intel Get Started with VTune Profiler User Guide

intel Get Started with VTune Profiler

Get Started with Intel® VTune™ Profiler

Use Intel VTune Profiler to analyze local and remote target systems from Windows, macOS, and Linux* hosts. Improve application and system performance through these operations:

• Analyze algorithm choices.

• Find serial and parallel code bottlenecks.

• Understand where and how your application can benefit from available hardware resources.

• Speed up the execution of your application.
  Download Intel VTune Profiler on your system through one of these ways:

• Download the Standalone version.

• Get Intel VTune Profiler as part of the Intel® oneAPI Base Toolkit.
  See the VTune Profiler training page for videos, webinars, and more material to help you get started.

NOTE
Documentation for versions of Intel® VTune™ Profiler prior to the 2021 release are available for download only. For a list of available documentation downloads by product version, see these pages:

• Download Documentation for Intel Parallel Studio XE
• Download Documentation for Intel System Studio

Understand the Workflow
Use Intel VTune Profiler to profile an application and analyze results for performance improvements.

The general workflow contains these steps:

Select Your Host System to Get Started
Learn more about system-specific workflows for Windows, Linux, or macOS*.

Get Started with Intel® VTune™ Profiler for Windows* OS

Before You Begin

1. Install Intel® VTune™ Profiler on your Windows* system.

2. Build your application with symbol information and in Release mode with all optimizations enabled. For detailed information on compiler settings, see the VTune Profiler online user guide.
   You can also use the matrix sample application available in

   \VTune\Samples\matrix. You can see corresponding sample results in \VTune\Projects\sample (matrix).

3. Set up the environment variables: Run the \setvars.bat script.
   By default, the for oneAPI components is Program Files (x86)\Intel\oneAPI.
   NOTE You do not need to run setvars.bat when using Intel® VTune™ Profiler within Microsoft Visual Studio.

Step 1: Start Intel® VTune™ Profiler
Start Intel VTune Profiler through one of these ways and set up a project. A project is a container for the application you want to analyze, the type of analysis, and data collection results.

Source / Start VTune Profiler

Standalone (GUI)

1. Run the vtune-gui command or run Intel® VTune™ Profiler from the Start menu.
2. When the GUI opens, click in the Welcome screen.
3. In the Create Project dialog box, specify the project name and location.
4. Click Create Project.

Standalone (Command line)
Run the vtune command.

Microsoft Visual Studio IDE
Open your solution in Visual Studio. The VTune Profiler toolbar is automatically enabled and your Visual Studio project is set as an analysis target.

NOTE
You do not need to create a project when running Intel® VTune™ Profiler from the command line or within Microsoft* Visual Studio.

Step 2: Configure and Run Analysis
After creating a new project, the Configure Analysis window opens with these default values:

1. In the Launch Application section, browse to the location of your application executable file.
2. Click Start to run Performance Snapshot on your application. This analysis presents a general overview of issues affecting the performance of your application on the target system.

Step 3: View and Analyze Performance Data
When data collection completes, VTune Profiler displays analysis results in the Summary window. Here, you see a performance overview of your application.
The overview typically includes several metrics along with their descriptions.

• A Expand each metric for detailed information about contributing factors.
• B A flagged metric indicates a value outside acceptable/normal operating range. Use tool tips to understand how to improve a flagged metric.
• C See guidance on other analyses you should consider running next. The Analysis Tree highlights these recommendations.

Next Steps
Performance Snapshot is a good starting point to get an overall assessment of application performance with VTune Profiler. Next, check if your algorithm requires tuning.

1. Follow a tutorial to analyze common performance bottlenecks.
2. Once your algorithm is well-tuned, run Performance Snapshot again to calibrate results and identify potential performance improvements in other areas.

See Also
Microarchitecture Exploration

VTune Profiler Help Tour

Example: Profile an OpenMP Application on Windows
Use Intel VTune Profiler on a Windows machine to profile a sample iso3dfd_omp_offload OpenMP application offloaded onto an Intel GPU. Learn how to run a GPU analysis and examine results.

