GPU Node

View GPU nodes

Click the GPU Node menu in the left navigation bar to access the GPU Node page, as shown below:

The GPU Node page is divided into three sections:

1. Cluster Selection Area: Select the desired cluster.
2. Query Area: Filter GPU nodes using various criteria.
3. List Area: Displays information about GPU nodes, their owners, configurations, and utilization metrics.

After selecting a cluster, the List Area will show GPU nodes that match the specified criteria within the cluster, as illustrated below:

The List Area is divided into five columns:

1. Node: Displays the GPU node name, instance ID, and region.
2. Owner: Shows the department, group, and assigned personnel responsible for the node.
3. Settings: Displays the instance type, number of CPU cores, memory size, number of GPUs, and GPU type for the node.
4. Details: Provides strategy-related information. For instance, in low-utilization strategies, average CPU and memory utilization are compared against platform thresholds to assist users in optimizing node instance types.
5. Actions: Provides options for managing the node.

Adjust node

Click the Adjust link in the Action Column to open the Adjust Node dialog box, as shown below:

Note: Currently, only manual operations for adjusting GPU nodes are supported. Automated actions will be available in future releases, allowing for automatic adjustments once user confirmation is received.