On Monday during the SC13 conference in Denver Colorado, Nvidia introduced what it calls the world's highest performance accelerator ever built, the Tesla K40 GPU accelerator, for supercomputing and high performance computing (HPC) applications. This new entry provides double the memory and up to 40 percent higher performance than its predecessor, the Tesla K20X GPU accelerator. The K40 also provides 10 times higher performance than today's fastest CPU.
"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."
The Tesla K40 is based on Nvidia's Kepler graphics processing architecture. Compared to the previous generation Tesla K20X accelerator, the K40 has 2880 CUDA cores at a base clock of 745 MHz with GPU Boost clocks of 810 and 875 MHz; whereas the K20X has 2688 cores and a base clock of 732 MHz. The K40 also uses a faster 3 GHz GDDR5 memory interface, enabling 12 GB on-board compared to the 6 GB used on the previous generation.
"The Tesla K40 GPU accelerator surpasses all other accelerators on two common measures of computational performance: 4.29 teraflops single-precision and 1.43 teraflops double-precision peak floating point performance," reads the company's press release.
One of the first Nvidia partners to take advantage of the new K40 is the Texas Advanced Computing Center (TACC) at the University of Texas at Austin. This center plans to deploy "Maverick", a new interactive, remote visualization and data analysis system powered by Tesla K40 GPU accelerators. Maverick is expected to be fully operational January 2014.
"The Tesla K40 GPU accelerators will help researchers crunch through massive volumes of big data and gain new insights through large-scale, sophisticated visualizations," said Kelly Gaither, director of Visualization at TACC. "With Nvidia GPUs, Maverick will provide researchers powerful interactive capabilities to advance their most complex scientific challenges."
Nvidia also revealed on Monday that it plans to collaborate with IBM on GPU-accelerated versions of IBM's wide portfolio of enterprise software applications on IBM Power Systems. The company announced that this will be the first time GPU accelerator technology will move beyond the realm of supercomputing and into the heart of enterprise-scale data centers.
About the Author
Kevin Parrish is a contributing editor and writer for Tom's Hardware, Tom's Games, Tom's Guide and Tom’s IT Pro. He's also a graphic artist, CAD operator and network administrator.
Both companies also announced plans to integrate the joint-processing capabilities of Nvidia Tesla GPUs with IBM POWER processors. This partnership will make it easier and more efficient for a wider range of companies to employ a style of supercomputing hardware used primarily by the scientific and technical communities for computing tasks like space exploration, decoding the human genome and speeding new products to market.
"IBM Power Systems will fully support existing scientific, engineering and visualization applications developed with the Nvidia CUDA programming model, allowing supercomputing centers and HPC customers to immediately take advantage of groundbreaking performance advantages," reads the press release. "IBM also plans to make its Rational brand of enterprise software development tools available to supercomputing developers, making it easier for programmers to develop cutting-edge applications."
The partnership between NVIDIA and IBM builds on the August announcement of the OpenPOWER Consortium.