Built the computer a week ago, was rock solid for 1 straight week. Ran all tests fine.
I was getting memory errors earlier in the day, very consistently. Memtest, the windows app, would give an error instantly during pretest when testing above 2000mb. Prime95 would give rounding errors on 2-3 workers instantly. Oddly, testing outside of Windows produced no errors, both Memtest86+ boot CD and the built in Windows 7 memory checker. Multiple soft restarts and the same behavior was seen.
After powering down the system completely and restarting about 10 minutes later, errors are gone no matter how hard I push the tests. Been solid for hours no matter what I throw at it.
Basically, it's doing the complete opposite of what it was doing earlier today, where it would crash on anything when above 2gb, instantly.
I assumed it was a bad stick of memory, but now I have no idea what to believe. Was it just an isolated bad boot I had this morning?
Windows 7 Pro 64 bit
Antec NEO ECO 520C 520W
Intel Core i7-860 Lynnfield 2.8GHz (Stock speed)
CORSAIR DDR3 1600 CMX4GX3M2A1600C9 (Running at 1333, SPD settings.)
BFG 8800 GTS OC (Stock Factory OC speed)
Antec 902 Case.
I have seen other reports of similar problems when people turn their PCs off at night and then first boot up the next day. This is sometimes referred to as a cold boot. When you have already booted the computer, then shut it down and restart it, that's typically not called a 'cold restart'.
What do you have the V on your RAM set to? Try testing it at 1.6V.
When I cold boot it, it doesn't work. It just doesn't work at all. Anything testing over 2gb, or any prime95 test, errors out before it even starts.
I have to warm restart it, then it works fine. I've since switched to the 1.65v XMP profile, as directed by Corsair tech support, but the machine hasn't gotten cold enough to do a real test. It is extremely stable at this voltage when warm right now though. Will check tomorrow.