|SC07 SuperComputing Coverage - NVIDIA|
|News - Featured Website News|
|Written by Olin Coles|
|Tuesday, 13 November 2007|
On day number one of SC07, the SuperComputing exposition hall was officially opened to the media and select visitors for open coverage. Many of those in attendance gravitated to some of the largest names in the high performance computing industry such as Sun Microsystems, Microsoft, and Intel. However Benchmark Reviews spent some quality time at a few exhibits by manufacturers with roots to the enthusiast segment. In this article we brief you on the Tesla project from NVIDIA.
The CPU and operating system powering the modern PC solve an incredibly difficult problem in computing. As you use the computer, the operating system tracks all your activities, communicates in the background, and organizes the information you use while you're listening to music, browsing the Web, and reading e-mail. Even though the CPU works on separate tasks one at a time, it has enough speed so these serial tasks appear to operate simultaneously. With new multi-core CPUs, each core can handle an additional task with true simultaneity. A different class of computing problem, parallel computing, has until recently remained the realm of large server clusters and exotic supercomputers. Standard CPU architecture excels at managing many discrete tasks, but is not particularly efficient at processing tasks that can be divided into many smaller elements and analyzed in parallel. This is exactly the type of problem solved by graphics processing units (GPUs).
The GPU has great potential for solving such problems quickly and inexpensively. GPU computing makes supercomputing possible with any PC or workstation and expands the power of server clusters to solve problems that were previously not possible with existing CPU clusters. The goal of computing with GPUs is to apply the tremendous computational power inherent in the GPU to solve some of the most difficult and important problems in high performance computing.
Revolutionary NVIDIA Tesla high performance computing (HPC) solutions arm scientists, engineers and other technical professionals with the power to solve previously unsolvable problems. A dedicated, high performance GPU computing solution, Tesla brings supercomputing power to any workstation or server and to standard, CPU-based server clusters.
Key elements include:
Power to Solve Complex Problems
Massively Parallel Performance
C for the GPU