Intel VTune Profiler Performance Analysis Cookbook

Download PDF
ID 766316
Date 5/19/2023
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents
Document Table of Contents x
Intel® VTune™ Profiler Performance Analysis Cookbook
Intel® VTune™ Profiler Performance Analysis Cookbook x
Methodologies Configuration Recipes Tuning Recipes Notices and Disclaimers
Methodologies x
Top-down Microarchitecture Analysis Method OpenMP* Code Analysis Method Software Optimization for Intel® GPUs (NEW) Core Utilization in DPDK Apps PCIe Traffic in DPDK Apps DPDK Event Device Profiling Effective Utilization of Intel® Data Direct I/O Technology Compile a Portable Optimized Binary with the Latest Instruction Set
Configuration Recipes x
Profiling Machine Learning Applications (NEW) Profiling Single-Node Kubernetes* Applications (NEW) Analyzing Hot Code Paths Using Flame Graphs (NEW) Improving Hotspot Observability in a C++ Application Using Flame Graphs Measuring Performance Impact of NUMA in Multi-Processor Systems (NEW) Profiling Games built with Unity* (NEW) Ingredients Build the Game in the Unity Editor Configure Intel® VTune™ Profiler and Run Hotspots Analysis Review Results Profiling Games built with Unreal Engine* (NEW) Profiling Java Applications as a Remote User (NEW) Profiling JavaScript* Code in Node.js* Analyzing CPU and FPGA (Intel® Arria® 10 GX) Interaction Profiling a .NET* Core Application Profiling Applications in Amazon Web Services* (AWS) EC2 Instances Enabling Performance Profiling in GitLab* CI Configuring a Hyper-V* Virtual Machine for Hardware-Based Hotspots Analysis Profiling an Application for Performance Anomalies (NEW) Profiling an OpenMP* Offload Application running on a GPU (NEW) Profiling a SYCL* Application running on a GPU Profiling an FPGA-driven SYCL* Application Profiling Hardware Without Intel Sampling Drivers Profiling MPI Applications Profiling Docker* Containers Profiling a Remote Target Through a Proxy Server (NEW) Profiling in a Singularity* Container Profiling Linux*, Android*, and QNX* System Boot Time Using Intel® VTune™ Profiler Server with Visual Studio Code and Intel® DevCloud for oneAPI (NEW) Using Intel® VTune™ Profiler Server in HPC Clusters Using the Command-Line Interface to Analyze the Performance of a SYCL* Application running on a GPU (NEW)
Tuning Recipes x
Cache-Related Latency Issues in Segmented Cache Environment False Sharing Frequent DRAM Accesses Poor Port Utilization Page Faults Instruction Cache Misses Inefficient Synchronization Inefficient TCP/IP Synchronization OS Thread Migration OpenMP* Imbalance and Scheduling Overhead Processor Cores Underutilization: OpenMP* Serial Time Scheduling Overhead in Intel® Threading Building Blocks (Intel® TBB) Apps PMDK Application Overhead
Intel® VTune™ Profiler Performance Analysis Cookbook

Profiling Games built with Unity* (NEW)

Use this recipe to profile a game built with the Unity game engine. See how you can run Intel® VTune™ Profiler within the Unity environment to profile your game.

Often, the most important factor that affects the performance of a game is the frame rate. This is the speed with which the GPU renders game graphics. However, the CPU can also impact game performance in several ways:

  • Slow transfer of data to the GPU
  • Slow or unnecessary operations
  • Poor parallelism
With games that use the Unity engine, much of the optimization takes place in the Unity editor. Therefore, it is critical to understand the performance of Unity-defined tasks. The 2019.2 and newer versions of Unity have the Intel® VTune™ Profiler Instrumentation and Tracing Technology (ITT) API built into the Unity editor. This recipe demonstrates how you can run VTune Profiler to highlight Unity tasks in the editor.

Ingredients

Here are the hardware and software requirements for this performance recipe.

  • Application: Unity 2022.1.16. The sample game in this version of Unity uses the free asset 3D Game Kit.

  • Tools: Intel® VTune™ Profiler version 2022 - Hotspots Analysis (using User-Mode Sampling)

    NOTE:

    • Starting with the 2020 release, Intel® VTune™ Amplifier has been renamed to Intel® VTune™ Profiler.

    • Most recipes in the Intel® VTune™ Profiler Performance Analysis Cookbook are flexible. You can apply them to different versions of Intel® VTune™ Profiler. In some cases, minor adjustments may be required.

    • Get the latest version of Intel® VTune™ Profiler:

  • CPU/GPU: Intel® Core™ i7-8665U CPU @ 1.90GHz with integrated GPU

  • Operating system: Windows* 10 Enterprise

Build the Game in the Unity Editor

  1. Open the game in the Unity editor.

  2. Build the game. Make sure to select the Copy PDB Files and Development Build options.

Configure Intel® VTune™ Profiler and Run Hotspots Analysis

  1. Open Intel® VTune™ Profiler and click New Project on the Welcome screen.

  2. Specify a project name and a location for your project.

  3. Click Create Project.
  4. In the Configure Analysis window, set these options:

    • In the WHERE pane, select Local Host.

    • In the Application field of the WHAT pane, enter the path to the game executable.

    • In the Application parameters field, enter -profiler-enable-vtune-markers.

  5. (Optional) If you want to skip profiling the start/loading phase of the game,
    • In the WHAT pane, open the Advanced section.
    • Set Automatically resume collection after to the number of seconds that Intel® VTune™ Profiler should wait for profiling to begin.

  6. In the HOW pane, select the Hotspots analysis type and enable user-mode sampling.

  7. Click Start to run the analysis.

NOTE:
If you set the Automatically resume collection after option in step 5, only the Start Paused button ( ) is available.

Review Results

After the data collection runs for about 30 seconds, click Stop to exit the game and finalize the VTune results. This process may take a few minutes as Intel® VTune™ Profiler finds and resolves debug symbols.

Once results have been finalized, the Summary tab displays information about:

  • Elapsed time
  • Top hotspots
  • Top Unity tasks
  • Additional insights and guidance

The Frame Rate Histogram shows the frequency at which frames executed during the collection.

In this example, most frames executed between 13-14 frames per second (FPS). This is much slower than the recommended minimum of 30 FPS.

Switch to the Bottom-up window to see a list of functions. The default sorting is by descending order of CPU time.

Change the grouping from Function / Call Stack to Task Domain / Task Type / Function / Call Stack to focus on Unity tasks which were identified by the Intel® VTune™ Profiler Instrumentation and Tracing Technology API (ITT API).

The collection was started in a paused state, and this is indicated in the timeline view as Intel® VTune™ Profiler was still actively running.

Select the time period when collection starts at 60 seconds and zoom in.

In this example, the frame rate is low and there is considerable spin time. In general, a game should not consume a high percentage of the CPU capacity of the system. If the worker threads have low parallelism, they may be adding too much overhead.

In one of the RunThreadWrapper threads, all of the CPU time is spin or overhead. Reducing the number of threads can improve performance here.

For additional insights into hotspots, open the Flame Graph view to see a graphical representation of call stacks from the top down.

Parent topic: Configuration Recipes

See Also

Analyzing Hot Code Paths Using Flame Graphs Follow this recipe to understand how you can use Flame Graphs to detect hotspots and hot code paths in Java workloads.
Level Two Title