Short answer: heat rarely kills a CPU outright
Modern CPUs throttle and shut down to protect themselves well before heat causes physical damage. A chip that has been run hot usually shows instability, not sudden death. Long-term wear comes from high voltage, high current and high temperature together, not from heat alone.
Most crashes that look like a dead CPU are caused by RAM, the power supply, a graphics driver or the motherboard. Work through the checks below before you assume the processor is at fault.
Signs that point to CPU damage
- Boot failures. The PC does not power on, or stops at CPU initialization, or shows CPU debug LEDs or error codes.
- Crashes at light load. Freezes, blue screens or restarts during basic tasks or at idle, not only in heavy work.
- Instability that persists. Errors continue after you have cleaned the cooler, repasted it and reset the BIOS.
- Lower clocks or higher voltage needs. The CPU cannot hold its normal boost clocks, or needs more voltage than before to stay stable.
- Machine check errors. In Windows Event Viewer, WHEA-Logger events (such as IDs 18 and 19) or “Machine Check Exception” blue screens point to CPU or memory instability. They do not prove the CPU is damaged on their own.
Signs that point elsewhere
- RAM problems. RAM and CPU errors look alike. Test the memory with MemTest86 or the memory test in OCCT.
- Power supply problems. A failing PSU can cause random reboots that look like CPU faults.
- GPU and driver problems. Crashes that only happen in games, or after a driver update, often come from the graphics driver. A clean driver install can fix them.
- Overclocking or memory profiles. XMP or EXPO profiles and manual overclocks can cause instability on a healthy CPU.
Step-by-step: test a suspect CPU
- Fix the cooling first. Clean dust, check the fans, reseat the cooler and repaste if needed. For more detail, see how to stop a CPU from overheating.
- Reset to stock settings. Load the BIOS defaults or clear the CMOS. Disable XMP or EXPO and any manual overclock.
- Update the BIOS. Current versions can fix stability and voltage problems.
- Rule out the other parts. Test with known-good RAM, or run a memory test. If you can, swap in a known-good power supply or GPU.
- Run a stress test with a tool such as OCCT, Prime95, y-cruncher, Cinebench in a loop, or Intel’s Processor Diagnostic Tool. Watch for errors and temperatures. The stress test guide explains the settings.
- Check for per-core failures. If errors only happen on certain cores, a single core may be failing. Ryzen Master or HWiNFO64 can help you see which core it is.
If the system passes all of these at stock settings, the CPU is probably fine.
Check the socket with the power off
Power off and unplug the PC before you look at the socket. Look for:
- Burnt or discoloured pins, or a scorched socket area
- Bulging or burnt pads on an AM5 board
LGA pins bend easily. Do not force the processor, and do not try to straighten pins with metal tools. If you find burn marks, stop and get help rather than running the system.
Intel 13th and 14th gen instability
Intel has documented instability in some 13th and 14th gen desktop chips. The cause was elevated voltage that shifted the minimum operating voltage over time. Intel has released microcode updates, including revision 0x12B, to stop further damage, and it has extended the warranty on affected boxed desktop CPUs. The updates do not repair chips that are already degraded.
Common signs on these chips include shader compile or “out of video memory” crashes in Unreal Engine games, and decompression errors. Check Intel’s current support pages for warranty terms and the latest microcode before you assume the chip is damaged.
If you think the CPU is damaged
- Gather evidence. Note the symptoms, the temperatures, the stress test results and what you changed.
- Contact support. For a chip under warranty, contact the CPU maker or the seller. AMD support handles cases linked to the AM5 SoC voltage issue.
- Avoid long runs. An unstable system can corrupt data, so do not keep a suspect CPU running for long periods while you wait.
For cooling fixes and when a heat problem is still ongoing, see how to stop a CPU from overheating.