Folding@home BSOD

[Skiiwee29]

Alright friends,

 

Need some help as I can't seem to figure this out. When Folding@home, I get BSOD's with the Whea Uncorrectable Error message. I know this is (usually) related to overclocking so I set out to dig in.

 

Things I have tried:

 

BIOS flashing, went from 3601 to 3801, reflashed twice, then back down to 3601.

With and without XMP/DOCP

Disabled PBO in BIOS

Reset BIOS to optimized Defaults

Upped DRAM voltage to 1.4v and set SOC to 1.15.

Ran Memtest86 with 2 full passed with XMP enabled with no errors found (Before DRAM/SOC Voltage changes)

System can run any stability stress test or benchmark out there with zero issues... Prime95, Aida64, Cinebench on loop, etc... 

Uninstalled and reinstalled Folding@home and tried both 7.5.1 as well as 7.6.21. I have 2 other machines running 7.6.21 with zero issues. 

 

Kind of at a loss here. Any help/suggestions you got, throw them at me. 

 

System Specs in my Signature. 

[svmlegacy]

14 minutes ago, Skiiwee29 said:

Alright friends,

 

Need some help as I can't seem to figure this out. When Folding@home, I get BSOD's with the Whea Uncorrectable Error message. I know this is (usually) related to overclocking so I set out to dig in.

 

Things I have tried:

 

BIOS flashing, went from 3601 to 3801, reflashed twice, then back down to 3601.

With and without XMP/DOCP

Disabled PBO in BIOS

Reset BIOS to optimized Defaults

Upped DRAM voltage to 1.4v and set SOC to 1.15.

Ran Memtest86 with 2 full passed with XMP enabled with no errors found (Before DRAM/SOC Voltage changes)

System can run any stability stress test or benchmark out there with zero issues... Prime95, Aida64, Cinebench on loop, etc... 

Uninstalled and reinstalled Folding@home and tried both 7.5.1 as well as 7.6.21. I have 2 other machines running 7.6.21 with zero issues. 

 

Kind of at a loss here. Any help/suggestions you got, throw them at me. 

 

System Specs in my Signature. 

Could try the full nuclear approach and see if it persists in Linux. That way you know if it's truely a hardware issue or some bug in the Windows OS. Possibly it's a hardware issue that the Linux kernel has worked around.

 

Other options to try: Disabling SMT, SVM, and IOMMU features.

[Moonzy]

can u try setting PCIE signaling to 3.0?

[Skiiwee29]

2 minutes ago, svmlegacy said:

Could try the full nuclear approach and see if it persists in Linux. That way you know if it's truely a hardware issue or some bug in the Windows OS. Possibly it's a hardware issue that the Linux kernel has worked around.

 

Other options to try: Disabling SMT, SVM, and IOMMU features.

I could, but I don't really want to do that as this machine is mostly for gaming, I just like to fold when I go to bed or when im not doing anything on it, so would like to avoid disabling features such as those. I do have the Virtual stuff enabled as I was recently testing out some Linux distros in VM, so ill disable SVM and see if that makes a difference. 

 

3 minutes ago, Moonzy said:

can u try setting PCIE signaling to 3.0?

ill give it a go after disabling SVM to see if that helps. 

[Skiiwee29]

@svmlegacy so SVM was already disabled actually from the BIOS flashing and setting optimized defaults which is SVM disabled. 

[Helpful Tech Witch]

2 minutes ago, Skiiwee29 said:

I could, but I don't really want to do that as this machine is mostly for gaming, I just like to fold when I go to bed or when im not doing anything on it, so would like to avoid disabling features such as those. I do have the Virtual stuff enabled as I was recently testing out some Linux distros in VM, so ill disable SVM and see if that makes a difference. 

 

ill give it a go after disabling SVM to see if that helps. 

you could live boot linux to test if tis a hardware or software bug still right?

[svmlegacy]

Just now, Skiiwee29 said:

@svmlegacy so SVM was already disabled actually from the BIOS flashing and setting optimized defaults which is SVM disabled. 

 

3 minutes ago, Skiiwee29 said:

I could, but I don't really want to do that as this machine is mostly for gaming, I just like to fold when I go to bed or when im not doing anything on it, so would like to avoid disabling features such as those. I do have the Virtual stuff enabled as I was recently testing out some Linux distros in VM, so ill disable SVM and see if that makes a difference.

Another option is to run F@H under WSL2, which should in theory use the Linux kernel at the same level as Windows'....

[Skiiwee29]

38 minutes ago, svmlegacy said:

 

Another option is to run F@H under WSL2, which should in theory use the Linux kernel at the same level as Windows'....

@HelpfulTechWizard creating a bootable Ubuntu to test on...

 

 

I may have accidentally tripped my circuit breaker for my office having my AC running and 3 machines folding at the same time... 

[Helpful Tech Witch]

Just now, Skiiwee29 said:

@HelpfulTechWizard creating a bootable Ubuntu to test on...

