What is missing or unclear:
I would like to help provide a complete guide for using NVIDIA GPUs with HAMi. Today, the required information is spread across installation guides, the FAQ, and troubleshooting pages.
The guide should explain the HAMi Helm chart values needed for these choices:
Driver and Container Toolkit management should be explained separately. For example, a machine may already have a driver, while GPU Operator installs and manages the Container Toolkit. The NVIDIA GPU Operator is becoming increasingly important, yet our installation documentation still instructs users to manually install the nvidia-container-toolkit and configure containerd—I believe fewer and fewer users are doing this now.
Where it should go (existing page, or a new page):
This is where we need your feedback the most, and it just so happens that we’re currently reorganizing the document structure.
I would like to discuss the scope and location first with you, including how this content should connect to Tutorials, labs or docs. Related discussions: #689.
Why this would help:
When I introduce HAMi installation to someone, I need to share many URLs. Users then need to work out which settings apply to their environment.
A clear starting point would help users find the right configuration, follow the installation steps, and verify the result. It would also make the documentation easier to maintain by reducing repeated instructions.
What is missing or unclear:
I would like to help provide a complete guide for using NVIDIA GPUs with HAMi. Today, the required information is spread across installation guides, the FAQ, and troubleshooting pages.
The guide should explain the HAMi Helm chart values needed for these choices:
Driver and Container Toolkit management should be explained separately. For example, a machine may already have a driver, while GPU Operator installs and manages the Container Toolkit. The NVIDIA GPU Operator is becoming increasingly important, yet our installation documentation still instructs users to manually install the nvidia-container-toolkit and configure containerd—I believe fewer and fewer users are doing this now.
Where it should go (existing page, or a new page):
This is where we need your feedback the most, and it just so happens that we’re currently reorganizing the document structure.
I would like to discuss the scope and location first with you, including how this content should connect to Tutorials, labs or docs. Related discussions: #689.
Why this would help:
When I introduce HAMi installation to someone, I need to share many URLs. Users then need to work out which settings apply to their environment.
A clear starting point would help users find the right configuration, follow the installation steps, and verify the result. It would also make the documentation easier to maintain by reducing repeated instructions.