Create and purchase Lingjun resources | PAI - Platform For AI

Overview

Lingjun intelligent computing resources are Alibaba Cloud's next-generation compute platform with RDMA high-speed networking, high-performance communication libraries, acceleration software, and GPU virtualization.

PAI offers fully managed Lingjun resources that you purchase and use within the PAI console. If you already own standalone Lingjun hardware, you can add it as semi-managed resources for training jobs.

Limitations

Supported regions

Available only in the following regions: China (Ulanqab), Singapore, China (Shenzhen), China (Beijing), China (Shanghai), and China (Hangzhou).
Supported users

Available with subscription and pay-as-you-go billing. Whitelisted users only. To request access, contact your business manager.
Supported job types

Supported training job types: TensorFlow, PyTorch, ElasticBatch, XGBoost, OneFlow, MPIJob, Slurm, and Ray.

Accounts and permissions

Alibaba Cloud account (Recommended): Full access to all operations without additional authorization.
RAM user: Your Alibaba Cloud account must grant permissions to manage resource pools or grant the AliyunPAIFullAccess permission to the RAM user.

Important
The AliyunPAIFullAccess permission grants full control over all PAI resources and features. Grant it with caution. We recommend using your Alibaba Cloud account for these operations.

Dependencies

Lingjun resources depend on other cloud products. Prepare the following before you begin.

(Required) Virtual Private Cloud (VPC)

Lingjun resources must be associated with a Virtual Private Cloud (VPC) in the same region. Configure a vSwitch and security group to ensure network connectivity with other cloud products.

(Optional) Internet NAT Gateway and EIP

To access the internet (for example, to pull custom images from public repositories), enable the SNAT feature of an Internet NAT Gateway and bind an Elastic IP Address (EIP) to it.

Use the SNAT feature of an Internet NAT Gateway to access the internet.

(Optional) OSS, NAS, or CPFS for AI

DLC jobs require a dataset. PAI supports datasets from Object Storage Service (OSS), Network Attached Storage (NAS), and CPFS for AI. Prepare storage resources and create datasets based on your needs. Prepare a dataset.

Procedure

Lingjun resources support subscription and pay-as-you-go billing. After creating a resource group, purchase resources based on your needs. Billing of AI computing resources.

Note

Lingjun resources divide compute capacity into high-speed interconnect zones (hz). Resources with the same hz ID communicate over a high-speed network. You can customize network affinity when purchasing, allocating, and using resources.

Create a resource group

Log on to the PAI console and go to the AI Computing Resources > Resource Pool page.
On the Lingjun Intelligent Computing resources tab, click Create Resource Group.

In the Create Resource Group dialog box, configure the parameters and click OK.

Parameter	Description
Type	Select Dedicated Resource Group.
Resource Group Name	Enter a resource group name that follows the on-screen naming conventions.

Purchase subscription resources

On the Intelligent Computing LINGJUN Resources tab, find the resource group that you created. In the Actions column, choose Create Order > Create Subscription Order.

Alternatively, on the resource group details page, in the upper-right corner, choose Create Order > Create Subscription Order.
On the purchase page, the resource group ID and region are preselected. Select the Node Specification, Quantity, and Duration based on your requirements, and then click Buy Now.

Common issues:
- The order does not contain information for the current resource group.
  - Cause: The selected region does not match the region of the resource group.
  - Solution: Switch to the default region of the resource group.
- The specified instance type is out of stock in zone.
  - Cause: The selected node specification is out of stock in the current zone.
  - Solution: Select a different node specification.
- The current kind of instance is temporarily not supported. Please choose another kind of ecs to purchase.
  - Cause: The selected node specification is not available for purchase in the current region.
  - Solution: Select a different node specification.
After payment, view purchased resources on the Order Information tab of the resource group details page. Details include resource specifications, quantity, running status, GPU type, CPU cores, memory, creation time, and expiration time. Click Renew to renew an order. The system splits each purchase into per-node orders, allowing you to renew or unsubscribe each node individually.

Purchase pay-as-you-go resources

Note

Available to whitelisted users only. Contact your business manager to request access.

On the Lingjun Intelligent Computing resources tab, find the resource group that you created. In the Actions column, choose Create Order > Create Pay-As-You-Go Order.

Alternatively, on the resource group details page, in the upper-right corner, choose Create Order > Create Pay-As-You-Go Order.
On the Create Pay-As-You-Go Order page, the Region, Resource Type, and Resource Group are preselected. Select a Resource Specification and specify the Quantity of nodes based on your requirements, then click OK.
After purchase, filter the order list by billing method to view subscription or pay-as-you-go orders.

Important
To avoid unnecessary charges, release unused pay-as-you-go resources from the Order Information page by clicking Release in the Actions column. Releasing a node causes running jobs on it to fail.

Next steps

After purchasing resources:

View and manage your resources on the resource group page. Manage resources.
Allocate resources to training jobs by configuring resource quotas. Create a resource quota.