Overview

You can use the rank table file to configure the NPU resources involved in collective communication. This file helps configure cluster information in the following scenarios:

• Initializing the communicator through the C API HcclCommInitClusterInfo or HcclCommInitClusterInfoConfig
• Initializing the TensorFlow distributed network communicator

For TensorFlow networks, you can also configure resource information using environment variables. However, only one method works: either the rank table or the environment variables. The following products support configuration using environment variables:

• Atlas A2 training product/Atlas A2 inference product
• Atlas training product