Prerequisites

• Make sure your system is running Microsoft* Windows 10 or a newer version.
• Use one of these versions of Intel Processor Graphics:
  • Gen 8
  • Gen 9
  • Gen 11
• Your system should be running on one of these Intel processors:
  • 7th Generation Intel® Core™ i7 Processors (code name Kaby Lake)
  • 8th Generation Intel® Core™ i7 Processors (code name Coffee Lake)
  • 10th Generation Intel® Core™ i7 Processors (code name Ice Lake)
• Install Intel VTune Profiler from one of these sources:
  • Standalone product download
  • Intel® oneAPI Base Toolkit
  • Intel® System Bring-up Toolkit
• Download the Intel® oneAPI HPC Toolkit which contains the Intel® oneAPI DPC++/C++ Compiler(icx/icpx) that you need to profile OpenMP applications.
• Set up environment variables. Execute the vars.bat script located in the \env directory.
• Set up your system for GPU analysis.

NOTE
To install Intel VTune Profiler in the Microsoft* Visual Studio environment, see the VTune Profiler User Guide.

Build and Compile the OpenMP Offload Application

1. Download the iso3dfd_omp_offload OpenMP Offload sample.

2. Open to the sample directory.
   cd /DirectProgramming/C++/StructuredGrids/iso3dfd_omp_offload

3. Compile the OpenMP Offload application.

mkdir build
cd build
icx /std:c++17 /EHsc /Qiopenmp /I../include\ /Qopenmp-targets:
spir64 /DUSE_BASELINE /DEBUG ..\src\iso3dfd.cpp ..\src\iso3dfd_verify.cpp ..\src\utils.cpp

Run a GPU Analysis on the OpenMP Offload Application
You are now ready to run the GPU Offload Analysis on the OpenMP application you compiled.

1. Open VTune Profiler and click on New Project to create a project.

2. On the welcome page, click on Configure Analysis to set up your analysis.

3. Select these settings for your analysis.

   • In the WHERE pane, select Local Host.
   • In the WHAT pane, select Launch Application and specify the iso3dfd_omp_offload binary as the application to profile.
   • In the HOW pane, select the GPU Offload analysis type from the Accelerators group in the Analysis Tree.
     

4. Click the Start button to run the analysis.

VTune Profiler collects data and displays analysis results in the GPU Offload viewpoint.

• In the Summary window, see statistics on CPU and GPU resource usage. Use this data to determine if your application is:
  • GPU-bound
  • CPU-bound
  • Utilizing the compute resources of your system inefficiently
• Use the information in the Platform window to see basic CPU and GPU metrics.
• Investigate specific computing tasks in the Graphics window.

For a deeper analysis, see a related recipe in the VTune Profiler Performance Analysis Cookbook. You can also continue your profiling with the GPU Compute/Media Hotspots analysis.

Example: Profile a SYCL Application on Windows
Profile a sample matrix_multiply SYCL application with Intel® VTune™ Profiler. Get familiar with the product and understand the statistics collected for GPU- bound applications.

Prerequisites

• Make sure you have Microsoft* Visual Studio (v2017 or newer) installed on your system.
• Install Intel VTune Profiler from the Intel® oneAPI Base Toolkit or the Intel® System Bring-up Toolkit. These toolkits contain the Intel® oneAPI DPC++/C++ Compiler(icpx -fsycl) compiler required for the profiling process.
• Set up environment variables. Execute the vars.bat script located in the \env directory.
• Ensure that the Intel oneAPI DPC++ Compiler (installed with the Intel oneAPI Base toolkit) is integrated into Microsoft Visual Studio.
• Compile the code using the -gline-tables-only and -fdebug-info-for-profiling options for Intel oneAPI DPC++ Compiler.
• Set up your system for GPU analysis.

For information on installing Intel VTune Profiler in the Microsoft* Visual Studio environment, see VTune Profiler User Guide.

Build the Matrix App
Download the matrix_multiply_vtune code sample package for Intel oneAPI toolkits. This contains the sample which you can use to build and profile a SYCL application.

1. Open Microsoft* Visual Studio.
2. Click File > Open > Project/Solution. Find the matrix_multiply_vtune folder and select matrix_multiply.sln.
3. Build this configuration (Project > Build).
4. Run the program (Debug > Start Without Debugging).
5. To choose a DPC++ or threaded version of the sample, use preprocessor definitions.
6. Go to Project Properties > DPC++ > Preprocessor > Preprocessor Definition.
7. Define icpx -fsycl or USE_THR.