 

 

I may have accidentally tripped my circuit breaker for my office having my AC running and 3 machines folding at the same time... 

um... well... dont do that

[Skiiwee29]

20 hours ago, svmlegacy said:

 

Another option is to run F@H under WSL2, which should in theory use the Linux kernel at the same level as Windows'....

 

19 hours ago, HelpfulTechWizard said:

um... well... dont do that

 

soooo, I dont know why... or how... but it seems to be running fine now... only change I made between yesterday and now is enabling.. not disabling.. SVM so I could fire up some Linux VMs....

[Helpful Tech Witch]

2 hours ago, Skiiwee29 said:

 

 

soooo, I dont know why... or how... but it seems to be running fine now... only change I made between yesterday and now is enabling.. not disabling.. SVM so I could fire up some Linux VMs....

huh, thats weird

well happy folding

[Skiiwee29]

3 minutes ago, HelpfulTechWizard said:

huh, thats weird

well happy folding

well, it ran fine for like 10-12hrs.. then it BSOD'd on me again lol 

[Helpful Tech Witch]

33 minutes ago, Skiiwee29 said:

well, it ran fine for like 10-12hrs.. then it BSOD'd on me again lol 

god damnit.

Well stot talking about it and it stops having bsod issues it seems

[Skiiwee29]

11 minutes ago, HelpfulTechWizard said:

god damnit.

Well stot talking about it and it stops having bsod issues it seems

Only change I can think I made was changing the curve optimizer from negative 25 to negative 20.... so maybe a Vdroop issue? Maybe I need to set a LLC on my Vcore... 

[svmlegacy]

2 hours ago, Skiiwee29 said:

Only change I can think I made was changing the curve optimizer from negative 25 to negative 20.... so maybe a Vdroop issue? Maybe I need to set a LLC on my Vcore... 

Very much could be the issue. In fact, I'd suggest running a pure stock config as a test... I had assumed you did that already though!

[Skiiwee29]

1 minute ago, svmlegacy said:

Very much could be the issue. In fact, I'd suggest running a pure stock config as a test... I had assumed you did that already though!

I did, and was still getting BSOD's. 

[Skiiwee29]

Troubles continue...  Next step that I just completed was removal of my cable extensions to make sure its not an issue from them and added the 2nd 4pin EPS in to have 8+4 now. 

[Skiiwee29]

@svmlegacy @HelpfulTechWizard @Moonzy

 

 

So after digging piece by piece... I'm 99.9999% sure I finally found the cause... it appears it was my CableMod Pro ModMesh sleeved extensions I've been using for like 3 or 4 years now. No problems up until now. I suspect that my 5900x may draw to much power/current and the resistance change was enough to cause voltage drops and the BSOD since the code was Whea_Uncorrectable_Error which is most common with bad voltages with Overclocking. I tore those out and also went from a single 8 pin EPS to the 8+4 my board has, though that was likely more placebo than actually beneficial. Never had an issue with my 3700x when it was in the system, but its a lower wattage part so makes sense. 

 

I have been folding all day without a BSOD since doing this, so 10+hrs at this point? Usually would BSOD within 10 minutes before so I'm growing more and more confident. 

 

Now to decide if I want to just stick with stock cables, order a full on replacement set from CableMod, or get a new PSU and sleeved cables for it... 

[Moonzy]

28 minutes ago, Skiiwee29 said:

I tore those out and also went from a single 8 pin EPS to the 8+4 my board has, though that was likely more placebo than actually beneficial.

perhaps the extra power PLUS removing potential voltage drop helped

 

anyhow, glad it's resolved

[svmlegacy]

17 hours ago, Skiiwee29 said:

@svmlegacy @HelpfulTechWizard @Moonzy

 

 

So after digging piece by piece... I'm 99.9999% sure I finally found the cause... it appears it was my CableMod Pro ModMesh sleeved extensions I've been using for like 3 or 4 years now. No problems up until now. I suspect that my 5900x may draw to much power/current and the resistance change was enough to cause voltage drops and the BSOD since the code was Whea_Uncorrectable_Error which is most common with bad voltages with Overclocking. I tore those out and also went from a single 8 pin EPS to the 8+4 my board has, though that was likely more placebo than actually beneficial. Never had an issue with my 3700x when it was in the system, but its a lower wattage part so makes sense. 

 

I have been folding all day without a BSOD since doing this, so 10+hrs at this point? Usually would BSOD within 10 minutes before so I'm growing more and more confident. 

 

Now to decide if I want to just stick with stock cables, order a full on replacement set from CableMod, or get a new PSU and sleeved cables for it... 

 

16 hours ago, Moonzy said:

perhaps the extra power PLUS removing potential voltage drop helped

 

anyhow, glad it's resolved

Glad to hear it's resolved!

 

More copper and connections always helps with voltage drop issues. I'm surprised this vdroop wasn't picked up by the motherboards SuperIO though..