How to fix CRITICAL_PROCESS_DIED in Windows 11/10

Last updated on 

CRITICAL_PROCESS_DIED is a blue screen (BSOD) that shows up when a system process that Windows can't run without crashes or stops responding. Windows kills everything and reboots rather than risk data corruption. You might see it once and never again, or it might start happening every few minutes until you can barely stay on the desktop long enough to fix it.

Why this happens

  • A driver is crashing, usually something recently installed or updated
  • System files are corrupted
  • Hard drive or SSD is failing
  • A Windows update broke something
  • Malware messed with system processes
  • RAM is faulty

If you can't stay in Windows long enough to fix anything

Boot into Safe Mode. Restart your PC and interrupt the boot three times in a row (hold the power button as soon as you see the Windows logo, wait for it to shut off, turn it back on, repeat). After the third time, Windows will launch the recovery environment. Go to Troubleshoot → Advanced options → Startup Settings → Restart, then press 4 or F4 for Safe Mode, or 5 / F5 for Safe Mode with Networking.

If even that doesn't work, boot from Windows installation media and click Repair your computer to get to the recovery tools.

1. Uninstall recent drivers

A bad driver is the most common cause. Think about what changed right before the crashes started. New GPU driver? Printer? USB device?

In Safe Mode, open Device Manager (Win+X, pick Device Manager). Look for any device with a yellow warning icon. Right-click it, Properties → Driver tab, click Roll Back Driver if the option is available. If it's grayed out, click Uninstall device instead and let Windows reinstall a generic driver on reboot.

If you recently updated your graphics driver, that's the first thing to roll back. GPU drivers cause more BSODs than everything else combined.

2. Run SFC and DISM

Corrupted system files will keep crashing whichever process depends on them. Open Command Prompt as admin (in Safe Mode if needed):

sfc /scannow

If SFC finds problems it can't fix, hit it with DISM:

DISM /Online /Cleanup-Image /CheckHealth
DISM /Online /Cleanup-Image /ScanHealth
DISM /Online /Cleanup-Image /RestoreHealth

RestoreHealth downloads clean copies from Microsoft, so you need Safe Mode with Networking or normal mode for this. After it finishes, run sfc /scannow one more time and reboot.

3. Check the drive

A drive with bad sectors can corrupt system files in ways that SFC can't keep up with. Open Command Prompt as admin:

chkdsk C: /f /r

It'll ask to schedule the scan for next reboot. Type Y and restart. The /r flag scans for physical bad sectors, which takes a while but that's the whole point.

If chkdsk finds bad sectors, that usually means the drive is on its way out. It'll patch around them for now, but start thinking about a replacement.

4. Uninstall recent Windows updates

If the BSOD started right after an update, that update is the likely problem.

Settings → Windows Update → Update history → Uninstall updates. Sort by date and remove the most recent one. Reboot and see if the crashes stop.

If you can't get into Windows at all, you can do this from the recovery environment: Troubleshoot → Advanced options → Uninstall Updates. You'll get the choice between uninstalling the latest quality update or the latest feature update. Try quality first since those are more frequent.

5. Check for malware

Malware that hooks into system processes can trigger this exact BSOD. Worth ruling out, especially if you've been downloading things from sketchy sources. Run a full scan, not a quick one.

If you're using Windows Defender, open Windows Security → Virus & threat protection → Scan options, pick Full scan, and let it run. It'll take a while.

For a second opinion, download Malwarebytes and run a scan with that too. It catches stuff that Defender misses, and the free version works fine for one-off scans.

6. Check your RAM

Bad memory will cause random crashes, corrupted files, and BSODs that seem to have no pattern. Windows has a built-in tool for this.

Win+R, type mdsched.exe, hit Enter. Pick Restart now and check for problems. The test runs before Windows boots and takes 10-20 minutes.

Windows Memory Diagnostic is decent but basic. If you want a more thorough test, download MemTest86 and run it from a USB stick. Let it do at least two full passes. If either tool finds errors, you've got a bad stick and need to replace it.

If you have multiple RAM sticks, try removing them one at a time and running with just one to figure out which stick is the problem.

7. Use Event Viewer to find the actual cause

If none of the above helped, you need to figure out exactly which process is dying. Open Event Viewer (Win+R, type eventvwr.msc). Go to Windows Logs → System and look for Critical and Error events around the time of the last crash.

You're looking for entries that mention a specific process or driver name. Something like csrss.exe or svchost.exe crashing, or a driver file like nvlddmkm.sys (Nvidia) or atikmdag.sys (AMD). Once you know the name, you can search for that specific crash and find a more targeted fix.

The Details tab on each event usually has the faulting module name buried in the data. That's the piece you want.

Last resort: in-place upgrade

Download the Windows Media Creation Tool from Microsoft. Run it, pick "Upgrade this PC now", keep your files and apps. This replaces all Windows system files while leaving your stuff alone. If the crash was caused by corrupted system files or a broken update, this fixes it without making you set up your PC from scratch.