HPC veteran and former Cray Research alum Steve Oberlin has joined NVIDIA to help lead the graphic chipmakers’s high-performance computing initiative. Oberlin’s new role as chief technology officer for NVIDIA’s Tesla business unit puts him in charge of NVIDIA’s Tesla roadmap and architecture.
“Steve has decades of experience in high performance computing, going back to his early contributions to Cray’s vector and parallel supercomputers,” observed Bill Dally, NVIDIA’s chief scientist, in a company blog post. “His extensive background and forward-looking vision will be of great value as he works to refine our HPC strategy and helps execute our roadmap to achieve exascale computing in the years ahead.”
Early on in his career, Oberlin worked at Cray Research with the father of supercomputing, Seymour Cray. During his two decade tenure, Oberlin held many positions including vice president of hardware and chief architect responsible for Cray’s first massively parallel processors.
More recently, Oberlin served as a distinguished engineer at CA Technologies, co-founder and chief scientist at Cassatt Corporation and vice president of software at Silicon Graphics. He holds more than a dozen US patents.
Tesla GPUs are NVIDIA’s most performant parts. Over the last few years, these highly parallel accelerator chips have become a fixture in some of the world’s most powerful supercomputers. According to the latest TOP500 list released this week at SC13, fifty three systems derive their FLOPS from a combination of GPUs and CPUs. Of those, thirty eight used Tesla graphics chips, two employed AMD GPUs, and thirteen were equipped with Intel’s Xeon Phi coprocessors.
NVIDIA also used the show as a launch pad for its newest GPU accelerator, called the Tesla K40, which has more processing capacity and more memory than the current K20X accelerator.