Run GPU Analysis
Run a GPU analysis on the Matrix sample.

1. From the Visual Studio toolbar, click the Configure Analysis button.
   The Configure Analysis window opens. By default, it inherits your VS project settings and specifies the matrix_multiply.exe as an application to profile.

2. In the Configure Analysis window, click the Browse button in the HOW pane.

3. Select the GPU Compute/Media Hotspots analysis type from the Accelerators group in the Analysis Tree.
   

4. Click the Start button to launch the analysis with the predefined options.

Run GPU Analysis from Command Line:

1. Open the sample directory:

   \VtuneProfiler\matrix_multiply_vtune

2. In this directory, open a Visual Studio* project file named matrix_multiply.sln

3. The multiply.cpp file contains several versions of matrix multiplication. Select a version by editing the corresponding #define MULTIPLY line in multiply.hpp

4. Build the entire project with a Release configuration.
   This generates an executable called matrix_multiply.exe.

5. Prepare the system to run a GPU analysis. See Set Up System for GPU Analysis.

6. Set VTune Profiler environment variables by running the batch file: export \env\vars.bat

7. Run the analysis command:
   vtune.exe -collect gpu-offload — matrix_multiply.exe

VTune Profiler collects data and displays analysis results in the GPU Compute/Media Hotspots viewpoint. In the Summary window, see statistics on CPU and GPU resource usage to understand if your application is GPU-bound. Switch to the Graphics window to see basic CPU and GPU metrics representing code execution over time.

Get Started with Intel® VTune™ Profiler for Linux* OS

Before You Begin

1. Install Intel® VTune™ Profiler on your Linux* system.

2. Build your application with symbol information and in Release mode with all optimizations enabled. For detailed information on compiler settings, see the VTune Profiler online user guide.
   You can also use the matrix sample application available in

   \sample\matrix. You can see sample results in \sample (matrix).

3. Set up the environment variables: source /setvars.sh
   By default, the is:

   • $HOME/intel/oneapi/ when installed with user permissions;
   • /opt/intel/oneapi/ when installed with root permissions.

Step 1: Start VTune Profiler
Start VTune Profiler through one of these ways:

Source / Start VTune Profiler
Standalone/IDE (GUI)

1. Run the vtunegui command. To start VTune Profiler from the Intel System Studio IDE, select Tools > VTune Profiler > Launch VTune Profiler. This sets all appropriate environment variables and launches a standalone interface of the product.
2. When the GUI opens, click NEW PROJECT in the Welcome screen.
3. In the Create Project dialog box, specify the project name and location.
4. Click Create Project.

Standalone (Command line)

• Run the vtune command.

Step 2: Configure and Run Analysis
After creating a new project, the Configure Analysis window opens with these default values:

1. In the Launch Application section, browse to the location of your application.
2. Click the Start to run Performance Snapshot on your application. This analysis presents a general overview of issues affecting the performance of your application on the target system.

Step 3: View and Analyze Performance Data
When data collection completes, VTune Profiler displays analysis results in the Summary window. Here, you see a performance overview of your application.
The overview typically includes several metrics along with their descriptions.

• A Expand each metric for detailed information about contributing factors.
• B A flagged metric indicates a value outside acceptable/normal operating range. Use tool tips to understand how to improve a flagged metric.
• C See guidance on other analyses you should consider running next. The Analysis Tree highlights these recommendations.

Next Steps
Performance Snapshot is a good starting point to get an overall assessment of application performance with VTune Profiler. Next, check if your algorithm requires tuning.

1. Follow a tutorial to analyze common performance bottlenecks.
2. Once your algorithm is well-tuned, run Performance Snapshot again to calibrate results and identify potential performance improvements in other areas.

See Also
Microarchitecture Exploration

VTune Profiler Help Tour

**Example: Profile an OpenMP Application on Linux***
Use Intel VTune Profiler on a Linux machine to profile a sample iso3dfd_omp_offload OpenMP application offloaded onto an Intel GPU. Learn how to run a GPU analysis and examine results.

