New Core2 Duo system stalls, need advice!

[FIZZ3]

Hi guys, I finally assembled my new PC yesterday and it's an awesome system to use... except that it seems to stall every now and then. I'll try to explain:

At random intervals, typically 10 or more minutes from each other, the system will suddenly freeze. Sound will stop playing, video and input devices freeze. Brief, everything and there is no detectable (harddisk) atcivity. It's not a crash though, the system will resume in a few seconds (~5). In game I noticed slight visual corruption after recovery a few times, but it goes away pretty quickly too. There's no slowdowns otherwise, it's smooth and fast in 3D games.

I'd suspect software, but it can't be: it even does this when I'm in the BIOS menus! I'm really at a loss here. I've run 3Dmark and Orthos and they seem fine. I've updated to the latest BIOS and reset to defaults (took battery out as well). Memory is only running at PC5300 speed atm.
One other odd thing is that setting the SATA mode to AHCI causes a reboot after reaching the windows screen (running xp x64).

System specs are in my signature... got the 4 pin and the big plug connected solidly, the 6pin to the video is secure as well. Temps with the stock cooler in BIOS seem ok: 40-smt for board and 30-smt for CPU. Board is passively cooled, so a higher temp should be expected I guess.

Help would be much appreciated!

 

[jcw122]

I'm assuming you've tried it, but have you tried putting the speeds back to stock?

Maybe ghetto-mod to board so it's actively cooled?

 

[FIZZ3]

It's running stock speed, I've not even tried clocking up yet.

I've got the air scoop in the V2000 installed so the 120mm case fan blows air onto the NB and mosfet sinks. I also removed the 'stickers' on the sinks. The NB sink felt quite hot under the sticker, but since I removed it and installed the air scoop, it feels only warm. Board temp (SB temp I believe) also dropped a lot after sticker removal.

It doesn't really crash, it just stalls... which is very annoying nonetheless. I'm thinking there must be some resource conflict... but I don't see it.

 

[orion456]

Sounds like a hareware problem.

Touch your HS, is it hot or cold? Cold might mean poor CPU contact. Reseat the HS making sure nothing is on the pad like a hair.

Could be a memory problem. Try taking out all memory cards except one. If that doesn't fix it, swap memory cards and try again. Try some memory from a computer known to be working. My 2gigs of OCZ memory never did work properly with the P5W DH but one gig worked perfectly.

Change video cards.

Disconnect everything but the CD drive, and boot from it. Then slowly add back in components one at a time.

 

[monkeyBite]

Sounds like maybe a GPU recovery or something. I'm only familiar with ATI cards, but if I have mine (X1900XTX) overclocked too high and it fails out in a game, the GPU recovery will run, reset the card and back to playing, however there is about a 5 second lag (which you describe), as well as some graphic corruption.

I don't know if you have any option in the video card config, but try turning your GPU and graphics memory clocks back to stock and see if you are still having problems.....

 

[FIZZ3]

Thanks for your thoughts, guys.

Yes, I've tried to run the system with as few components as possible (using only harddisks and videocard, and ram and cpu of course).

I tested the RAM now, running memtest86+ from the CD for 4 hours. It completed more than 6 full passes in that time and returned 0 errors. This was with both sticks in.

I will run Orthos this afternoon. Perhaps there is a heat issue. I ran orthos for 30 minutes before and the Intel TAT didn't show crazy high temps, but who knows what happens over time. The CPU heatsink isn't cold to the touch, but not warm either... the sink on the northbridge is warmer.

If this works out fine I'll continue with 3D testing. Card should be at stock speed though. Control panel doesn't seem to show the clocks. Does that coolbits thing still enable this? I've got to find how that worked again.
I used to run 3Dmark for 3D testing, but after seeing the 03 version I'm not sure that's convenient. It only allowed me a demo run at low res, coated in "Futuremark ftw!" screens and didn't even give me a bench result. Are the newer versions better? Can you loop (and bench) them if you register for free?

 

[FIZZ3]

Update: Ran Orthos for 8 hours straight now. Again no errors! Intel TAT shows about 60 degrees C for both cores. Not horrid with the stock cooler, is it?

I'm beginning to suspect the videocard. I guess I'll try throwing 3DMark at it tomorrow. If you have a better suggestion, I'd love to hear it.

 

[TheCheat]

I've got the same board... I didn't have the problems you're having now, but I did have shutdown issues... try updating the bios.

