←back to thread

AMD funded a drop-in CUDA implementation built on ROCm: It's now open-source

(www.phoronix.com)
1045 points mfiguiere | 3 comments | 12 Feb 24 14:00 UTC | HN request time: 0.908s | source
Show context
wheybags ◴[12 Feb 24 14:33 UTC] No.39345186[source]▶
Cannot understand why AMD would stop funding this. It seems like this should have a whole team allocated to it.
replies(1): >>39345254 #
otoburb ◴[12 Feb 24 14:41 UTC] No.39345254[source]▶
They would always be at the mercy of NVIDIA's API. Without knowing the inner workings, perhaps a major concern with this approach is the need to implement on NVIDIA's schedule instead of AMD's which is a very reactive stance.

This approach actually would make sense if AMD felt, like most of us perhaps, that the NVIDIA ecosystem is too entrenched, but perhaps they made the decision recently to discontinue funding because they (now?) feel otherwise.

replies(2): >>39345418 #>>39347490 #
blagie ◴[12 Feb 24 14:55 UTC] No.39345418[source]▶
They've been at mercy of Intel x86 APIs for a long time. Didn't kill them.

What happens here is that the original vendor loses control of the API once there are multiple implementations. That's the best possible outcome for AMD.

In either case, they have a limited window to be adopted, and that's more important. The abstraction layer here helps too. AMD code is !@#$%. If this were adopted, it makes it easier to fix things underneath. All that is a lot more important than a dream of disrupting CUDA.

replies(3): >>39345550 #>>39345891 #>>39346080 #
rubatuga ◴[12 Feb 24 15:06 UTC] No.39345550[source]▶
x86 is not the same, the courts forced the release of x86 architecture to AMD during an antitrust lawsuit
replies(2): >>39345691 #>>39345787 #
anon291 ◴[12 Feb 24 15:16 UTC] No.39345691[source]▶
You don't think the courts would force the opening of CUDA? Didn't a court already rule that API cannot be patented. I believe it was a Google case. As long as no implementation was stolen, the API itself is not able to be copyrighted.

Here it is: https://arstechnica.com/tech-policy/2021/04/how-the-supreme-...

replies(2): >>39345827 #>>39346343 #
1. Symmetry ◴[12 Feb 24 15:24 UTC] No.39345827[source]▶
Regardless of the legal status of APIs, this Phoronix article is about AMD providing a replacement ABI and I wouldn't assume the legal issues are necessarily the same. But because this is a case where AMD is following a software target there's the possibility, if AMD starts to succeed, that NVidia might change their ABI in ways that deliberatly hurt AMD's compatibility efforts in ways that would be much more difficult for APIs or hardware. That's, presumably, why AMD is going forward with their API emulation effort instead.
replies(1): >>39345855 #
2. anon291 ◴[12 Feb 24 15:26 UTC] No.39345855[source]▶
If you read the article, it's about Google's re-implementation of the Java API and runtime. Thus, yes, Google was providing both API and ABI compatibility.
replies(1): >>39346214 #
3. Symmetry ◴[12 Feb 24 15:50 UTC] No.39346214[source]▶
I read the article when it came out and re-scimmed it just now. My understanding at the time and still was that the legal case revolved around the API and the exhibits entered into evidence I saw were all Java function names with their arguments and things of that sort. And I'm given to understand that the Dalvik Java implementation Google was using with Android was register based rather than than the stack based standard Java, which sounds to me like it would make actual binary compatibility impossible.