Frame Analysis: Analyze Frames for Performance Bottlenecks
Use Intel® GPA Graphics Frame Analyzer to identify bottlenecks in the graphics pipeline. Capture a stream (a sequence of frames), focus on a particular frame, and explore API calls and graphics resources of this frame.
To start frame analysis:
Run Graphics Monitor by double-clicking the
Graphics Monitor icon in the taskbar notification area or on the desktop.Follow Step 1 to collect data (
):A.1. Select a data collection type:
Select the Frame [DX11] startup mode for DirectX 11 applications.
Select the Frame [DX12] startup mode for DirectX 12 applications.
Select the Stream startup mode to capture a sequence of frames until the application is stopped
A.2. Select an application for analysis:
Select an application to analyze. Use the Browse
button or select from the application history.A.3. Run the application by clicking the Run Application button.
The application launches in a separate window. By default, stream capture starts immediately when the application is started.
NOTE:To get additional information, click
.
Follow Step 2 to analyze data (
):B.1. Select the Frames and Streams tab
B.2. Open the collected data. Double click the Preview image or click Open Stream/Frame.
Multiframe View of Frame Analyzer opens.
For detailed analysis, open a frame that appears to perform poorly. For example, identify the longest frame by analyzing the GPU Time Elapsed metric, which shows the amount of time spent on frame rendering:
On the GPU Time Elapsed track, click the longest frame (
).Click the Open button on the lower left (
).
Profiling View window of Frame Analyzer opens.
Start analysis with the Advanced Profiling Mode, which allows you to identify potential bottlenecks. Click the
Advanced Profiling Mode button on the top left.The API calls are grouped by bottlenecks on the Visualization (
) and API Log ( ) panes. The first group of draw calls is the most impactful and time-consuming.
Start resolving issues with the most impactful bottleneck. Click the first group of draw calls either in the Visualization or API Log pane.
The resources associated with this group of calls appear in the Resources pane. The Bottlenecks tab presents the graphics pipeline color-coded by bottleneck distribution. Red portions of the pipeline indicate main bottlenecks, and orange indicate secondary bottlenecks.
In the Bottlenecks tab, click the red bottleneck, for example, Local/Host Memory: Graphics-to-Memory Interface. The bottleneck cause, hints, and related metrics appear.
Once step 10 is done, close Graphics Frame Analyzer window.
As a result, you can identify potential bottlenecks of your application and get hints on how to resolve issues.
Next Steps
To take a closer look at the features of Graphics Frame Analyzer, including working with single and multiple frames, watch short video series In Depth: Graphics Frame Analyzer.
To learn more about frame analysis workflow, refer to the user guide section Analyze GPU-bound Applications with Graphics Frame Analyzer.
To learn about bottlenecks identified by Graphics Frame Analyzer, refer to the cookbook topic Performance Optimization for Intel® Processor Graphics.