Question Freeze then restart?
Question Freeze then restart?
Hello there! Recently I've been having some trouble with my rig. Suddenly it freezes for about 5-10 seconds and then restarts. It
only happens during gaming i.e when the system is under load
. Rarely theres nothing else to it but once or twice this was accompanied by the common
watchdog_timout
error code. I've done some research and
tried updating my bios
as well as turning off any cpu enhancements and xmp profile but it still persists.
Temps are also good
as far as I can see. I'm writing here to see if any of you are familiar with this problem or have any tips on troubleshooting or thoughts, it would suck if I have to replace a component.
My specs:
OS: Windows 11 Pro 64 bit
GPU: Gigabyte RTX 4080 Eagle OC 16gb
MOBO: Gigabyte z690 Aorus Elite AX
RAM: 32gb 2400Mhz DDR5
PSU: Corsair RM1000X
What is the BIOS version for your motherboard?
RAM details: 32gb at 2400Mhz DDR5.
Do you have a link to the RAM kit you're using?
Processor information was omitted—please specify.
Cooling solution for the processor: How are you managing the heat?
Power supply unit model: Corsair RM1000X.
Age of the PSU in this build?
BIOS version is the latest, F31
RAM part number is "CMK32GX5M2A4800C40", Corsair
CPU: intel i9 12900k 3.2Ghz, water cooled with Corsair H100i Elite
All components are roughly four years old. Appreciate the update, no clear cause for this issue.
Reset the CMOS and observe the change. Check if the problem continues when using one stick of RAM in slot A2.
I completed three passes on memtest86 with no errors. During stress tests and benchmarks, the system consistently crashes under CPU load. The Windows Event Viewer displays a critical entry indicating kernel-power issues, with event ID 41 and a log noting 12 occurrences in the last week—possibly matching the number of crashes I’ve experienced. This suggests the CPU is likely the main problem. There are questions about what could be causing it, the meaning of kernel-power, and possible solutions like adjusting the vcore in BIOS or restricting CPU power to prevent turbo boost in Windows Power Options.
I completed three passes on memtest86 without any errors.
When using memtest, it's recommended to run it for 10 passes.
This suggests the CPU might be the main issue or my assumption is incorrect?
Have you considered that the CPU could be overheating because of a bad or failing cooler?
Perhaps I should adjust the vcore in BIOS or set the CPU power to 99% to turn off turbo boost in Windows Power Options?
If your BIOS was reset after clearing CMOS, then the CPU power settings should default to AUTO.
If you think your CPU is likely causing the problem, try moving it onto another stable motherboard with proper cooling and observe if the issue continues.
I might also borrow a solid PSU instead of buying one and test whether the problem persists on your current system.
Old firmware on the motherboard might lead to CPU microcode or timing problems, occasionally causing CLOCK_WATCHDOG_TIMEOUT. When did you upgrade the BIOS to F31? Before that, there were microcode updates that could harm a 12th generation processor. You might still be covered under warranty. I had a comparable issue with a 12900K that was replaced under warranty. It could be necessary to purchase a cheaper processor to confirm whether the fault lies with your motherboard or your CPU.
Updated several months ago and when this began happening, temperatures are normal with CPU peaks around 65 during the prime95 blend stress test. I've experimented with the BIOS and Prime95, trying different settings. It seems the issue lies with "Intel® Turbo Boost Technology." When enabled, the system crashes after a few minutes of testing, and I received a machine_check_error. However, when Turbo Boost is disabled, there are no crashes during stress tests or gaming, regardless of other adjustments. It appears that something happens when the CPU exceeds its base speed, causing instability. As long as the base clock remains the limit with Turbo off, the system stays stable. This suggests that keeping Turbo disabled might be the best option despite the performance loss. It's a temporary fix, not a permanent solution, and it doesn't resolve the underlying problem. The lack of warranty and receipt makes it hard to proceed further. What are your thoughts on this?