Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

The big difference is hardware access.

I used to do VFIO with hardware passthrough so I could have linux but still run windows software like CAD that takes advantage of the gfx card. That was a pain to set up and use.

The other way, its very simple. WSL2 can run ML tasks with just a tiny bit of overhead in moving the data to the card.



PyTorch and most of other ML stuff have native Windows ports.


Yes but the point is to not do dev on windows, because window "terminal" sucks




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: