Upgraded i7 Build Random Hardlocks

J

johnnow

n00b
Joined
Feb 17, 2006
Messages
13
I upgraded my computer recently to an i7 920 build that is experiencing random hardlocks. When it happens, the screen just freezes. There is no keyboard, mouse or network response, so I have to hit the reset button to get it going again. At first, it would happen every few days, but it started happening a couple times a day at the end of last week. The OS is Windows 7 Ultimate x64, clean install. There is nothing in the event viewer that points to any error.

I have run memtest on all 3 memory modules, both individually and as a set overnight. So, I'm fairly confident that the memory is good. I've reinstalled the OS and tried running Prime95. The small FFT test runs ok. The large FFT and Blend tests both cause the PC to hardlock within 30 minutes to an hour. I don't think its heat related as Realtemp shows my maximum temp at 67C under load. What should I look at next?

The system specs are as follows.

P6T Deluxe V2
Corsair XMS3 TR3X6G1600C8
i7 920
Ultra-120 extreme-1366 Rev C
XFX 8800GT 512MB reused from my old system
Corsair HX620 PSU reused from my old system
Nvidia Dual TV reused from my old system - has been removed while testing stability
Western Digital Caviar Black WD6401AALS
Samsung DVD Burner SH-S223L
P180 Case - I spent a lot of time making sure the cabling was as neat as possible, so airflow is good

The parts from my old system were working without issue at the time of the upgrade.
 
Last edited:
Do you have another PSU you could try?

If Prime95 hardlocks in SmallFFTs, there's a good chance its a CPU and/or mobo issue. To eliminate software/driver issues, use a LiveCD OS, like BartPE or UBCD4Win -- something with a CPU stress test included in the image.
 
I tried using a new Antec 500W Earth Watts from another system. It still hardlocked after 30 minutes while running the large FFT test. However, when I ran the UBCD4Win it passed the large FFT test for 5 hours. With that said, I don't think the live CD was stressing the system to it's full potential. It only recognized one core, and half the memory. I've got 6GB, but the live CD was running a 32 bit OS.

I suppose I can try installing Vista x64, and see if that makes a difference.
 
I had the same issue with the same memory on my Asrock X58 Extreme board. I was getting random hardlocks and would have to reboot to recover. I tried a bunch of different things and finally bumped my memory down to 1333mhz. I had read where other people were having problems running memory at 1600mhz with my particular board. I have had zero issues since then and hopefully a new BIOS update in the future will correct this problem.

You might want to try this just to see if this helps your problem.
 
DillnutMcGee said:
I had the same issue with the same memory on my Asrock X58 Extreme board. I was getting random hardlocks and would have to reboot to recover. I tried a bunch of different things and finally bumped my memory down to 1333mhz. I had read where other people were having problems running memory at 1600mhz with my particular board. I have had zero issues since then and hopefully a new BIOS update in the future will correct this problem.

You might want to try this just to see if this helps your problem.
Click to expand...

Ding, Ding, Ding, we may have a winner. I tried setting the memory for 1333 and it's been running the large FFT test for two hours now. I'll toss in the tv tuner and see if I can get the test to run overnight.
 
That memory currently has issues with Corei7 motherboards. In fact, alot of DDR3-1600 memory has problems right now with Corei7. The latest RAM chips that both Corsair and Kingston (and possibly other brands) uses has an issue relating to the QPI. I have only spoken to Kingston about this, but the symptoms I've seen with Corsair are nearly identical, and the solutions are the same as well.

If you lucky, turning up your QPI/DRAM voltage will get you completely stable. 1.32500 would be a number I would start at. Its low enough that you shouldn't need any active cooling for the chipset, but high enough that it has fixed most issues I've seen.

What I would suggest doing is trying the voltage tweak I suggested (maybe trying a bit higher voltage if your still unstable), and in a few months RMA'ing the sticks to Corsair. By then, they should be using newer RAM chips which hopefully won't have this issue.
 
Xiun, thanks for the heads up on the DDR 1600 memory issues. I'll test that out at some point.

But, for now, it seems like setting the memory at 1333 has made the system stable. It has been running a Prime95 blend test for over 8 hours with no issues. I'll check it again when I get home tonight. If it's stable, I'll leave it alone for now. I've just got too much going on at home to spend more time messing around with it.:(
 
You must log in or register to reply here.
Back
Top