Visit the purchase page:JD.com NVIDIA Series Product Summary
For a long time, Nvidia has been known as a supplier of personal computer graphics chips that make video games look more realistic, but today researchers are also using their chips in data centers to speed up artificial intelligence computing, such as training computer-identified images. .
To this end, NVIDIA's so-called accelerator chip is also used toIntelThe central processors of companies such as IBM work together.
On Monday, Nvidia said at its supercomputing conference in Germany that its accelerator chip will be compatible with the Arm processor by the end of this year.
The Arm of Japan Softbank Group is the majorityMobile phoneThe chip provides the underlying processor technology. But companies such as Ampree Computing, led by Intel's former president, have been trying to bring these chips to the data center. Intel's chips dominate the data center.
But the Arm processor is different from Intel or IBM chips, and Arm itself does not produce chips. Instead, it authorizes the underlying technology so that others can use it to make chips.
Ian Barker, vice president of Nvidia accelerated computing, said the project to make supercomputers with Arm would be a "heavy burden" from a technical point of view.
But he said that Nvidia is responsible for this task because European and Japanese researchers hope to use Arm's technology to develop supercomputing chips, which actually gives them a third option beyond IBM and Intel. In this regard, they can Have more control.
In an interview before the meeting, Buck talked about Arm's technology. He said: "This kind of opening makes it very attractive. The interesting part of Arm and the reason we announced support for it is that it can provide an open for supercomputing. Architecture."
Before working with Arm to develop supercomputers, Nvidia acquired Israeli company Mellanox Technologies for $6.8 billion. Mellanox manufactures high-speed network chips that connect many small computers into large computers and can be found in some of the world's most powerful supercomputers.