NVIDIA Brings CUDA to Arm

Auer

Auer

[H]ard|Gawd
Joined
Nov 2, 2018
Messages
1,972
"International Supercomputing Conference -- NVIDIA today announced its support for Arm CPUs, providing the high performance computing industry a new path to build extremely energy-efficient, AI-enabled exascale supercomputers.

NVIDIA is making available to the Arm® ecosystem its full stack of AI and HPC software — which accelerates more than 600 HPC applications and all AI frameworks — by year’s end. The stack 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."

https://nvidianews.nvidia.com/news/...-enabling-new-path-to-exascale-supercomputing
 
Hmm... I bet this is a result of nVidia working on the next Tegra. Probably terms in some cross-license deal, so ARM could try to push into datacenter harder.

I don't know enough about it to be knowledgeable, but I thought CUDA and other APIs ran on the GPU, I don't know how a deal with ARM would do anything differently. It even mentions in the PR here that the libraries are GPU-accelerated.
 
I hope I'm not too off the mark here. But this allows arm. Based systems to make appropriate api calls to Nvidia hardware when before this wasn't something it could do. The api probably needed some licensed libraries to function.
 
From reading it, NVidia GPUs still do the real CUDA work, this is just ARM CPU code so they can drive NVidia GPUs.

"Once stack optimization is complete, NVIDIA will accelerate all major CPU architectures, including x86, POWER and Arm.
....
NVIDIA’s support for Arm complements our latest developments on the HPE Apollo 70, an Arm-based, purpose-built HPC system, and now, NVIDIA GPU-enabled. With the HPE Apollo 70 supporting a 2U GPU tray and multiple energy-efficient cooling options, we can further help the HPC industry address increasingly unsustainable levels of power consumption.”
"
 
You must log in or register to reply here.
Back
Top