Data Corruption and BSOD

[RMSe17]

A little history:
Got my new system back in November, after finishing OC optimization, the computer was rock (and prime) stable. Except that any drive connected to my PCI-IDE adapter recieved delayed write transaction errors and data corruption, so then I just put my DVD drives on it, and problem went away.
Then I tried connecting my DVD burner (Plextor 16x IDE) to a SATA slot with a IDE-SATA adapter that I bought. It never got detected in BIOS. Following this, board started to not see SATA drives correctly(lockup on POST) when one was plugged in to the RED (Master) SATA slot. Tried BIOS update to 1707. No SATA resolution.
Bought another PCI-IDE adapter, still data corruption, sounds like PCI slot?! Switching to another PCI slot makes the board lock up in POST if SATA drives present.. ?!
One of the drives connected to the Intel IDE seemed to die, no partition information found even with recovery tools.
Data corruption now apparent on other drives, half the stuff written to the drive fails checksum. Downgraded to 1305 BIOS revision, somehow the "dead" IDE drive got it's paritions back.
Seems like data corruption is affecting all drives now, SATA, IDE, everything.
Random BSOD's with stuff about NON_PAGED_AREA (makes me think either memory or page file) or something, or IRQ something LESS THAN OR EQUAL... etc.. some are even weirder.
Tried Prime 95 test, immediate failure on torture test, both cores.
So, I turn off all OC, set everything to default in BIOS,
Prime 95 still failed immediately on both cores. Except the test that doesn't really use much memory (first test), that one lasts for couple minutes, then fails too.
Tried memtest86+, got no errors. (so... ram is good?)

Any idea?

I'm thinking Motherboard?

 

[Matthias99]

Got my new system back in November, after finishing OC optimization, the computer was rock (and prime) stable. Except that any drive connected to my PCI-IDE adapter recieved delayed write transaction errors and data corruption, so then I just put my DVD drives on it, and problem went away.

...so it was "rock stable" other than that it randomly corrupted your data?

Prime 95 still failed immediately on both cores. Except the test that doesn't really use much memory (first test), that one lasts for couple minutes, then fails too.
Tried memtest86+, got no errors. (so... ram is good?)

Memtest86+ is very good at detecting totally garbage RAM (like a bad chip or address pin on the memory module). It's exhaustive, but not particularly stressful. I've seen situations where the memory is obviously at fault (I OCed it too far or lowered timings too much), but memtest will pass and prime95 will not. Likely this is because Prime95 is also pushing the CPU 100% and causing a lot more bus traffic, which is pushing the memory controller over the edge. memtest86 doesn't stress the CPU much at all.

If prime95 is failing, either your hardware is bad or your OS/drivers are corrupted (likely due to bad hardware).

The FAULT_IN_NON_PAGED_AREA and IRQL_LESS_THAN_OR_EQUAL errors are, in my experience, almost always due to 'bad memory' -- but it might be the motherboard causing the fault and not the actual RAM.

From what you're describing, I would lean towards a bad motherboard (northbridge or IDE controller not working right). But it could be screwy memory as well. Try underclocking your CPU/RAM (and using very relaxed memory timings) and see if the problems persist. If you have multiple DIMMs, you might try them one at a time as well. If possible, try testing the RAM in another system. If those don't fix your problem, it's very likely to be your motherboard.

 

[imported_AndyT]

I would test each stick of memory seperately using Prime95 to see if you have a errors on one stick. Only one stick will go bad in a kit, so you can go from there.

Just for kicks try disconnecting all hard drive excepet for your OS. You might have aPSU loading issue with all of those drive. Just a thought.

 

[RMSe17]

Originally posted by: Matthias99

Got my new system back in November, after finishing OC optimization, the computer was rock (and prime) stable. Except that any drive connected to my PCI-IDE adapter recieved delayed write transaction errors and data corruption, so then I just put my DVD drives on it, and problem went away.

...so it was "rock stable" other than that it randomly corrupted your data?

Well, other than the PCI-IDE controller (At the time I thought I just damaged that controller during move from my other mobo.

Thanks for suggestions guys, will try them out.

 

[RMSe17]

So yea, memory was bad, got it replaced, finally... still hasnt stress tested new system, but I really hope it will work now.

 

[Roguestar]

What kind of temps are you getting? Use the Intel TAT. Any chance your heatsink isn't seated right and it's overheating like crazy?

 

[SuperFreaky]

Originally posted by: RMSe17
So yea, memory was bad, got it replaced, finally... still hasnt stress tested new system, but I really hope it will work now.

I actually guessed that from your title.... had same issue a few month ago

 

[RMSe17]

Originally posted by: Roguestar
What kind of temps are you getting? Use the Intel TAT. Any chance your heatsink isn't seated right and it's overheating like crazy?

Heatsink on my CPU? Could be, I mean, I dont know.. it's Big Typhoon, and I got AS5 under all heatsinks on the mobo...

ASUS Probe reports 60C max on Prim95 (two threads).
From what I remember reading a while back, when I was determining the highest my system would go, C2D starts to mulfunction at around 65C analog (not the coretemp digital).

CoreTemp maxed out at 79C (85C highest safe opeartional temp?).

I do know that I had my ram at 2.35V with no active cooling for few months, and from what I was told at Mushkin forums, that's a big no-no, I just had no idea.

 

[Roguestar]

Screw motherboard utility temp sensors, use this:
http://www.techpowerup.com/downloads/392/mirrors.php

PS: That's pretty high for coretemp. I'd trust it and ITAT more because they read from the sensor inside the CPU die and not the motherboard socket. Try removing and reseating your HSF?