Affinity Scheduling Policies

The table below describes the characteristics and resource usage rules of an inference server (with Atlas 300I Duo inference cards).

**Table 1** Affinity policies
Policy Name	Policy Description
Affinity scheduling by inference card	The Ascend AI processors on one Atlas 300I Duo inference card are preferred. If one or two Ascend AI processors need to be allocated, ensure that the processor(s) is (are) selected from one Atlas 300I Duo inference card. The node with one available Atlas 300I Duo inference card is the best and then two.
Distributed inference scheduling by Ascend AI processor	The job must be scheduled to the entire Atlas 300I Duo inference card. If the number of Ascend AI processors required by the job is an odd number, the job is preferentially scheduled to the Atlas 300I Duo inference card with one remaining AscendAI processor.

Parent topic: Inference Server (with Atlas 300I Duo Inference Cards)