SERVER VENDOR SGI has announced that it will put Nvidia's latest and greatest Tesla 20-series GPGPU boards into its high performance computing (HPC) servers.
SGI, once of Silicon Graphics Inc fame and now part of Rackable Systems, has been a staunch supporter of GPGPU computing on its servers for some time, and its announcement to use Nvidia's latest Tesla 20-series boards in all of its server lines shows great commitment to Nvidia. SGI's move follows that of HP, which was the first major computer vendor to ship servers with Nvidia's Telsa 20-series boards.
Last year SGI made comments to The INQUIRER that gave the distinct impression that the firm was falling out of love with Nvidia's GPGPU boards. SGI also announced accelerator cards from other vendors. However now it seems that SGI still has a soft spot for Nvidia's Tesla boards and it's not surprising to see why. Both AMD and Nvidia agree that GPGPU computing is needed to hit the next level in high performance computing - exascale computing - and customers have started buying into that idea as well.
Bill Mannel, VP of product marketing at SGI said, "Interest in GPU computing is growing fast, and the latest Nvidia Tesla M2090 GPU coupled with their CUDA Toolkit 4.0 software release brings superior capability into the hands of developers."
SGI's Altix UV, Altix ICE 8400, Rackable and Prism XL servers can all be fitted with Nvidia's Tesla M2090 boards as of today. SGI also said that it has updated its Accelerator Execution Environment for the Prism XL server to Nvidia's Cuda Toolkit 4.0. µ
Uses 20 percent less power than traditional systems
It's becoming more prevalent in car research and development
Sign up for INQbot – a weekly roundup of the best from the INQ