WHEA need some Advice

[Redoitall]

Since I delidded my 3770k, I have a few overcloks saved with some pretty descent voltages. ran Prime 95 with amall, large, blend 24 hours each . LInx for 12 hours straight, IBT FOR another 12 hours, memtest overnight. I have also been using my 4.7 ghz for FSX and used the game for 24 hours straight meaning plane on auto with ultra settings. ran cine bench, ran vantage 06. absolutely no errors not BSOD's .
I have been reading about the whea Errors and decided to check it out . well anything above 4.6 ghz will show errors. how many for each overclock I do not know. When I ran all of those test at 4.7 and 4.8 I ran them with low voltages for example my 4.7 was at 1.315 offset. I decided to do a quick 12 hour test prime small fft at 1.344 It showed no errors on event manager. Now I forgot to mention that there are no errors on all the stressers except Prime 95 with the lower voltages. My question is does this mean I am not stable unless I raise the voltages? even though it only occurs on Prime. Or is it a software bug with Prime because I get highter temps with IBT than Prime and I was confused how I did not get any errors with IBT with the same 1.315 volts V core offset. Does ignoring the Whea errors cause system malfunction eventually? I never got any of these errors when I ran my 2700 k rig. Am I fooling myself with lower voltages? ......I know it is a lot of questions but I really need some sound advice on whea

 

[Idontcare]

You need more voltage. Stress tester programs like Prime95 and IBT only test a small subset of the total instructions contained in the ISA.

So when you run Prime95 and declare to yourself "I'm Prime stable! W00t!" all that really means is that your computer is stable for maybe 50 or 60 instructions out of the thousand or so instructions the CPU could be asked to process by any number of software programs (like the OS).

Getting a WHEA means there are instructions for which your CPU is not stable, it needs more voltage, or lower temperature, or lower clockspeed or possibly all of the above.

As end-users we do not have access to any programs that will stress test each and every instruction that our CPU's ISA supports.

Stress tester programs give us a rough indication as to how unstable our processors are but they can never give us any indication as to how stable the processor is.

Get an error in prime95 and you can safely conclude your CPU is unstable. Get no errors in prime95 and you cannot conclude your processor is stable, merely that it appears to be stable with whatever handful of instructions prime95 uses. It may well be unstable with another application that uses different instructions (like the OS or a web browser or a game, etc).

 

[Redoitall]

You need more voltage. Stress tester programs like Prime95 and IBT only test a small subset of the total instructions contained in the ISA.

So when you run Prime95 and declare to yourself "I'm Prime stable! W00t!" all that really means is that your computer is stable for maybe 50 or 60 instructions out of the thousand or so instructions the CPU could be asked to process by any number of software programs (like the OS).

Getting a WHEA means there are instructions for which your CPU is not stable, it needs more voltage, or lower temperature, or lower clockspeed or possibly all of the above.

As end-users we do not have access to any programs that will stress test each and every instruction that our CPU's ISA supports.

Stress tester programs give us a rough indication as to how unstable our processors are but they can never give us any indication as to how stable the processor is.

Get an error in prime95 and you can safely conclude your CPU is unstable. Get no errors in prime95 and you cannot conclude your processor is stable, merely that it appears to be stable with whatever handful of instructions prime95 uses. It may well be unstable with another application that uses different instructions (like the OS or a web browser or a game, etc).

Thanks again, I will raise the voltages.

 

[Idontcare]

Thanks again, I will raise the voltages.

If all else fails you can always just Redoitall or not, Idontcare

 

[thekimbobjones]

Since on topic
WHEA - Where is this information found?

 

[thekimbobjones]

As I know it is Windows Hardware Error Architechure, but how are these errors monitored, noticed or explored? A pop-up, utility?

 

[Kenmitch]

Since on topic
WHEA - Where is this information found?

Control Panel > System and Security > Administrative tools >Event Viewer

As I know it is Windows Hardware Error Architechure, but how are these errors monitored, noticed or explored? A pop-up, utility?

http://msdn.microsoft.com/en-us/library/windows/hardware/ff559509%28v=vs.85%29.aspx

 

[MrSquished]

As I know it is Windows Hardware Error Architechure, but how are these errors monitored, noticed or explored? A pop-up, utility?

Administrative Options>Event Viewer in windows 7

You can setup a popup notification for specific errors, including WHEA, so you don't have to check it all the time.

