Nvidia has hired former IBM Power Systems Chief Engineer Steve Fields to work on system architecture and product road map for the chipmaker’s data center systems.
The Santa Clara, Calif.-based company confirmed to CRN Thursday that it hired Fields, who worked at IBM for 30 years, as vice president of data center system architecture in November, underscoring the chipmaker’s continued transformation into what Nvidia CEO Jensen Huang has described as a “data center-scale company” and “full-stack computing company.”
Fields shared the job update on LinkedIn earlier this week, saying that the new opportunity at Nvidia brought him out of retirement after ending his tenure at IBM in April.
“In the end I wanted to create some new work memories, make some new friends, and work with some new technologies, so last week I started a new job with NVIDIA,” he said in a LinkedIn post. “Lots to learn but looking forward to having a new kind of fun.”
Fields was critical to the development of IBM’s Power Systems servers over the last several years. In 2015, he was named an IBM fellow, the company’s highest technical honor, for his leadership on the architecture and design of Power Systems. At the time, he was focused on evolving the Power Systems platform for cloud computing, analytics and other emerging workloads.
The Armonk, N.Y.-based company said Fields was also “essential” to the success of IBM’s Power processors in the UNIX and Linux server markers.
“His ability to solve problems that span multiple hardware and software disciplines has contributed to the improved performance and quality of IBM Power Systems,” IBM said in a 2015 press release. “He‘s also a key contributor to IBM’s OpenPOWER initiative, which has opened the chip and system architectures for partners to build products on.”
The hiring of Fields happened as Nvidia continues to gain traction with its data center systems products, which includes the newly announced DGX Station A100 that packs four A100 GPUs into a workstation for AI applications. The company’s DGX A100 system serves as the basis for its DGX SuperPod high-performance computing clusters, two of which are among the world’s 500 fastest supercomputers, with Nvidia’s Selene cluster currently in the No. 5 spot.
Dominic Daninger, vice president of engineering at Burnsville, Minn.-based Nor-Tech, an Nvidia high-performance computing system builder partner, said the hiring of Fields reflects the importance of the data center market to Nvidia, which is becoming increasingly vertically integrated with its systems approach, its acquisition of Mellanox and its pending acquisition of Arm.
“It can tie a customer in more tightly to them,” Daninger said. “It gets more and more difficult for a customer to move away from that.”
With Fields’ experience in both silicon and systems, the veteran engineer will help Nvidia continue down that path in more tightly integrating products into systems, according to Daninger.
“If you‘re looking at it from that standpoint, it makes a lot of sense,” he said.