Both PyTorch and TensorFlow support ROCm, but there are still bugs that prevent ... | Hacker News

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

		bitL on Oct 28, 2020 \| parent \| context \| favorite \| on: AMD Reveals the Radeon RX 6000 Series, Coming Nove... Both PyTorch and TensorFlow support ROCm, but there are still bugs that prevent running all CUDA code on AMD. Now that AMD has money, they can ramp up SW development there.

dogma1138 on Oct 28, 2020 [–]

The issue is that AMD doesn't support ROCm on it's newer consumer GPUs. It also doesn't want to release a ROCm compatible driver for Windows, and you can't run ROCm in WSL/WSL2.

Direct ML is still in its infancy and is quite slow right now, AMD really needs to step up their game if they want to compete.

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