Feature: Support single op profiling

ge/single_op/single_op.cc

@@ -349,9 +349,17 @@ Status DynamicSingleOp::ExecuteAsync(const vector<GeTensorDesc> &input_desc,
   }
 
   if (op_task_->GetOpTaskType() == OP_TASK_TBE) {
-    return ExecuteTbeTask(input_desc, inputs, output_desc, outputs);
+    auto ret = ExecuteTbeTask(input_desc, inputs, output_desc, outputs);
+    if (ret == SUCCESS) {
+      GE_CHK_STATUS_RET_NOLOG(ProfilingTaskInfo());
+    }
+    return ret;
   } else if (op_task_->GetOpTaskType() == OP_TASK_AICPU || op_task_->GetOpTaskType() == OP_TASK_AICPUCC) {
-    return op_task_->LaunchKernel(input_desc, input_buffers, output_desc, output_buffers, stream_);
+    auto aicpu_ret = op_task_->LaunchKernel(input_desc, input_buffers, output_desc, output_buffers, stream_);
+    if (aicpu_ret == SUCCESS) {
+      GE_CHK_STATUS_RET_NOLOG(ProfilingTaskInfo());
+    }
+    return aicpu_ret;
   } else {
     GELOGE(ACL_ERROR_GE_OP_TASK_TYPE_INVALID,
            "Only TBE_Task, AI_CPU_Task and AI_CPUCC_Task are supported, but got %u",