Display driver amdkmdap stopped responding...

xeon84

Honorable
Jun 11, 2012
11
0
10,510
Hi,

I’m new to the forum. I’ve seen a lot of good advice handed out here, hopefully someone will be able to assist me with my issue. My hair is starting to recede on its own, me pulling it out isn’t helping! Thanks in advance.

I built a new system 5 weeks ago with the following spec:
• Intel i5 - 3570K (IvyBridge)
• MSI Z77A-G43
• 8Gb (2x4Gb) Corsair Vengeance Black LP 1600Mhz
• XFX ATI/AMD Radeon 6870 1Gb
• 550W OCZ ZT PSU
• Arctic Cooling Freezer i30
• Crucial m4 128Gb SATA-III (SSD)
• Asus Xonar DG
• Fractal Design R3
• Windows 7 Home Premium x64

NO overclock (only Intel Turbo thingy), all latest drivers and BIOS.

I ran Passmark Burn-In-Test on it when first built and all seemed ok (although I now have some doubts over whether the graphics part may have stopped…).
Anyway, the system was used for about 4 weeks for non-intensive tasks (web-browsing, music, excel, work, pictures etc.) no issues at all.

Until… I tried to run Crysis (original – part of maximum edition 3 part set), one of two games (other was Fallout 3 GOTY edition, see later) I bought to try out on my new PC.

I encountered many stability issues with Crysis (which I believe is buggy at the best of times). In the initial cut-scene on the plane (about 1 minute in) it kept crashing to desktop (went black for a few seconds then desktop message that Crysis has stopped working blah blah. It really loved to crash when about to jump from the plane. I tried lowering settings (initially these were set to 1920x1080, all high, 2x AA, vsync). I tried 32bit exe, 64bit exe, -dx9, Vista compatibility, XP compatibility, lowering settings/res, removing AA and vsync) nothing worked. I patched to v 1.2 and downloaded DirectX web installer/updater and this seemed to improve things (not sure which) and for the first time it would get to the parachute jump part (without skipping using spacebar). Now it started gameplay but anything from a few seconds to a minute the screen would freeze and go black, then recover (sometimes 3 or 4 few times with some gameplay in between) then eventually crash completely to desktop with Crysis has stopped working...

Anyway during a freeze up I Alt-Esc’d back to desktop and caught my first glimpse of the apparently infamous ‘Display driver stopped responding and has recovered’. I checked event viewer and there are numerous occurrences of this event 'Display driver amdkmdap stopped responding and has successfully recovered', corresponding with the freeze ups, each followed eventually by an application error from Crysis (various, depending on how the game was launched I think, see below).

Anyway I felt it may just be Crysis being buggy and decided to try Fallout 3 instead.

Around 30-40 minutes in and after the initial cut-scene, character creation and early part of tutorial (movement as a 1 year old bit) which I guess aren’t too intensive, I get to the 10 year old party bit and FREEZE, exactly the same, screen goes black, recovers, play, black, recovers, play. I got fed up and exited after around 10-15 minutes at which point it crashed and I got a Fallout 3 has stopped working (and log in event viewer - see below)…

Ok so, I’ve tried messing with Crysis settings before eventually giving up and trying another game to see if issues arise and they do, much the same ones.

I’ve tried uninstalling graphics driver/CCC, using Driver Sweeper to remove remnants and reinstalling driver/CCC.

I decided to try and eliminate various factors. I’ve run MemTest86 for 10 hours, 8 passes, no errors. I’ve run Prime95 (blend) for 6 hours twice getting to ‘FATAL ERROR rounding…’ both times after around 1.5 hours on one core; I think both times were core 3. However when trying to confirm if it was core 3 I have since run Prime95 (blend) for 23 hours non-stop with NO issues! So I’m a little perplexed. I ran FurMark 15 minute benchmark a couple of times no issues. Then this morning (after the 23 hour Prime95 marathon) I ran OCCT GPU test twice (DX10, Shader 8, 1920x1080, Full Screen, Error Check). Both times, after just a couple of minutes, screen freezes and OCCT starts reporting hundreds of thousands of errors and event viewer has several ‘Display driver stopped responding and has recovered’ errors, same as in game.

I’ve checked temps (motherboard/card readings via software) CPU maxes out around 56C during Prime95, GPU during gameplay maxes around 68C, 80C in OCCT, 85C in FurMark). Temps look OK (maybe a little high on stress tests, but then issues occur at lower temps with normal gameplay so...). Voltages, although from motherboard/card readings via software look fairly stable to me (although +12V rail reads as 8.18V! Fluctuation is only between 8.10V and 8.27V as reported by HWMonitor, so if you assume 8.18V is actually 12V ish then fluctuations look OK?). Likewise +5V reports as 5.76V, fluctuation from 5.68V to 5.80V. CPU VCore goes from .99V to 1.14V which I believe is increased/decreased automatically by motherboard dependent on CPU loading/clock (throttled or turbo). GPU reports 1.17V without fluctuation.

I’m really a little confused what is at fault here. I want to isolate the issue better to ensure, if it is hardware related, I return the correct component(s) first time to avoid unnecessary cost and delay.

