Seeing Assistance With Core i7 Extreme 2011 system - Weird Problem

Brad McHenry

Honorable
Aug 12, 2013
2
0
10,510
I'm having a lock up problem with my system which persists even when all components have been replaced (except the processor). I'm running an i7-3960x Extreme Edition 2011 processor.

My issue repeats itself even when using different motherboard, video card, hard drive, power supply, memory and so on which have all been replaced. The problem has persisted thru at least 15 rebuilds and about 50 reinstalls of Windows thru the last nine months.

Basically what happens is that I'll be in Windows and my computer will lock doing the simplest things, opening a new browser tap, scrolling a webpage, opening a folder etc. But if I'm running in either safe mode or with just a standard VGA drive all the issues go away. This has happend on three different video cards and two different motherboards.

These video cards work flawlessly in other systems so it leads me to wonder.. Could there be an issue in the PCE-E controller built into the CPU? This is the one component which has stayed the same thru all the rebuilds and part swap outs. CPU stress test with no problem though none of them tax the PCI-E controller in that processor. I hate to spend $300+ dollars buying another 2011 processor to test my theory. It almost feels like the PCI-E controller can handle it in safe mode or vga driver mode but as soon as any correct driver is installed it pushes the video card past what the controller can handle.

Basically wondering if anyone feels that an i7 Extreme processor could have a defect in the PCI-E controller but all else work flawlessly. I've been building my own PC's for about 15 years and never run into anything like this. When I check the event viewer there is no mention of anything crashing, just about the system shutting down unexpectedly. The only way to shut down when it freezes is to hold down power button, never even get a blue screen.

Intel i7-3960x Extreme Edition Processor
32GB Corsair Dominator Platinum Memory (4 x 8GB)
Intel DX79SR Motherboard (Also tried DX79SI) both brand new out of box
860W Corsair AX860i PS
256GB Samsung 840 Pro SSD
4GB GeForce 680 GTX Video Card


Thanks for any assistance.
 
Solution
Sounds like you have tried everything you can with all the resources you have. Everything points to a problematic cpu, so I think you are pretty much done. BUT just to feel 110% sure about it (to justify chunking a $1000 cpu), I would still try 1 more mobo & 1 more gpu. If I were you, I would try another brand (preferably ASUS, Gigabyte, etc) mobo + AMD gpu (with latest driver, of course). If possible, may be someone you know has 2011 mobo that you can borrow for a couple of days. Or you could buy one at a local store and return once troubleshooting is done (Frys, microcenter, etc), just make sure to check return policy.

I have friends who troubleshoot and repair mobos (all brands), and they have shown me (with data) that more...

goonbar79

Distinguished
Dec 31, 2008
213
0
18,860
Why not RMA you cpu to Intel since you have tried replacing all components multiple times EXCEPT the cpu. Is it over 3 yrs since purchase date? I haven't heard of 3960X having any sort of PCI-E controller problem. If that was the case, it would be pretty much widely known by now.
 

goonbar79

Distinguished
Dec 31, 2008
213
0
18,860
You want to isolate the problem with only the primary components installed, and run stability test. If the issue is in fact related to PCI-E, the failure will occur when PCI-E is utilized. So try to replicate the problem with the least number of components (cpu, cooler, 1 ram, 1 hdd, mobo, psu), and remove all other stuff. Run all sorts of test to see if you can get the same error, and try to go deeper from there. If problem doesn't occur, it could you the remaining components (i.e. gpu), even if they seem normal on other machines.
 

Brad McHenry

Honorable
Aug 12, 2013
2
0
10,510
Hi, thanks for all the feedback. I'm going to respond to each of your questions/tips below.

This is a pre-production CPU I received from Intel. I worked for 13 years with a white box manufacturer until earlier this year. As such no warranty. I've RMA'd every single component besides the CPU and the problem persists. Two SSD's, two different Intel motherboards, 4 different sets of qualified memory running one module at a time even at lowest 1333MHz downclock, three different video cards, two power supplies and so on. Everytime the same exact problem comes back where the system will freeze up requiring a hold down the power to restart. This happens even with a brand new reinstall with only the INF and graphic driver loaded going back 4 or 5 versions. I've even used the Windows update Nvidia driver with exact same results.

As mentioned if I don't load a video card driver or go into safe mode it will run for days. As soon as a load an Nvidia driver and do something as simple as move a browser window or file explorer around the screen with any speed it locks. The lockup is the same each time no matter what hardware is running which tells me I RMA'd a ton of components the last several months likely all for nothing.

The computer passes Prime95 everytime on every test CPU or memory as well as MemTest 86+ without a single error. I have a Corsair H100i water cooler with temps in the 32c range at idle.. No more than 45c at 100%. It seems like everything in my system is rock solid but doing something which shouldn't even strain my computer brings it down each time.

My thinking is the system is able to handle the low performance non-driver driven video without a problem with standard vga drivers. As soon as any correct video driver gets loaded the system crashes. All of the video cards doing this exhibit the same failure when drivers loaded but run smooth otherwise. When I put any of them in another system I could replicate what causes the crashes and none failed in the other PC even once. I believe that the Nvidia drivers cause the video cards to ramp up their performance in a way that the CPU can't handle over the PCI-E express. I've got 100's of hours of testing into it and wouldn't have thought a CPU would fail in this way but my experience is showing me otherwise. I hate to go out and buy a new 2011 processors especially as they are now a generation behind with no Haswell yet but thinking I may have no other choice to at least see if this is the problem.

I appreciate any feedback or pointing out something I might be missing.

Thank you
 

goonbar79

Distinguished
Dec 31, 2008
213
0
18,860
Sounds like you have tried everything you can with all the resources you have. Everything points to a problematic cpu, so I think you are pretty much done. BUT just to feel 110% sure about it (to justify chunking a $1000 cpu), I would still try 1 more mobo & 1 more gpu. If I were you, I would try another brand (preferably ASUS, Gigabyte, etc) mobo + AMD gpu (with latest driver, of course). If possible, may be someone you know has 2011 mobo that you can borrow for a couple of days. Or you could buy one at a local store and return once troubleshooting is done (Frys, microcenter, etc), just make sure to check return policy.

I have friends who troubleshoot and repair mobos (all brands), and they have shown me (with data) that more problems (defective or compatibility related) occur due to mobo than anything, and repair can never 100% confident. It is simply impossible, they say. I personally had experiences (not with Intel mobo, though) that many times, I RMA the board with detail description of the problem, and the mobo comes back "no problem found" or "fixed" but the same problem persists. I am not sure about Intel's RMA process, but if anything other than the cpu has an issue, it is usually the mobo.

So if I can't RMA the cpu, I would try this one last time, before getting tears in my eyes and saying goodbye to our expensive, yet defective friend..
 
Solution