CUMBERNAULT, U.K.—Computing specialist Nallatech recently announced a strategic collaborative agreement that will see the company work with California-based Silicon Graphics Inc. (SGI) to develop high-performance computing (HPC) solutions for a variety of fields, including bioinformatics, medical imaging, and visualization systems. The new product and service offerings will combine SGI's expertise in Reconfigurable Application-Specific Computing (RASC) and Nallatech's field-programmable gate array (FPGA) technology.
"SGI recognized Nallatech's technical breadth, depth and success in the FPGA technology space and Nallatech recognized SGI's expertise and brand in the HPC market space together with their vision of how FPGAs could be integrated into HPC products that would be accepted by the HPC user community," says Mr. Allan Cantle, Nallatech president and CEO. "Therefore, it made sense for both companies to collaborate as this would give customers the most complete offering in the shortest timeframe."
According to SGI's Dr. Dan Stevens, chemical and life sciences business manager, the announcement is the latest step in SGI's plans, announced in November 2003, to deliver a revolutionary class of supercomputers.
"Using a new concept SGI calls 'multi-paradigm computing', the groundbreaking machines will unite previously disparate computing architectures with SGI's scalable shared-memory technology," he explains. "SGI's long-term vision is to improve productivity by creating the first supercomputers capable of supporting and combining different computational approaches, providing optimal performance specifically for technical applications, regardless of programming model or computational balance."
The two companies will develop integrated high-performance FPGA-based computing solutions to facilitate large-scale data analysis projects, such as those found in bioinformatics initiatives.
"The collaboration will bring a deeper understanding of HPC users needs together with a technological lead in computing with FPGAs and the tools for making FPGA more accessible," Cantle says. "Data analysis, bioinformatics, and visualization enhancements all have the capability of benefiting from FPGA computing."
SGI has been testing early-phase technology for the past six months and anticipates the first public iteration by early next year.
"It has been widely accepted for some time that FPGAs can significantly impact the processing performance of drug discovery," explains Cantle. "This is primarily because the architecture of the FPGA more readily lends itself to drug discovery models. In this sector in particular, existing processor architectures are particularly bad at processing these models."
For this reason, he adds, the drug discovery market provides an opportunity to clearly demonstrate the benefits of FPGA computing. The challenge to date has been in the programmability of these systems. This, he says, is where the collaboration with SGI plays such an important role.
According to Stevens, RASC will enable customers to program reconfigurable chips to optimize their specific tasks in co-processing systems. This optimization process can offer significant—potentially orders of magnitude—performance improvements at lower power and heat than conventional microprocessors alone.
"Now that the techniques and system requirements for computational drug discovery are defined and routinely administered, to remain competitive, R&D organizations must push the envelope in two directions: increased speed and increased capability," he adds "The SGI and Nallatech partnership offers state-of-the-art computational data processing capability as an option to the SGI Altix, a capability computing standard architecture for drug discovery research. Now, customers can add this device to their Altix to either decrease the time to result by an order of magnitude or ask questions with a virtually unlimited scope. Either way, results that answer larger questions can now be obtained that, ultimately, will unleash the power of R&D organizations and drive a stronger and more competitive organization."