Quadro K2000 |
July 2013 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The Quadro K2000 is nVidia’s latest iteration in the middle-range CAD and digital content creation graphics cards. It provides a rounded 30% increase in performance in the same platform. This performance improvement can be seen in the improved specifications of the K2000 which features many more CUDA cores and an improved memory bandwidth.
We built a Compucon Superhawk system for testing K2000: Z77
motherboard; Intel i3-3220 CPU; 2x 2GB DDR3 RAM modules; Windows 8; and
nVidia Quadro driver version 311.15. This model was the platform for some tests
previously done on Quadro 2000, and we were able to make a direct comparison of the
speed of the nVidia graphics hardware. The first set of tests are the SPECviewperf
11 suite of CAD benchmarks, done at a resolution of 1280x960 with 8x FSAA.
Results are in the table below. The Quadro K2000 shows an all-round improvement
with the exception of LightWave, where the test frame rate may have saturated.
The next set of tests which we performed is an assessment of the card’s CUDA processing ability. To do this we use the CUDA SDK N-Body simulation as a benchmark of the processing power of the card. We used the Superhawk-based platform previously mentioned for the SPECviewperf benchmarks. Using this tool we measured the time required to perform a simulation of 10,000 bodies with single precision (SP) and double precision (DP), this produces time in the order of tens to thousands of milliseconds. The graph expresses the time on a 10-logarithmic scale to enable the breadth of values to be witnessed. Hover over image to enlarge.
For Compucon Design Station information and pricing, please contact us. |