Intel® Advisor Help

Launch Intel® Advisor

This topic provides overview and simple steps for running Intel® Advisor graphical user interface (GUI) and command line interface (CLI).

Launch the Intel® Advisor GUI

The Intel® Advisor GUI is available:

Note

By default, Intel Advisor opens a new graphical user interface (GUI). To switch back to an old GUI, set ADVISOR_EXPERIMENTAL=advixe-gui variable and re-open the Intel Advisor.

After opening the Intel Advisor, continue to create a project or run a perspective and view the results if you already have a project.

Launch the Intel Advisor CLI

Prerequisite: Set up environment variables to enable the command line interface.

To run the advisor command line interface, use the following syntax: 

advisor <--action> [--action-options] [--global-options] -- <target-application> [target options]

where:

The advisor command line interface supports all Intel Advisor perspective and is the recommended method to run the Intel Advisor from command line.

You can also run the Offload Modeling perspective using Python* scripts as follows: 

advisor-python <APM>/<offload-script>.py <project-dir> [--options] [-- <target-application> [target-options]]

where:

When you run the first Intel Advisor analysis to a target application from the command line, it also creates a new project for the target.

Review the typical workflows for the Intel Advisor CLI in the dedicated topics for each perspective.

For details about the Intel Advisor command line syntax and options, see the advisor Command Line Interface Reference.

Launch the Intel Advisor from a Docker* Container on Linux* OS

This section contains steps to run Intel® Advisor in a Docker* container. Containers allow you to set up and configure environments and distribute them using images:

Set up the Docker container

  1. Pull the Docker image from the oneAPI Containers Repository with the following commands: 

    image=amr-registry.caas.intel.com/oneapi/oneapi:base-dev-ubuntu18.04
docker pull "$image"

  2. Run the Docker container using the following command: 

    docker run --cap-add=SYS_PTRACE -it "$image"

    Note

    • The --device=/dev/dri option enables the gpu (if available).

    • You can specify proxy information using options as follows: -e http_proxy="$http_proxy" -e https_proxy="$https_proxy"

  3. For the rest of the steps in this section, run any commands from the command line prompt inside the Docker container.

    For example, to set up the Mandelbrot sample, you can run: 

    cd /one-api-code-samples/HPC/mandelbrot
make
./main -d1
./main -t gpu # run on gpu
./main -t cpu # run on cpu
make clean

  4. Run the following commands to source Intel Advisor variables: 

    source /opt/intel/oneapi/setvars.sh

    Note

    This step is not required, but allows you to run tools from any directory, rather than using absolute file paths.

  5. Now that your Docker container is running, you can run Advisor from the command line as you would without a container. For example: 

    advisor --collect=survey /bin/ls

When you run the first Intel Advisor analysis to a target application from the command line, it also creates a new project for the target.

For details about the Intel Advisor command line syntax and options, see the advisor Command Line Interface Reference. Review the typical workflows for the Intel Advisor CLI in the dedicated topics for each perspective.

Parent topic: Install and Launch Intel® Advisor

See Also

Set Up a Project
Analyze Vectorization Perspective Improve your application performance, get code-specific recommendations for how to fix vectorization issues and quick visibility into source code and assembly code by running the Vectorization and Code Insights perspective.
Analyze CPU Roofline Visualize actual performance against hardware-imposed performance ceilings by running the CPU / Memory Roofline Insights perspective. It helps you determine the main limiting factor (memory bandwidth or compute capacity) and provides an ideal roadmap of potential optimization steps.
Model Threading Designs Analyze, design, tune, and check threading design options without disrupting your normal development by running the Threading Perspective.
Model Offloading to a GPU Find high-impact opportunities to offload/run your code and identify potential performance bottlenecks on a target graphics processing unit (GPU) by running the Offload Modeling perspective.
Analyze GPU Roofline Measure and visualize the actual performance of GPU kernels using benchmarks and hardware metric profiling against hardware-imposed performance ceilings, as well as determine the main limiting factor, by running the GPU Roofline Insights perspective.