How to raise FPS on Windows: what really works and what is a myth

October 16, 2026

There are hundreds of “FPS tweaks” online, but most change nothing or stay within the margin of error. Here are the Windows settings that actually do something, and how to check the result on your own PC.

First, honest expectations

Windows settings will not turn a weak graphics card into a strong one. FPS depends mostly on the game itself, its graphics settings, the graphics driver and what runs in the background. Windows settings have a humbler job: get out of the game's way and make frames more even. So treat promises like “+40% FPS in one click” with healthy scepticism.

What is actually worth doing

Windows Game Mode. While a game is in the foreground, Windows gives it priority and postpones update installs and notifications. Average FPS usually does not change, but drops may happen less often.
Xbox Game Bar background recording. “Record what happened” keeps encoding the last minutes of your game. If you don't use it, turn it off: it frees the video encoder and some GPU resources, most noticeably on weaker cards.
Hardware-accelerated GPU scheduling. May slightly lower latency and CPU load, and is required for DLSS Frame Generation. Without support from the card and its driver it changes nothing. More in a separate article.
The “High performance” power plan. The CPU holds high clocks longer. On a desktop the difference is usually small; on a laptop the battery drains noticeably faster.

What helps only sometimes

Turning off fullscreen optimizations lowers latency in some older DirectX 9–11 games. In newer games there is usually no effect, while Alt+Tab gets slower and overlays or HDR may stop working.
Game priority in the multimedia scheduler mostly shows in games that register with that service. Often the effect is within the margin of error.

What not to do for FPS

Don't turn off Microsoft Defender, the firewall or Windows Update. The gain in games is tiny and the risk is real. Be careful with “tweak packs” from the internet too: many change dozens of values without explanation and without a way back.

How to check the result

The only honest way is to measure. Run the same scene in the same game before and after a change and compare not just average FPS but also 1% lows — that is where stutters show. NyxBoost has FPS Lab for this, built on Intel PresentMon, and every change is shown in advance and undone with one click. If a setting makes things worse, just revert it.