Fix NVIDIA Display Driver Stopped Responding (TDR)

Quick answer: For most users, bumping the TDR timeout via registry and switching your power plan to High Performance solves this within minutes. If not, do a clean driver install with DDU.

You're in the middle of a game, screen goes black, then Windows pops up that dreaded bubble: "Display driver stopped responding and has recovered." Happens on Windows 10 and 11, mostly with NVIDIA cards. The culprit here is almost always the Timeout Detection and Recovery (TDR) mechanism. Windows waits a certain time for your GPU to respond, and if it takes longer (often because a game is hammering the card), it kills the driver and restarts it. It's a protection thing, but it's overzealous. Sometimes it's also a power delivery issue — the GPU isn't getting stable voltage. Let's fix both.

Increase the TDR timeout. Open Registry Editor (Win+R, type regedit). Go to HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\GraphicsDrivers. Create a new DWORD (32-bit) named TdrDelay. Set its value to 10 (decimal). That gives your GPU 10 seconds instead of the default 2. Reboot. Done.
Check your power plan. Go to Control Panel → Power Options. Select High Performance. If you're on a laptop, also go into NVIDIA Control Panel → Manage 3D Settings → Power Management Mode → Prefer Maximum Performance. This stops the GPU from downclocking mid-render.
Update or rollback drivers. Sometimes the latest driver is the problem. Head to NVIDIA's driver page, grab the most recent Game Ready driver, but if the issue started after an update, roll back to the previous one. Use Device Manager or just reinstall the older driver.
Clean install with DDU. If you've got leftover driver files, they cause conflicts. Download Display Driver Uninstaller (DDU), boot into Safe Mode, run DDU, then install the fresh driver. This nukes all remnants. I've seen this fix issues that nothing else could.
Check your PSU and temps. If it's an older card, a failing power supply or overheating GPU can trigger TDR. Use HWMonitor to check temps. If your GPU idles above 50°C or hits 90°C under load, clean the fans and make sure the case has airflow.

Alternative fixes if the main ones fail

Disable TDR entirely — set TdrLevel to 0 in the same registry key. This stops Windows from checking, but you'll get freezes instead of recovery. Only do this if you know your hardware is stable.
Turn off hardware acceleration in your browser or Discord. That can cause TDR in some edge cases, especially with multiple monitors.
Use NVIDIA Inspector to lock your GPU clocks slightly below factory. Overclocking (even factory OC) can be unstable. Drop core clock by 50MHz and see if that helps.
Check your RAM. Yes, RAM issues can crash the GPU driver. Run mdsched.exe to test memory.

Prevention tip

Don't chase every driver update. NVIDIA's new releases month, but you only need them for new game optimizations. If you're stable, wait a week and read forums before updating. Also, if you're on a laptop, keep it on a cooling pad — thermal throttling is a sneaky TDR trigger. And always keep your BIOS updated; motherboard vendors fix PCIe stability issues that cause driver crashes.

Remember, the TDR timeout fix is a bandage. If it happens constantly, something deeper is wrong — hardware failure, unstable OC, or a dying PSU. Don't ignore it forever.