Activate PAI and purchase PPU resources

1. Activate PAI

Log on to the PAI console using your main account. In the upper-left corner, select a region that supports PPU resources. This topic uses China (Ulanqab) as an example. After you select the region, click Authorize and Activate Product with One Click.

2. Create a resource group and buy Lingjun resources

1. Create a resource group. In the left-side navigation pane of the PAI console, choose AI Computing Resources > Resource Pool > Lingjun Resources > Create Resource Group. Enter a Resource Group Name and click Confirm.

2. Purchase Lingjun nodes. On the resource group list page, choose Create Order > Create Subscription Order.

   Note

   Before you set the resource type to Lingjun Intelligent Computing Resources and purchase nodes, confirm the following specification and billing method limits so that you do not mistake them for a permission or region issue:

   • Each Zhenwu 810E node is physically designed with 16 PPU cards, and no separate 8-card specification exists. As a result, neither the purchase page nor the DLC job creation page lists an 8-card Zhenwu 810E specification.

   • Container Compute Service (ACS) can allocate 8 PPU cards because its Kubernetes scheduling layer supports logical, partial allocation of PPU cards. This is a scheduling-layer capability, not a change to the physical design of Zhenwu 810E.

   On the purchase page, configure the following parameters and then click Buy Now.

   • Region and Zone: Select China (Ulanqab).

   • Resource Type: Select Lingjun Resources.

   • Node Specification: Select ml.gp7vf.16.40xlarge.

   • Set the number of nodes and the subscription duration.

For more information, see Create a resource group and purchase Lingjun Intelligent Computing Resources.

3. Create a resource quota

Create a resource quota to allocate resources from your resource group. Follow these steps:

1. In the left-side navigation pane, choose AI Computing Resources > Resource Quota > Lingjun Resources > Add Resource Quota.

2. Configure the following key parameters.

   Parameter	Description
   Name	Enter a name for the resource quota.
   Associate Workspace	Select a workspace to associate with the resource quota. You must be in this workspace to use this resource quota for DSW, EAS, or DLC resources.
   Source Type	Select one of the following options: Dedicated Resource Group: Allocate resources from a resource group. You can select the ml.gp7vf.16.40xlarge resource group that you created in the previous step. Existing Resource Quota: Allocate resources from an existing resource quota.
   Source	Select the resource group or resource quota that you created.
   Node/Instance Type	Click Add to add node specifications and specify the number of nodes as needed.
   VPC	Select an existing Virtual Private Cloud (VPC), vSwitch, and security group. Note If you need to access the internet, you must configure an Internet NAT gateway and bind an EIP to the associated VPC. For this reason, select the VPC that you plan to use for internet access. For information about how to configure the SNAT feature of an Internet NAT gateway, see Use the SNAT feature of an Internet NAT gateway to access the internet.
   Security Group
   vSwitch

   Then, click OK to create the quota.

After you create a resource quota, you can click its name to view details, check usage levels, and manage the quota. You can also scale the quota or add child quotas to optimize resource allocation. For more information, see Manage resource quotas.

4. (Optional) Unsubscribe from resources

For PAI Serverless, we recommend that you also use products such as CPFS Intelligent Computing Edition and Container Registry (ACR) Enterprise Edition.

• CPFS: You can use CPFS to create datasets or mount it on an instance to read data or use it as persistent storage.

• Container Registry (ACR): Use it to host and distribute your custom images.