Repeated crashes occurred without blue screens, likely due to insufficient CPU power.
Repeated crashes occurred without blue screens, likely due to insufficient CPU power.
Hello,
I've experienced some problems recently where my computer would shut down unexpectedly, especially while gaming. There were no blue screens or warnings—just a sudden hard restart. In Event Viewer, the only entries were Event 41 Kernel Power and Event 6008, which didn’t provide clear solutions. Over time, the problem became more frequent, making it nearly impossible to turn on the machine before it crashed again.
I ran diagnostics on all components: everything seemed normal—Memtest86 reported no errors, the CPU passed the test for five hours, and the GPU was functioning properly. The boot drive had been replaced, and self-tests came back clean. At this point, I suspect a CPU power issue, though it doesn’t seem logical since the system handled the tests well and the PSU could easily support it.
I tried changing the CPU’s power settings from Auto to AMD Overclock, which improved stability enough to play less demanding games. However, crashes still occurred under heavy load. I then switched from AMD Overclocking to Override mode, adjusting the voltage to 1.335 and 1.35, but the instability persisted. My CPU is now running at its default speed, and I haven’t attempted overclocking yet.
The remaining possibilities point to the CPU, motherboard, or power supply unit. Any suggestions on what to test next would be appreciated to help pinpoint the cause?
Specs:
- 5800X processor
- Noctua NHD-15 cooling fan
- MSI B550 A Pro motherboard
- Patriot Viper Steel DDR4 RAM (4x8GB, 32GB total)
- Gigabyte 3070 OC motherboard
- WD Blue SN580 NVMe 2TB HDD
- 2 TB Seagate Barracuda SSD
- 8 TB Seagate Barracuda SSD
- EVGA 850 B5 graphics card
EVGA 850 B5 - What is the age of the PSU in your system?
Patriot Viper Steel DDR4, 4x8GB (32 GB total)
Do you have a link to this RAM kit?
MSI B550 A Pro
What BIOS version applies to your motherboard?
The remaining possibilities for the problem seem to be the CPU, the motherboard, and the PSU.
Fixing involves narrowing down options, so you might need to find or obtain a well-made PSU with extra capacity, then check if the issue continues. EVGA has experienced inconsistent quality in some of their premium PSUs.
Now that you reference the motherboard BIOS version, I assume it hasn't been updated recently. I'll try updating it first.
The RAM consists of two identical sets of 2x8GB modules,
link
The power supply is only three years old.
I'll provide an update after the BIOS change and verify that all drivers are functioning properly.
Stress tests indicate a very immediate shutdown. Games usually crash during loading or after a short delay. The system also tends to stop even if nothing is open for a few minutes. Also, I have confirmed all drivers are up-to-date. I just updated the BIOS, and it is now running the latest version from MSI's site, which was released two months ago.
The system is currently unresponsive, and I have a possible clue. It appears that while I was pressing the lever on the GPU slot, something slipped and damaged one of the components. If anyone recognizes this part or knows if it’s causing the problem, please let me know.
Pic 1
Pic 2
I disassembled everything thoroughly, cleaned it completely, reworked it, and tested to check its performance. So far, the results look very promising. It has been running Prime95 for around 30 minutes without any signs of trouble, and I've noticed power usage, clock speeds, and temperatures all increasing, indicating significant stress. I'll follow up later to confirm if this truly resolved the problem after more testing, though it's possible that some conductive dust or loose connections were the cause.
Confirmed, system is back to normal operation. CPU completed Prime95 tests for several hours before switching to Linpack Xtreme, which ran without any issues until I powered it down after about 14 hours. Intensive games like Elden Ring are now functioning properly. I believe the problem has been fully resolved.