docs: Add section on viewing topology #8638
Conversation
✅ Deploy Preview for determined-ui ready!
To edit notification comments on pull requests, go to your Netlify site configuration. |
docs/tools/webui-if.rst
Outdated
| ************************** | ||
|
|
||
| To view a resource pool's node and GPU distribution, as well as check which GPUs are currently in | ||
| use, start by ensuring there's an active experiment running. Then, follow these steps: |
There was a problem hiding this comment.
Does there have to be an active experiment running? Having that will ensure that there is some compute resource active and available but if you're in an on-prem situation or if your autoscaler hasn't scaled down your instances yet, you can still view topology whether tasks are running or not.
There was a problem hiding this comment.
resolved by removing the phrase about ensuring there's an active experiment running
docs/tools/webui-if.rst
Outdated
|
|
||
| #. View the Topology. | ||
|
|
||
| Under the **compute-pool** section, select the **Active slots** hyperlink to access the topology |
There was a problem hiding this comment.
You don't need to select anything, whether you're in Active slots or Queued slots or any other view within the resource pool the topology will persist.
There was a problem hiding this comment.
having two ways to view the topology might be confusing to users: they can select a resource pool but there is also a hyperlink. i think the hyperlink is obvious while selecting a resource pool is less obvious.
There was a problem hiding this comment.
resolved by avoiding mention of the hyperlink
docs/tools/webui-if.rst
Outdated
| Viewing Cluster Topology | ||
| ************************** | ||
|
|
||
| To view a resource pool's node and GPU distribution, as well as check which GPUs are currently in |
There was a problem hiding this comment.
Technically it doesn't tell you which GPUs are in use, just how many
tara-det-ai
force-pushed
the
docs/Add-section-on-viewing-topology
branch
from
97ed121
to
b0be950
Compare
## Description TECHWR-369 As an MLE, I would like to have a macro understanding of how the GPUs and nodes in my cluster are distributed within Determined and which slots on which GPUs are occupied, enabling me to know if my job will run and/or if there are sufficient resources for it to do so. To visualize each node and the number of slots available and which slots are active vs used, visit the Topology section in the resource pools' details page.
there is also a hyperlink but the docs will avoid mentioning this in favor of just selecting a resource pool to view its details
tara-det-ai
force-pushed
the
docs/Add-section-on-viewing-topology
branch
from
9076f28
to
1a02296
Compare
Ticket
TECHWR-369
Description
As an MLE, I would like to have a macro understanding of how the GPUs and nodes in my cluster are distributed within Determined and which slots on which GPUs are occupied, enabling me to know if my job will run and/or if there are sufficient resources for it to do so.
To visualize each node and the number of slots available and which slots are active vs used, visit the Topology section in the resource pools' details page.