Nvidia to adjust to Arm's chips for supercomputers

Monday, 17. June 2019 10:07

Nvidia Corp. acknowledged on Monday during the International Supercomputing Conference in Germany that it would enable support for Arm Holdings Plc's CPUs or central processing units, expanding cooperation. The aim is to provide "extremely energy efficient" exascale supercomputers with artificial intelligence features, according to the company based in Santa Clara, California. Exascale is the category of at least one exaflop – a quintillion calculations per second or one thousand times more than at petascale systems.

Arm has headquarters in Cambridge, England. It is owned by SoftBank Group Corp. and its SoftBank Vision Fund. The firm designs microchips and licenses them, while it also builds software development solutions. Nvidia claimed the project for so-called accelerator chips would be implemented by the end of the year.

The entire portfolio of AI and high performance computing (HPC) software and libraries will be accessed including CUDA-X, GPU-accelerated AI frameworks and software development tools such as PGI compilers with OpenACC support and profilers, according to the announcement.

Related Links: Softbank Group CorporationNVIDIA Corporation
Breaking the News / IT