What should I do about a blue screen or a kernel panic?
A crash is the system stopping deliberately because continuing would corrupt something. The message names a culprit — here is how to read it and what the pattern tells you.
- Difficulty
- intermediate
- Time
- 1 hr
- Read
- 4 min
- Safety
- warning
Short answer
Note the stop code or the panic report — that text is the diagnosis. Then ask what changed: new hardware, a new driver, a recent update, or nothing. Crashes that started after a change point at that change; crashes that appear from nowhere point at hardware, most often memory, storage or overheating. Test memory, check drive health, check temperatures, and update or roll back the driver the message names. Back up before any of it.
The crash screen looks like a catastrophe and is actually a safety feature: the system detected an inconsistency it could not continue past, and stopped rather than write nonsense to your files. The useful information is on the screen and in the log, and most people never read either.
Safety
Step by step
- Write down the code or take a photograph.On Windows, the stop code and any named file; on a Mac, the panic report offered at the next start. Searching that exact text plus your machine's model is the single most productive thing you can do, because most codes point at a category — memory, driver, storage, power.
- Ask what changed.New memory, a new drive, a new peripheral, a driver update, a system update, a new program that installs a driver such as security or virtualisation software. If the crashes began within days of a change, treat that change as the prime suspect regardless of what the code says.
- Note the pattern.Crashes always during the same activity point at the driver or hardware that activity uses — graphics during games, network during video calls, storage during large copies. Random crashes at idle point more strongly at memory, power or heat.
- Test the memory.Both platforms can run a memory test, and free bootable testers exist that are more thorough. Run it for a full pass, ideally overnight. Faulty memory is the most common cause of apparently random crashes and the easiest to confirm.
- Check the drive.Read the self-monitoring health data and run the file system check. A failing system drive produces crashes, freezes and corrupted files, and it will only get worse. If the health data is poor, stop diagnosing and get the data off.
- Check temperatures and dust.A machine that crashes under load and never at idle is often overheating. Listen for fans, feel for heat, and clean the vents and heatsink. Laptops that are years old and have never been opened are usually thick with dust in exactly the wrong place.
- Update or roll back the driver.If the message names a driver file, search it — it will usually identify graphics, storage, network or a security product. Update it from the manufacturer, and if the crashes began after an update, roll back to the previous version.
- Use safe mode to separate software from hardware.If the machine is stable in safe mode with a minimal set of drivers, the fault is in software or a driver. If it crashes in safe mode too, suspect hardware. This single test saves hours.
- Read the log if it is still not clear.Windows' reliability monitor gives a timeline of crashes against installed updates and software, which is unusually readable for a system tool. macOS's crash reports name the extension involved. Both connect the crash to what was installed the day before it started.
If it doesn't work
Crashes only when playing games or using video software
Cause: Graphics driver, graphics hardware, or power supply under load — Fix: Update the graphics driver from the chip maker rather than through the general update mechanism. If crashes persist, check temperatures under load; on a desktop, an ageing power supply is a strong candidate.
Started immediately after adding memory
Cause: Incompatible, faulty, or not fully seated modules — Fix: Reseat, then test one module at a time. Check the speed and type against the machine's supported list. Mismatched modules can pass a short test and fail under real load.
Crashes at random with no pattern, machine otherwise fine
Cause: Memory faults are the leading candidate; overheating and power supply next — Fix: Run a full overnight memory test. It either identifies the fault or eliminates the most likely cause.
Machine crashes during start and loops
Cause: A driver or update installed at the last shutdown — Fix: Boot into the recovery options and use safe mode or the system's undo-update facility. Modern Windows offers recovery automatically after repeated failed starts.
Mac restarts unexpectedly with a report mentioning an extension
Cause: A third-party system extension, often from security, virtualisation or peripheral software — Fix: Remove the software that owns that extension, or update it. Older extensions frequently break after a major system update.
Questions people ask
Is a single crash something to worry about?
No. One crash in a year is noise. A crash a week, or a cluster after a change, is a problem to work on.
Will reinstalling the system fix it?
It fixes software causes and does nothing for hardware ones, at the cost of a day. Test memory and drive health first — if either is faulty, the reinstall will crash too.
The screen appears and restarts too fast to read. What can I do?
Windows can be set not to restart automatically on failure, which leaves the message on screen. The reliability monitor and the crash dumps also record the code after the fact.
Should I use a driver updater tool for this?
No. Get the specific driver from the machine or component manufacturer. Generic updater tools are a common cause of exactly this problem.