Platform For AI:AI computing resources

PAI AI computing resources deliver high-reliability, high-performance GPU computing power for training and inference scenarios that demand low latency and consistent stability.

Resource types

• Cloud-native resources

  These resources support PAI-DSW, PAI-EAS, and PAI-DLC.

  • General-purpose computing resources

    General-purpose computing resources, such as ECS, ECI, and EGS, provide a flexible and stable deep learning training environment. After you activate PAI, a public resource quota for general-purpose computing resources is created by default. You can associate this quota with a workspace.

  • Lingjun intelligent computing resources

    Lingjun intelligent computing resources are designed for large-scale deep learning. With integrated software and hardware optimization, they deliver high performance, efficiency, and utilization for large model training, inference, and development.

• Big data engine resources

  These resources support PAI-Designer.

  • MaxCompute

    MaxCompute is an enterprise-grade, SaaS-based cloud data warehouse that provides fully managed data analysis and processing at scale. For details, see What is MaxCompute.

  • Fully-managed Flink resources

    Alibaba Cloud Realtime Compute for Apache Flink is a one-stop, real-time big data analytics platform built on Apache Flink, delivering sub-second data analysis. For details, see What is Alibaba Cloud Realtime Compute for Apache Flink.

Get started

Go to AI Computing Resources > Resource Pools to create a resource group and purchase computing resources:

• General-purpose computing resources: For details, see Create a resource group and purchase general-purpose computing resources.

• Lingjun intelligent computing resources: For details, see Create a resource group and purchase Lingjun intelligent computing resources.

Go to AI Computing Resources > Quotas to create a resource quota and allocate computing power across teams. After you bind the quota to a workspace, you can use its computing power for AI development and training.

• Cloud-native resource quota

  After purchasing computing resources in a resource pool, you can allocate computing power from one or more resource groups to a resource quota. You can also create a sub-level resource quota to form a quota tree, enabling more flexible queuing and scheduling.

  • Lingjun intelligent computing resources: For details, see Create a resource quota.

  • general-purpose computing resources: For details, see General-purpose computing resource quotas.

• Big data resource quota

  For details on activating, purchasing, and using big data resource quotas, see:

  • MaxCompute: For details, see MaxCompute resource quotas.

  • fully-managed Flink resources: For details, see Manage fully-managed Flink resources.