GPU Pods¶

Overview¶

The GPU Pods Management section provides a comprehensive guide on creating, managing, and scaling GPU Pods within the TurboAI Cloud platform. This section is designed to assist users in optimizing their AI applications by fully utilizing the capabilities of GPU resources.

Accessing GPU Pods¶

To access the GPU Pods management interface, navigate to the sidebar menu and click on GPU Pod.

Creating New GPU Pod¶

Step 1: Click the New Button¶

Click the New button on the GPU Pod page to start the creation process.

Step 2: Fill in Pod Details¶

Enter a name for your Pod and select the region resources. Choose the image type (Public or Private) and set the image pull policy (Always or IfNotPresent).

Summary¶

Before confirming, review the summary which includes details like GPU Type, GPU Memory, CPU, vCPU, Memory, Name, Instance Count, and Service Region.

Advanced Settings¶

Configure advanced settings such as Start Command, File system mount, Environment Variable, Ports & Protocol, and Lifecycle scripts.

Managing GPU Pods¶

You can manage your GPU Pods by performing various tasks such as viewing their status, analyzing logs, monitoring performance metrics, and troubleshooting.

Viewing Pod Status¶

Check the operational status of your GPU Pods, including startup status, runtime, and resource usage.

Log Analysis¶

Use log analysis tools to query logs based on keywords and time ranges.

Performance Monitoring¶

Monitor performance metrics like network traffic and request response times.

Troubleshooting¶

Use diagnostic tools to locate and resolve faults.

Viewing Pod Details¶

For detailed information about a specific Pod, click on the Detail button.

Job Info¶

Resource Information¶

Next Steps¶

For more information on how to optimize your GPU Pods, proceed to the Bare Metal section.