I was gone this weekend and I came home to my computer and logged in, thinking everything is normal because I usually lock my computer when I'm gone.  I noticed my PPD was a little down on Saturday so I think that's when the problem started but when I logged in, it said that windows recovered from an unexpected shutdown or something like that.  Then it said it was a blue screen and gave me this info.  

Problem signature:
  Problem Event Name: BlueScreen
  OS Version: 6.1.7600.2.0.0.768.3
  Locale ID: 1033

Additional information about the problem:
  BCCode: 116
  BCP1: FFFFFA8009115330
  BCP2: FFFFF8800FFB4D50
  BCP3: 0000000000000000
  BCP4: 0000000000000002
  OS Version: 6_1_7600
  Service Pack: 0_0
  Product: 768_1

Files that help describe the problem:
  C:\Windows\Minidump\052210-36312-01.dmp
  C:\Users\Justin\AppData\Local\Temp\WER-143489968-0.sysdata.xml

Read our privacy statement online:
  http://go.microsoft.com/f...04288&clcid=0x0409

If the online privacy statement is not available, please read our privacy statement offline:
  C:\Windows\system32\en-US\erofflps.txt




I searched this a tiny bit and somebody said that nvidia drivers could have caused this.  Is this what you guys mean by unstable drivers for folding?  I don't know if it's from having fermi and maybe those aren't quite ready or if it's from the 250 drivers I downloaded, version 197.45

techmasterjoe

VTT a tad low or ram had a hiccup

Thanks, my VTT is at 1.15000, do you think I should bump it up to 1.17500?  It's been stable since December without any changes except adding a 470.  

whats your power supply is it the same as your mod rigs if so i bet your 5v rail is dropping under load to 4.90 or lower a 470 is a BIG DRAW
my system is eating 640watts folding and that just a 260 216 core

if not then bump it that 1 tad and i bet it wont die anymore

side note:
i have killed 4 sets of OCZ in my board 3-5months then rams toast.
i switched to G.skill not a problem yet and ram runs cooler to

but ya 1.175 vtt is fine just keep it with in + or - 0.5v of ram

I forgot, I need to update my mods rigs, I'll do that now.  I have a Corsair 850 watt modular.  I didn't have any problems before the 470 and I'm pretty sure it's not because of my PSU being not enough.  

I just updated my drivers to the new ones that came out today, both on my 250 and 470 and then I bumped up my VTT one notch via eleet and I'll do that with my bios later today.  We'll see if I have any more problems when I get home.  

I came home and everything was fine except that when I left today I turned my 470 fan to 100% while running boinc and when I came home it was only at 53% and running at 100C, I almost pissed my pants.  That's a different discussion though.  Hopefully these new drivers will work.  

But I still have one question from the top, what do you guys mean when you say that drivers are unstable for folding?  Do you mean that they computer crashes like this or does the GPU just get NAN's and EUE's and down clock?  

Davabled

700_up

...
But I still have one question from the top, what do you guys mean when you say that drivers are unstable for folding?  Do you mean that they computer crashes like this or does the GPU just get NAN's and EUE's and down clock?  

"Unstable for folding" could mean NAN, EUE's, Down Clocks, lock ups, strange behavior.
 
As far as crashing to a BSOD though, drivers are kept in separate memory spaces... hrm, let me see if I can explain this.
 
Windows is designed to keep it's kernel (the core of the operating system) protected.  One of the ways it does this is by putting it's code into a protected area of memory, and putting the code for video (and other) drivers in a separate area of memory.
 
That way, if a driver goes on a memory-hogging rampage, it won't be able to damage the kernel.
 
But, just protecting the kernel in it's own memory space isn't always enough.  So, when the operating system detects that a driver is trying to trespass into the protected property of the kernel, it puts up a big stop sign and makes everyone stop and get off the road.
 
If you're following my metaphor/analogy, the stop sign is the BSOD.
 
The memory dump files you listed above could be analyzed to try and pinpoint the exact problem, if you want to learn more I could take a look at them or point you in the direction on how to examine memory dumps.
 
... Did I make any sense?

That kind of makes sense.  How hard would it be to look at them?  I don't want to give you a ton of work.  

Davabled

700_up
That kind of makes sense.  How hard would it be to look at them?  I don't want to give you a ton of work.  

 
You'd have to send me the files or upload them somewhere I could download them.  Then I'd run debugging tools for windows to analyze them.  It can sometimes take me an hour or so to sort through the output and get something meaningful out of it, so it might take me a day or to to get too it depending on how my days go.  I'm not nearly as expert at it as some are (or as I used to be), but I enjoy doing it to keep my skills sharp now that I'm no longer doing it as a day job anymore. 

I don't know if I have any files from it, I don't know if it ever game me any.  Where would they be at?  

Hey 700_up, that BSOD is due to NVIDIA driver crash and that probably happened because the NB voltage is insufficient. With multiple cards and especially when overclocked CPU and/or memory and/or GPU, NB chokes easier. Up the NB voltage by 50mV in the BIOS and try; if you still get it increase again, etc. I'm 99% certain this is your remedy...

No, sorry, it's the IOH voltage, not QPI PLL!

700_up
what is the QPI PLL, I thought the QPI was like the Northbridge.  

That's the Phase Lock Loop for the Quick Path Interface.  It makes sure the clock signal sent along the QPI link stays in-phase.  Somewhere around here is a nice diagram indicating what's connected to the QPI.... Here we go:

 
Above you can see the QPI connecting the CPU to the Northbridge.  With Xeon CPUs it's also used as a bridge between multiple CPUs.

700_up

Did you just read up a lot on this stuff or is there some kind of class you can take to learn this kind of stuff?  

If you're referring to debugging crash dumps, the answer is "all of the above".  Learned some of it through my own research, and some through custom seminars (Microsoft would sometimes come out and give special custom-tailored classes to our I.T. departments).

As far as learning OverClocking and stuff, it's mostly been from almost 30 years of trial and error.  Started taking things apart and flipping switches until something broke, then made myself learn how to fix it.  When I started overclocking, the Internet wasn't open to the public yet.  I exchanged experiences with friends and colleagues and local computer clubs.  Having degrees in electronics engineering helped a fair bit as well by providing a deeper understanding of voltages, currents, components, etc.

Simplest answer: Years of trial & error.

So taking a computer engineering class in college might help?  

I've another question, I see IOH CCore and IOH/ICH I/O Voltage, which one should I increase?  

Thanks Gryphon, I bumped it up, hopefully that'll be the end of all my problems.  

I bumped it up .025 volts so I guess 25mV.  I'm not having any more BSOD's any more, I really only had one and that was the end of it but right no the major problem is that my monitor just starts going blank and I have to hard restart my computer.  I watched it happen in front of me this time.  

I walked into my room and felt the air coming out of my 470 and it was extremely hot and didn't seem to be coming out very fast.  I got on my computer and everything was fine, didn't look at temps  but I'm sure they were hot because of the air coming out.  Anyways, I put on some music and went on the internet and when I was looking at the webpage, my monitor just went blank and the air coming out of the card got cool really fast.  I also couldn't pause my music with my keyboard anymore like I usually do.  And the fan didn't sound very loud either.  


So basically, I think my GPU is getting too hot because my fan is slowing down.  I don't know if you know what to do about that or not but it's really making me get frustrated.