I did that and i had no WHEA errors for like 2 weeks. I got one, then I bumped the PLL Level 1x and haven't gotten one since.

 

[Idontcare]

Administrative Options>Event Viewer in windows 7

You can setup a popup notification for specific errors, including WHEA, so you don't have to check it all the time.

I did that and i had no WHEA errors for like 2 weeks. I got one, then I bumped the PLL Level 1x and haven't gotten one since.

I look forward to the day where this would all be controlled in an active feedback loop transparent to the end-user.

Windows detects a whea, knows it needs to tell the bios to bump up the PLL level 1x, does it and all is good.

Set some thresholds to prevent run-away loop control (don't want CPU voltage going to 11 ) and just sit back and let the magic happen.

 

[Redoitall]

Since on topic
WHEA - Where is this information found?

windows 7 type on the search window event viewer
-expand on windows logs
-expand on applications and services logs
-expand microsoft
-expand windows
-expand Kernell-Whea
-click on errors
on the lower right hand corner you can click on attach a task to this log. the rest is up to you just follow the steps but you can have windows notify you with a message whenever there is a kernell-whea error.

 

[Redoitall]

If all else fails you can always just Redoitall or not, Idontcare

LOL

 

[thekimbobjones]

Thanks Guys.

 

[coffeejunkee]

windows 7 type on the search window event viewer
-expand on windows logs
-expand on applications and services logs
-expand microsoft
-expand windows
-expand Kernell-Whea
-click on errors
on the lower right hand corner you can click on attach a task to this log. the rest is up to you just follow the steps but you can have windows notify you with a message whenever there is a kernell-whea error.

I found it easiest to: create custom view > by source > whea-logger. By clearing windows logs > system you can remove old errors.

As for stable or not, no you're not stable because when running stock there are no whea errors. On the other hand, they are corrected errors, so I guess you could ignore them but correcting errors also uses cpu power. My feeling though is that every corrected whea error might/will eventually lead to an uncorrectable whea error bsod.

What I find interesting is that you didn't have these errors with a 2700K. Also, do you have C3/C6 states enabled? I found after 4.5GHz it was basically impossible to not have whea-errors, no matter how much vcore.

 

[wand3r3r]

You need more voltage. Stress tester programs like Prime95 and IBT only test a small subset of the total instructions contained in the ISA.

So when you run Prime95 and declare to yourself "I'm Prime stable! W00t!" all that really means is that your computer is stable for maybe 50 or 60 instructions out of the thousand or so instructions the CPU could be asked to process by any number of software programs (like the OS).

Getting a WHEA means there are instructions for which your CPU is not stable, it needs more voltage, or lower temperature, or lower clockspeed or possibly all of the above.

As end-users we do not have access to any programs that will stress test each and every instruction that our CPU's ISA supports.

Stress tester programs give us a rough indication as to how unstable our processors are but they can never give us any indication as to how stable the processor is.

Get an error in prime95 and you can safely conclude your CPU is unstable. Get no errors in prime95 and you cannot conclude your processor is stable, merely that it appears to be stable with whatever handful of instructions prime95 uses. It may well be unstable with another application that uses different instructions (like the OS or a web browser or a game, etc).

Another solid post from you. This makes sense.

 

[Idontcare]

I found after 4.5GHz it was basically impossible to not have whea-errors, no matter how much vcore.

The source of the error may well be outside the core logic then. You may need to boost CPU PLL or Vdimm or something.

 

[Redoitall]

I found it easiest to: create custom view > by source > whea-logger. By clearing windows logs > system you can remove old errors.

As for stable or not, no you're not stable because when running stock there are no whea errors. On the other hand, they are corrected errors, so I guess you could ignore them but correcting errors also uses cpu power. My feeling though is that every corrected whea error might/will eventually lead to an uncorrectable whea error bsod.

What I find interesting is that you didn't have these errors with a 2700K. Also, do you have C3/C6 states enabled? I found after 4.5GHz it was basically impossible to not have whea-errors, no matter how much vcore.

Should I disable c3/c6 would that help with the errors or core voltage is the only way

 

[Redoitall]

The source of the error may well be outside the core logic then. You may need to boost CPU PLL or Vdimm or something.

