DXGI_ERROR_DEVICE_REMOVED

What is this error?

The GPU device was removed or stopped responding during a DirectX operation. This is the most common DirectX crash in modern games.

Common causes

  • GPU driver crash or timeout
  • overclocked GPU instability
  • outdated or corrupted GPU driver
  • GPU hardware failure or overheating
  • conflicting overlay software

How to fix it

  1. Update GPU driver via DDU clean install
  2. disable GPU overclock
  3. lower in-game settings
  4. check GPU temps

Too many steps? Crashless can diagnose this automatically — checks your drivers, temps, VRAM, and 400+ known error patterns.

Get free AI diagnosis

Detailed analysis

DXGI_ERROR_DEVICE_REMOVED means your GPU crashed mid-game. The fastest fix for most people: DDU clean install your GPU driver, disable any overclock, and close Discord overlay. If it keeps happening after that, read on — there's a specific cause and this guide will help you find it.

What's actually happening

Your GPU stopped talking to Windows. The DirectX graphics layer tried to send a command to your graphics card and got no response — the card either crashed, timed out, or physically disconnected from the PCIe bus for a split second. Windows kills the connection and your game dies.

The frustrating part: this error is generic. NVIDIA, AMD, and Intel GPUs all throw the same 0x887A0005 code, but the actual cause varies wildly. An RTX 4070 throwing this error at stock clocks is a completely different problem than an overclocked RX 7900 XTX doing the same thing.

The most common causes (in order of likelihood)

If you just want to shotgun fixes, start with #1 and work down. But knowing which category your crash falls into saves you hours.

Unstable GPU overclock — This includes factory overclocks from MSI, EVGA, or Gigabyte that push the card beyond its comfortable limits. Even an overclock that's been "stable" for months can start crashing as temperatures change seasonally or the silicon degrades over time.

Outdated or corrupted GPU driver — NVIDIA and AMD push driver updates that fix DXGI crashes in specific games. The driver version matters. For example, NVIDIA drivers in the 546.x-551.x range had known DXGI issues with Unreal Engine 5 games that were fixed in 555.85+.

GPU overheating — When your GPU thermal throttles above 83-90C (depending on model), it can slow down so much that Windows thinks it's unresponsive. The card isn't broken — it's just too hot to keep up.

Overlay and capture software conflicts — Discord overlay, GeForce Experience ShadowPlay, Steam overlay, MSI Afterburner's RTSS, and OBS game capture all inject DLLs into the game process. Sometimes two overlays fight over the same DirectX hook and one of them crashes the graphics pipeline.

VRAM exhaustion — Your GPU ran out of video memory. If you're running Ultra textures on a 6GB card with Chrome eating 500MB of VRAM in the background, the game has nowhere to put its texture data and the GPU crashes.

How to fix it

  1. DDU clean install your GPU driver. Download DDU (Display Driver Uninstaller) from guru3d.com. Boot into Safe Mode (hold Shift while clicking Restart > Troubleshoot > Startup Settings > Safe Mode). Run DDU, select 'Clean and restart'. Once rebooted, immediately install the latest GPU driver from nvidia.com or amd.com. Don't let Windows Update install its own generic driver first — it's always outdated. This single step fixes DXGI crashes for about 40% of people.
  2. Remove any GPU overclock. Open MSI Afterburner (or whatever you use) and click the reset button to return to stock clocks. If you don't have overclocking software installed but your GPU is a factory-overclocked model (like an 'OC Edition'), this probably isn't your issue — factory OCs are validated. But if you've manually touched core clock, memory clock, or voltage, reset everything.
  3. Check your GPU temperature during gaming. Download HWiNFO64 (free), run it in sensors-only mode. Play your game until it crashes (or for 30+ minutes if it doesn't). Check 'GPU Temperature' maximum. NVIDIA cards throttle starting at 83C and shut down at 92C. AMD cards throttle at 90-110C junction temperature. If you're hitting these numbers, your GPU needs better cooling — clean the fans, reapply thermal paste, or improve your case airflow.
  4. Disable overlay software one at a time. Turn off Discord overlay (User Settings > Game Overlay > off). Turn off GeForce Experience overlay (Settings > In-Game Overlay > off). Turn off Steam overlay (Steam > Settings > In-Game > uncheck overlay). Test after disabling each one. When the crash stops, you've found your culprit. You can leave the others enabled.
  5. Check your VRAM usage. Open Task Manager > Performance > GPU while gaming. Look at 'Dedicated GPU Memory'. If it's above 90% of your total VRAM, you need to lower texture quality, disable ray tracing, or close Chrome and other background apps that eat VRAM. VRAM-heavy settings to reduce first: Texture Quality, Shadow Quality, Ray Tracing, Render Resolution.
  6. Increase TDR timeout. Windows gives your GPU 2 seconds to respond before killing it. Some demanding scenes need more time. Open Registry Editor, navigate to HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\GraphicsDrivers, create a new DWORD called 'TdrDelay' set to
  7. This gives the GPU 10 seconds instead of
  8. This doesn't fix the root cause but stops the crash while you troubleshoot.
  9. Disable Hardware-Accelerated GPU Scheduling (HAGS). Go to Windows Settings > System > Display > Graphics > Change default graphics settings > turn off Hardware-accelerated GPU scheduling. HAGS can cause DXGI errors on some GPU/driver combinations. Restart after changing this.

