Been fighting this problem for about a year on and off, my system will simply shut down and reboot from time to time almost always when running full screen graphics (FSX, Movies etc.)
Spec is as follows:
Intel DX48BT2 Motherboard
4GB DDR3 Corsair Ram (Dual, Symmetric)
Intel E8600 Core 2 Duo
NVIDIA GeForce 9600GT w/ 1024MBytes
Dual Monitors DELL E248WFP (Digital) 1920x1024 and Acer G195HQ (Analog) (extended mode)
Earthwatt 650W PSU
Operating System Windows7 64bit
Two Harddrives Western Digital SATA 500GB and 250GB
I have all the latest BIOS, video and chipset drivers (I think)
My first thought was the graphics card, swapped that out - no change
Turning off the second monitor makes no difference.
I recently "borrowed" another CPU from work - no change.
I am not an overclocker, BIOS settings are "optimal defaults" when its working it runs fast enough running FSX.
I have run several diagnostics including the Intel Desktop Control center temps and voltages appear OK.
I am testing with OCCT but nothing conclusive so far.
Is it likely there's a problem with the motherboard? The Intel CPU heatsink is the older type (push/turn) and was I suspicious that it wasn't making good contact - it definitely is now but the problem persists.
My other thought was the Power Supply but I think the Earthwatt 650 should be OK?
1. Ran the built-in Windows 7 Mem Diagnostic, it ran and reported no errors but I wasn't convinced - it was supposed to give a report when Windows re-started and it didn't.
2. Downloaded WMD it gave a message that my system had more than 4GB of memory (?) and that he results would not be reliable - but it ran overnight (116 passes) and seemed OK.
3. Downloaded Memtest86 v3.5. v3.5 would not run - the system crashed halfway through the first test, but v3.4 was fine - ran several complete passes.
4. Downloaded Memtest86+ v4.1 - it seemed to detect everything correctly and ran one pass without problem. I am going to let that run overnight tonight.
One thing more, it seems that this problem can occur at any time when running any application but it is much more likely to happen when running full-screen video on the larger (1920 x 1200) monitor.
The main hard-drive just failed - I had noticed that things were getting a little slow, looking back in the event viewer there had been occasional "Bad Block" messages going back several months until the other night when I suddenly realized things were getting really bad, rushed out bought a replacement, cloned it (with a few errors) and am now back running again.
Credit to Windows7 to live with the problem for so long and to keep going albeit very slowly, also the very good software that came with the Seagate replacement - very easy to use, ran a full disc repair and so far everything is cool, and surprisingly fast!
But you have to wonder if some slightly more obvious message would be a good idea? I mean if the drive is failing its unlikely to get better all by itself I think a warning like "Your drive is failing replace it soon" would have gotten my attention sooner.
The question is - could this be the cause of the original problem?
So I have now run every combination of single 2GB memory stick for an extended period, one stick, then the other, then in a different slot etc etc.
It never crashes with a single stick of memory, as soon as I put 2 x 2GB of RAM it will eventually crash - almost always when playing full screen video, so its not a bad memory stick and its not a bad memory slot.
Is it possible that this motherboard doesn't really support 4GB of RAM with a 64bit operation system? (I never had any problems under 32bit XP)
In a desperate attempt to get the system a bit more reliable, and since the problem does seem to be memory-related I am going to try tweaking the memory timings, I have never attempted this and would certainly appreciate any advice...
All I have done initially is to increase everything by 1 (so 7,7,7,20 is now 8,8,8,21) - so far nothing bad has happened.