BSOD when overclocking.

[MccGarthur]

Hey everyone, I was hoping someone could help me figure out why I keep crashing.

My rig:

Motherboard: GA-Z77X-UD3H
CPU: Intel i5-3570 @3.4, OCed to 4.5
Heatsink: CM Hyper 212+
GPU: EVGA GTX670 FTW 2048MG GDDR5 256bit
Ram: Patriot extreme masters 2x 4GB
PSU: Corsair TX 750 W
Storage: ADATA Sx900 256 GB SSD
Secondary storage: Seagate Barracuda 1TB
OS: Windows 7 64-bit.

Decided to boot from my OCed profile today and make sure everything was running fine, ran Prime95 but I keep getting BSOD. The first one was this (sorry for the flash) - http://imgur.com/QQCZF After doing some research, I updated my graphics driver and ran Prime95 again. I got to about test 12 then I received this one - http://imgur.com/KcVrg At this point I ran verifier.exe and rebooted, and received this one. http://imgur.com/VW7tv

I seem to be getting these a lot lately. Event viewer has seven instances of kernal-power id 41 (63) crashes for this week. It will sometimes crash when not OCed and on default settings, so I'm not sure what's causing it. It's a brand new computer I built about a month ago.

Here are my temps for Prime95- http://i.imgur.com/SS1jY.jpg The highest the max has gotten is 79, 86, 82, 82, respectively.


Here are the minidump files from today's crashes. http://www.mediafire.com/?8ilaxe9bm2rjn72#!

Any help is greatly appreciated!

 

[MccGarthur]

I'd also like to mention that I have ran memtest before when I first crashed and didn't receive any errors.

 

[steddora]

Did you notice CPU-Z is showing there is 1.320v going into that CPU?

Those temperatures are pretty warm as well.

If you're having bsods at stock with that setup you're getting instability from somewhere.

I'd first run memtest386 just to verify the ram is operating properly.

If memtest passed a few hours of cycles I'd go and check the GPU and do a perfectly clean install of the nVidia drivers.

Your overclock bsods may just be the chip being unstable. I'd look into the power saving features and see if disabling them helps stabilize the chip during overclock.

Personally I'm going to point at your ram for right now as if you're still getting random blue screens from stock, there's an issue. So you need to download memtest386 and burn the CD and boot onto it so it can properly test the ram. If they pass a few hours to an overnight test; you can probably eliminate that being an issue.

 

[MccGarthur]

I'm pretty new to OCing, is 1.320v bad for 4.5? I just ran GW2 and received this error:

*--> Crash <--*
Exception: c0000005
Memory at address 0000006f could not be written
App: Gw2.exe
Pid: 1640
Cmdline:
BaseAddr: 00400000
ProgramId: 101
Build: 16293
When: 2012-12-16T00:45:35Z 2012-12-15T16:45:35-08:00
Uptime: 0 days 0:00:26
Flags: 0


I usually have no problems when playing on stock settings. Also, I recently did a clean install of all my nVidia drivers, and have already disabled all power saving features. I'll try running memtest again, though.

 

[MccGarthur]

Did you notice CPU-Z is showing there is 1.320v going into that CPU?

Those temperatures are pretty warm as well.

If you're having bsods at stock with that setup you're getting instability from somewhere.

I'd first run memtest386 just to verify the ram is operating properly.

If memtest passed a few hours of cycles I'd go and check the GPU and do a perfectly clean install of the nVidia drivers.

Your overclock bsods may just be the chip being unstable. I'd look into the power saving features and see if disabling them helps stabilize the chip during overclock.

Personally I'm going to point at your ram for right now as if you're still getting random blue screens from stock, there's an issue. So you need to download memtest386 and burn the CD and boot onto it so it can properly test the ram. If they pass a few hours to an overnight test; you can probably eliminate that being an issue.

When I do memtest, should I do both sticks at the same time, or each separately?

 

[steddora]

It's best to test them separately if you're trying to find a hard to find quirk with them. Another way I've learned to overcome finding if it's memory (as testing memory is a pain in the butt), I'll loosen the timings on a machine that's suffering from RAM problem symptoms. Say the timings were 9-9-9-24, I'd back off to 10's and maybe even back it off on the frequency from 1600 to 1333mhz. If the problem eliminates itself; I'd then test the crap out of the sticks and find the problem while they were at stock settings.

