News for hackers. Stuff that matters.

A comment on Why Nvidia Keeps Winning: The Rise of an AI Giant. Parent comment

Re:Why Nvidia Keeps Winning: The Rise of an AI Giant by thou4o2i34u234 on Friday July 07, 2023 @03:42AM (#36628503)
OpenCL is deprecated.

Apple has ditched it for metal. AMD has ditched it for ROCm which is just a AMD compiler for CUDA PTX.

[ Parent | on HN ]
Re:Why Nvidia Keeps Winning: The Rise of an AI Giant by synergy20 on Friday July 07, 2023 @09:55AM (#36631811)
it's not, Nvidia just added support to OpenCL 3.0, AMD also continues to support it. Apple always does things its own way.
[ Parent | on HN ]
Re:Why Nvidia Keeps Winning: The Rise of an AI Giant by pjmlp on Friday July 07, 2023 @10:16AM (#36632149)
OpenCL 3.0 is basically OpenCL 1.0 rebranded, hence why NVidia didn't had any issue adding support for what they already had anyway.
[ Parent | on HN ]
Re:Why Nvidia Keeps Winning: The Rise of an AI Giant by zorgmonkey on Friday July 07, 2023 @10:42PM (#36640864)
ROCm does not compile PTX IR to AMD assembly, if it did you'd be able to run nearly any compiled CUDA program on and AMD GPU without the source. ROCm is a source level compiler for CUDA (technically HIP is the actual compiler, but whatever), it allows you to a substantial fraction of the CUDA APIs. This notably also means that any Nvidia open source library that uses inline PTX assembly won't work, but fortunately AMD does have alternatives to many of the Nvidia libraries.
[ Parent | on HN ]