The Tesla K20X GPU accelerator is said to provide a tenfold application speed-up when paired with leading CPUs and the card provides 3.95 teraflops single-precision and 1.31 teraflops double-precision peak floating-point performance. The implementation adds Dynamic Parallelism capability that enables GPU threads to automatically spawn new threads and Hyper-Q that enables multiple CPU cores to simultaneously use the cores on a single Kepler GPU.
The K20X has 2688 CUDA cores, 6 Gbytes of GDDR5 memory, and 250 GBytes/sec off board memory bandwidth. The family also includes the Tesla K20 accelerator, with 3.52 teraflops of single-precision and 1.17 teraflops of double-precision peak performance. Tesla K20 and K20X both feature a single GK110 Kepler GPU with the Dynamic Parallelism and Hyper-Q features. The K20X takes about 235 W. (Price not given — available now.)
By Jim Harrison
NVIDIA , Santa Clara , CA
Sales 408-486-2000
www.nvidia.com
Advertisement
Learn more about NVIDIA
