Various BSOD errors during normal use and sleep

I got two more BSOD's, and one of them created a dump file, so I'm attaching that to this post. I'm also attaching the screenshots you requested (the HDD has two screenshots to show all the values at the bottom of the program).

I ran all the tests you posted and none of them gave me any errors.

I'm not sure if it BSOD's in safe mode because it doesn't happen all that often and not in response to anything that I do (it is quite random). I ran the computer in safe mode for a day and no BSOD, but I haven't had one for a few days anyway.

Unfortunately the last dump is of 9th Jan'13. Configure the system for minidumps again.

PS: Have you reinstalled the OS after making hardware changes? Not sure if I've asked this earlier.

Does this zip have the correct dump? I just got another BSOD (while logging in) and then went into Safe Mode to grab the dump. I did the reconfiguring after, though.

Yes, I reinstalled Windows to the SSD. I actually had to do it several times, as I couldn't install SP1 because of a general error. I found that the only way to install it was to only install the networking driver onto my computer and then immediately install SP1 (and then install everything else).

Just so you know OCZ brand SSDs have a very high ROF (rate of failure). Even though tests do not show up problems they often don't because the tests aren't very well structured on testing SSD drives. Keep that in mind as we approach this.

I recommend you remove the SSD drive, install Windows on one of the HDDs, and test stability. If things stabilize, you can return the SSD drive for a replacement. You may have to go through a few before you find a stable OCZ drive - you are paying for pure performance, not reliability. If you want reliability, you're going to need to go with a non-Sandforce controller Samsung, Intel or Crucial drive (Samsung being top-notch).

The crashdumps so far aren't showing a reliable pattern. You can try turning on Driver Verifier and let the system crash some more then send us the resulting crashdumps, as those may prove more reliable. However, if this is certainly been going on shortly after you installed the new hardware, then most likely one of them is bad, in which case these minidumps won't prove worth much in telling us what. Still worth a try.

Yes, it's a hardware issue. We won't be able to ascertain cause from crashdumps. You've already done thorough memory testing so the RAM seems fine. If you haven't already, it's time to test the CPU. Run Prime95 on Torture Test on Large FFT settings for a few hours (make sure CPU cooling is stable since this test produces tons of heat). Follow up with another run on Blend settings instead for a few hours, regardless if something bad happens with the first one.

If Large FFT runs fine for a while but Blend croaks, the CPU is more likely at fault. If it's vice versa, the motherboard is more likely at fault. If both are very bad in state, it could be either or, but more likely CPU. If your system BSOD's during a run, just do the other run and then send us the crashdumps so we can ascertain whether the BSOD was from Prime95 or from something else.

I ran both Large FFT and Blend. The Large FFT test lasted 2 hours until all cores failed, with the first core failing at ~1 minute. With Blend, the cores took quite a lot longer to fail, with the first core failing at 30+ minutes and the total test lasted >4 hours. I got a BSOD after 4 hours, so I'm attaching the crash dump.

I'd start with replacing the CPU then, based on those results. It may be that there's some issues with the internal CPU caches that aren't getting hit with Blend settings until much later. Of course, mobo is still potential suspect here, but I'd still go with CPU first.

I just ran a longer Blend test. 1 core failed after 2 hours and the other three lasted >9 hours (I had to stop it after that).

Is there way to be more sure whether it is the CPU or motherboard? I don't have a spare CPU on hand (my last one was a Core 2 Duo, which doesn't fit), so I would have to go through Intel's warranty service since it's been over 30 days since I bought the hardware from Newegg. It's not much of a hassle to do this, but if it turns out it's the motherboard that is failing, then I would be left without the computer for several weeks while both items were shipped there and back. It's possible, as I can use the university's computers during that time, but it would be nice if I didn't have to.

Nope. That's the problem with the CPU/Mobo/PSU trio (I call them the Trio of Trouble). All three are involved in nearly every activity on a system and all three share a very close relationship, so if one fails, it'll often show symptoms that makes it appear like any of the other two parts (or any other PC part for that matter). That's why in a hardware diagnostic lab, they will have a motherboard testing kit which is a diagnostic version motherboard that the tech will connect the potentially suspect PSU/CPU onto and run diagnostics. If they come up clean, the original motherboard is the offender. They also have PCI POST cards (punny) that can run diagnostics on a potentially suspect motherboard, but those things are quite old, and I'm not sure how they operate with today's mobos (if they even do). Either way, you're looking at a few hundred to a thousand bucks for the kits. Best to just do it the old fashioned way and swap hardware and cross fingers.