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

Wouldnt this be fixed by CUDA_LAUNCH_BLOCKING=1? Or putting a bunch of torch.cuda.synchronizes in the suspected lines.


lol whoops yeah that would definitely solve the problem. I wasn't familiar with `CUDA_LAUNCH_BLOCKING` but `torch.cuda.synchronizes` does work.




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

Search: