of extremely large data sets. Large shared memory should enable us to handle such data at a much higher speed and with a greater focus on the biological questions at hand,” added Peter Rigby, ICR’s chief executive professor. The ICR (affiliated with the University of London) is one of Europe’s leading cancer research centers, with a formidable record in pioneering research into cancer genetics, cell biology and epidemiology.
Noer said that SGI is in a strong position moving forward, and that computational bioinformatics remains one of its key HPC markets. “We’ll be introducing more products over time and Altix UV is already grabbing a lot of attention. Bioinformatics is a core market for us and one in which we plan to be more visible in the future,” he said.