Intel® VTune™ Profiler User Guide

User Guide

  • 2020
  • 06/18/2020
  • Public Content
Contents

Stitch Stacks for Intel® Threading Building Blocks or OpenMP* Analysis

Use the
Stitch stacks
 option to restore a logical call tree for Intel® TBB or OpenMP* applications by catching notifications from the runtime and attach stacks to a point introducing a parallel workload.
Typically the real execution flow in the applications based on Intel Threading Building Blocks (Intel TBB) or OpenMP is very different from the code flow. During the user-mode sampling and tracing analysis of an Intel TBB-based application or an OpenMP application using Intel runtime libraries, the
Intel® VTune™
Profiler
 automatically enables the
Stitch stacks
 option. To view the OpenMP or Intel TBB objects hierarchy, explore the data provided in the
Top-down Tree
 pane.
  • To analyze a logically structured OpenMP call flow, make sure to compile and run your code with the Intel® Compiler 13.1 Update 3 or higher (part of the Intel Composer XE 2013 Update 3).
  • Stack stitching is available when you run the application from the
    VTune
    Profiler
     (the
    Launch Application
    target type). It does not work when attaching to the application (the
    Attach to Process
     target type).
You may want to disable stack stitching, for example, to minimize the collection overhead. To do this for your predefined user-mode sampling and tracing analysis type (for example, Hotspots or Threading), you need to create a new custom analysis configuration and deselect the
Stitch stacks
 option in the Custom Analysis configuration. You may use the same modified GUI analysis configuration for command line analysis. For this, just click the
Command Line…
 button in the
Configure Analysis
 window and copy the generated command line to run it from the terminal window. Alternatively, you can manually configure the command line for a custom
runss
 analysis using the
knob stack-stitching=false
 option like this:
>
vtune
 -collect-with runss -knob cpu-samples-mode=stack -knob stack-stitching=false -knob mrte-type=java,dotnet,python -app-working-dir <
path
 > -- <
application
 >
In this case, the
Top-down Tree
 pane (or report) displays separate entries for OpenMP worker threads.
Examples
Call stack in the
Top-down Tree
 pane with the
Stitch stacks
 option disabled:
Call stack in the
Top-down Tree
 pane with the
Stitch stacks
 option enabled (default behavior):

Product and Performance Information

1

Intel's compilers may or may not optimize to the same degree for non-Intel microprocessors for optimizations that are not unique to Intel microprocessors. These optimizations include SSE2, SSE3, and SSSE3 instruction sets and other optimizations. Intel does not guarantee the availability, functionality, or effectiveness of any optimization on microprocessors not manufactured by Intel. Microprocessor-dependent optimizations in this product are intended for use with Intel microprocessors. Certain optimizations not specific to Intel microarchitecture are reserved for Intel microprocessors. Please refer to the applicable product User and Reference Guides for more information regarding the specific instruction sets covered by this notice.

Notice revision #20110804