Supergroup: Nvidia Announces Support for Arm Processors

Credit: ShutterstockCredit: Shutterstock

Nvidia made a slew of supercomputer-related announcements today at the International Supercomputing Conference in Germany. They announced that Japan's fastest supercomputer adopted its Nvidia GPU Cloud offering, that it developed the world's 22nd-fastest supercomputer to help it make self-driving cars and plans to add support for Arm processors to its software by the end of the year.

That last announcement could be the most important. The company said it's bringing the "full stack" of artificial intelligence (AI) and high-performance computing (HPC) software, which "accelerates more than 600 HPC applications and all AI frameworks" to the Arm ecosystem. That means Nvidia's effectively making its software hardware-agnostic by letting supercomputer makers use x86 offerings from Intel and AMD or Arm processors at their discretion.

“Supercomputers are the essential instruments of scientific discovery, and achieving exascale supercomputing will dramatically expand the frontier of human knowledge,"  Nvidia CEO Jensen Huang said in the company's announcement. "As traditional compute scaling ends, power will limit all supercomputers. The combination of Nvidia's CUDA-accelerated computing and Arm’s energy-efficient CPU architecture will give the HPC community a boost to exascale.”

That may sound like a bunch of gobbledygook, but it signals Nvidia's intention of competing against Intel and AMD--as well as benefiting from their success--when it comes to exascale supercomputers. Intel announced in March that its Xe graphics would power the first exascale supercomputer; AMD followed that announcement in May by saying its chips would power the fastest exascale supercomputer.

What's an Exascale Computer?

Exascale computers are devices capable of processing at least one exaFLOP per second. That's a quintillion -- or a billion billion -- operations. That much power requires advanced hardware, sure, but it also relies on software designed specifically to handle such a massive undertaking. Nvidia wants to offer both via its GPUs and the software it's expanding to Arm.

Nvidia said that software includes "all Nvidia CUDA-X AI and HPC libraries, GPU-accelerated AI frameworks and software development tools such as PGI compilers with OpenACC support and profilers." The breadth of the company's software offering combined with the performance offered by its GPUs has reportedly led Nvidia to power 22 of the world's 25 most energy-efficient supercomputers.

12 comments
    Your comment
  • setx
    Is anyone even interested in ARMs for supercomputers?
    Quick search in Top500 shows only 1 ARM entry on 156'th place.
  • remixislandmusic
    I dont known why you would want arm and nvidia. ARM processors are generally not going to be able to utilize the graphics to do much
  • bit_user
    Quote:
    Nvidia's effectively making its software hardware-agnostic by letting supercomputer makers use x86 offerings from Intel and AMD or Arm processors at their discretion.

    Hmmm... what about POWER? Not long ago, Nvidia had partnership with IBM, whose latest POWER CPUs are the first (and AFAIK only) CPU to natively implement NVLink.

    Quote:
    "As traditional compute scaling ends, power will limit all supercomputers. The combination of Nvidia's CUDA-accelerated computing and Arm’s energy-efficient CPU architecture will give the HPC community a boost to exascale.” That may sound like a bunch of gobbledygook

    a bunch of Gobbledygook? Are you trying to write for like 6th-graders? It's fine if you want to break it down, but maybe don't demean readers.

    And as for breaking it down, you skipped what they meant by "traditional compute scaling" - the idea that legacy CPU architectures will naturally get more efficient over time.