Nvidia announced on Monday that it is collaborating with IBM on next-generation supercomputers for the corporate data center. Under the deal, IBM Power processors will make use of Nvidia's GPU technology. On Monday, Nvidia also launched its fastest-ever GPU accelerator for supercomputing and high-performance computing (HPC) applications.
What does all this mean to planners and companies working with huge amounts of data in science, engineering and technical fields? Plenty.
Nvidia's Tesla K40 GPU accelerator has been optimized for big data analytics and large-scale scientific workloads. The payoff for customers is breakthrough performance and higher memory capacity. The Tesla K40 has double the memory of the Tesla K20X, which, until Monday, has been known as Nvidia's top GPU accelerator. The Tesla K40 also shows a 40 percent performance boost over its predecessor.
One player eager to step up to the Tesla K40 is the Texas Advanced Computing Center (TACC) at the University of Texas at Austin. The center's team have their eyes on the new accelerator for their upcoming system, Maverick, targeted to be fully operational in January. Specifically, Maverick is an interactive, remote visualization and data analysis system.
Crunching Through Volumes
Kelly Gaither, director of Visualization at TACC, said the Tesla K40 GPU accelerators will help researchers crunch through massive volumes of big data and gain new insights through large-scale, sophisticated visualizations.
"GPU accelerators have gone mainstream in the HPC and supercomputing industries," said Sumit Gupta, general manager of Tesla Accelerated Computing products at Nvidia.
One attention-drawing feature of the Tesla K40 is its GPU Boost, a user-controlled performance enhancer.
Boston Limited, a company focused on performance, server, and workstation solutions that can be tailored for every client, on Monday announced support of the Nvidia Tesla K40 GPU. The company pointed to the "Boost" technology, saying it can deliver dramatic performance improvements to HPC workloads, such as AMBER for computational chemistry and ANSYS for engineering.
Also on Monday, Penguin Computing, a provider of HPC solutions, announced it will offer Tesla K40 GPU accelerators across a range of GPU computing platforms, including Relion 2808GT servers.
Boost for Data Centers
AMAX, which specializes in server and storage solutions, similarly announced on Monday its support for the new Tesla K40 accelerators. "The new accelerators enable premier system integrators like AMAX to provide cluster and server solutions with previously unattainable levels of computing power," said Nvidia's Gupta.
Meanwhile, IBM would never ignore a good thing to advance its vision of next-generation computing for big data and analytics. IBM and Nvidia jointly announced on Monday that they plan to collaborate on GPU-accelerated versions of IBM applications on IBM Power systems.
Specifically, Nvidia and IBM plan to integrate the joint-processing capabilities of Nvidia Tesla GPUs with IBM Power processors. Both companies are also referring to the move as a significant sign that corporate data center applications are in for a supercharge.
"The move marks the first time that GPU accelerator technology will move beyond the realm of supercomputing and into the heart of enterprise-scale data centers," according to the Monday announcement from Nvidia and IBM. "The collaboration aims to enable IBM customers to more rapidly process, secure and analyze massive volumes of streaming data."