VBS and Game Mode: What Costs FPS in Windows 11
Measure what actually costs frames on Windows 11: memory integrity's 5-15% hit, Game DVR background recording, and the settings that are safe to leave alone.
Windows 11 ships with several features that quietly cost frames, and most guides either tell you to turn everything off or claim "Windows 11 is fine, don't touch it". Both are wrong. The honest answer: two of these features - memory integrity (VBS/HVCI) and background game recording - have measurable costs in CPU-bound games, while Game Mode and GPU scheduling are usually neutral or helpful. This guide shows you how to measure each one on your own machine, so you can turn off only what actually costs you frames and keep the security features you should keep.
Start by finding out whether you are CPU-bound
Every one of these settings matters most when the CPU is the bottleneck - competitive shooters, simulation games, dense open-world cities at 1080p with a fast GPU. If your GPU is at 97 to 100 percent utilization during the slow parts, these tweaks will do almost nothing because the graphics card, not Windows, is the limit.
Check it in two minutes: press Ctrl+Shift+Esc to open Task Manager, go to the Performance tab, launch your game's built-in benchmark or a demanding scene, then watch GPU and CPU utilization. GPU pegged near 100 percent with headroom on the CPU means you are GPU-bound; skip to the Game DVR section and stop there. CPU pegged (or single cores maxed) while the GPU idles below 90 percent means the rest of this guide is for you.
Memory integrity: the biggest single cost
Virtualization-based security (VBS) runs part of Windows in an isolated environment that protects against certain kernel-level attacks; memory integrity (HVCI) is the piece that checks drivers. The price is that some memory operations that used to be direct now go through the isolation layer. In CPU-bound games the commonly measured cost is roughly 5 to 15 percent of average FPS, with 1% lows often hit harder than averages. The impact has shrunk across Windows versions and newer CPUs with faster virtualization support, but on most systems it has not reached zero.
Check whether it is on
- Press Win+R, type msinfo32, press Enter.
- Scroll to the bottom of System Summary and find Virtualization-based security.
- If it says Running, VBS is active. To see the HVCI state specifically: Settings > Privacy & security > Windows Security > Device security > Core isolation, and look at the Memory integrity toggle.
Measure before you decide
Do not flip the toggle on faith. Run your game's benchmark or a repeatable route three times with memory integrity on, note average FPS and 1% lows, then toggle it off in Core isolation, restart, and run the same three tests. On a CPU-bound game you will often see something like 144 average dropping to 128 with lows falling from 110 to 92 - or, on a newer machine with a GPU-bound workload, you may see no difference worth having, in which case turn it back on and keep the protection.
The tradeoff, stated plainly: memory integrity blocks a real class of driver-based attacks. Turning it off makes your system more vulnerable to exploits that abuse vulnerable signed drivers. If this PC holds banking sessions, work documents, or anything you cannot afford to lose to malware, and your FPS gain measured under 5 percent, leave it on. If the machine is a dedicated gaming rig and you measured a double-digit gain, the tradeoff is yours to make - that is exactly why you measured first.
Game DVR and the Xbox Game Bar
Background recording is the most underrated FPS thief, because it costs you even when you never press a record button. Windows can continuously buffer the last 30 to 60 seconds of gameplay "just in case" you want to clip something.
- Open Settings > Gaming > Captures (on older builds: Game DVR).
- Turn off "Record what happened" (background recording). This is the expensive one.
- Also check Settings > Gaming > Game Bar: if you never use the overlay for captures, disable the controller shortcut for opening it, since accidental overlays mid-match cause stutter.
Typical cost of active background recording: 3 to 10 percent of FPS plus noticeably worse 1% lows, depending on encoder. With it off, the Game Bar itself is harmless - it only activates when summoned.
Game Mode: leave it on
Despite its reputation, Game Mode does something modest and useful: it suspends most Windows Update activity and background app priority while a fullscreen game is in the foreground. It does not overclock anything and it does not degrade anything else on modern builds. Find it at Settings > Gaming > Game Mode. The old advice to disable it dates from Windows 10's 2017-2018 builds. On Windows 11 the measurable effect in most tests is neutral-to-positive; the people who report harm are usually describing a specific bug from a specific build two years ago. Check it is on and move on.