CPU PLL I have it at 1.8 should I bump it to 1.9 ? and ram I am running it at 2133 rated at 2400 @ 1.65

 

[Idontcare]

CPU PLL I have it at 1.8 should I bump it to 1.9 ? and ram I am running it at 2133 rated at 2400 @ 1.65

I wasn't thinking you needed to change anything per se, not until you verify for certain the core voltage bumps are ineffective.

My comments was specific to coffeejunkie who had already verified core voltage tweaks for his processor did not eliminate the WHEA issues.

WHEA won't have a single common source or cause within the CPU, it can come from any component within the entire hardware loop. If the cpu core is the weakest link then bumping Vcore will help. If the weakest link for your specific CPU is outside the core logic then you need to address that weakest link at its source, bumping Vcore in that case won't help because Vcore isn't the problem to begin with.

But each situation is not going to be the same, that is why you must do a little bit of debugging on your own to isolate the source of the WHEA. Once you know the source then we can assist you in bolstering the weakness.

Until then all we can give you is the equivalent of shotgun solutions which hit everything but may still miss the mark.

 

[coffeejunkee]

Should I disable c3/c6 would that help with the errors or core voltage is the only way

Well, see if disabling them solves the whea errors. I couldn't get 4.5/C-states enbabled without whea errors while using a higher vcore than 4.6/C-states disabled. I figured that was a pretty good indication that more vcore wasn't going to help.

But that doesn't mean that what Idontcare says couldn't be the problem. It might indeed be c-states disabling other parts of the cpu that are the cause of my whea errors.

 

[Kenmitch]

If anybody wants to reset there event viewer during their overclocking adventure I found the following script that will let you do it. Just tried it on one of my rigs and it looks to clear out everything.

Paste the following into notepad and save as clear events.bat or something similar. Right click run as admin and it'll clear all the event logs out.

@echo off
FOR /F "tokens=1,2*" %%V IN ('bcdedit') DO SET adminTest=%%V
IF (%adminTest%)==(Access) goto noAdmin
for /F "tokens=*" %%G in ('wevtutil.exe el') DO (call :do_clear "%%G")
echo.
echo goto theEnd
:do_clear
echo clearing %1
wevtutil.exe cl %1
goto :eof
:noAdmin
exit

 

[Redoitall]

I wasn't thinking you needed to change anything per se, not until you verify for certain the core voltage bumps are ineffective.

My comments was specific to coffeejunkie who had already verified core voltage tweaks for his processor did not eliminate the WHEA issues.

WHEA won't have a single common source or cause within the CPU, it can come from any component within the entire hardware loop. If the cpu core is the weakest link then bumping Vcore will help. If the weakest link for your specific CPU is outside the core logic then you need to address that weakest link at its source, bumping Vcore in that case won't help because Vcore isn't the problem to begin with.

But each situation is not going to be the same, that is why you must do a little bit of debugging on your own to isolate the source of the WHEA. Once you know the source then we can assist you in bolstering the weakness.

Until then all we can give you is the equivalent of shotgun solutions which hit everything but may still miss the mark.

Will test again and try to isolate the errors

 

[Redoitall]

So I increased the volts in all my saved profiles and no long getting WHEA Errors thanks for all your help went from 1.315 volts to 1.344 @4.7 24/7 small, larger and blend prime all the other testers passed accordingly

 

[Idontcare]

So I increased the volts in all my saved profiles and no long getting WHEA Errors thanks for all your help went from 1.315 volts to 1.344 @4.7 24/7 small, larger and blend prime all the other testers passed accordingly

That is good news, glad you were able to resolve it.

If anybody wants to reset there event viewer during their overclocking adventure I found the following script that will let you do it. Just tried it on one of my rigs and it looks to clear out everything.

Paste the following into notepad and save as clear events.bat or something similar. Right click run as admin and it'll clear all the event logs out.

@echo off
FOR /F "tokens=1,2*" %%V IN ('bcdedit') DO SET adminTest=%%V
IF (%adminTest%)==(Access) goto noAdmin
for /F "tokens=*" %%G in ('wevtutil.exe el') DO (call :do_clear "%%G")
echo.
echo goto theEnd
:do_clear
echo clearing %1
wevtutil.exe cl %1
goto :eof
:noAdmin
exit

Thanks for the code Kenmitch :thumbsup: Just used it, worked like a charm +1 rep pts for you