Prerequisites

• Make sure your system is running Linux* OS kernel 4.14 or a newer version.
• Use one of these versions of Intel Processor Graphics:
  • Gen 8
  • Gen 9
  • Gen 11
• Your system should be running on one of these Intel processors:
  • 7th Generation Intel® Core™ i7 Processors (code name Kaby Lake)
  • 8th Generation Intel® Core™ i7 Processors (code name Coffee Lake)
  • 10th Generation Intel® Core™ i7 Processors (code name Ice Lake)
• For the Linux GUI, use:
  • GTK+ version 2.10 or newer (2.18 and newer versions are recommended)
  • Pango version 1.14 or newer
  • X.Org version 1.0 or newer (1.7 and newer versions are recommended)
• Install Intel VTune Profiler from one of these sources:
  • Standalone product download
  • Intel® oneAPI Base Toolkit
  • Intel® System Bring-up Toolkit
• Download the Intel® oneAPI HPC Toolkit which contains the Intel® oneAPI DPC++/C++ Compiler(icx/icpx) that you need to profile OpenMP applications.
• Set up environment variables. Execute the vars.sh script.
• Set up your system for GPU analysis.

Build and Compile the OpenMP Offload Application

1. Download the iso3dfd_omp_offload OpenMP Offload sample.

2. Open to the sample directory.
   cd /DirectProgramming/C++/StructuredGrids/iso3dfd_omp_offload

3. Compile the OpenMP Offload application.

mkdir build;
cmake -DVERIFY_RESULTS=0 ..
make -j

This generates a src/iso3dfd executable.

To delete the program, type:
make clean

This removes the executable and object files that you created with the make command.

Run a GPU Analysis on the OpenMP Offload Application
You are now ready to run the GPU Offload Analysis on the OpenMP application you compiled.

1. Open VTune Profiler and click on New Project to create a project.

2. On the welcome page, click on Configure Analysis to set up your analysis.

3. Select these settings for your analysis.

   • In the WHERE pane, select Local Host.
   • In the WHAT pane, select Launch Application and specify the iso3dfd_omp_offload binary as the application to profile.
   • In the HOW pane, select the GPU Offload analysis type from the Accelerators group in the Analysis Tree.
     

4. Click the Start button to run the analysis.

VTune Profiler collects data and displays analysis results in the GPU Offload viewpoint.

• In the Summary window, see statistics on CPU and GPU resource usage. Use this data to determine if your application is:
  • GPU-bound
  • CPU-bound
  • Utilizing the compute resources of your system inefficiently
• Use the information in the Platform window to see basic CPU and GPU metrics.
• Investigate specific computing tasks in the Graphics window.

For a deeper analysis, see a related recipe in the VTune Profiler Performance Analysis Cookbook. You can also continue your profiling with the GPU Compute/Media Hotspots analysis.

Example: Profile a SYCL Application on Linux
Use VTune Profiler with a sample matrix_multiply SYCL application to quickly get familiar with the product and statistics collected for GPU-bound applications.

Prerequisites

• Install VTune Profiler and Intel® oneAPI DPC++/C++ Compiler from the Intel® oneAPI Base Toolkit or the Intel® System Bring-up Toolkit.
• Set up environment variables by executing the vars.sh script.
• Set up your system for GPU analysis.

Build the Matrix Application
Download the matrix_multiply_vtune code sample package for Intel oneAPI toolkits. This contains the sample which you can use to build and profile a SYCL application.

To profile a SYCL application, make sure to compile the code using the -gline- tables-only and -fdebug-info-for-profiling Intel oneAPI DPC++ Compiler options.

To compile this sample application, do the following:

1. Go to the sample directory.
   cd <sample_dir/VtuneProfiler/matrix_multiply>

2. The multiply.cpp file in the src folder contains several versions of matrix multiplication. Select a version by editing the corresponding #define MULTIPLY line in multiply.h.

3. Build the app using the existing Makefile:
   cmake .
   make
   This should generate a matrix.icpx -fsycl executable.
   To delete the program, type:
   make clean
   This removes the executable and object files that were created by the make command.

Run GPU Analysis
Run a GPU analysis on the Matrix sample.

1. Launch VTune Profiler with the vtune-gui command.

2. Click New Project from the Welcome page.

3. Specify a name and location for your sample project and click Create Project.

4. In the WHAT pane, browse to the matrix.icpx-fsycl file.

5. In the HOW pane, click the Browse button and select GPU Compute/Media Hotspots analysis from the Accelerators group in the Analysis Tree.
   

