Configure the plugin through Project Settings

In Development Editor, open Edit > Project Settings > Plugins > Arm Neural Graphics Plugin 1.1.0 to configure the NFRU settings stored with the project.

After you save a setting that affects shaders or assets, cook the Windows content again. Then close the editor and run a Development, DebugGame, or Debug non-editor executable to observe the result. NFRU does not run in an Editor build.

Configure the running build with console variables

The following table summarizes key console variables for Neural Frame Rate Upscaling (NFRU) in Unreal Engine. These variables control enablement, debugging, performance tuning, and frame generation modes. Adjust them to optimize NFRU behavior for your development, testing, and performance needs.

Console variableTypeDefaultDescriptionNotes
r.NFRU.Enableint0Enables NFRU when set to 1.
r.NFRU.CaptureDebugUIint1Captures debug UI rendered on the first Slate DrawWindow call.Non-shipping builds only
r.NFRU.UpdateGlobalFrameTimeint0Includes interpolated frames in global frame-time and FPS calculations.
r.NFRU.ModifySlateDeltaTimeint1Sets the Slate delta time to zero during UI redraws to avoid NativeTick side effects.Slate Redraw UI mode
r.NFRU.PaceAdjusterint0Enables dynamic target-FPS adjustment.Set to 1 to enable
r.NFRU.UpAdjustFrameCountint40Sets how many frames above the target are required before the Pace Adjuster increases the target FPS.Requires r.NFRU.PaceAdjuster 1
r.NFRU.DownAdjustFrameCountint20Sets how many frames below the target are required before the Pace Adjuster decreases the target FPS.Requires r.NFRU.PaceAdjuster 1
r.NFRU.DataGraphOpticalFlowint1Selects the optical-flow method: 0 = Data Graph preferred, 1 = Data Graph only, 2 = shader-based.Availability depends on the plugin build
r.NFRU.DataGraphFrameGenerationint1Selects frame generation: 0 = NFRU preferred, 1 = neural, 2 = shader-based.Availability depends on the plugin build
r.NFRU.OnlyInterpolatedFramesint0Presents only interpolated frames for debugging.Development and test builds only
r.NFRU.ShowDebugViewint0Displays the NFRU intermediate-buffer grid.Development and test builds only

For example, enable NFRU and display its debug view with:

    

        
        
r.NFRU.Enable 1
r.NFRU.ShowDebugView 1

    

Tune frame pacing

Enable the Pace Adjuster when you want NFRU to change its target FPS in response to sustained performance changes:

    

        
        
r.NFRU.PaceAdjuster 1
r.NFRU.UpAdjustFrameCount 40
r.NFRU.DownAdjustFrameCount 20

    

r.NFRU.UpAdjustFrameCount controls how long performance must remain above the current target before the target increases. r.NFRU.DownAdjustFrameCount controls how long performance must remain below the target before it decreases. Larger values make the adjustment less reactive; smaller values make it respond sooner.

Use stat FrameGen while you tune these values. Compare the generated and paced statistics to determine whether the application produces frames consistently and whether presentation pacing matches the intended cadence.

Monitor frame generation performance with STATGROUP_FrameGen

NFRU provides the STATGROUP_FrameGen statistics group to monitor frame generation performance at runtime.

Run the following command in the Unreal Engine console:

    

        
        
stat FrameGen

    

Image Alt Text:Unreal console output from the stat FrameGen commandMonitor NFRU frame generation and pacing with stat FrameGen

This command displays runtime metrics such as generated frame rate, generation time, and pacing information. Use these values to assess both internal frame generation performance and final frame pacing.

StatStat IDDescription
Generated FPSSTAT_FrameGen_GeneratedFPSFrames per second produced by the frame generator. Compare this value with the engine FPS from stat FPS.
Generated MsSTAT_FrameGen_GeneratedMsAverage time in milliseconds to generate an interpolated frame.
Paced FPSSTAT_FrameGen_PacedFPSPresentation-paced FPS after swap-chain pacing is applied.
Paced MsSTAT_FrameGen_PacedMsAverage wall-clock time in milliseconds between paced frame presentations.

Interpret frame generation statistics

When reviewing statistics, consider the following:

  • Generated FPS and Generated Ms show internal frame generation performance.
  • Paced FPS and Paced Ms reflect actual presentation cadence after pacing or synchronization.

Typically, Generated FPS matches the engine FPS, while Paced FPS might differ if pacing or synchronization limits the display rate.

What you’ve accomplished and what’s next

You’ve now learned how to adjust and optimize the main console variables that control NFRU features in Unreal Engine, including enablement, debugging, and performance tuning. You’ve also used STATGROUP_FrameGen to monitor frame generation performance.

Next, you’ll explore how to use the debug view to visualize frame interpolation and validate NFRU output during development and testing.

Back
Next