Nvidia launches Tesla "personal supercomputing" GPU

Stan Beer
Thursday, 21 June 2007 08:17

Your IT - Home IT

Graphics chip maker Nvidia has launched a new high performance computing processor series based on its graphics processing unit (GPU) technology which the company claims will make parallel computing power more pervasive and affordable.

The Nvidia Tesla family of GPU computing products span PCs to large scale server clusters and are aimed at the physical sciences, geophysical and biosciences markets. Nvidia claims that the Tesla family of GPU products will transorm workstations into personal supercomputers.

Key elements of the new system are:

    * A multi-threaded architecture with a 128-processor computing core
    * A C-language development environment for the GPU
    * A suite of developer tools (C-compiler, debugger, performance profiler, optimized libraries)

The Tesla family of products includes:

    * Nvidia Tesla GPU Computing Processor, a dedicated computing board that scales to multiple Tesla GPUs inside a single PC or workstation. The Tesla GPU 
features 128 parallel processors, and delivers up to 518 gigaflops of parallel computation. The GPU Computing processor can be used in existing systems partnered with high-performance CPUs.

    * Nvidia Tesla Deskside Supercomputer, a scalable computing system that includes two Nvidia Tesla GPUs and attaches to a PC or workstation through an industry-standard PCI-Express connection. With multiple deskside systems, a standard PC or workstation is transformed into a personal supercomputer,  delivering up to 8 teraflops of compute power to the desktop.

    * Nvidia Tesla GPU Computing Server, a 1U server housing up to eight Nvidia Tesla GPUs, containing more than 1000 parallel processors that add teraflops of parallel processing to clusters. The Tesla GPU Server is the first server system of its kind to bring GPU computing to the datacenter.

By comparison, the world's most powerful supercomputer, the IBM BlueGene/L system at DOE’s Lawrence Livermore National Laboratory has performance of 280.6 teraflops.

Software developers will be able to build high performance computing applications using the Nvidia CUDA C development environment, which includes a C-compiler for the GPU, debugger/profiler, dedicated driver, and standard libraries. CUDA aims to simplify parallel computing development by using standard C to create programs that process large quantities of data in parallel. Nvidia CUDA is currently supported on Linux and Windows XP.

Scientists and academics have appeared to be willing to lend their names in publicly supporting the new Nvidia platform.

"Many of the molecular structures we analyze are so large that they can take weeks of processing time to run the calculations required for their physical simulation," said John Stone, senior research programmer at the University of Illinois Urbana-Champaign. "Nvidia's GPU computing technology has given us a 100-fold increase in some of our programs, and this is on desktop machines where previously we would have had to run these calculations to a cluster."