Re:Run CUDA, unmodified, on AMD GPUs
by selimnairb on Monday July 15, 2024 @07:28PM
(#40972385)
Patented API? I thought Google v. Oracle settled this? Making an implementation of an API spec is fair use, is it not?
Re:Run CUDA, unmodified, on AMD GPUs
by fngjdflmdflg on Monday July 15, 2024 @09:18PM
(#40972915)
My understanding is that Google v. Oracle only applies to copyright.
Well you can't patent an API so....
Re:Run CUDA, unmodified, on AMD GPUs
by fngjdflmdflg on Tuesday July 16, 2024 @01:22AM
(#40973833)
You can patent the implementation. You can't patent the API name DecodeH265Video() but you can still sue someone for implementing that function correctly.
Re:Run CUDA, unmodified, on AMD GPUs
by anticensor on Tuesday July 16, 2024 @03:40PM
(#40979595)
If there is only one way to solve a problem, there is nothing to invent, just discover, and discoveries are decidedly not patentable.
Re:Run CUDA, unmodified, on AMD GPUs
by fngjdflmdflg on Wednesday July 17, 2024 @11:50AM
(#40987202)
>If there is only one way to solve a problem
h.265 is one way, av1 is another way.
I disagree. AMD can simply not implement those APIs, similar to how game emulators implement the most used APIs first and sometimes never bother implementing obscure ones. It would only matter that NVIDIA added eg. patented APIs to CUDA if those APIs were useful. In which case AMD should have a way to do them anyway. Unless NVIDIA comes up with a new patented API which is both useful and impossible to implement in any other way, which would be bad for AMD in any event. On the other hand, if AMD start supporting CUDA and people start using AMD cards, then developers will be hesitant to use APIs that only work on NVIDIA cards. Right now they are losing billions of dollars on this. Then again they barely seem capable of supporting RocM on their cards, much less CUDA.
You have a fair point in terms of cuDNN and cuBLAS but I don't know that that kind of ToS is actually binding.