Why does my PC keep crashing? How to find the cause
“Crashing” covers several different failures, and each points somewhere different. Windows records nearly all of them, usually with a code that narrows the cause down. Here is how to tell which kind you have, where Windows keeps the record, and where to start with each.
Updated 10 November 2026
First: which kind of crash is it?
- A blue screen, then a restart. A stop error: Windows found something so wrong that it
stopped to protect your files. The screen shows a stop code such as
WHEA_UNCORRECTABLE_ERRORorVIDEO_TDR_FAILURE, and that code is the best clue you will get. - It switches off or restarts instantly, with no blue screen. The power went, or the PC froze too hard to write anything down. Windows notes it at the next start as Kernel-Power 41.
- It freezes solid and you hold the power button. A hard freeze. Windows notes this as Kernel-Power 41 too, with a sign that the power button was held.
- The screen goes black for a few seconds, then comes back. The graphics driver stopped responding and Windows reset it: “Display driver stopped responding and has recovered”. VIDEO_TDR_FAILURE explains it.
- A game or program closes to the desktop. Only that program crashed. Windows records it as an Application Error naming the program and the part of it that failed.
Where Windows records it
Reliability Monitor is the easiest way in. Type reliability in Start and open View reliability history. Every day with a red cross had a crash; click it to see what it was — “Windows was not properly shut down”, a hardware error, or a program that stopped working — and open the details.
Event Viewer holds the same records in more detail. Under Windows Logs, System has blue screens and unexpected shutdowns (Kernel-Power 41, BugCheck 1001, WHEA-Logger), and Application has programs that crashed (Application Error 1000).
One thing to know: Windows writes some crash reports again every few hours until it manages to send them to Microsoft. One crash can appear dozens or hundreds of times in these lists, dated when the report was queued again rather than when the crash happened. Count the crashes by their dump files or their times, not by the number of entries.
The usual causes, most common first
- An unstable setting. An overclock or undervolt on the processor or graphics card, or a memory profile (XMP or EXPO) faster than this PC can hold. These pass light use and fail under a game, which is why switching them off is the first test.
- A driver. Most often the graphics driver, then the chipset, storage, network and sound drivers, and programs that install drivers of their own: fan and lighting control, anti-cheat, VPNs.
- Power. A power supply too small for the parts, or failing, or a cable not fully home — the graphics card’s above all. Crashes that come with load, and shutdowns with no blue screen, point here.
- Heat. A processor or card that gets too hot slows itself down, and if it gets hot enough, the PC switches itself off.
- A known fault in a part. Intel 13th and 14th gen desktop processors without Intel’s microcode fix (how to check), or an SSD on firmware with a known bug.
- A failing part. A memory stick, a drive, the processor or the graphics card itself — usually what is left once the rest is ruled out.
What each stop code points at
| Stop code | Points at |
|---|---|
| WHEA_UNCORRECTABLE_ERROR (0x124) | A hardware error the processor, memory or a PCIe device could not correct |
| CLOCK_WATCHDOG_TIMEOUT (0x101) | A processor core stopped responding |
| VIDEO_TDR_FAILURE (0x116) | The graphics driver hung and could not be reset |
| MEMORY_MANAGEMENT (0x1A) | Memory holding the wrong data, or a driver writing where it should not |
| DPC_WATCHDOG_VIOLATION (0x133) | A driver kept the processor busy too long, often storage |
| IRQL_NOT_LESS_OR_EQUAL (0xA), DRIVER_IRQL_NOT_LESS_OR_EQUAL (0xD1) | A driver reached for memory it should not — or unstable memory |
| KERNEL_DATA_INPAGE_ERROR (0x7A) | Windows could not read what it needed from the drive (check its health) |
| Kernel-Power 41, no code | Power, a hard freeze or overheating |
What to do, in order
- Write down the stop code, if there was one, and what you were doing when it crashed.
- Update the BIOS, and install the newest graphics driver with its clean install option.
- For a few days, switch off any overclock, undervolt and the memory profile (or load the BIOS defaults). If the crashes stop, bring the settings back one at a time to find the one the PC cannot hold.
- Test it under load: a stress test shows whether it holds at full speed and full heat.
- Check the power supply against the parts with the PSU calculator, and every power cable at both ends.
- If a crash names a driver, update it, or uninstall the program that brought it.
The dump file names the driver
After a blue screen Windows normally saves a small dump file in C:\Windows\Minidump.
Microsoft’s free WinDbg (from the Microsoft Store) opens it, and the command !analyze -v
shows the module that was running when it stopped — often the driver at fault. If that folder is
empty, check that Windows is set to keep dumps: System, About,
Advanced system settings, Startup and Recovery, Settings,
and under Write debugging information choose a small or automatic memory dump.