Tesla K10 hybrid architecture server
19 October 2012
Silicon Mechanics, which manufactures rackmount servers, storage, and high-performance computing hardware, is now offering the Nvidia Tesla K10 GPU accelerator in its hybrid architecture server product line. The new Kepler-based GPU is designed for high throughput of large, single-precision workloads, and is ideal for seismic processing, signal and image processing, and video analytics.
With a Silicon Mechanics hybrid architecture server equipped with the Tesla K10 GPU, applications run significantly faster, as users can process large datasets with a massively multi-threaded computing architecture. The accelerator board features two GK104 GPUs and delivers up to twice the performance for single-precision applications compared to the previous generation Fermi-based Tesla M2090, in the same power envelope. With 4.58 teraflops peak single-precision performance and 320 gigabytes per second memory bandwidth, the Tesla K10 is intended for workloads where data reliability and overall performance are critical.