Re:Run CUDA, unmodified, on AMD GPUs
by jokethrowaway on Monday July 15, 2024 @08:50PM
(#40972780)
a lot of ML researchers stay pretty high level and reinstall conda when things stop working
and rightly so, they have more complicated issues to tackle
It's on developers to provide better infrastructure and solve these challenges
Not rightly. It'd be faster on the long term to address the issues.
Currently nobody think that long term. They just reinstall, that’s it.
I'm not really talking about companies here for the most part, I'm talking about academic ML researchers (or industry researchers whose role is primarily academic-style research). In companies there is more incentive for good software engineering practices.
I'm also speaking from personal experience: I once had to hand-write my own CUDA kernels (on official NVIDIA cards, not even this weird translation layer): it was useful and I figured it out, but everything was constantly breaking at first.
It was a drag on productivity and more importantly, it made it too difficult for other people to run my code (which means they are less likely to cite my work).
I don't think so. I agree it is too hard for the ML researches at the companies which will have their rear ends handed to them by the other companies whose ML researchers can be bothered to follow a blog post and prompt ChatGPT to resolve error messages.