New PC - random short hangs???

[smartech]

Hi. I recently finished setting up a new PC:

CPU: Intel i7 920 D0 (stock clock). Cooled by Coolermaster V8
Motherboard: ASUS P6T Deluxe V2
Memory: Corsair 3x2GB 1600Mhz cl 7-7-7-20 (run on 1066)
GPU: Sparkle nVidia 275 GTX (stock clocks)
PSU: Corsair 850 TX
Sound: onboard AD2000B.
HDD: system - 1TB Samsung spinpoint F3, and WD green 1.5 TB, and WD 320Gb KS
OS: Windows 7 64 bit

The problem is that the PC have random short (about 1 - 0.5 sec) hangs, sound is stopped, nothing moves (no mouse too), just all stuck?!?!?! than it back to normal. no reset or something.
I ran Prime95 blend for 2+ hours no errors found.

I listen to the music non stop so it's really annoying , also in games.
Any idea on what it can be? Its a brand new PC, and its really annoying

 

[Snake91839]

How often does it do this? You said randomly, but how frequent is it? Also, I'm assuming you built this yourself?

This is a problem that I've seen happen when a hard drive needs to be defragged, but obviously that is not the case in this situation. The next thing I would do is check your memory using UBCD. Other than that I can't really think of anything that would cause that unless maybe some of your hardware is defective.

 

[smartech]

Well, sometimes it happens every 10 min, sometimes twice in a minute sometimes i ran a pc for 6 hours and all good.
But most of the time it appears at leas once in a 5 minutes.

All my drives were defragmented a day ago, and yes, I bpught and built this by myself. When i was researching about the parts for it, I never heard of any conflicts in such a configuration.

I'm downloading the UBCD now, will post the results when its done.

Just finished running full passes of Memtest86 and Memtest85+. Both are good no errors. same with few othe memory tests supplied by the UBCD.

Any Ideas???

Btw, just finished running S&M tests (it tests CPU, L2 &L1 cache, mmx, sse, also somehow tests the Power Supply). During the tests being running the "Freezing" appeared several times. However the tests passed all good.

what can it be???

 

[Slaymate]

It sounds like a Northbridge/Southbridge type of issue to me.

Can you post a screenshot of Hardware Monitor running?

http://www.cpuid.com/download/hwmonitor/HWMonitor_115_setup.exe

You can try bumping up the IOH VCore (Northbridge Voltage) a little and see if that works. Next I would bump up the ICH VCore (Southbridge Voltage), then I would try it with both bumped up a little at the same time. If it's still acting badly, while both IOH and ICH VCores are bumped up, try also bumping up the IOH/ICH I/O Voltage (IOH to ICH Termination Voltage).

I'm running these settings at:
IOH VCore - 1.375v (default - 1.1v)
ICH VCore - 1.25v (default - 1.05v)
IOH/ICH/ I/O Voltage - 1.625v (default - 1.5v)

My system is Overclocked (a lot) and I'm running SLI so you shouldn't need to apply as much voltage as I am. Use my settings as a MAXIMUM LIMIT ONLY.

 

[smartech]

I read also some advices on Tom'sHardware to enable XMP on my RAM.

Done so. Now Ram runs on 1600Mhz cl7-7-7-20 DRAM voltage set to 1.66
(intel warns about above 1.65, but it's an XMP value, (Is it really SAFE???)).
QPI/DRAM voltage is 1.35 by BIOS but equals 1.375 by Turbo V (ASUS tool).

All sounds nice but it still didn't solve the problem.

Here some screenshots:
Hardware Monitor before XMP
(In BIOS, the CPU and PCI spread spectrum disabled, Speed Step and Virtualization disabled, 133 BCLCK 100 PCI CLCK).
Vcore 1.225. All the memory voltages and speeds left on auto:

And here is a shots of some monitoring tools. With XMP turned on with voltages as stated at the start.
And all CPU stuff is back to defaults (I mean spread spectrum, speed step, turbo, virtualization)
so the Vcore is varied as u see on a screenshot due to speed step:

By the way the CPU runs extremely hot i mean on regular load, kinda idle to 50% (internet, music, video, most of the games) it runs at 40-50 C Tj and on full load (Prime95) it jumps to 80 C in a moments. And its all on stock CPU clocks.
Also the difference between core 1 and core 4 is almost 10 C degrees.

I use Coolermaster V8 cooler, and Element S case, which both must be very good.
So I think its because of wrong thermal paste appliance.
I ordered Arctic Cooling MX-5 and will try to reassemble the cooler properly.

But can it be related to some wrong voltages that r set on auto in bios, but don't "set themselves" right?
Or somehow related to a hanging issue?

 

[Slaymate]

You can raise the DRAM Voltage above 1.65v so long as the QPI/DRAM Voltage is within .5v of the DRAM Voltage.

If your DRAM Voltage is at 1.65v then the QPI/DRAM Voltage must be at least 1.15 volts.
If your DRAM Voltage is at 1.70v then the QPI/DRAM Voltage must be at least 1.20 volts.
If your DRAM Voltage is at 1.8v then the QPI/DRAM Voltage must be at least 1.3 volts.
ETC....

Have you tried adjusting the voltages I suggested earlier?

And your cpu is running way to hot considering your at stock speeds with a Cooler Master V8. Try reseating the heatsink and set the fans to run at 100%.

 

[smartech]

HI again everyone.

I monitored the task manager for a while and noticed a litle spikes on all cores during the hangs.
So, I spent a lot (really a lot) of time trying to catch what process caused it (spike was small, so the load of its process dropped immediately).

But after a while I noticed that Net Limiter 3 client CPU usage rising a bit on a spikes.
So I uninstalled it and - WOOHOO - everything is fine, tested it for two days now, seems good.

It was Net Limiter 3 Beta 6, seems to be a beta bug.
Weird that it turned out to be a software problem, but its always better to uninstall than RMA.
And I don't need to modify the IOH voltages. But I'll stay on a XMP (1600 memory is better anyway).
Also I'll reinstall the CPU cooler gain as soon as i get the thermal paste.

Thanks for assistance everyone.
-AndreiD