GPU Benchmark

AccelerEyes Releases GPU Benchmarking Suite

A suite of tests for CPU vs GPU comparisons

ATLANTA, Nov. 17 — AccelerEyes is pleased to announce the release of the HPC industry’s first GPU performance benchmark: Jacket GBENCH. GBENCH allows users to gauge the GPU performance of their computer relative to equivalent benchmarks obtained from a variety of other computers, including the CPU of the same computer. Benchmarks include six different tasks, common to the technical computing community:

LU – LU decomposition of 1024 x 1024 matrix

FFT – FFT of a 2^20 x 1 vector

BLAS – Matrix multiplication of two 1024×1024 matrices

3D-CONV - Convolution of 64x64x64 matrix with 3x3x3 kernel

Loops - Multiplication of 1024x1024x32 matrix with a vector

Equations - Solution of a system of 1024 equations

GBENCH is available for Windows (32 and 64 bit platforms) and Linux (32 and 64 bit platforms), freely downloadable from the following link: http://www.accelereyes.com/gbench.

GBENCH feedback forums are also available and can be used to post benchmark results and feedback for on the GBENCH test suite: http://www.accelereyes.com/forums/viewforum.php?f=12.

The benchmark will be released in conjunction with start of Supercomputing 2009 (SC09) conference in Portland, Ore., Nov. 14-19, 2009.

“The introduction of “Jacket GBENCH” brings a much needed GPU benchmark to the industry”, said Jeff Fettig, Director of Sales for James River Technical, Inc.,an AccelerEyes reseller and HPC Solution provider specializing in GPU Accelerated Computing Solutions. “Much like LINPACK, GBENCH establishes an industry standard for measuring the incredible computational horsepower of GPU accelerated solutions like our 8 GPU Velocity Micro VSC455 V8.”

About AccelerEyes

AccelerEyes launched in 2007 to commercialize Jacket, the first software platform to deliver productivity in GPU computing. With advanced language processing and runtime technology to transform CPU applications to high performance GPU codes, Jacket extends from desktop workstation performance to also fully leverage GPU clusters. Based in Atlanta, GA., the privately held company markets Jacket for a range of defense, intelligence, biomedical, financial, research, and academic applications. Additional information is available at www.accelereyes.com.

—–
Źródło: http://www.hpcwire.com/offthewire/AccelerEyes-Releases-GPU-Benchmarking-Suite-70431137.html
Source: AccelerEyes