Posted on Monday, May 04 2009 @ 15:17 CEST by Thomas De Maesschalck
NVIDIA rolled out the Tesla GPU Preconfigured Cluster, a new GPU-based computing solution for researchers:
Tesla GPU Preconfigured Clusters provide up to 30 times the performance of CPU-only solutions, enabling IT managers to deploy smaller, more computationally dense systems. These consume substantially less power -- which creates substantially lower costs -- while still meeting the increasing demands of such applications as computational fluid dynamics, molecular dynamics, seismic processing and financial computing.
French banking giant BNP Paribas' Corporate and Investment Banking division recently replaced 500 traditional CPU cores, consuming 25 kW, with a smaller cluster consisting of CPU servers and two Tesla S1070 1U systems, which consume only 2kW. Taking into account the dramatic acceleration achieved using Tesla GPUs, the lower power cluster enabled BNP Paribas to consume 190-times less electricity than before.
The increasing momentum being experienced by GPUs in the high performance computing segment today demonstrates a clear industry need for accessible computing solutions.
"There are 15 to 20 million engineers, scientists and researchers around the world struggling for time on supercomputers, which has led to a huge pent-up demand for computation," said Andy Keane, general manager of the Tesla business at NVIDIA. "With the launch of the Tesla Preconfigured Cluster, every one of them can easily deploy a GPU-powered supercomputing cluster that dramatically reduces their power consumption while still advancing the pace of their work."
"Time on a supercomputer can be extremely difficult to get, especially since some of our computations run for weeks to months. Also, buying a supercomputer is expensive for every university research group," said Axel Kohlmeyer, associate director at the Center for Molecular Modeling at the University of Pennsylvania. "Since we got access to a Tesla GPU enabled cluster, we can run our molecular dynamics algorithms up to 100X faster and more importantly run bigger and more complex simulations and do research that was impossible to do before - this is game changing for us."
Tesla Preconfigured Clusters consist of x86 CPU servers coupled with Tesla S1070 1U GPU systems. Configurations start at 16 Teraflops of performance, delivered by four Tesla S1070s, each containing four Tesla 10-series GPUs. All systems include host servers, infiniband switches, cabling and are fully customizable to suit individual customer needs.
Tesla GPU Preferred Partners around the world including AMAX, Appro, CADNetwork, Colfax, Cray, Depo, FluiDyna, HPC Technologies, Inspur, JRTI, Megware, Microway, NetWeb, PCPC, Penguin, Silicon Mechanics, Sprinx, T-Platforms, Viglen and Xenon are offering Tesla Preconfigured Clusters today.
