I'm running an Intel Core i7-14700K on Windows 11, with Arch Linux installed on a separate drive as well. Recently, demanding games such as Cyberpunk 2077 have been crashing during loading, while lighter games like Slay the Spire work normally. The applications close without producing a blue screen or shutting down the entire computer. Cinebench shows a similar pattern: the single-core and single-threaded tests complete, but the multithreaded test crashes almost immediately. Temperatures remain reasonable during testing, and I haven't found problems with my other components. XMP is enabled, but I haven't manually overclocked the system. Is there a reliable way to determine whether the processor is degraded without having another CPU available for comparison?
4 Answers
First establish a completely stock baseline: load BIOS defaults, disable XMP, remove any undervolt or overclock, and run the multithreaded test again. Also try temporarily lowering the CPU multiplier by a small amount. If the system becomes stable only when the CPU is slowed down or memory runs without XMP, that strongly suggests a CPU or memory-platform stability problem, although it isn’t definitive proof by itself. Run a dedicated memory test as well, since unstable RAM can look like a CPU failure.
A crash that occurs only in demanding software can also come from the graphics driver, GPU, power delivery, or corrupted game files, so it’s worth reinstalling the graphics driver and checking system logs. However, the fact that the multithreaded CPU benchmark fails almost immediately makes CPU, motherboard firmware, or memory stability more suspicious than a problem limited to one game. If it still fails with BIOS defaults and known-good memory settings, document the tests and pursue an RMA rather than continuing to run it outside specification.
The 13th- and 14th-generation desktop Intel processors have had documented stability and degradation issues. Since this appeared gradually and now happens under heavy all-core loads, update the motherboard BIOS to the newest version containing the latest Intel microcode and use the manufacturer’s recommended Intel default power limits. If the problem remains at stock settings, contact Intel or the system vendor about warranty service.
Because there is no blue screen, check Windows Event Viewer or Reliability Monitor for the exact failure. Look for WHEA hardware errors, application crash entries, or corrected CPU errors around the time of the failure. Also check whether any files exist under C:WindowsMinidump, though an application crash may not create a dump. Testing with XMP disabled, updated BIOS firmware, and Intel’s processor diagnostic tools can help separate memory, firmware, and CPU instability.
I haven’t manually overclocked the processor, but XMP is enabled. The applications close to the desktop rather than powering off, and temperatures stay within normal limits during the crashes.