Supermicro, a leading maker of high-performance servers, unveiled its range of GPU optimized platforms at the 2013 Nvidia GPU technology conference (GTC 2013) i. At the show, Supermicro will exhibit its extensive line of FatTwin and SuperServer GPU platforms, including the industry’s only 1U 3x Nvidia Grid K2 qualified SuperServer.
“Supermicro is a leader in GPU computing solutions. We truly focus on hardware optimization to maximize performance per watt, per dollar, per square foot. With platforms ranging from 12 GPUs in our 4U power saving FatTwin to the widest variety of 3U, 2U and 1U SuperServer solutions, Supermicro has the reach to provide the absolute best GPU cluster solutions quickly for any customer requirement. With virtualization at the forefront of this year’s GTC, Supermicro will be showing our expertise and readiness for this new revolution in GPU computing,” said Charles Liang, president and CEO of Supermicro.
Live systems featuring Nvidia Grid K1 and K2 boards will demonstrate how Nvidia Kepler architecture-based technology maximizes server utilization and, when scaled out, can support over 600 users per 42U rack in existing reference designs, and with a forthcoming 4U FatTwin configuration can support 1,800 users per rack. Combined with increased security and system manageability, Nvidia Grid delivers the most flexible, accelerated VDI experience for end users. In addition to the demos, Supermicro executives will present strategies and server architecture designs to consider when selecting GPU platforms for HPC, cloud and workstation applications.
“Nvidia and Supermicro are bringing to market a high-density server platform for virtual desktops and applications. The Supermicro servers set a new standard for high-density Nvidia Grid deployments, delivering high user densities with maximized performance, power conservation and ROI, thus removing conventional barriers to VDI deployments,” said Ed Ellett, senior vice president of Nvidia’s professional solutions business.