Processor · critical

How to fix Machine Check Exception (fatal hardware error)

Processor troubleshooting guide

What this error means

The processor detected a hardware error it could not correct and stopped Windows to prevent data corruption. On current Windows versions the same event usually appears as WHEA_UNCORRECTABLE_ERROR (0x124); the older stop code 0x9C appears only when Windows' hardware error handling is not fully started, for example very early during boot. Either way, Event Viewer often records the details as "A fatal hardware error has occurred" with "Error Source: Machine Check Exception". The error is reported by the CPU, but the cause can be the processor, memory, power delivery, firmware or overclocking settings.

Updated

Applies to
Windows 11 and Windows 10 desktops and laptops with Intel or AMD processors
Difficulty
Intermediate
Before changes
BIOS changes and updates restart the PC; a failed BIOS update can stop it from starting

Reviewed by Matyáš Rada · Documentation checked: 2026-10-30 · Checked against Microsoft's bug check documentation; not tested on every hardware configuration.

Before you begin

Back up important files now: repeated hardware crashes can corrupt data. Photograph or write down the exact stop code and any text shown on the blue screen. If Windows cannot start, see the startup guide.

Work through these checks

0 of 7 complete
01Read the hardware error record in Event Viewer

The blue screen only says that a hardware error happened. Windows usually saves which component reported it, and that decides where to look next.

  1. Press Win + R, type eventvwr.msc and press Enter.
  2. Open Windows Logs > System, then choose Filter Current Log on the right.
  3. In Event sources select WHEA-Logger and click OK.
  4. Open the newest event with ID 18 (fatal error). Write down Reported by component, Error Source, Error Type and Processor APIC ID.
  5. Also note any events with ID 19 (corrected errors) shortly before the crash.
What to expect

You have a record such as "Component: Processor Core, Error Source: Machine Check Exception, Error Type: Cache Hierarchy Error". Cache Hierarchy, Internal or Bus/Interconnect errors point first to CPU, memory controller or overclock stability; Memory Hierarchy errors point toward RAM.

If it does not help

If there is no WHEA-Logger event, the crash may have happened before Windows could save it. Continue with the next steps and check again after the next crash.

Browse related errors
02Return the BIOS to default settings

Overclocks, undervolts and memory profiles are the most common reason a healthy CPU reports machine check errors, and they can become unstable months after they seemed fine.

  1. Photograph your current BIOS settings so you can restore them later.
  2. Restart, enter the BIOS/UEFI (usually Del, F2 or F10 during startup) and choose Load Optimized Defaults or Load Setup Defaults.
  3. Make sure XMP/EXPO memory profiles, Precision Boost Overdrive, Curve Optimizer, manual voltages and multipliers are off.
  4. Save, restart and use the PC normally for a few days.

Do not raise CPU voltage to stop the crashes. Extra voltage can permanently damage the processor.

What to expect

If the crashes stop at default settings, the hardware is probably fine and one of the tuning settings was unstable.

If it does not help

If you want the performance back, re-enable one setting at a time, starting with the memory profile, and keep the last stable combination. If crashes continue at defaults, go to the next step.

Browse related errors
03Update the BIOS/UEFI

BIOS updates include new CPU microcode and memory compatibility fixes. Several processor families have had stability problems that were fixed or reduced only through microcode updates.

  1. Find your exact motherboard or laptop model: press Win + R, type msinfo32 and look at BaseBoard Product or System Model.
  2. Download the latest BIOS only from the manufacturer's support page for that exact model, and read its release notes.
  3. Follow the manufacturer's update instructions. Connect a laptop to its charger and do not switch the PC off during the update.
  4. After updating, load defaults again, because some updates reset or change settings.

A BIOS update that is interrupted can leave the PC unable to start. Skip this step if you cannot guarantee stable power.

What to expect

The system runs the newest microcode and memory support for your board.

If it does not help

If you have an Intel 13th or 14th generation Core desktop processor, see the separate Intel 13th/14th gen instability guide on this site: microcode updates prevent further degradation but cannot repair a processor that is already affected.

Browse related errors
04Check temperatures under load

An overheating processor throttles first, but a badly seated cooler or dried-out paste can still cause errors during sudden load.

  1. Install a hardware monitor such as HWiNFO (free) and watch the CPU package temperature while playing a game or running the task that usually crashes.
  2. Compare the maximum with the limit published by Intel or AMD for your processor model.
  3. Check that all fans spin, the vents are not blocked and, on desktops, the cooler is firmly mounted.

Switch off the PC and unplug it before opening the case.

What to expect

Temperatures stay below the manufacturer's limit with some margin.

If it does not help

If temperatures reach the limit, clean the dust, check the cooler mounting and replace the thermal paste, or have it done by a service. Then test again.

Browse related errors
05Test the memory

The memory controller sits inside the processor, so faulty RAM or a memory profile the controller cannot handle often shows up as a CPU machine check error.

  1. Save your work, press Win + R, type mdsched.exe and choose Restart now and check for problems.
  2. For a more thorough test, run MemTest86 from a USB stick for at least four passes.
  3. If you have more than one RAM module and errors appear, test the modules one at a time in the slot the motherboard manual recommends.
What to expect

The test finishes with no errors, or it identifies a faulty module.

If it does not help

Replace a module that fails. If all modules pass but the errors started after adding RAM, try the original amount or the modules on the motherboard's supported memory list.

Browse related errors
06Look for a pattern in the processor core

The Processor APIC ID in the WHEA record tells you which CPU core reported the error.

  1. Compare the Processor APIC ID across several WHEA-Logger events with ID 18 and 19.
  2. Note whether the errors happen at idle, during light use, or only under heavy load.
What to expect

If the same APIC ID appears every time at default settings, one core is likely faulty. If the ID changes, power delivery, memory or settings are more likely than a single bad core.

If it does not help

If you can borrow a compatible processor, power supply or RAM kit, swap one part at a time to confirm the cause before buying anything.

Browse related errors
07Contact the manufacturer with the evidence

A processor that keeps reporting machine check errors at default settings, with an up-to-date BIOS and normal temperatures, is likely faulty and may be covered by warranty.

  1. Collect the stop code, the WHEA-Logger event details, your BIOS version, temperatures and memory test results.
  2. Contact the PC maker for a laptop or prebuilt PC, or Intel or AMD support for a boxed processor, and ask about a warranty replacement.
What to expect

Support can decide quickly whether the processor, motherboard or another part needs replacing.

If it does not help

Until it is resolved, avoid heavy workloads and keep backups current.

Browse related errors

Possible causes

  • Unstable overclock, undervolt or memory profile (XMP/EXPO, PBO, Curve Optimizer)
  • Outdated BIOS/UEFI or CPU microcode
  • CPU overheating or a poorly seated cooler
  • Faulty or incompatible RAM, or an overloaded memory controller
  • Unstable power from the power supply or motherboard
  • Degraded or defective processor

Useful Windows tools

These commands open Windows tools. Review changes before applying them.

eventvwr.msc
mdsched.exe
msinfo32

Sources & further reading