Nvidia on Monday took the wraps off of its Tesla K40 GPU accelerator, supposedly the world's highest performance accelerator ever built. The card is intended for extreme performance applications in the fields of scientific research, engineering, high performance computing (HPC), and enterprise applications. For heavy duty tasks, the Tesla K40 GPU boasts twice as much memory as its predecessor (Tesla K20x) and up to 40 percent higher performance, Nvidia says.
It also features 10 times higher performance than today's fastest CPU. Such is the strength of Nvidia' Kepler's compute architecture. In this case, the Tesla K40 offers 4.29 teraflops of single-precision and 1.43 teraflops of double-precision peak floating point performance.
"GPU accelerators have gone mainstream in the HPC and supercomputing industries, enabling engineers and researchers to consistently drive innovation and scientific discovery," said Sumit Gupta, general manager of Tesla Accelerated Computing products at Nvidia. "With the breakthrough performance and higher memory capacity of the Tesla K40 GPU, enterprise customers can quickly crunch through massive volumes of data generated by their big data analytics applications."
Nvidia equipped the card with 12GB of GDDR5 memory and 2,880 CUDA parallel processing cores; memory bandwidth is rated at 288GB/s. On the software side, Nvidia said that there are over 240 professional software applications that take advantage of GPU acceleration.
The Tesla K40 is available now and in the coming months from a variety of server manufacturers, Nvidia says.