User Guide

Contents

Generate Pre-configured Command Lines

You can use
Intel® Advisor
to generate command lines for perspective or analysis configuration and copy the lines to a clipboard to run from a terminal/command prompt. If you use pre-configured command lines, you do not need to provide analysis configuration, project directory, target application, and application options manually for each analysis you run.
To generate the command lines, you can do one of the following:
  • Generate the command lines from
    Intel Advisor
    graphical user interface (GUI).
  • For the
    Offload Modeling
    perspective only: Generate the command lines from
    Intel Advisor
    command line interface (CLI).

Generate Command Lines from GUI

Prerequisite
: Set up a project.
  1. Select a perspective from a
    Perspective Selector
    or a drop-down menu in the
    Analysis Workflow
    pane.
    The
    Analysis Workflow
    pane opens for a selected perspective.
  2. In the
    Analysis Workflow
    pane, do one of the following:
    • To use a pre-defined set of analyses and properties, select an accuracy level.
    • Select analyses and options manually from the
      Analysis Workflow
      tab. To set additional options, go to
      File
      Project Properties
      and set the properties for desired analyses in the respective left-side sections. .
  3. From the
    Analysis Workflow
    pane, generate the command lines:
    • For all selected analysis: Click the
      Get Command Line
      button.
    • For a specific analysis: Expand the analysis controls and click the
      Command Line
      button.
    The
    Copy Command Line to Clipboard
    dialog box opens, which provides commands to launch the perspective or the analysis with selected configuration. Options with default values are hidden by default.
  4. Click the
    Copy
    button to copy the generated command lines to the clipboard.
  5. Paste the copied command lines to a terminal/command prompt one by one and run them.

Generate Command Lines from GUI for an MPI Application

Prerequisite
: Set up a project.
  1. Go to
    File
    Project Properties
    and select the analysis you want to generate the command line for from the left-side pane. For example, go to
    Survey Hotspots Analysis
    to generate command line for the Survey analysis.
  2. Set properties to configure the analysis, if required.
  3. Select the
    Use MPI Launcher
    checkbox.
  4. Specify MPI run parameters:
    • If you select the
      Intel MPI
      as the MPI launcher: specify the number of ranks to profile, select
      Selective
      and specify ranks to profile, if required.
    • If you select
      Other
      as the MPI launcher: specify an MPI launcher executable and the number of ranks to profile. Make sure the
      All
      is selected under
      Profile ranks
      .
    • Click
      OK
      to save the parameters.
    If you need to generate commands for more than one analysis or for the whole perspective, specify the MPI run parameters for
    each
    analysis you want to run. Otherwise,
    Intel Advisor
    uses default parameters for all analyses you do not configure properly.
  5. Open the
    Analysis Workflow
    pane and select a perspective to run.
  6. Select an accuracy level to preconfigure the perspective or configure it manually using the checkboxes. Make sure to select only analyses that you have specified the MPI run parameters for.
    Do not
    change the accuracy level if you specified additional parameters in the
    Project Properties
    . Preconfigured accuracy levels reset your configuration.
  7. Generate the command lines:
    • For all selected analysis: Click the
      Get Command Line
      button.
    • For a specific analysis: Expand the analysis controls and click the
      Command Line
      button.
    The
    Copy Command Line to Clipboard
    dialog box opens, which provides commands to launch the perspective or the analysis with selected configuration. Options with default values are hidden by default.
  8. In the
    Copy Command Line to Clipboard
    dialog box, select the
    Generate command line for MPI
    checkbox to apply the MPI syntax you configured.
  9. Click the
    Copy
    button to copy the generated command lines to the clipboard.
  10. Paste the copied command lines to a terminal/command prompt one by one and run them.

Generate Command Lines for
Offload Modeling
from CLI

You can generate pre-configured command lines to analyze your application with
Offload Modeling
. Use this feature if you want to:
  • Analyze an MPI application
  • Customize preset
    Offload Modeling
    commands
Note
: In the commands below, make sure to replace the
myApplication
with your application executable path and name
before
executing a command. If your application requires additional command line options, add them
after
the executable name.
Prerequisite
: Set up environment variables to enable
Intel Advisor
CLI.
  1. Generate pre-configured command lines using the option of the
    --collect=projection
    analysis or
    collect,py
    script. Specify accuracy level and paths to your project directory and application executable. For example, to generate commands for the default medium accuracy with
    --collect=projection
    :
    advisor --collect=offload --dry-run --project-dir=./advi_results -- ./myApplication
    If you want to scale parameters, for example, to change the number of GPU tiles per MPI process, generate the commands using the
    collect.py
    script as follows:
    advisor-python $APM/collect.py ./advi_results --set-parameter <parameter-to-modify> --dry-run -- ./myApplication
    Families of Intel® X
    e
    graphics products starting with Intel® Arc™ Alchemist (formerly DG2) and newer generations feature GPU architecture terminology that shifts from legacy terms. For more information on the terminology changes and to understand their mapping with legacy content, see GPU Architecture Terminology for Intel® X
    e
    Graphics
    .
    After you run the command, a list of analysis commands to run the Offload Modeling with the specified accuracy level is printed to the terminal/command prompt.
  2. If you analyze an MPI application
    : Copy the generated commands to your preferred text editor and modify each command to use an MPI launcher: add
    mpirun
    or
    mpiexec
    and a number of MPI processed to launch to each command. For details about the syntax, see Analyze MPI Applications.
  3. Run the generated commands one by one from a command prompt or a terminal. You can add more options or modify the commands to better fit your goal. See Command Line Option Reference for available options.
  4. If you generated commands with the script: Run Performance Modeling step with the scaled parameter:
    advisor --collect=projection --project-dir=./advi_results --set-parameter <parameter-to-modify>
For details about MPI application analysis with Offload Modeling, see Model MPI Application Performance on GPU.

Product and Performance Information

1

Performance varies by use, configuration and other factors. Learn more at www.Intel.com/PerformanceIndex.