User Guide

Intel® VTune™ Profiler User Guide

ID 766319
Date 11/07/2023
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents

Top-down Report

Similar to the Top-down window, available in GUI, the Top-down represents call sequences (stacks) detected during collection phase starting from the application root. Use the top-down report to explore the call sequence flow of the application and analyze the time spent in each program unit and on its callees.

NOTE:

Intel® VTune™ Profiler collects information about program unit callees only during User-Mode Sampling and Tracing Collection or Hardware Event-based Sampling Collection with stack collection enabled.

Examples

Example 1: Hotspots Top-down Report

This example displays the report for the specified Hotspots analysis in the user-mode sampling mode with functions stacks limited to 5 elements.

vtune -report top-down -r r001hs -limit 5
Function Stack CPU Time:Total CPU Time:Effective Time:Total CPU Time:Spin Time:Total CPU Time:Overhead Time:Total ---------------------- -------------- ----------------------------- ------------------------ ---------------------------- Total 100.000% 100.000% 100.000% 100.000% func@0x6b2daccf 99.853% 99.835% 100.000% 100.000% func@0x6b2dacf0 99.853% 99.835% 100.000% 100.000% BaseThreadInitThunk 99.853% 99.835% 100.000% 100.000% thread_video 95.614% 97.876% 78.195% 0.0%

Example 2: Hotspots Report with Enabled Call Stack Collection (Linux*)

This command runs the Hotspots analysis in the hardware event-based sampling mode with enabled call stack collection.

vtune -collect hotspots -knob sampling-mode=hw -knob enable-stack-collection=true -- /home/tachyon

The following command generates the top-down report for the previously collected result and shows the result for columns with the time:total strings in the title.

vtune -report top-down -r r001hs -column=time:total

Function Stack CPU Time: CPU Time: CPU Time: Context Switch Time: Context Switch Time: Context Switch Time: Total Effective Time:Total Spin Time:Total Total Wait Time:Total Inactive Time:Total ---------------------- --------- -------------------- --------------- -------------------- -------------------- -------------------- Total 100.000% 100.000% 100.000% 100.000% 100.000% 100.000% func@0x6b2daccf 97.595% 97.704% 89.202% 65.777% 90.121% 62.893% func@0x6b2dacf0 97.595% 97.704% 89.202% 65.777% 90.121% 62.893% BaseThreadInitThunk 97.595% 97.704% 89.202% 65.777% 90.121% 62.893% threadstartex 67.091% 67.855% 8.335% 29.825% 9.027% 32.289% ...

Example 3: Hotspots Report with Disabled Stack Collection (Windows*)

This command runs the Hotspots analysis in the hardware event-based sampling mode with disabled call stack collection.

vtune -collect hotspots -knob sampling-mode=hw -knob enable-stack-collection=false -- C:\tachyon\tachyon.exe

This command generates the top-down report for the previously collected result, and shows the result for columns with the time:total string in the title. The report does not include information about program unit callees, as it was not collected during the analysis.

vtune -report top-down -r r001hs -column=time:total

Function Stack CPU Time:Total CPU Time:Effective Time:Total CPU Time:Spin Time:Total ---------------------- -------------- ----------------------------- ------------------------ Total 100.000% 100.000% 100.000% grid_intersect 50.172% 50.213% 0.0% sphere_intersect 31.740% 31.766% 0.0% grid_bounds_intersect 3.766% 3.769% 0.0% pos2grid 0.778% 0.778% 0.0% ...