News for hackers. Stuff that matters.

A comment on Llama.cpp: Full CUDA GPU Acceleration. Parent comment

Re:Llama.cpp: Full CUDA GPU Acceleration by rapsey on Tuesday June 13, 2023 @04:23AM (#36307226)
If AMD was software competent the frameworks would support their drivers just as well. No one wants a monopoly.
[ Parent | on HN ]
Re:Llama.cpp: Full CUDA GPU Acceleration by marcyb5st on Tuesday June 13, 2023 @04:30AM (#36307302)
Agreed. Sorry if I gave the impression of being pro-nVidia. I am not.

But the reality is that when Tensorflow and Pytorch came to be there was no alternative. Now you need to jump through hoops to make it work with non CUDA hardware.

Additionally, while drivers play a role, I think the main difference is in the computing libraries (CUDA vs ROCm)

[ Parent | on HN ]