Join GitHub today
GitHub is home to over 50 million developers working together to host and review code, manage projects, and build software together.
Sign upGitHub is where the world builds software
Millions of developers and companies build, ship, and maintain their software on GitHub — the largest and most advanced development platform in the world.
Create an example of how to use nodeSelector and nodeAffinity to request specific GPU types #598
Comments
|
Some examples of NodeSelectors can now be found here: https://github.com/NVIDIA/deepops/blob/master/docs/k8s-cluster/nvidia-mig.md. We could use a few more examples of NodeAffinity with minimal CUDA versions, etc. |
Since we are planning to include the GPU Feature Discovery in a future release of DeepOps it might be useful to include a few examples of how to use GPU Node Labels to request specific GPU types, CUDA versions, etc.
I haven't found any solid documentation or examples on how to do this, so if they exist, perhaps simply linking to them would be good enough in the README, if not we should create:
This might be a better feature request for the actuo GPU Feature Discovery repo (https://github.com/NVIDIA/gpu-feature-discovery).