Tencent Technology News, June 17 news, according to foreign media reports, chip giant Nvidia has just announced that it will use its chips and ARM's processor to build supercomputers, further promoting NVIDIA to develop systems for climate change prediction and nuclear weapon modeling .
For a long time, Nvidia has always been known for providing graphics chips for personal computers (PCs) to make video games look more realistic, but researchers are now using NVIDIA chips in data centers to speed up artificial intelligence (AI) computing. Speed, such as training a computer to recognize an image.
To this end, NVIDIA's so-called accelerator chip has always worked with the central processors of companies such as Intel and IBM. At the Supercomputing Conference in Germany on Monday, Nvidia announced that its accelerator chip will be used with ARM processors by the end of this year.
ARM has been acquired by SoftBank, which provides the underlying processor technology for most mobile phone chips. However, companies such as Ampere Computing, formerly known as Intel's former president, have been working hard to apply these chips to data centers. In this field, Intel's chips have always dominated.
But the difference between an ARM processor and an Intel or IBM chip is that ARM itself does not manufacture chips. Instead, it authorizes the underlying technology so that others can make their own chips.
Ian ·, vice president of accelerated computing at NVIDIA; Ian Buck said that from a technical point of view, building a supercomputer with an ARM processor would be a “difficult task”.
But he said that NVIDIA did this because European and Japanese researchers wanted to use ARM technology to develop supercomputing chips, which actually provided them with a third option beyond IBM and Intel. Have more control.
When Buck mentioned ARM's technology in an interview, he said: "The openness makes ARM's technology more attractive. More interestingly, the reason we announced support for ARM is that it provides an open architecture for supercomputing. ”
After Nvidia spent $6.8 billion to acquire Israeli company Mellanox Technologies, it began working with ARM in the supercomputer field. Mellanox manufactures high-speed network chips that help stitch many smaller computers into large computers. The shadow of the company's chips can also be found in many of the world's most powerful supercomputers. (Tencent Technology Review / Jinlu)