--ac_parallel_enable

Applicable Products

• Ascend 950PR/Ascend 950DT: supported
• Atlas A3 training products/Atlas A3 inference products: supported
• Atlas A2 training products/Atlas A2 inference products: supported
• Atlas 200I/500 A2 inference products: not supported
• Atlas inference products: supported
• Atlas training products: supported

Description

Specifies whether to enable parallel execution between AI CPU operators and AI Core operators in dynamic-shape graphs.

When this function is enabled for dynamic-shape graphs, the system automatically identifies AI CPU operators that can run concurrently with AI Core operators. Operators from different engines are dispatched to different streams to achieve parallel execution across multiple engines, improving resource utilization and execution performance for dynamic-shape graphs.

Arguments

• 1: AI CPU operators and AI Core operators are allowed to run in parallel.
• 0 (default): AI CPU operators are not separately distributed.

Suggestions and Benefits

None

Example

--ac_parallel_enable=1