6. Click the Start button at the bottom to launch the analysis with the pre-selected options.

Run GPU Analysis from Command Line:

1. Prepare the system to run a GPU analysis. See Set Up System for GPU Analysis.

2. Set up environment variables for Intel software tools:
   source $ONEAPI_ROOT/setvars.sh

3. Run the GPU Compute/Media Hotspots analysis:
   vtune -collect gpu-hotspots -r ./result_gpu-hotspots — ./matrix.icpx -fsycl
   To see the summary report, type:
   vtune -report summary -r ./result_gpu-hotspots

VTune Profiler collects data and displays analysis results in the GPU Compute/Media Hotspots viewpoint. In the Summary window, see statistics on CPU and GPU resource usage to understand if your application is GPU-bound. Switch to the Graphics window to see basic CPU and GPU metrics representing code execution over time.

Get Started with Intel® VTune™ Profiler for macOS*

Use VTune Profiler on a macOS system to perform remote target analysis on a non-macOS system (Linux or Android only) .

You cannot use VTune Profiler in a macOS environment for these purposes:

• Profile the macOS system on which it is installed.
• Collect data on a remote macOS system.

To analyze performance of a remote Linux or Android target from the macOS host, do one of these steps:

• Run a VTune Profiler analysis on the macOS system with a remote system specified as the target. When analysis begins, VTune Profiler connects to the remote system to collect data, then brings the results back to the macOS host for viewing.
• Run an analysis on the target system locally and copy the results to a macOS system for viewing in VTune Profiler.

The steps in this document assume a remote Linux target system and collect performance data using SSH access from VTune Profiler on a macOS host system.

Before You Begin

1. Install Intel® VTune™ Profiler on your macOS* system.
2. Build your Linux application with symbol information and in Release mode with all optimizations enabled. For detailed information, see the compiler settings in the VTune Profiler help.
3. Set up SSH access from the host macOS system to the target Linux system to work in the password-less mode.

Step 1: Start VTune Profiler

1. Launch VTune Profiler with the vtune-gui command.
   By default, the is /opt/intel/oneapi/.

2. When the GUI opens, click NEW PROJECT in the Welcome screen.

3. In the Create Project dialog box, specify the project name and location.

4. Click Create Project.

Step 2: Configure and Run Analysis
After you create a new project, the Configure Analysis window opens with the Performance Snapshot analysis type.
This analysis presents an overview of issues that affect the performance of your application on the target system.

1. In the WHERE pane, select Remote Linux (SSH) and specify the target Linux system using username@ hostname[:port].
   VTune Profiler connects to the Linux system and installs the target package.

2. In the WHAT pane, provide the path to your application on the target Linux system.

3. Click the Start button to run Performance Snapshot on the application.

Step 3: View and Analyze Performance Data
When data collection completes, VTune Profiler displays analysis results on the macOS system. Start your analysis in the Summary window. Here, you see a performance overview of your application.

The overview typically includes several metrics along with their descriptions.

• A Expand each metric for detailed information about contributing factors.
• B A flagged metric indicates a value outside acceptable/normal operating range. Use tool tips to understand how to improve a flagged metric.
• C See guidance on other analyses you should consider running next. The Analysis Tree highlights these recommendations.

Next Steps
Performance Snapshot is a good starting point to get an overall assessment of application performance with VTune Profiler.
Next, check if your algorithm requires tuning.

1. Run Hotspots Analysis on your application.
2. Follow a Hotspots tutorial. Learn techniques to get the most out of your Hotspots analysis.
3. Once your algorithm is well-tuned, run Performance Snapshot again to calibrate results and identify potential performance improvements in other areas.

See Also
Microarchitecture Exploration

VTune Profiler Help Tour

Learn More
Document / Description

• User Guide
  The User Guide is the primary documentation for VTune Profiler.
  NOTE
  You can also download an offline version of the VTune Profiler documentation.

• Online Training
  The online training site is an excellent resource to learn the basics of VTune Profiler with Getting Started guides, videos, tutorials, webinars, and technical articles.

• Cookbook
  Performance analysis cookbook that contains recipes to identify and solve popular performance problems using analysis types in VTune Profiler.

• Installation Guide for Windows | Linux | macOS hosts
  The Installation Guide contains basic installation instructions for VTune Profiler and post-installation configuration instructions for the various drivers and collectors.

