Okay, I never got around to posting back because I had to move house, and we're still getting things settled -- I'm staying in temporary digs while our new house is being built, and that's taking longer than it should have, so my net access is limited.
I haven't had a BSOD since the last one I reported -- I'll try and get that report information for you when I have time, koolkat -- but shortly after getting the computer set up in our temporary digs, I began to experience frequent freezes and audio looping (where the sound loops into a buzzing). Sometimes the system will manage to stabilize and continue on its way, sometimes it has to be reset by holding the power button down.
This lock-up only seems to happen when I'm on the desktop or trying to browsing the internet. It never happens when I'm in a game or watching a video file, but it will happen when I'm just looking at icons on the desktop or when I'm using Firefox. Usually I have music playing on Winamp. The audio loop only seems to happen if the sound is outputting through my HDMI cable. If I'm listening to music through USB headphones, the music stops immediately.
During one of these lock ups, the screen turned pink, the color looked like lines running up the screen. On one occasion, the mouse turned into a cursor and the cursor was a weird rainbow color. These particular symptoms have only happened once each, they haven't occurred during any of the other lock ups.
So those are the symptoms I'm currently experiencing. Sometimes the computer locks up almost immediately after booting Windows, sometimes it takes hours (I used the computer from around midnight to 9AM without incident, and then it locked up). I've tried scrubbing and replacing video drivers, but that had no effect on the freezing. I also scanned my system for malware or viruses and found nothing.
So here's what I've tried in my search to figure out what in the hell has gone wrong: the first thing I did after fooling with drivers was run a scan of my HDD. I right clicked on the HDD and told it to scan for errors and bad sectors. The computer rebooted and started scanning, which took about four hours. The scan found no errors and no bad sectors.
I then moved on to the RAM. I'm running 2x4GB sticks of Corsair Vengeance. There was a mix up with my roommate's RAM while we were working on our PCs, and so I'm running different sticks from different kits, but they're the same brand and have identical specs (1600mhz, 1.50v, 9-9-9-24 timing). My former roommate is running the other two sticks from these kits in his PC, and he's had no problems whatsoever, so I don't think there's a compatibility issue (more on that later).
So it was at this point that I broke out memtest86+. The first time I ran the program, it found 18 errors within minutes (although none of them came up in red). Given what I've read about RAM errors (if you get even one, replace the RAM) I immediately stopped the test, powered down the computer, and stripped out one of the two DIMMs. I wanted to test them separately to see if one or both had errors.
I ran memtest on the first DIMM for 5 hours, did 5 passes, and found no errors. I removed the first DIMM, inserted the second DIMM, and ran the same 5 hour/5 pass test. No errors.
At this point, I was confused. I put both sticks back in the computer and began another test of both DIMMs together. At around the 8 hour mark, on the 5th pass, I found a single error.
Because I had found no errors on the sticks individually, I wondered if something might be wrong with one or more of the ports on my motherboard. I left the first stick in the port it was in (both sticks had shown clean in that port) and moved the second stick down to the port at the far end. I then ran memtest for a whopping 20 hours. 10 passes. Found zilch for errors.
Feeling stumped, I gave up on memtest and rebooted my computer. It took a little while, but eventually it froze up again. I decided to remove one of the DIMMs and see if that would solve the problem. While I waited for my next freeze, I downloaded Prime95 so that I could use the stress test feature to put my CPU through the ringer. The test I picked also stressed the RAM, so I thought it'd be a good way to kill two birds with one stone.
While running the stress test, the computer locked up. I turned it off and swapped the first DIMM for the second DIMM. When I rebooted, I got an odd message saying I'd had a boot failure due to overclocking. My CPU has not been overclocked, so I wasn't sure what the message was on about. All I could assume was that it was referring to my RAM: I had gone into the BIOS and manually set the RAM up, because the auto settings wanted to run my 1600mhz ram at 1333mhz, and the timing was all wrong (11-11-11-29 as opposed to 9-9-9-24). This had not been a problem previously, but I booted into the BIOS to reset it to AUTO, only to find it had already been reset to auto.
I tried rebooting and got the same message again, so I swapped the RAM again to see if the other DIMM would let me in. Same message again. The only way I ever got past it was by removing the battery on my motherboard to reset to CMOS. I put the second DIMM back in and the computer booted fine with the Default motherboard settings.
I then set about stress testing my CPU and RAM again with Prime95. I let it run for about an hour, with the CPU pushed all the way to 100%, and there were no errors and the system never crashed, so I assumed the first DIMM was faulty. I left the second DIMM in the machine, turned it off for a while to let it rest, and then proceeded to use the computer without incident from around midnight to 9AM, when it suddenly froze again. All I was doing was writing a post on a message board and listening to music, there was nothing processor intensive going on. Browsing the internet was pretty much all I did last night, without incident. I'm back to the other DIMM now, in a port on the board I haven't been using.
I'm stumped as to what to do next. It's hard for me to believe that two DIMMs from two different kits have suddenly gone bad -- the DIMM that was already in my machine prior to the mix-up has been working A-okay for over a year, but the machine locks up even when that's the only stick present. I would just swap RAM with my roommate and see what happens, but he's not living with me anymore and we're two far apart for that little experiment now.
The HDD appears to be fine, I can put ample amounts of stress and heat on the CPU without getting errors or locking the system... the RAM seems like the most likely culprit, but I'd hate to shell out $80 to replace it and have the issue continue. I'm not sure how to test things like the video card or the motherboard itself, but I'd like to rule them out before I start throwing money around.
Any advice? The biggest pain here is that, as far as I can tell, there are no error reports of any kind after these crashes.
