What this section changes
The display-driver timeout back at the Windows default, and the window and taskbar animations turned off. What either does to the frame rate has not been measured.
| Setting | What it does | Registry path |
|---|---|---|
| Restore the display-driver timeout to the Windows default | Deletes TdrDelay so Windows uses its own 2-second timeout. That timeout is how long Windows waits before rescuing a graphics card that stopped responding — raising it does not make anything faster, it only makes a stall last longer, with no picture, cursor or input for the whole time. Earlier builds of w0a set 10 s here; this puts it back. | HKLM\...\GraphicsDrivers\TdrDelay |
| Turn off window animations | Minimize, maximize and taskbar animations stop, so windows minimize and restore at once. They are drawn on the same GPU the game uses; what turning them off gives back in a game has not been measured (transparency is its own tweak on the Optimizations page) | HKCU\...\WindowMetrics\MinAnimate + Explorer\Advanced\TaskbarAnimations |
Applying it
- Open w0a
- Go to the Regedit Tweaks page in the left menu
- Find this section and press Apply on the entries you want
Every entry has its own button, so you are not required to apply a whole section — Apply All at the top of the page is there for when you do want all of them.
Undoing it
Two ways, both of which work:
- Backups & Restore — puts back the checkpoint written before the change was applied. It
puts back everything applied after that point, later checkpoints included — the confirmation says exactly that before it starts.
- The Fix page — searched by symptom, writes the Windows default unconditionally. It does not
read a backup, so it works even if w0a's own backup file is gone.
The undo log lives at %ProgramData%\w0a\backup.json, outside your user profile.
The risk
Everything here is a documented value Windows supports; none of it is kernel-level and none of it is irreversible. The honest sentence is still this: a setting may have no measurable benefit for you. If you apply a change and feel no difference, you are probably right — these pages are not selling you a feeling, they are telling you what was changed.
If you want to measure it, the rule from reading the benchmark applies: change one thing at a time, and read the lows rather than the average.
Did this page help?
What was missing, or what did not work? We read every one of these.
If this page did not solve it, write to us — say which article you read, and we will fix the article.