cancel
Showing results for 
Search instead for 
Did you mean: 

Cold reboot Kernel-Power 41 (63)

everdrone
Level 7
Hi all, I'm new here and been working on this problem for almost a week now.

My setup for context:
HX1000i
7950X
X670E Hero
MSI 4090 Gaming Trio
EK AIO 360
2x16GB G.Skill DDR5 6000MHz CL36 (QVL)

Basically sometimes my PC reboots, no BSOD, no minidump. just reboot.
I tried installing a clean windows 11 multiple times to do some tests and the problem occurred with every install.

I even managed to find a "manual" trigger for the crash by opening (with any software, quicklook, windows photos, photoshop) a very large 20000x13000@24bit jpg image. Every time I open that one or a similar one, the PC reboots (i can provide urls if you have a similar system and are willing to help out)

Other things I have tried that DID NOT WORK:
- memtest86 4/4 0 errors
- sfc /scannow
- chkdsk /f /x /r
- dism /Online /Cleanup-Image /RestoreHealth
- disabling EXPO
- enabling EXPO
- Resetting BIOS to default
- Updating windows
- Updating drivers
- Clean install windows
- Enabling ECO Mode

Things I tried that did work:
- Running in safe mode (image loads fine, no crash, no reboot)

At this point no BIOS updates have been arriving for a long time and I'm kind of losing hope.
There's also the PCIE slot issue that someone else mentioned in another thread about the GPU not fully clicking into the slot, probably because of the glass next to the SATA connectors.
8,717 Views
29 REPLIES 29

No idea. I am running 0923 with all default settings EXCEPT c-state.

Seems stable for the most part...

Just saw they published 0925 yesterday so I might give that a try.

Oddly, I was having exactly the same issues as OP.  Disabled C-states in the bios and haven't had a kernel crash since then.  Some random threads have commented that this could mean I have a mildly defective CPU that can't throttle back up out of C6, but without buying a second CPU, I can't test whether a new one fixes the issue.

Set minimum CPU usage to 5% in the power settings.

Silent_Scone
Super Moderator
Hello,

The HX1000i by default operates in multi-rail mode. Have you tried switching to a single rail to see if the issue persists? The switch is located on the port-facing side of the PSU.
13900KS / 8000 CAS36 / ROG APEX Z790 / ROG TUF RTX 4090

Darox
Level 9
sounds like power supply playing check your power supply

Kymer
Level 8

Hi all, also new here.  Having the exact same issue as everyone here.  My setup is as follows:

Windows 11, fully updated running High performance power plan.

AMD 7950X3D, overclocked using PBO and curve optimizer

ROG Crosshair  X670E Hero, Bios 1101

Dual Gskill 32GB RAM, f5-6400j3239g32gx2-tz5rk, running DOCP profile 1, overclocked to 6000

I can confirm it's been happening for about a week for me now, just random reboots with no BSOD.  I have none of these kernel events prior to about a week ago and had already been running overclocked for at least two weeks prior to that. 

I'm curious what everyone's power plan is because about that time, I believe I'd been messing with power plan options?

 

Shenny
Level 11

Just to share my experience about something similar that happened to me last year. I had EVGA RTX 3080 FTW Ultra and PSU Corsair RM 850x that were happily running together for two years, until all of a sudden BSODs and reboots started. Sometimes I managed to play a game at the lowest settings for an hour without an issue, sometimes the reboot happened immediately after launching the game. Once or twice the PC rebooted on the desktop. After I replaced the PSU with RM 1000x everything fell into place. Later I read that GPU and CPU may have power spikes for milliseconds that are enough to throw the PSU off the rails if it is not stable enough. And in my case I had used this PSU for two years with the same hardware before it started to fail me so there is no guarantee that something which has worked before will continue to do so.

So, in my opinion, it is possible that the reboots in your case are caused by similar power spikes from the GPU/CPU and the PSU cannot handle them. It may be also due to faulty components on the motherboard that must deliver stable power supply to the components but I have no experience with such issues, thank God.

Kymer
Level 8

I may have accidentally fixed this issue by explicity enabling c states in bios. In my case, it appears the event would occur under minimal load not under high load as some have mentioned.  Give it a shot.  I went from 4ish events a day to none.

C states should be enabled by default.

Yes agreed, it was set to auto previously but I haven't changed anything else and I've now had zero reboots in the last day, normally hitting 4 or more.  Then after rereading this thread saw other people had seen similar results from enabling or disabling so maybe this wasn't quite implemented correctly by the bios team.