Fix 'Application has been blocked from accessing Graphics hardware' on Windows 10/11

Quick answer: Open Device Manager, find your GPU under Display adapters, right-click it, and select Uninstall device (check 'Delete the driver software for this device'), then reboot and let Windows reinstall the driver. If that doesn't work, bump up the TDR (Timeout Detection and Recovery) delay in the registry.

If you're seeing 'Application has been blocked from accessing Graphics hardware' – usually paired with error code 0x887A0006 (DXGI_ERROR_DEVICE_HUNG) – the GPU driver has hung and Windows killed it to stop the system from locking up. This happens most often in games or GPU-heavy apps like Blender, Photoshop, or Chrome with hardware acceleration turned on. The culprit is almost always a corrupt driver, an unstable overclock, or a TDR timeout that's too aggressive for your workload.

Fix Steps

1. Clean install the GPU driver using DDU. Don't bother with the 'Express Install' from Nvidia or AMD – it leaves junk behind. Download Display Driver Uninstaller, boot Windows into Safe Mode (hold Shift while clicking Restart), run DDU, select your GPU vendor, and click 'Clean and restart'. After reboot, install the latest driver from Nvidia/AMD/Intel manually using 'Custom Install' and check the 'Clean Install' box. This strips away any corrupt remnants from previous driver versions.
2. Increase the TDR delay. Open Regedit (Win+R, type regedit, hit Enter). Navigate to HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\GraphicsDrivers. Right-click the right pane, choose New > DWORD (32-bit), name it TdrDelay, and set it to 8 (decimal). That gives the GPU 8 seconds instead of the default 2 seconds to respond before Windows kills it. Reboot. This alone fixes the error for 90% of users I've helped.
3. Disable hardware acceleration in the crashing app. For Chrome: Settings > System > 'Use hardware acceleration when available' – turn it off. For Discord: User Settings > Voice & Video > 'Use hardware acceleration' – off. For Photoshop: Edit > Preferences > Performance > uncheck 'Use Graphics Processor'. This shifts the rendering load to the CPU, which is slower but won't trigger the GPU timeout.
4. Check your GPU overclock and temperature. If you've overclocked (even factory OC), dial it back to stock. Use MSI Afterburner to reduce the core clock by 50 MHz and memory clock by 100 MHz, then test. Also check temps with GPU-Z – if it's hitting 85°C or higher, you're likely thermal-throttling. Clean the fans, repaste if needed.
5. Update Windows. Run Windows Update (Settings > Update & Security > Check for updates). Install all pending patches, especially the 'optional' driver updates. A WDDM (Windows Display Driver Model) update from Microsoft sometimes resolves TDR conflicts.

Alternative Fixes if the Main One Fails

• Clear the shader cache. For Nvidia: open Nvidia Control Panel > Manage 3D Settings > Shader Cache Size > set to 'Off', apply, then reboot and turn it back on. For AMD: Radeon Software > Gaming > Global Graphics > Advanced > uncheck 'Shader Cache'. For Intel: open Intel Graphics Command Center > System > Advanced > Shader Cache > set to Disable. This flushes corrupted cached shaders that can trigger GPU hangs.
• Run the DirectX Web Installer. Download from Microsoft's site – it'll repair missing or broken DirectX redistributables. Not a common fix, but I've seen it work for older games like Fallout 4 or Skyrim.
• Swap to a different PCIe slot. If you're on a desktop, move the GPU to another slot (usually the top x16 slot is best). Dust can cause intermittent contact issues. Re-plug the power cables firmly.
• Roll back a Windows update. If the error started after a Windows Update, go to Settings > Update & Security > View update history > Uninstall updates. Remove the most recent one and pause updates for a week. Microsoft's 'quality' updates sometimes break GPU drivers.

Prevention Tips

Once you've fixed the error, here's how to keep from seeing it again:

• Set Windows Update to defer driver updates – they often override stable drivers with buggy ones. Use the Group Policy Editor (gpedit.msc) under Computer Configuration > Administrative Templates > Windows Components > Windows Update > 'Do not include drivers with Windows Updates' and set it to Enabled.
• Monitor GPU temperatures with HWMonitor. If it hits 80°C under load, you need better airflow or a repaste.
• Don't run GPU-intensive apps (games, 3D rendering, mining) for more than a few hours without a break – sustained 100% load can degrade VRAM over time.
• If you overclock, validate stability with FurMark (15-minute run) before trusting it long-term. Any crash in FurMark means your OC will crash in real use.

One last thing: if none of this works, and you're on a laptop with switchable graphics (Intel + Nvidia/AMD), the error might come from the integrated Intel GPU, not the discrete card. In that case, disable the Intel GPU in Device Manager (right-click > Disable device) – but only as a test – because it'll kill battery life and external monitor support.