Visible to Intel only — GUID: GUID-0098BCAE-58B9-4CB7-A55E-EC46A89999DD
Visible to Intel only — GUID: GUID-0098BCAE-58B9-4CB7-A55E-EC46A89999DD
Offload Modeling Command Line Reference
This reference section describes the command line options available for each of the Python* scripts that you can use to run the Offload Modeling perspective.
To use the Offload Modeling, run one or two of the following scripts, depending on a method you chose:
- Use run_oa.py script to collect performance data and model performance on a target device using a single command with a set of default recommended settings.
- Use collect.py to collect baseline performance data for your application on a host device.
- Use analyze.py to model your application performance on a target device.
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.
Command Syntax
The syntax for Offload Modeling commands is as follows:
advisor-python <APM>/<script-name>.py <project-dir> [--options] [-- <target> [target-options]]
where:
advisor-python |
A call to the Intel® Advisor Python* command line tool. advisor-python is recommended to run the scripts. The main advantage of using this command line tool is that it does not require you to install a specific Python version on your system because it calls to an internal Python version of the Intel Advisor. |
<APM> |
The environment variable that points to the directory with the scripts. Replace it with:
|
<script-name> |
A script name to run: run_oa.py, collect.py, or analyze.py. |
<project-dir> |
The path to a project directory to save collection results. |
<--options> |
Options to modify behavior specific to the script. You can specify several options per script. Using an option not supported by the script causes a usage error. |
<target> |
A target application to analyze.
IMPORTANT:
You do not need to specify a target executable and target options when running the analyze.py script. |
[target-options] |
Options to modify target application behavior. |
Syntax Rules and Alternatives
An option can be preceded by one or two dashes. This section uses two dashes before long version of options and one dash before short version of options. For example, the following commands are equivalent:
advisor-python $APM/run_oa.py -h
advisor-python $APM/run_oa.py --help
The path to a project directory must always follow after a script name. For example:
advisor-python $APM/analyze.py ./advi_results
If an option accepts values, they can be separated by a space or by an equal sign (=). This document uses space for all such options. For example, the following are equivalent:
advisor-python $APM/analyze.py ./advi_results --out-dir ./report
advisor-python $APM/analyze.py ./advi_results --out-dir=./report
The target executable must be preceded by two dashes and a space. For example:
advisor-python $APM/analyze.py ./advi_results --out-dir=./report
advisor-python $APM/collect.py ./advi_results -- myApplication
If you have Python 3.6 or 3.7 installed and it is the default Python version on your system, you can run Offload Modeling with your system Python instead of the advisor-python tool:
python $APM/run_oa.py ./advi_results -- ./myApplication
python3.6 $APM/run_oa.py ./advi_results -- ./myApplication
python3.7 $APM/run_oa.py ./advi_results -- ./myApplication