Most times when it IS a ram error, loosening the timings and slowing it down will eliminate the issue while overclocking will show the RAM's faults a lot faster. HOWEVER, on these new chipsets of today; I don't suggest overclocking through the FSB/BCLK settings.

Sometimes finding the issue just takes swapping parts out until the problem is eliminated. That's the worst thing about getting blue screens that are pretty much random. However, there is an instability in your rig somewhere if you're getting a BSOD at stock settings. My system is at 4.4Ghz from 3.4Ghz and I NEVER have a blue screen unless I drop the voltage on my processor.

 

[steddora]

I forgot, most of the users here who have Ivy Bridge CPU's don't recommend going over 1.3v at all. Intel hasn't specified a maximum voltage for the IB chips; but that seems to be the point to stay away from. I would guess it would cause sped up degradation of the CPU which could result in a quick shortening of life.

 

[MccGarthur]

I forgot, most of the users here who have Ivy Bridge CPU's don't recommend going over 1.3v at all. Intel hasn't specified a maximum voltage for the IB chips; but that seems to be the point to stay away from. I would guess it would cause sped up degradation of the CPU which could result in a quick shortening of life.

Alright, I tried turning down the voltage to match what someone on Youtube had for the same processor, and now it crashes before I can boot to windows, and freezes in the BIOS before I can change the settings back. Any idea what I can do?

 

[steddora]

What did you set it to? You'll have to clear the cmos with the switch or button for your motherboard.

Just because someone else can push that voltage at X frequency doesn't mean your chip will handle it. Each chip is different, so go back into the BIOS after clearing it and set up the overclock and drop the frequency and start at say 1.2v and test where it will boot stable at. What you're experiencing is an undervolt failure where there's enough to turn on the processor but not enough to keep it stable at all.

 

[MccGarthur]

What did you set it to? You'll have to clear the cmos with the switch or button for your motherboard.

Just because someone else can push that voltage at X frequency doesn't mean your chip will handle it. Each chip is different, so go back into the BIOS after clearing it and set up the overclock and drop the frequency and start at say 1.2v and test where it will boot stable at. What you're experiencing is an undervolt failure where there's enough to turn on the processor but not enough to keep it stable at all.

I set it to 1.15v. I'm gonna do a bit more research and wait to overclock until I know more about what I'm doing. I had a friend help set up the original overclock, but he wasn't with me in person so I may have done some things wrong.

Also, I ran memtest on one of my sticks and it passed 3 times, ran for an hour and fifty minutes, then just stopped in the middle of a test, time and everything, and was unresponsive to my keyboard. I waited about 15-20 minutes and rebooted and am about to try the other stick. Does that mean that the first one is bad?

 

[steddora]

That definitely can. But if your CPU isn't stable now; memtest could even cause the instability to "mess" up. I'd go into your BIOS and set your processor to auto on frequency and on voltage. That way you can be assured it's stable as well.

But yeah, I've had a bad stick do that to me before; so it could be evidence of a bad stick. Make sure that processor is stable and run it again brother. And no worries. I'll keep up with this thread here and we'll get you in the know and you'll know all about overclocking before you know it.

 

[MccGarthur]

That definitely can. But if your CPU isn't stable now; memtest could even cause the instability to "mess" up. I'd go into your BIOS and set your processor to auto on frequency and on voltage. That way you can be assured it's stable as well.

But yeah, I've had a bad stick do that to me before; so it could be evidence of a bad stick. Make sure that processor is stable and run it again brother. And no worries. I'll keep up with this thread here and we'll get you in the know and you'll know all about overclocking before you know it.

Alright, awesome. I ended up clearing the CMOS before running memtest, so it's at optimized defaults now. I ended up contacting Amazon and they're sending me a relacement, but I won't be able to install it until the 18th. I'll give you an update then and let you know how everything's running. Thanks for all your help, by the way!

 

[steddora]

Not a problem! I'll be here, and I always follow up on the threads I participate in; so just drop a line and let me know how it goes from there!

 

[MccGarthur]

Not a problem! I'll be here, and I always follow up on the threads I participate in; so just drop a line and let me know how it goes from there!

Sorry for the super late reply! I got the replacement RAM in and everything is working perfectly!