Fullscreen Optimizations: Impact on Hardware Benchmarking

This article explores the impact of disabling Windows Fullscreen Optimizations (FSO) when benchmarking computer hardware. It covers what the feature does, how turning it off alters key performance metrics such as average frame rates, 1% lows, and input latency, and why the specific graphics API being tested—such as DirectX 11 versus DirectX 12—dictates whether disabling this setting yields more accurate and consistent hardware benchmark results.

Understanding Fullscreen Optimizations

Fullscreen Optimizations is a Windows feature enabled by default. When a game or benchmark is set to run in "Exclusive Fullscreen" mode, Windows intercepts this and forces the application into a hybrid "Borderless Windowed" mode instead. This allows the Desktop Window Manager (DWM) to render overlays (like the Xbox Game Bar or volume sliders) and enables instantaneous Alt-Tabbing. However, keeping the DWM active in the background introduces a layer of operating system overhead between the hardware and the display.

Impact on Average FPS and 1% Lows

When you disable Fullscreen Optimizations in an application's properties, you force Windows to surrender control of the display directly to the graphics driver, restoring true Exclusive Fullscreen mode.

  • Average FPS: In most hardware benchmarks, disabling FSO changes average frame rates by less than 1% to 2%, which typically falls within the margin of error.
  • 1% and 0.1% Lows: This is where the biggest difference occurs. Bypassing the DWM frees up slight CPU and GPU resources and prevents background Windows processes from interrupting the render queue. Consequently, disabling FSO often yields higher 1% and 0.1% low framerates, translating to fewer micro-stutters and a cleaner benchmark run.

Frame Pacing and System Latency

Hardware benchmarking relies heavily on frame time consistency (measured in milliseconds). When FSO is enabled, the DWM's compositing process can occasionally cause erratic frame pacing spikes, especially on multi-monitor setups with different refresh rates.

Disabling FSO results in flatter, more consistent frame time graphs in diagnostic tools like PresentMon or CapFrameX. Furthermore, because the GPU sends frames directly to the monitor without waiting for the Windows compositor, overall system latency (input lag) is measurably reduced.

The Role of Graphics APIs

The effect of disabling FSO depends entirely on the graphics API the benchmarking software uses:

  • DirectX 9, 10, and 11: Disabling FSO has a noticeable impact. These older APIs rely on traditional Exclusive Fullscreen to achieve peak performance. Turning off FSO removes the Windows wrapper and allows the hardware to perform without OS-level bottlenecks.
  • DirectX 12 and Vulkan: Disabling FSO has virtually no effect on benchmarking results. These modern APIs do not use traditional Exclusive Fullscreen; instead, they utilize advanced low-level flip models that naturally bypass DWM overhead even in borderless modes.

Best Practices for Hardware Benchmarking

To get the most accurate reading of raw CPU and GPU capability, you want to eliminate as many operating system variables as possible.

If you are benchmarking hardware using older games or synthetic tests running on DirectX 11 or earlier, disabling Fullscreen Optimizations is recommended. It isolates the hardware's performance by preventing the Windows Desktop Window Manager from skewing your 1% lows and frame time metrics. Conversely, if you are benchmarking modern DirectX 12 or Vulkan titles, changing this setting is unnecessary, as the API already manages display output at optimal efficiency. Regardless of whether you disable or enable it, the golden rule of benchmarking applies: keep the setting consistent across all hardware configurations being compared.