Function: get_step_timestamp

Applicability

Product	Supported
Atlas 350 Accelerator Card	√
Atlas A3 training product / Atlas A3 inference product	√
Atlas A2 training product / Atlas A2 inference product	√
Atlas training product	√
Atlas inference product	√
Atlas 200I/500 A2 inference product	√

Description

Sets the start timestamp and end timestamp of an iteration when single-operator loading and execution APIs are used for training, to display the profile data by iteration.

Prototype

C Prototype

        
             aclError aclprofGetStepTimestamp(aclprofStepInfo* stepInfo, aclprofStepTag tag, aclrtStream stream)

Python Function

        
             ret =acl.prof.get_step_timestamp(stepinfo, tag, stream)

Parameters

Parameter	Description
stepinfo	Int, step information. Call the acl.prof.create_step_info API to create data of the aclprofStepInfo type in advance.
tag	Int, step start timestamp and end timestamp. Pass the value ACL_STEP_START at the step start and the value ACL_STEP_END at the step end.
stream	Int, stream.

Return Value

Return Value	Description
ret	Int, error code. 0 indicates success. Other values indicate failure.

Parent topic: Iteration Time in PyTorch Scenarios