Function: get_op_end

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	√

Obtains the operator execution end time (ns).

You are advised to create a thread and call this API in the new thread. Otherwise, this API call will block the main thread.

C Prototype

uint64_t aclprofGetOpEnd(const void *opInfo, size_t opInfoLen, uint32_t index)

Python Function

op_end = acl.prof.get_op_end(op_info, op_info_len, index)

Parameter	Description
op_info	Int, specified memory address.
op_info_len	Int, length of the operator information.
index	Int, operator index. The value range is [0, Operator count – 1]. Call acl.prof.get_op_num to obtain the operator count.

Parameter

Description

op_info

Int, specified memory address.

op_info_len

Int, length of the operator information.

index

Int, operator index.

The value range is [0, Operator count – 1]. Call acl.prof.get_op_num to obtain the operator count.

Return Value	Description
op_end	Int, execution end time of the operator.

Parent topic: Subscribing to Operator Information