• Tutorials
  VTune Profiler tutorials guide a new user through basic features with a short sample application.

• Release Notes
  Find information about the latest version of VTune Profiler, including a comprehensive description of new features, system requirements, and technical issues that were resolved.
  For the standalone and toolkit versions of VTune Profiler, understand the current System Requirements.

Notices and Disclaimers
Intel technologies may require enabled hardware, software or service activation.
No product or component can be absolutely secure.
Your costs and results may vary.
© Intel Corporation. Intel, the Intel logo, and other Intel marks are trademarks of Intel Corporation or its subsidiaries. Other names and brands may be claimed as the property of others.
Intel, the Intel logo, Intel Atom, Intel Core, Intel Xeon Phi, VTune and Xeon are trademarks of Intel Corporation in the U.S. and/or other countries.
*Other names and brands may be claimed as the property of others.
Microsoft, Windows, and the Windows logo are trademarks, or registered trademarks of Microsoft Corporation in the United States and/or other countries.
Java is a registered trademark of Oracle and/or its affiliates.
OpenCL and the OpenCL logo are trademarks of Apple Inc. used by permission by Khronos.

Intel technologies may require enabled hardware, software or service activation.
No product or component can be absolutely secure.
Your costs and results may vary.
© Intel Corporation. Intel, the Intel logo, and other Intel marks are trademarks of Intel Corporation or its subsidiaries. Other names and brands may be claimed as the property of others.
Intel, the Intel logo, Intel Atom, Intel Core, Intel Xeon Phi, VTune and Xeon are trademarks of Intel Corporation in the U.S. and/or other countries.
*Other names and brands may be claimed as the property of others.
Microsoft, Windows, and the Windows logo are trademarks, or registered trademarks of Microsoft Corporation in the United States and/or other countries.
Java is a registered trademark of Oracle and/or its affiliates.
OpenCL and the OpenCL logo are trademarks of Apple Inc. used by permission by Khronos.

References

• Fix Performance Bottlenecks with Intel® VTune™ Profiler
• porkbun.com | parked domain
• d1hdbi2t0py8f.cloudfront.net/index.html?prefix=vtune-docs/
• oneAPI-samples/DirectProgramming/C++/StructuredGrids/iso3dfd_omp_offload at master · oneapi-src/oneAPI-samples · GitHub
• oneAPI-samples/Tools/VTuneProfiler/matrix_multiply_vtune at master · oneapi-src/oneAPI-samples · GitHub
• Download Documentation: Intel® Parallel Studio XE (Current and...
• Download Documentation: Intel® System Studio (Current and Previous)
• Intel® VTune™ Profiler Release Notes and New Features
• Intel® VTune™ Profiler Tutorials
• Get Started with Intel® VTune™ Profiler for Linux* OS
• Intel® VTune™ Profiler Performance Analysis Cookbook
• Intel | Data Center Solutions, IoT, and PC Innovation
• Intel® VTune™ Profiler User Guide
• GPU Compute/Media Hotspots Analysis (Preview)
• Hotspots Analysis for CPU Usage Issues
• Microarchitecture Exploration Analysis for Hardware Issues
• Performance Snapshot
• vtune Command Syntax
• Set Up System for GPU Analysis
• Get Help
• Open Intel® VTune™ Profiler
• Intel® VTune™ Profiler Installation Guide
• Intel® VTune™ Profiler Installation Guide
• Intel® VTune™ Profiler Installation Guide
• Intel® oneAPI Toolkits
• Compile Cross-Architecture: Intel® oneAPI DPC++/C++ Compiler
• Download the Intel® VTune™ Profiler
• Intel® oneAPI HPC Toolkit: Cluster & HPC Development Tools
• Intel® System Bring-up Toolkit, Powered by oneAPI
• Intel® System Bring-up Toolkit, Powered by oneAPI
• Intel® oneAPI Base Toolkit: Essential oneAPI Tools & Libraries
• Compiler Switches for Performance Analysis on Linux* Targets
• Compiler Switches for Performance Analysis on Windows* Targets
• Configure SSH Access for Remote Collection
• Microsoft Visual Studio* Integration

Read User Manual Online (PDF format)

Read User Manual Online (PDF format)  >>

Download This Manual (PDF format)

Download this manual  >>