Nvidia: New Tesla P100 GPU will be available in servers in 2017

During the course of a keynote address at the GPU Technology Conference in San Jose, California on Tuesday, chip maker Nvidia’s CEO Jen-Hsun Huang said that servers equipped with the company’s fastest GPU -- the Tesla P100 GPU -- will start shipping in 2017.

Announcing the imminent availability of its new Tesla P100 GPU in servers next year, Huang said that orders for the servers will be opened by Dell, Hewlett Packard Enterprise, IBM and Cray, in the fourth quarter of this year.

According to the details shared by Huang, the shipments of the Tesla P100 GPU-equipped servers will start in the 2017 first quarter.

Huang also revealed that the Tesla P100 GPU will also be shipped to companies which design hyperscale servers in-house; and, subsequently, to outsourced manufacturing shops. The GPU will also be available for in-house "cloud servers" by 2017-end.

Touting the new Tesla P100 GPU as “the largest chip that has ever been made,” Huang said that the GPU – based on Nvidia’s new Pascal architecture – has essentially been designed for use in corporate data centers, and has the capability to achieve very high performance; thanks to the packing of 15 billion transistors on a piece of silicon.