Hardware-accelerated GPU scheduling
HAGS lets the GPU manage its own video memory queues instead of the CPU doing it. Find it in Settings > System > Display > Graphics > Change default graphics settings. Effects are small either way - typically within 1 to 3 percent - but it is required for NVIDIA's DLSS 3 frame generation on RTX 40/50-series cards, and it can slightly reduce latency on newer GPUs. Practical advice: leave it on unless you have an older GPU (GTX 10/16 series) where it occasionally caused stutters, in which case test it off with your usual three-run benchmark.
The windowed-games optimization
Under the same Default graphics settings page, "Optimizations for windowed games" routes borderless-window games through the modern flip model that fullscreen games use. Borderless windowed is the default mode in most modern games, and on older titles this toggle can reduce latency and occasionally improve FPS for free. There is no downside on current builds; turn it on and re-test a couple of older games.
Power plan and the "Balanced" trap
On desktops, Settings > System > Power should be set to Best performance while gaming sessions matter; the default Balanced plan can let cores park or drop clocks during light load, which shows up as inconsistent 1% lows rather than lower averages. On laptops, plug in first - every power-related tweak in this guide is meaningless on battery, where thermal and discharge limits dominate everything.
Why 1% lows matter more than the average
Averages hide the damage these settings do. Memory integrity might cost 8 percent of average FPS while costing 15 percent of 1% lows, and 1% lows are what your hands feel - they are the stutters, the hitch when you round a corner, the missed frame during a fight. This is why "it costs nothing on my machine" arguments online are worthless: someone with a GPU-bound setup sees nothing, someone CPU-bound loses both average and stability. Track both numbers from your first benchmark run. If a change drops your average by 2 percent but your 1% lows by 12, that change hurt you far more than the average suggests, and you should revert it.
For repeatable testing, a free capture tool such as CapFrameX (built on Intel's PresentMon) records frame times per run and computes the same percentiles every time, which removes most of the human error from "does this feel faster". If you use one, lock every other variable: same game version, same save file or benchmark scene, same background apps, and ideally similar room temperature - thermal throttling on a dusty desktop can shift results by more than the settings you are testing.
The overlay and app traps that look like Windows settings
Three non-Windows programs produce FPS costs indistinguishable from OS overhead, and they should be cleared before you blame any system setting:
- GPU vendor overlays (the performance/statistics layers from driver software) each add a rendering hook; two overlays on top of each other are worse than either alone. Keep one active at most.
- Discord, browser hardware acceleration, and Spotify all render on your GPU. A hardware-accelerated browser with 20 tabs open on a second monitor can consume 3 to 6 percent of frame budget in CPU-bound games - close or suspend them for ranked sessions rather than tweaking Windows.
- RGB and peripheral suites with per-frame lighting effects occasionally leak a polling loop onto the CPU; if you see a background process holding a consistent few percent in Task Manager during gameplay, that is your suspect, not memory integrity.
A tested, numbered sequence
- Check CPU vs GPU utilization in Task Manager during a demanding scene.
- Run your game's benchmark 3 times; write down average FPS and 1% lows.
- Turn off background recording in Settings > Gaming > Captures. Re-run the benchmark. Keep it off - the cost is real on nearly every machine.
- Check VBS status in msinfo32. Toggle memory integrity off, restart, re-run the benchmark. If the gain is under about 5 percent, toggle it back on and keep the protection.
- Enable Optimizations for windowed games and confirm HAGS state; re-test only games where you actually notice something.
- Set the power plan to Best performance and re-run once more.
Write the numbers down at each step. Ten minutes of measurement beats every settings list on the internet, because every answer in this area depends on your CPU generation, your GPU, and whether the game you play is CPU-bound at your resolution.
If your stutter problem survived all of this, the likely culprit is not Windows at all - see fixing game stuttering on Windows 11 for the driver-and-background-process checklist, and reducing input lag if the issue is feel rather than frames.
By the SavePoint editorial team · How we check our guides
