nccl default for backend=any (#2528)

* nccl default for backend=any

* check num gpus + ensure row contiguous for all reduce

* comment
This commit is contained in:
Awni Hannun
2025-08-22 12:24:27 -07:00
committed by GitHub
parent 5722c147de
commit 068a4612e9
5 changed files with 68 additions and 31 deletions

View File

@@ -405,6 +405,7 @@ jobs:
sudo dpkg -i cuda-keyring_1.1-1_all.deb
sudo apt-get update
sudo apt-get install cuda-toolkit-12-9 libcudnn9-dev-cuda-12
sudo apt-get install libnccl2 libnccl-dev
sudo apt-get install libblas-dev liblapack-dev liblapacke-dev
sudo apt-get install zip
pip install auditwheel