 

[orion456]

At random intervals, typically 10 or more minutes from each other, the system will suddenly freeze. Sound will stop playing, video and input devices freeze. Brief, everything and there is no detectable (harddisk) atcivity. It's not a crash though, the system will resume in a few seconds (~5). .....I'd suspect software, but it can't be: it even does this when I'm in the BIOS menus! I'm really at a loss here.

Microsoft claims that 80% of freezes are caused by driver issues. In this case it sounds like a driver loaded by the bios. Have you tried setting your OS to be plug and play and using windows to install drivers instead of the bios? Also try updating your bios and check that your other drivers are up to date as well.

 

[FIZZ3]

The latest official BIOS is loaded. All the drivers I am using have been freshly downloaded (Intel INF, Nvidia gfx, Creative snd, Asus marvell LAN) from their respective sites. I must say I don't think it is a windows driver issue, since I've observed freezing whilst in the BIOS menus-- which is before anything driver-related is loaded.

I've had the BIOS set to "PnP OS" at some point, but it was still having issues then. It may have reverted to non-PnP now, after the battery-out reset. I will try setting it like back to PnP again though, thanks for the tip.

I'm quite puzzled by the system passing memory and CPU ("blend" mode Orthos) tests for hours. The tests I have done clearly show it has no problems there. Yet, the symptoms I was having were quite apparent and noticable within <30 minutes.
One other difference between the problematic setup and the test runs now is that I've switched power plugs around. I have some hope that the issues may have been due to the previously used power delivery. I had some multi-plug strips branching off each other, even though I was using only the one for the new PC at any one time. But I guess the 3D test will give me the answer...

 

[RJARRRPCGP]

video and input devices freeze. Brief, everything and there is no detectable (harddisk) atcivity. It's not a crash though, the system will resume in a few seconds (~5).

With some video cards, you can get this symptom if the video card is OC'ed too much.

My eVGA GeForce 4 Ti 4200 64 MB has that problem in 3D Mark 2001 SE if OC'ed too much.

And that wasn't because of thermal throttling, AFAIK. GeForce 4 Tis probably didn't have thermal throttling.

It would freeze for at least approximately 10 seconds then unfreeze in the Car Chase test when the above problem occured.

Because you said that it's fine in 3D Mark, it's more likely because of the processor doing the dreaded thermal throttling.

 

[FIZZ3]

With some video cards, you can get this symptom if the video card is OC'ed too much.

My eVGA GeForce 4 Ti 4200 64 MB has that problem in 3D Mark 2001 SE if OC'ed too much.

And that wasn't because of thermal throttling, AFAIK. GeForce 4 Tis probably didn't have thermal throttling.

It would freeze for at least approximately 10 seconds then unfreeze in the Car Chase test when the above problem occured.

Because you said that it's fine in 3D Mark, it's more likely because of the processor doing the dreaded thermal throttling.

No no, I only did the 3DMark test just now. It was memtest86+ and orthos prime that were fine. However, I observed the freezing in 3DMark05, and it seemed to do it during the 'nature' part of it (at least it happened there twice while I watched). It was easily found too, 2nd/3rd run and it showed up, unlike the hours of stable testing in memtest and orthos.

So yes: everything is pointing at the videocard. I'm still a bit scared about the system acting up in the BIOS, but perhaps it's somehow related to a faulty card. I will return the card. In the mean time I could maybe borrow another PCI-E card to check for mainboard issues. I just hope it's the videocard at this point as I fear the RMA people will give me grief about removing the stickers from the heatsinks on the board. It's packaging material in my view... but they may be looking for an excuse to void warranty.

To be continued...

 

[FIZZ3]

Update:
I ran tests using an old 6600GT card and it was absolutely fine. So that settles it; the videocard was defective. I've sent it back. I'll update this thread if new problems arise. Thanks for the help!

 

[dylskee]

Glad to hear you got it worked out FIZZ3 Happy Thanksgiving.

 

[FIZZ3]

Thanks, same to you!

 

[FIZZ3]

Update: Resolved. Got the 8800 GTS in and it works like a charm. (Well, after some searching to get x64 drivers that would recognize the card!)

This new card gets HOT by the way! The 7900GTX was much more subdued and even then I thought it was rather warm when touching the heatpipes. The 8800 is in a whole different class though. Looks like I'll be using that sidepanel fan over the PCIE slots after all...
I feel sorry for anyone cramming two of these or their bigger brothers in their system, that's gonna be a real furnace.