[SOLVED] Crashing: Kernel-Power 41, Bug Check Code 0

[bprescot]

Built a new system with the below specs.

CPU: 5600x
Mobo: Aorus b550i Pro Ax
SSD: Intel 1TB 665p
RAM: 2x 8Gb Ballistix 3600 MHz (BL2K8G36C16U4B)
PSU: Corsair SF750
GPU: EVGA 3060ti xc
Case: Velka 7, aka 'The Toaster Oven'

When the system runs, it runs great. While the case itself can radiate some heat, the components stay cool. With HWinfo running at max load, CPU tops out at 71, and GPU at 68. VMOS is a nice and chilly 52.

The problem is the system periodically crashes. Event Viewer lists Event ID: 41 "Kernel Power", and doesn't provide a bug-check code. It will crash several times in an evening, sometimes while under load, other times just browsing the web, trying to research why the system is crashing.

Solutions/Diagnostics tried:
1.) Memtest86 - 4 clean passes
2.) New GPU Drivers - Cleaned with DDU in safe mode, no joy.
3.) Swapped to new PSU - No luck
4.) Re-seated... everything.
5.) Uninstalled and re-installed the latest windows updates.

I'm honestly a bit lost on this one, since I don't seem to have a lot of information. The internet says it could be anything from the MoBo's power delivery, to bad power delivery from the home electric grid, to bad RAM, to evil sprites angry at me for not believing in angry sprites.

ANY guidance would be greatly appreciated.

 

[popatim]

Tried using Motherboard drivers from Asus?

 

[bprescot]

Wait, what? On a gigabyte board? I didn’t know that was a thing.

 

[bprescot]

I did not try to use another manufacturer's Bios, but I did update to Gigabyte's latest Bios.

Issue still occurs, sadly.

Are any of the below worth trying?
1.) Alternate Ram
2.) Alternate GPU
3.) Alternate Gen 3 PCIE riser cable
4.) Alternate SSD
5.) Clean Windows Install

Wait, what? On a gigabyte board? I didn’t know that was a thing.

 

[SuperIronOut]

Microsoft support for WHEA codes is terrible so it'll be a challenge to find anything on their support page about it. This is a shot in the dark, but it could be ASPM Link State Power Management causing the crash.
Right click the Windows icon on the far left corner of your task bar and at the top of the menu click: Power Options. On the far right you will see Additional Power Settings.
Once you have the Edit Plan Settings window open hit Change Advanced Power Settings.

The recommend plan for Ryzen processors is Balanced; the AMD Ryzen Balanced Power Plan is obsolete and has already been integrated into the default Windows Balanced Plan so make sure this is the one you have selected; you should not lose any performance from the change.
Once you've chosen the Windows Balanced Plan from the drop down if it wasn't set to this already, scroll down to PCI Express and hit the + to expand, and do the same for the Link State Power Management option that generates below it. It's usually at Moderate Power Savings by default, but try switching it to Off.

If the option is not available and you don't see it, close the Power Options window after hitting Apply (if you had to change it to the Windows Balanced Plan), then open up an elevated command prompt and type:

powercfg -attributes SUB_PCIEXPRESS ee12f906-d277-404b-b6da-e5fa1a576df5 -ATTRIB_HIDE

Once you've done this, return to the Power Options window and set ASPM Link State Power Management to Off and hit Apply. Watch your temps after making these changes and report back with your results.

 

[bprescot]

The good news: I was able to do as you recommended, and after the change, temps looked much as they did before.

The bad news: It had no effect.

I've got the case open, so I'm going to try a RAM swap since it's easy to do. My hopes are low that RAM is the cause.

 

[SuperIronOut]

Disabling Fast Startup might do the trick. Also, in Power Options is the Hard Disk option set to sleep after a certain period? You could try setting it 0 if it is not already. This seems to be something I've seen somewhere.
When does the issue normally happen? Is it during gaming or when idle?

 

[bprescot]

Both of those were actually already done. I should have mentioned that. Apologies.

Disabled XMP (DOCP), then swapped in a kit of Ripjaws V, but started seeing some odd behavior with the monitor randomly going dark with that kit. I remember having an issue with this kit before where it didn't play nice with a 3600 system, so swapped in some Trident Z Unicorn Vomit.

Been stable for the last 30 minutes, so that's a good start, but the old memory was also solid for long lengths of time... I'll just continue to use it through the evening and report back. If I go 72 hours without a crash, I'll post back, then re-enable xmp, and see if it goes another 72 hours... I hate these issues with what seem to be random occurences.

 

[LebronLopez]

Hey bro, did you solve the problem? I have the same problem, with the same error code, everything exactly the same, I changed PSU, I did ram memory tests, I reinstalled operating system, I made sure everything is compatible, etc. But it still doesn't work, how did you fix it?

 

[bprescot]

I'm not calling this one completely solved, but I haven't crashed in a while. The problem is that I did two things simultaneously; Swapped into yet another ram kit and disabled xmp.

Since then, it has been working perfectly. I turned xmp back on last night so we will see if the issue recurs. If it does, I'll turn it off again and switch back to the old ram kit to see if the issue was just xmp. If it doesn't, then I think the mobo or 5600x just didn't play nicely with the Ballistix or G.Skill ram kits that I had given to it before.

Hope that helps.

 

[bprescot]

Since re-enabling xmp I've had two kernel 41 crashes, where with it disabled, I hadn't had any in almost a week.

How much performance am i losing if I leave xmp disabled?
If the error is tied to xmp being enabled (and this is tested with two different ram kits), is that indicative of a bad board? Bad RAM compatibility? Does it lend itself to any solution ideas other than just leaving it disabled?
Please note this board is running the latest Bios version but that this issue also happened with the original bios version the board came with.

 

[LebronLopez]

I still have the same problem, I recommend that you try each component on another computer, you did it?

 

[LebronLopez]

I have 16GB of RAM, 4GB each module, I had done many tests at 8GB and 16GB, and I passed all the tests without problems, so it could only be the processor, the hard disk or the motherboard. However, I again removed 8gb of ram, and everything has been working fine so far, did you try doing that?

 

[bprescot]

This is an itx motherboard, so only two dimm slots. I can try pairing back and trying each dimm individually.

But for now, I'm running without xmp enabled. System is stable without xmp, so... actually let me triple check that various kits I have are officially supported by Gigabyte.

Edit: All the kits I have are officially supported. No idea why enabling XMP causes this issue.

 

[bprescot]

Manually set memory timings and dram voltage. After quick test, looks promising, but will need to test more tomorrow.

 

[LebronLopez]

This problem is a headache, keep trying, most likely the problem is in the ram, in my case it was the ram (I think)

 

[LebronLopez]

and although all the modules seem to be compatible, for some reason when I removed 2 memory modules and left only 2, the problem disappeared (so far)

 

[bprescot]

Manually setting the memory timings and voltage rather than simply enabling xmp seems to have solved the issue. Advice came from another forum, so will mark this answer as the "winner" even though that feels odd to do.