Display driver stopped responding and has recovered
What this error means
TDR is a protective mechanism: without it a hung GPU would freeze the whole computer. The driver restart itself is not the error but the reaction to it. nvlddmkm is NVIDIA’s kernel driver; AMD logs similar events as amdkmdag/amdwddmg. If failures happen in the browser and video playback too, the cause is more likely hardware or the driver than a particular game.
Where exactly the chain breaks
Causes — most common first
Step-by-step fix: from simple to advanced
Test the game after each step — only move on if the error is still there.
-
Do a clean driver reinstall medium ≈20 min
Profiles and caches left from the old driver keep failing. NVIDIA: pick custom install and tick “Perform a clean installation”. AMD: the Adrenalin installer has a “Factory Reset” option, or remove the driver with the official AMD Cleanup Utility. If the error appeared right after an update, roll back to the previous stable version from the same site instead.
-
Return to stock clocks (overclock, XMP/EXPO, undervolt) medium ≈15 min
An overclock that passes stress tests can still fail in a specific game or shader load. Reset the profile in MSI Afterburner / Adrenalin, remove the undervolt and, to test memory, temporarily disable XMP/EXPO in BIOS. If the error disappears at stock, lower clocks or improve stability.
-
Disable overlays and on-top tools easy ≈5 min
Overlays inject into the game process and hook frame presentation — a frequent cause of launch crashes and black screens. For the test, turn off: Discord overlay, Steam overlay (game Properties → “Enable the Steam Overlay”), NVIDIA App / GeForce Experience, AMD Adrenalin, Xbox Game Bar, MSI Afterburner + RivaTuner, screen recorders and “FPS boosters”.
-
Check temperatures under load medium ≈20 min
GPU temperature is shown in Task Manager (Performance → GPU) and in the NVIDIA/AMD panels; CPU temperature — in the motherboard or CPU vendor’s utility. If after 10–20 minutes of play the values hit the limit (GPU and CPU start throttling, a laptop runs fans at maximum), clean the cooling, check the fans and do not put a laptop on soft surfaces.
-
Disable hardware acceleration in background apps easy ≈3 min
If failures happen while video plays in another window (browser, Discord), temporarily disable hardware acceleration there to rule out multiple apps conflicting on the GPU.
-
Check the power supply advanced ≈20 min
If the PC shuts down, reboots or the GPU “drops out” only under load, the power supply is the first suspect: not enough wattage, an aged unit, adapters instead of native PCIe cables, a 12VHPWR plug not fully seated. Compare your PSU wattage with the calculated need for your GPU and make sure each GPU power connector has its own cable.
-
Update the motherboard BIOS advanced ≈30 min
BIOS updates fix memory and CPU stability (including known voltage issues on some generations) and compatibility with new GPUs. Download firmware only from the motherboard maker’s site for your exact revision and never power off during flashing. This is the last step — do it only if everything else failed.
Diagnostics for your PC
Enter your GPU, RAM and game — we will compare them with the requirements and point out what matters for your PC. If you are signed in and saved a PC, it is filled in automatically.
This is a hint based on specs, not a measurement. Browsers cannot see your driver version or power supply model: the FPSHQ app for Windows shows your driver version and real FPS.
Key figures
Unique FPSHQ metrics from this page's data
How long each step takes
Time estimate from the guide: 7 steps, about 113 minutes. Steps go from simple to complex — the error usually disappears on the first ones.
Display driver stopped responding and has recovered
Windows restarts the driver if the GPU does not respond for about 2 seconds (the TDR mechanism — Timeout Detection and Recovery). The screen flickers, the game crashes or shows DXGI_ERROR_DEVICE_REMOVED, and an nvlddmkm or Display 4101 event appears in the log. Causes: driver, unstable overclock or undervolt, overheating, power. Clean-install the driver, return to stock clocks, check temperatures and the PSU.
- steps in the guide
- 7
- minutes for all steps
- 113 min
- guide verified
- Oct 2, 2026
FAQ
How often is this normal?
Never. A single failure after a driver update or in a game beta is not alarming, but regular TDRs are a signal to work through the steps above.
Will reinstalling Windows help?
Only if software is the cause. With a hardware cause (power, heat, overclock) failures return. A clean driver install fixes the software side without reinstalling the OS.
What is TDR?
Timeout Detection and Recovery is a Windows mechanism: if the GPU does not respond for about 2 seconds, the system restarts the driver instead of freezing the whole PC. The “display driver stopped responding” message is the result of such a reset.
Where can I see the history of these failures?
Event Viewer → Windows Logs → System: events with source nvlddmkm (NVIDIA), amdkmdag (AMD) or Display with ID 4101. The Reliability Monitor (perfmon /rel) is also handy — it shows failures by day.
🧭 Nothing helped?
Write down the exact error text and the faulting module from Event Viewer (Windows Logs → Application) and contact the game developer’s support — attach the crash report if the game created one. Also check our related errors.
Related errors
Official sources
Facts and links checked: 2026-10-02