Troubleshooting

[Crispy002]

Here's the story, I have:
i7 920
EVGA X58
6gb OCZ Gold 1600
2x 640gb WD Caviar Blacks in raid 0

About a month after I built this computer, I had a hard drive 'error' get flagged in the intel matrix storage manager. It was overclocked to 3.2ghz at the time. On reboot, HDD detection was slow and the onboard raid controller showed there was an error on one of the drives. After several a few restarts I was able to load into windows and under the intel raid software, flagged the drive as 'normal' and everything ran fine for another few months. Same problem, same solution.

A few months later, 3.4ghz, I had the same thing occur. After dozens of restarts I still could not boot on my raid 0 array. After an hour or so of having the computer off, I tried again and everything booted fine. I identified the drive with the error, assumed it was bad, RMAed the drive. I ran memtest for around 3 hours to determine if memory was somehow at fault.

About 2 months ago, my computer lagged, the intel storage matrix software flagged an error, and once again my computer wouldn't boot. This time it was saying the newly replaced drive was failing. Not believing the software, I moved my array from ports 0,1 on the motherboard to ports 2,3, changed the cables, and reduced my clock speed to 2.22ghz after a BIOS update. Everything ran fine....until....

About a week ago I overclocked back to 3.2ghz. I just now had another failure, this time with the error being on the remaining original drive. Windows 7 mentioned something about a failure to call data.

Since the first failure the following has been done
i7 C0 --> i7 D0 (changed for reasons unrelated to this problem)
Change of PSU 620hx --> 850HX (changed for reasons unrelated to this problem)
2 tests of the latest memtest for 2-3 hours no errors
2 BIOS updates
RMA 1 HDD
Change of SATA ports
Change of SATA cables

If it's possible the motherboard has a faulty ICH10R southbridge I can still replace it through warranty, though I'd really like to avoid RMAing it "just 'cause". Regardless, I'm running out of ideas. I'm debating switching to a SSD, the only thing holding me back is the potential of still having this issue.

The million dollar question? What's broked!?

Additional Info

RAM
8-8-8-24 timing set as per stock
1600 mhz always
1.65v

OC
I used 16x200 for 3.2ghz and 17x200 for 3.4ghz@ 1.05v
Northbridge/Southbridge voltage was/is stock.

 

[Dangman]

Ok, do you run into those issues at all when the CPU is at stock speeds?

 

[Crispy002]

Ok, do you run into those issues at all when the CPU is at stock speeds?

I went roughly the last 4 weeks at stock clocks without issue. It does seem noteworthy that once I increased the clocks again it started acting up hmm... thinking southbridge voltage?

 

[Dangman]

I went roughly the last 4 weeks at stock clocks without issue. It does seem noteworthy that once I increased the clocks again it started acting up hmm... thinking southbridge voltage?

Or an unstable OC. A RAID array created using the motherboard's onboard RAID array uses the CPU. If the OC is unstable, then you can get the issues you've described.

 

[PAo_ReVoLT]

he is doing raid-0 it is barely need any calculation, i was going to say PSU but your 850w should be sufficient. what is your ICH10R temp and Voltage?

 

[Crispy002]

Or an unstable OC. A RAID array created using the motherboard's onboard RAID array uses the CPU. If the OC is unstable, then you can get the issues you've described.

I've run 20+ hours of prime95 in-place FFT on BOTH CPUs I've had without issue. OC strategy was reach target speed, undervolt (since 3.2-3.4ghz is undervolt-capable) until the system would not run prime95 stable, then increase voltage by .02v.

he is doing raid-0 it is barely need any calculation, i was going to say PSU but your 850w should be sufficient. what is your ICH10R temp and Voltage?

Both north and southbridge are watercooled, I doubt heat is an issue.
Northbridge is stable at 1.10v 24c while idling at desktop via e-leet.
Southbridge (IOH/ICH/ I/O voltage right???) is at 1.5v
ICH vcore is 1.05v

I appreciate the help guys keep it coming!

 

[Crispy002]

Bumping because the problem has not been resolved.

Windows began to stutter like crazy after running some hard drive benchmarks (I'm thinking about grabbing an x25-m and wanted to compare speeds) and soon a raid 0 error was flagged. The voltage on the southbridge had been upped from 1.05v to 1.075. I bumped it to 1.1v with no effect, reduced back to 1.05v. Everything is on stock clocks and has been since the OP.

Ran memtest again just to be sure, no errors after a brief 2 passes. I'm making an image backup atm with the intention of breaking the array and testing each drive with western digital's diagnostic tool.

I still doubt it's a drive issue, every time there's been an issue a different drive is flagged with an error. I'm beginning to think it may just be a problem with the mobo's raid controller. RMA time?

What do you guys think?

Edit: Forgot to mention that during the stuttering (but not complete crash), my processor was running 7 degrees C hotter than normal idle while showing typical 0-1% load at desktop.

 

[Crispy002]

After separating the raid array and checking each disk individually, I have some promising results. Both drives passed SMART tests though WD's drive utility. However, HDtune found some pretty ugly problems.

The drive Intel Storage Matrix flagged as broken:

The drive that was still fine: