maxkraft
About
- Username
- maxkraft
- Joined
- Visits
- 8
- Last Active
- Roles
- member
- Points
- 7
- Badges
- 0
- Posts
- 6
Reactions
-
Apple Silicon machine learning code may become more easily portable to Nvidia hardware
I can see some use for this. But not a lot of good ones. If you want good CUDA code. Write it on a system that then runs it on an Nvidia GPU. Porting it from a non native system makes no sense. I am also sure you could just run your CUDA code remotely on a system with an Nvidia chip. The only good reason to do this is you like run stuff locally on your Mac and only sometimes need to export and run it on an Nvidia Chip.
