can an oced cpu cause BSOD at idle?

[toyota]

I have been at 4.2 using 1.23 for about 3 weeks with no problems until 2 days ago when I started getting BSODs. I have not had but 5 of them but obviously something is not right. it ONLY happens when the pc is just idling and I have zero issues with stress tests or any gaming.

to be clear I am at 1.23 and 4.2 all the time as all power saving features are disabled.



Problem signature:
Problem Event Name: BlueScreen
OS Version: 6.1.7601.2.1.0.256.48
Locale ID : 1033

Additional information about the problem:
BCCode: 124
BCP1: 0000000000000000
BCP2: FFFFFA80086A9028
BCP3: 00000000BE200000
BCP4: 000000000005110A
OS Version: 6_1_7601
Service Pack: 1_0
Product: 256_1

 

[RavenSEAL]

Yeap....I've had it happen to me.

 

[nanaki333]

absolutely.

 

[Brunnis]

Yep, it has happened to me a few times. Give it an extra 0.05V and see if it helps.

 

[toyota]

well what could have changed after 3 weeks to make it unstable when just idling? I could maybe understand that under load there might be an issue but to cause a BSOD idling is weird. all power saving features are disabled too.

and what would I need to change? the oc is so very mild that nothing but the voltage had to be changed. if it was stable under load for hours of gaming or stress tests then what can I possibly adjust to make it not BSOD at idle?

 

[aigomorla]

check RAM..

 

[Kenmitch]

well what could have changed after 3 weeks to make it unstable when just idling? I could maybe understand that under load there might be an issue but to cause a BSOD idling is weird. all power saving features are disabled too.

and what would I need to change? the oc is so very mild that nothing but the voltage had to be changed. if it was stable under load for hours of gaming or stress tests then what can I possibly adjust to make it not BSOD at idle?

Isn't your vcore hard locked?

This happens to SB's sometimes when using offset voltage. Can be stress tested forever but ransomed bsods under light loads. Fixed by upping vcore a tic or two.

 

[toyota]

check RAM..

I did using the Windows built in test.

 

[toyota]

Isn't your vcore hard locked?

This happens to SB's sometimes when using offset voltage. Can be stress tested forever but ransomed bsods under light loads. Fixed by upping vcore a tic or two.

its at 4.2 and 1.23 the whole time. I have no offset options for my mobo.

 

[Duvie]

Happened to me after I stress tested OCCT for 9 hours, prime 95 for 4hours and then BSOD at idle....

I think I isolated it to PLL overvoltage and hibernation sleep mode....check my thread out on my ocing and it is mentioned there.....

Mine happened within the first 2 days of my build....

I set bios settings back to default and only changed a few and have been rebuilding my oc. Back up to 4.5ghz 100% stable no BSOD now for a few days.

 

[Kenmitch]

its at 4.2 and 1.22 the whole time. I have no offset options for my mobo.

That's what I thought.

If it was me I'd just bump the vcore 2 tics and see how it goes.

Maybe use memtest to test your memory.

What's your memory voltage in BIOS? Possibly dropping below spec? I most of time set mine a couple tics above spec just in case.

 

[toyota]

well I am going to run cpu stock for a few days just to see what happens. and I run my memory at 1.52 in bios.

 

[Idontcare]

I did using the Windows built in test.

Try using HCI memtest, there's a link for it in the OC Stability sticky.

That said, my SB will do what you are reporting (random crashing at idle) when I have the Vcc too low.

 

[toyota]

Try using HCI memtest, there's a link for it in the OC Stability sticky.

That said, my SB will do what you are reporting (random crashing at idle) when I have the Vcc too low.

yeah I used that HCI test for over an hour and over 100% coverage and there were no errors there either.

so whats the point of passing intense cpu tests like IBT if just idling can make it BSOD anyway? not to mention what changed after 3 weeks to make it start having issues at idle?

 

[mrjoltcola]

Even at "idle", the CPU is never really idle except when it is suspended, or off.

All desktop OSes have constant processes in execution, albeit most of them are blocked on IO or IPC, but there is steady execution of code, even at idle frequency / voltage. Countless instructions are happening every second, even while throttled down.

Windows update could cause changes in idle state code paths, but there are so many other possibilities. Can you correlate a Windows update with the time that you started having the problem?

 

[toyota]

no it had been 3 weeks ago since my last Windows update so nothing there.

I will be staying at stock cpu speeds and auto voltage for a few days just to make sure that my oc was the cause.

 

[E6700]

more vcore, nuff said.

 

[BD231]

Just sounds like your voltage isn't high enough in downclocked states to me, as IDC pointed out with Vcc.

 

[toyota]

its at 4.2 and 1.23 ALL the time. so if that voltage is enough for heavy loads then why is not enough for idling?

 

[mrjoltcola]

its at 4.2 and 1.22 ALL the time. so if that voltage is enough for heavy loads then why is not enough for idling?

Load Line Calibration could explain that.

Either way, 1.22v is pretty low for 4.2Ghz.

 

[toyota]

Load Line Calibration could explain that.

Either way, 1.22v is pretty low for 4.2Ghz.

oops it was actually it was 1.23 not 1.22.

 

[Magic Carpet]

Can't you manually edit p-states?

Heck, you can do that on AMD ;-p

 

[Idontcare]

its at 4.2 and 1.23 ALL the time. so if that voltage is enough for heavy loads then why is not enough for idling?

Probably for the same reason voltage X.xx is enough for gaming but not enough for Prime95 Large, or is enough for Prime95 small but not enough for F@H, etc.

That reason being - different instruction mixes.

When you say "good enough for heavy load" that load can't possibly be an exhaustive check of all 700+ instructions in the ISA.

Likewise at idle, windows is doing all kinds of background service stuff, updates checks at random times, etc.

It is possible that at your voltage, with your specific CPU and its within-chip process-induced variations in the xtors and wires, there are some instructions that are so on the border of being instable that even at idle temps with idle voltage those instructions are occasionally unstable.

Imagine if you had a stress test program that would run those instructions while your CPU was all heated up with LinX it would probably bomb out in an instant unless you cranked the Vcc up 0.1V or more.

Another thing - you say you disabled your C-states and so on, but maybe your BIOS is corrupt and its not actually doing things correctly. You might try reflashing your BIOS or re-enabling those C-states to see if the problem goes away.

 

[sm625]

I only have limited experience with a 2x00k, but one thing I did notice was that it changes the voltages even when you have speedstep and turbo disabled. It just does it, using pstates I guess.

I seriously doubt you're staying at the same voltage for any load. When you load your cpu, you get a spike in current, which will drop the voltage for an instant no matter what kind of compensation the motherboard has. I bet that's where you get your bsod. Your cpu is sitting there idling, then the windows kernel makes the current draw spike and poof there it goes.

 

[Rvenger]

Toyota - Your BSOD code is related to cpu voltage. Increase it a hair and see what happens... or run at stock to see if you get BSOD.

I got the BSOD at idle and I disabled every power saving feature (C1E and All Cstates)EXCEPT Speedstep and it cured my issue.