Is this a hardware problem or a software problem?

Here's how to tell: If a DDU clean install at stock GPU clocks fixes it, it's software — you're done. If the crash persists with a clean driver at stock clocks across multiple games, your GPU might be dying.

To test GPU hardware: run FurMark or Unigine Heaven at stock clocks for 30 minutes. If it crashes or shows visual artifacts (colored pixels, flickering), your GPU VRAM may be failing. Run OCCT's VRAM test to confirm. If VRAM test shows errors, the GPU needs replacement — check your warranty.

One more thing to check: your PCIe power cables. Make sure you're using separate cables from the PSU to the GPU, not a single daisy-chained cable. Insufficient power delivery causes the GPU to brown out under load, which looks exactly like DXGI_ERROR_DEVICE_REMOVED.

Games where this is most common

Fortnite, Call of Duty: Warzone/MW3, Apex Legends, Hogwarts Legacy, Cyberpunk 2077, Baldur's Gate 3, Elden Ring, PUBG, Dead by Daylight, Palworld, ARK: Survival Ascended, and most Unreal Engine 5 games. If you're crashing in a UE5 game specifically, the Unreal Engine error message will say 'D3D Device Being Lost' — same issue, same fixes.

Frequently asked questions

Q: Does DXGI_ERROR_DEVICE_REMOVED mean my GPU is broken?
A: Usually no. About 90% of the time this is a driver or software issue. Try a DDU clean install first. Only suspect hardware failure if the crash persists at stock clocks with a clean driver across multiple different games.

Q: Why does this only happen in one specific game?
A: Different games stress different parts of the GPU. A game that maxes out VRAM will crash on a 6GB card while other games are fine. A game with aggressive shader compilation can trigger TDR timeouts that other games don't. Check that specific game's subreddit — there's often a known fix.

Q: I just got a new GPU and I'm getting this error. Is it defective?
A: Probably not. New GPU installations almost always need a DDU clean install to remove old driver remnants. If you swapped from AMD to NVIDIA (or vice versa), there are leftover driver files that cause exactly this error. DDU in Safe Mode is essential after any GPU swap.

When to seek help

If this error keeps happening after trying the fixes above, it may point to a deeper hardware or system issue. Consider professional help if:

  • The crash occurs across multiple games or applications
  • You see the same error after a clean Windows install
  • Your PC is less than a year old (could be a warranty issue)
  • You smell burning or hear unusual sounds from your PC

Or let Crashless do the deep analysis for you -- our AI checks drivers, temps, event logs, and 400+ known patterns automatically.

Chat with AI about DXGI_ERROR_DEVICE_REMOVED

Describe your setup and get a personalized diagnosis in seconds. Free, no signup needed.

Get AI diagnosis

Chat with AI about this error

Describe your setup and what you were doing when the crash happened. Our AI checks against 400+ known crash patterns — free, no download needed.

or type your own question below

Paste or drop screenshots for better diagnosis

Let Crashless handle it

The desktop app scans your drivers, temps, VRAM, event logs, and 400+ known patterns — then walks you through the fix step by step.