2011 MBP (10.8.4) freezes exactly one hour after reboot

Question

Hello,
I'm having an issue identical to this poster insofar as I can tell, but some test results are different for me. Exactly one hour after rebooting or starting the computer, applications become unresponsive, video stops playing, then a little while later sound stops playing, and the cursor becomes a beachball if interacting with an application. Some windows can still be dragged or brought to the front of the screen for a short time. Eventually everything (except sometimes the cursor) is frozen and the machine never recovers from this state (for at least 15 minutes, I haven't let it go longer).
Most of the time this is accompanied by a CPU use spike and the fans spinning up, but not always. I have trawled through the system logs and disabled lots of LaunchAgents to no avail, so let me just list a few things I've tried and their result.
Disabling wifi: no change
Removing RAM one slot at a time: no difference
Memtest: clean results, froze after 1 hour anyway
Apple Hardware test: clean result on extended test, progress *sometimes* halts at the end of the third or fourth pass after the "Main Logic Board" test had finished (during message "Testing in progress... Extended testing in progress").
System console/system.log shows no distinctive output immediately before freeze, unless I'm doing other things like playing audio or using the internet in which case the wifi interface has activity or CoreAudio turns outputs on, but this is different every time it crashes. I did find that GoogleSoftwareUpdate was implicated in one of the console logs, so I disabled and uninstalled that without effect. I recently started using Time Machine on an external drive, so I disabled that and removed the drive I'd selected from Time Machine preferences - no improvement.
While watching Activity Monitor during one freeze, I saw that cprefsd (root) had spiked to 260% CPU usage, fseventsd (root) to 128%, and several other processes to 100% (UserEventAgent, cprefsd, iTunes Helper, and apsd (root)). Everything else was at 1%.
Symptomatically, this appears to be a software issue (not an instant crash, elevated CPU usage, regular timed interval), but I've been unable to isolate the cause. I strongly suspect Time Machine is responsible (fseventsd is linked to it) but having disabled it in System Preferences I'm not sure if it should be eliminated. Alternately I believe this may be symptomatic of logic board, memory controller, or graphics card failure, but am at a loss to explain the regular (non-random, non-usage-dependent) occurrence. I performed the syslog test prescribed by Linc Davis in this thread and found only regular shutdown codes, so that also seems to point to software issues.
--- last message repeated 9 times ---
Sep 15 22:18:20 localhost kernel[0] <Debug>: Previous Shutdown Cause: 5
--- last message repeated 1 time ---
Sep 15 23:22:24 localhost kernel[0] <Debug>: Previous Shutdown Cause: 0
Sep 15 23:50:11 Solidspace kernel[0] <Debug>: Previous Shutdown Cause: 0
Sep 16 00:20:10 localhost kernel[0] <Debug>: Previous Shutdown Cause: 0
Sep 16 11:29:05 localhost kernel[0] <Debug>: Previous Shutdown Cause: 5
Sep 16 12:41:26 localhost kernel[0] <Debug>: Previous Shutdown Cause: 0
--- last message repeated 1 time ---
Sep 16 14:11:11 localhost kernel[0] <Debug>: Previous Shutdown Cause: 5
This morning I tried leaving the computer on in Safe Boot mode and the freeze occurred on schedule, with the windows becoming unresponsive at 60 minutes in, and the the beachball appearing as soon as I moved the cursor.
If I can get a full backup sorted (only essential stuff is cloned right now), I'll try a clean install of OS X from the recovery partition to see if it nixes whatever process is at fault. I have an appointment with the Genius Bar tomorrow evening, but if this is really a software problem then I'm not sure how they can help (except to replace it at cost). Are there any other tests or fixes I can try? What's the likely prognosis?
Thanks a lot for your help.

Answer 1

Apple replaced the logic board and the fix and it didn't fix the problem. To their credit, they refunded the repair fee. I ended up tracing the issue to a third-party drive I'd bought from NewEgg, a Crucial M4 SSD, which had a very peculiar firmware bug:
Correct a condition where an incorrect response to a SMART counter will cause the m4 drive to become unresponsive after 5184 hours of Power-on time. The drive will recover after a power cycle, however, this failure will repeat once per hour after reaching this point.
I would have probably gotten the update notice had I purchased direct from Crucial (or had it come pre-installed if I'd bought it a month later). If you have a similar issue, check your drive firmware!