Computer freezes randomly, may be hardware

chumchu

Honorable
Oct 17, 2013
1
0
10,510
Hello all, this is my first post and I hope it's okay. It might be a bit long so I'll start with a tl;dr in bold. If you're going to read the whole thing you can skip that.

TL,DR: The computer freezes, seemingly at random and can do most tasks with luck, but will always seems to freeze sooner during a game. I'm fairly certain that it's a hardware fault and suspect that it is either the CPU, GPU or RAM and have no spares to switch out, any way I can definitely test them to see witch is broken?



Long version:

So after having a cheap PSU blowout a few weeks back and replacing it with a new one I changed my motherboard as the PC wouldn't boot.

The tested refurbished motherboard finally arrived and everything seemed to work fine (New motherboards with LGA 1366 aren't produced anymore as far as I know). I bought an SSD and installed Windows 8 Pro N on it, and then Linux Mint alongside it. The old HDD is still there and just for storage, I haven't formatted it yet. They're plugged into SATA slots 2 and 3 by mistake, is that an issue?


The computer freezes seemingly at random, sometimes it might be ten minutes in or several hours. When not doing anything stressful, like general web browsing, or playing a game, or just finishing a game. I can't really notice any link between what's being done with the PC and its freezes. Sometimes it will freeze at login screen, sometimes after several hours. Though I have noticed that if it does work fine for several hours, playing a game for a while will always make it freeze within an hour, but usually much sooner.

When freezing the sound will usually cut out and the screen will remain on a single frame, uncorrupted. Just once has the sound gone on an endless repeating loop. The PC doesn't respond to any mouse or keyboard commands and I'm forced to turn it off with the case power switch.

This happens on both Windows and Linux so I'm fairly certain that it's a hardware fault.
I reinstalled graphics drivers on both systems and the problem persisted. It occurs with the xorg wrapper and proprietary AMD drivers on mint.

The temperatures all seem fine and I have an aftermarket CPU cooler. I ran memtest86 and it found one error as detailed on the bottom. After switching the RAM sticks of the outer slots around and running the test again it no errors were found. I made extra certain to populate the memory slots correctly. The PC won't boot with only one stick according to the motherboard manual, so I tried taking each one out in turns and a failure occurred every time just as before.
I only noticed memtest errors when the third RAM bay was filled (C1). The address is always in the form 0000813axxx with numbers in place of the xxx. The PC will still freeze if only two DIMM slots are occupied but no memory errors are detected.

Once after many attempts Windows failed to boot and produced a blue screen about clock_watchdog_timeout. So I thought it might be the CPU, I ran prime95, and sometimes the PC will freeze as before but the frozen screen doesn't show any errors on prime95 or it will go one for hours without error. Then if I decide to play a game it will freeze after a while whilst playing, so I think it could also be the graphics card. I have no onboard video so I can't think of anyway of testing the CPU only to see if it's that or the GPU. The freezes happen just as frequently when booting Windows into safe mode.

I tried tinkering with the CPU settings, upping the voltage, under-clocking, even overclocking and turning off some Intel power-saving features. None of these changes seem to affect the freezing. Maybe their frequency, but it's hard to tell as it's quite random and either way the computer always ended up freezing at some point.

With the HDD and other unnecessary hardware removed the problem persists, I tried different mice, the keyboard is analogue.

Also I didn't install a bracket to fit the SSD into the case properly yet. At the moment it's just loose on base of the large drive bays, I figured this wasn't an immediate problem as it can be used externally according to one of the illustrations and doesn't have moving parts, don't think it's relevant but will put it out there.

Any way to know for sure if it's the CPU or GPU? I can't really afford to replace both unless I absolutely have to. My PC will be used for playing computer games, CAD, and running CFD and numerical methods so I would intend to get higher performance components.


I'm not sure where to go from here for troubleshooting and would appreciate any advice.
Thanks in advance.



Computer specs:

Motherboard: ASUS P6T V2 Deluxe
PSU: Corsair TX750M 750W
Graphics card: AMD Radeon HD 5770
CPU: Intel i7 920 default clock
RAM: Corsair xms3 DDR3 in three 2GB sticks 1333MHz
SSD: Toshiba 256GB
HDD: Hitachi 1TB

Details of the memtest result
Test 7 (That's the one with random numbers)
Failing address: 0000813a804 129.2MB
Good f62d3566
Bad f62d3166
Err bits 00000400