News for hackers. Stuff that matters.

A comment on An Interview with Nvidia CEO Jensen Huang About AI’s iPhone Moment. Parent comment

Re:An Interview with Nvidia CEO Jensen Huang About AI’s iPho... by CuriouslyC on Saturday March 25, 2023 @09:38PM (#35309736)
Except that CUDA is low level, so it's not hard to shim above it and write interoperable code. There are too many players who don't want to play the Nvidia tax, this will play out like OpenGL vs Direct3d in reverse.
[ Parent | on HN ]
Re:An Interview with Nvidia CEO Jensen Huang About AI’s iPho... by ethbr0 on Saturday March 25, 2023 @10:12PM (#35309974)
> this will play out like OpenGL vs Direct3d in reverse

Is that also like OpenCL vs CUDA in reverse?

[ Parent | on HN ]
Re:An Interview with Nvidia CEO Jensen Huang About AI’s iPho... by WanderPanda on Saturday March 25, 2023 @11:29PM (#35310518)
I feel like abstractions don't work if we want to get the maximum performance. Afaik Tensor cores are not usable from opencl and on the other hand even in the CUDA universe cuBLAS (hand optimized) seems to outperform cutlass (using abstractions)
[ Parent | on HN ]