-
Notifications
You must be signed in to change notification settings - Fork 16
kuberesource: template GPU per platform #2039
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
Since the `gpu` just deploy target relies on the resourcegen tool, the verdict made in [1] was false. The GPU target is only used in development, so hardcoding the platforms should be fine. [1]: #2030 (comment)
burgerdev
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The other commit lgtm. Maybe it makes sense to separate them?
| case platforms.MetalQEMUTDXGPU: | ||
| deviceURI = "nvidia.com/GB100_B200" | ||
| case platforms.MetalQEMUSNPGPU: | ||
| deviceURI = "nvidia.com/GH100_H100_PCIE" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I still believe that this is not the right place to configure this. Rather, this should be modelled similar to the tcb-specs.json and be an input to resourcegen.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
How would that work with the just targets? Not saying that it isn't possible, I just wonder what the dataflow would look like.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
While not ideal, we could configure this with a justfile.env var for now, since we're already doing it for the nodeinstaller target config:
Lines 338 to 339 in fb4de83
| sed -i 's/^default_platform=.*/default_platform="Metal-QEMU-TDX"/' justfile.env | |
| sed -i 's/^node_installer_target_conf_type=.*/node_installer_target_conf_type="k3s"/' justfile.env |
Eventually, I'd prefer this to be configured in the cluster instead of in the repo, though.
00491da to
b5c9064
Compare
|
Obsolete with #2070 |
Since the
gpujust deploy target relies on the resourcegen tool, the verdict made in 1 was false. The GPU target is only used in development, so hardcoding the platforms should be fine.