BSOD on Boot - CPU or Memory Issue?

[Mchoekst]

I have a new(ish) system with the following configuration (all brand new):

Intel Core i7-2600K Sandy Bridge 3.4GHz
GIGABYTE GA-Z68X-UD4-B3
CORSAIR Vengeance 16GB (4 x 4GB) DDR3 1600
EVGA GeForce GTX 560
WDl Caviar 1TB 7200 RPM SATA 6.0Gb/s
CORSAIR Professional PSU 750W

Worked excellent for 2 months, and started giving BSODs (in Windows 7) while running applications. This slowly got worse until now as soon as it transitions to the "loading windows" screen, it crashes. I have tried bootable Knoppix from CD, and gives a boot error as well.

The most common BSOD is 'IRQL Not Less Or Equal/System Service Exception' although several other system service exceptions have been given as well. With 4 sticks of RAM, I have dropped it down to one and rotated them; not to mention memtest (from bootable USB) runs perfectly clean. I have changed the graphics card, hard drive, and motherboard with no success. Bios Flash and update to F9, tried fail-safe settings and confirmed RAM voltage @1.5 and tried both 1333 and 1600 MHz. All memory sticks and CPU are detected fine by BIOS, and I am running the CPU without overclocking.

There is a similar post (ironically with this motherboard) in which the guy ended up having a bad CPU http://www.techpowerup.com/forums/showthread.php?t=150791.

Any other suggestions? I would like to prove the CPU bad before shipping it back to Intel, and of course I am open for any/all suggestions.

Thanks all

 

[eidairaman1]

memtest your machine

 

[Mchoekst]

Memtest runs just fine with all sticks installed.

 

[sneekypeet]

memtest86+ in dos runs fine but you get a memory or controller error? How many passes are you completing?

 

[eidairaman1]

always run the max passes, even reseat ram too. sometimes windows gets so screwed up that a format is needed to fix any issues that truly are not hardware related.

aumha.org is a reference for any BSODs you will have

 

[Mchoekst]

@sneekypeet - only ran a few passes, since i had zero errors (will start it running again). Starting to doubt the RAM though since I tried the sticks individually so all 4 would have had to fail at the same time to cause my issue.

@eidairaman1 - I have already replaced the motherboard, and moved the RAM quite a few times so reseating is ruled out. Also have used bootable Linux distros and I get the same error, so it can't be windows or the hard drive.

Thanks for the help guys, let me know if you have any more ideas!

 

[Mchoekst]

A few shots of the Voltages, Memory and CPU from BIOS... plus the most common BSOD I am getting. Memtest is running again, we will see if it yields different results.

Again thanks for the help, open to any ideas at this point!

 

[OneMoar]

pull all but one stick of ram and then test each stick on its own
and then go back and test ALL the dimm slots one at a time ...
and memtest should be run for 8 hours or more to be sure

 

[Mchoekst]

The DIMM slots are eliminated since I changed the motherboard. Memtest has started to freeze occasionally, but I did get one error during test #10 that I am trying to duplicate. Going to pull 2 sticks and see if it comes back.

Thanks for the help, open for more ideas!

 

[OneMoar]

I hate to be a alarmist but I am starting to think based on what you have told us that the cpu might be at fault ...
what do you have for voltage controls
do you have VTT/QPI voltage adjustment ?
add +0.25mv to the vtt and qpi and see if that helps it

 

[Mchoekst]

After many memtest86 freezes and occasional errors, I turned the number of processor cores in BIOS down to 1. Everything has been running fine, including many error-free memtest passes. I guess I will ship the CPU back to Intel, thankfully its still under warranty!

Thanks for all the suggestions!

 

[Aquinus]

After many memtest86 freezes and occasional errors, I turned the number of processor cores in BIOS down to 1. Everything has been running fine, including many error-free memtest passes. I guess I will ship the CPU back to Intel, thankfully its still under warranty!

Thanks for all the suggestions!

Try removing two stick of your memory and downclocking it to 1333 and then run memtest. If it works, re-populate with all 4 and run at 1333. If it is still stable and 1600 isn't, try bumping your memory controller / qpi voltage (only by a little bit,) I would also make sure that your memory is properly reporting the timings. Also double check your memory voltage, it might want more for standard speeds.

I'm assuming that you're not overclocking, but if you are, reset all of your settings to defaults before you begin.

By only using 1 core you may not be taxing your memory controller (and memory) nearly as hard as it can handle. I suspect it's your memory and not your CPU. It's very uncommon that a CPU memory's controller is made to be slightly stable. More often than not, the CPU is either critically damaged or it works fine. 90% of BSOD errors I see are memory-related in computers that aren't overclocked.

Hope you're computer works well soon. :<

 

[Mchoekst]

Running stable 1600 MHz with correct timings, hyper threading but not overclocked as long as I run 1-2 cores. As soon as I switch to 3 or 4 cores I get freezing memtest and bsods.

Seems like a CPU core issue.

 

[OneMoar]

you never did tell us if you bumped your qpi/uncore,vtt voltage
its possible you might have gotten a chip that degrades easy ..

 

[Mchoekst]

you never did tell us if you bumped your qpi/uncore,vtt voltage
its possible you might have gotten a chip that degrades easy ..

You are awesome! Bumped QPI/VTT CPU voltage by .005v and it is working perfectly... Even over clocked!

Thanks for the help guys

 

[OneMoar]

You are awesome! Bumped QPI/VTT CPU voltage by .005v and it is working perfectly... Even over clocked!

Thanks for the help guys

I know I am awesome
if it crashes again give it a bit more voltage
not more then 1.25 on the vtt tho and no more then 1.32 on qpi