MSI P35 Neo-F (MS-7360): Computer randomly freezes while playing games

Question

Hi there,
I have reached the end of my technical knowledge right now and I'm not even sure if this area of the forum is appropriate. The problem:
While playing higher-end games like Crysis 2, Anno 1404, Mass Effect 2 etc. my computer randomly freezes and the last few milliseconds of sound are being looped on my speakers (kind of like a buzzing sound).
By "freezing" I mean that my computer becomes unresponsive to any action other than pressing the Reset-Button or holding down the Power-Up-Button (no reaction to Ctrl+Alt+Del, Alt+F4, Alt+Tab, no mouse movement... you get it).
There's no exact timing as to when these freezes occur. The fastest freeze happened after less than a minute of playing a game, slowes one (just today after switching to my newer case with better cooling) took about 2 hours.
This freeze only occurs while my System is under load. No freezes while browsing the web, using office programs, listening to music or watching videos.
System Specs:
    Mainboard: MSI P35 Neo-F (MSI-7360)
    CPU: Intel Core2Quad Q6600 @2.4GHz
    RAM: 4x 1GB DDR2-800 Corsair XMS2
    Graphics: Gainward GeForce GTX 560 Ti, 1GB GDDR5 VRAM
    PSU: Antec High Current Gamer 620W (upgraded from Xilence XP600, 600W)
    New CPU Cooler: Arctic Freezer 13 Pro (upgraded from OEM Intel Boxed Cooler)
    New Case: Zalman Z9 Plus w/ 4x 120 mm Fans (Upgraded from a Case with only 1x 250mm side Fan)
    2 Western Digital HDD (1x 2TB, 1x320 GB)
    1 DVD Reader/Burner (unimportant for that error I guess)
    OS: Windows 7 Ultimate 64-Bit
    2 Screens: 1x 24" Asus VS248H and 1x 21" LG Flatron Wide W2240T
Since this error is bugging me for quite a while now and since this is not the first forum I'm trying to get help from, here's a list of steps i already took to get rid of it, obviously to no good use:
    reinstalled games where the freeze occurred
    updated all Drivers (including cleaning old versions in safe mode; also tried older drivers)
    updated BIOS to v1.A0
    Reinstalled Windows
    bought new Case and CPU Cooler to improve Air Flow/Cooling
    checked Seating of all components (took out, cleaned, reapplied thermal grease where neccessary, re-inserted)
    cleaned old case of dust (obviously no problem in the new case, which is 2 days "old")
    ran 7 sequences of Memtes86+ without any errors
    ran Prime95 without any errors for about 45 Minutes
    ran Furmark for 15 Minutes under full load (burn-in test) without any problems, even though my graphics card hit the 100°C frequency-cut level
    ran Furmark and Prime95 simultaneously without any problems
    took records of component temperatures and system voltages without noticing anything unusual (at least not after i switched to the newer PSU, which is about 1 Month old - yet didn't resolve the problem)
    tested with cyclic removal of 1 RAM device to further rule out RAM malfunction
    tested with only 2 RAM devices in dual channel mode (Slots 1 and 3)
    tried to reduce RAM clock (800 MHz --> 667 MHz) to no avail
    checked S.M.A.R.T. values of HDDs, no issues reported
    tested without any background programs
    raised/lowered voltages to RAM, CPU and Chipset in BIOS to increase logic circuit stability/decrease heat production
I need to mention: After I transferred my system to the new case, i was able to play longer than average compared to my old case. So I am guessing that improper cooling was part of the problem, yet improving cooling obviously didn't resolve it.
Since these freezes are really annoying and nobody could help me resolve this so far, I'm hoping to find some new ways of narrowing down the problem or even getting rid of it here.
Regards
Sebb

Answer 1

Quote from: Bernhard on 10-July-12, 13:15:30
Lets be realistic. A system with a 560TI is recommended to have a around 31Amps under full load, a CX500 would be sufficient.
Have you monitored your CPU and memory utilisation ?
As suggested before, test the GPU in another computer to see if the behaviour is the same.
Whew, I thought for a second there that my common sense had left me when I saw that 60-Amp-Post ^^
Yes, I monitored CPU and RAM usage, there is no recognizeable pattern to be seen and it almost never happens under full load.
I will see if a friend of mine can test my GPU on his computer, and I will test it underclocked right away.
Though one more question regarding that: Should I only underclock GPU Core Clock or both Core and RAM clock?