Lingjun intelligent computing resources are high-performance computing resources provided by PAI for large-scale AI training and inference. They support DSW, DLC, and EAS modules. You must create a resource group before purchasing these resources.
Overview
Lingjun intelligent computing resources are Alibaba Cloud's next-generation intelligent computing resources. They provide RDMA high-speed networking, high-performance communication libraries, acceleration software, and GPU virtualization solutions for high-performance computing.
PAI provides fully managed Lingjun intelligent computing resources. You can purchase and use them entirely on the PAI console. If you have purchased standalone Lingjun hardware, you can also add it to the PAI console as semi-managed resources.
Limitations
Supported regions
Available only in China (Ulanqab), Singapore, China (Shenzhen), China (Beijing), China (Shanghai), and China (Hangzhou).
Supported users
Lingjun intelligent computing resources are available on subscription and pay-as-you-go billing models. Access is currently limited to whitelisted users. To request access, contact your business manager.
Supported job types
Lingjun intelligent computing resources support only TensorFlow, PyTorch, ElasticBatch, XGBoost, OneFlow, MPIJob, Slurm, and Ray training jobs.
Account and permissions
Alibaba Cloud account (Recommended): You can perform all operations with your main account without requiring additional authorization.
RAM user: Your Alibaba Cloud account must grant the RAM user either permissions to manage the resource pool or the AliyunPAIFullAccess policy.
ImportantThe AliyunPAIFullAccess policy grants full control over all PAI resources and features. Grant this policy with caution. We recommend using your main Alibaba Cloud account for these operations.
Dependencies
Creating, purchasing, and using Lingjun intelligent computing resources require other Alibaba Cloud services. Before you begin, review these dependencies and prepare the necessary services.
VPC (required)
When you allocate Lingjun intelligent computing resources, you must bind them to a VPC in the same region. You also need to configure a vSwitch and a security group to ensure network connectivity between the Lingjun resources and other cloud services.
Internet NAT Gateway and EIP (optional)
If you need to access the internet later, for example, to pull custom images, you must also enable the SNAT feature of an Internet NAT Gateway and bind an EIP to it.
For instructions on how to configure the SNAT feature, see Use the SNAT feature of an Internet NAT Gateway to access the internet.
OSS/NAS/CPFS for AI (optional)
To submit DLC jobs that use Lingjun intelligent computing resources, you must first create a dataset. PAI currently supports OSS, NAS, and CPFS for AI datasets. Prepare the required storage resources in advance. For more information, see Prepare datasets.
Create a resource group and purchase resources
Lingjun intelligent computing resources are available with subscription and pay-as-you-go billing models. After you create a resource group, you can purchase resources. For billing details, see Billing of AI computing resources.
Lingjun organizes underlying computing resources into High-speed Interconnect Zones, referred to as hz. Resources with the same hz ID can communicate over a high-speed network. You can customize this high-speed network affinity when purchasing, allocating, and using resources.
Create a resource group
Log on to the PAI console and go to AI Computing Resources > Resource Pool.
On the Lingjun Intelligent Computing resources tab, click Create Resource Group.
In the Create Resource Group dialog box, configure the following parameters and click OK.
Parameter
Description
Type
Select Dedicated Resource Group.
Resource Group Name
Enter a name for the resource group that follows the on-screen naming conventions.
Purchase subscription resources
In the resource group list on the Intelligent Computing LINGJUN Resources tab, in the Actions column, click Create Order > Create Subscription Order.
Alternatively, click the resource group name. In the upper-right corner of the details page, click Create Order > Create Subscription Order.
On the purchase page, the resource group ID and region are automatically selected. Select parameters such as Node Specification, Quantity, and Duration, and then click Buy Now.
Common issues and solutions:
The order does not contain information for the current resource group.
Cause: You switched regions, so the resource group ID does not match the region.
Solution: Switch back 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 supported in the region (not available for purchase).
Solution: Select a different node specification.
After you complete the payment, the purchased subscription-based Lingjun intelligent computing resources appear on the order list page. On the resource group details page, click the Order Information tab to view instance details, such as resource specifications, quantity, running status, GPU type, number of CPU cores, memory, creation time, and expiration time. You can click Renew to renew the subscription. The system splits the purchase into separate orders for each node instance, allowing you to manage each node individually (for example, by renewing or unsubscribing).
Purchase pay-as-you-go resources
This feature is available only to whitelisted users. To request access, contact your business manager.
In the resource group list on the Lingjun Intelligent Computing resources tab, in the Actions column, click Create Order > Create Pay-As-You-Go Order.
Alternatively, click the resource group name. In the upper-right corner of the details page, click Create Order > Create Pay-As-You-Go Order.
On the Create Pay-As-You-Go Order page, the system preselects the Region, Resource Type, and Resource Group. Select a Resource Specification, set the node Quantity, and then click OK.
After the purchase completes, you can filter the order list by billing method to distinguish between subscription and pay-as-you-go orders.
ImportantIf you no longer need a pay-as-you-go Lingjun intelligent computing resource, click Release in the Actions column on the Order Information tab to avoid ongoing charges. Releasing a node causes any jobs running on it to fail. Proceed with caution.
Next steps
After you create a resource group and purchase computing resources, you can perform the following operations:
View basic information and manage your purchased resources on the resource group page. For more information, see Manage resources.
Configure a resource quota to allocate resources to training jobs. For more information, see Create a resource quota.