I have so many questions and ideas for solutions:
• Why did Prime95 error at 1.5 hours twice then work fine for 23 hours?
• Is it possible for Prime95 not to have used all memory (system only reports around 3.5Gb in usage during test - system has 8Gb). Does the system keep cycling the same memory parts over and over and keep rest in reserve or constantly cycle reserve and active memory throughout test, using/testing all of it?
• Does working fine for 23 hours eliminate memory/cpu as issue or does previous failure override?
• If MemTest86 returns 0 errors after 10 hours/8 passes does that eliminate memory as the issue?
• Why did FurMark work fine for 15 minute benchmark, but OCCT GPU test faults within around 5 minutes?
• Would an older graphics driver or driver only install without CCC be a good move (currently running 12.4 with CCC – is this stable?)
• Do I need to try Windows reinstall (please no!)


Any suggestions on best way forward or different things to try would be much appreciated.

Please HELP!

Thanks,

Rob

Prime95 Error:

‘[Jun 11 00:24] Test 15, 5300 Lucas-Lehmer iterations of M14942209 using Core2 type-2 FFT length 960K, Pass1=256, Pass2=3840.
[Jun 11 00:25] FATAL ERROR: Rounding was 0.4992675781, expected less than 0.4
[Jun 11 00:25] Hardware failure detected, consult stress.txt file.
[Jun 11 00:25] Torture Test completed 122 tests in 1 hour, 46 minutes - 1 errors, 0 warnings.
[Jun 11 00:25] Worker stopped.’

Various application errors:

Error 10/06/2012 20:36:31 Application Error 1000 (100) "Faulting application name: Crysis.exe, version: 1.1.1.6115, time stamp: 0x47c58d57
Faulting module name: CryRenderD3D9.dll, version: 1.1.1.6115, time stamp: 0x47c58dbe
Exception code: 0xc0000005
Fault offset: 0x00021b8b
Faulting process id: 0xb34
Faulting application start time: 0x01cd473ff98d8d1e
Faulting application path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin32\Crysis.exe
Faulting module path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin32\CryRenderD3D9.dll
Report Id: 9427eb8c-b333-11e1-838b-8c89a5c3853d"

Error 10/06/2012 20:33:23 Application Error 1000 (100) "Faulting application name: Crysis64.exe, version: 1.1.1.6115, time stamp: 0x47c58f4f
Faulting module name: CrySystem.dll, version: 1.1.1.6115, time stamp: 0x47c58fb1
Exception code: 0xc000008f
Fault offset: 0x0000000000085c8c
Faulting process id: 0x1754
Faulting application start time: 0x01cd473fe4b5666c
Faulting application path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin64\Crysis64.exe
Faulting module path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin64\CrySystem.dll
Report Id: 23eff9c9-b333-11e1-838b-8c89a5c3853d"

Error 10/06/2012 20:32:41 Application Error 1000 (100) "Faulting application name: Crysis64.exe, version: 1.1.1.6115, time stamp: 0x47c58f4f
Faulting module name: CrySystem.dll, version: 1.1.1.6115, time stamp: 0x47c58fb1
Exception code: 0xc000008f
Fault offset: 0x0000000000085c8c
Faulting process id: 0x16a4
Faulting application start time: 0x01cd473fc8d5fb71
Faulting application path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin64\Crysis64.exe
Faulting module path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin64\CrySystem.dll
Report Id: 0ae404b6-b333-11e1-838b-8c89a5c3853d"

Error 10/06/2012 20:27:43 Application Error 1000 (100) "Faulting application name: Crysis.exe, version: 1.1.1.6115, time stamp: 0x47c58d57
Faulting module name: CryRenderD3D9.dll, version: 1.1.1.6115, time stamp: 0x47c58dbe
Exception code: 0xc0000005
Fault offset: 0x00021b8b
Faulting process id: 0xeec
Faulting application start time: 0x01cd473e66e74dc7
Faulting application path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin32\Crysis.exe
Faulting module path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin32\CryRenderD3D9.dll
Report Id: 59676f15-b332-11e1-af4f-8c89a5c3853d"

Error 10/06/2012 20:21:42 Application Error 1000 (100) "Faulting application name: SharedServiceHost.exe, version: 6.1.1839.0, time stamp: 0x4d38a956
Faulting module name: KERNELBASE.dll, version: 6.1.7601.17651, time stamp: 0x4e21213c
Exception code: 0xe0434352
Fault offset: 0x000000000000cacd
Faulting process id: 0xa34
Faulting application start time: 0x01cd473e0a65867a
Faulting application path: C:\Program Files\Windows Server\Bin\SharedServiceHost.exe
Faulting module path: C:\Windows\system32\KERNELBASE.dll
Report Id: 81e19041-b331-11e1-af4f-8c89a5c3853d"

Error 09/06/2012 22:55:24 Application Error 1000 (100) "Faulting application name: Fallout3.exe, version: 1.0.0.12, time stamp: 0x48d194b3
Faulting module name: ntdll.dll, version: 6.1.7601.17725, time stamp: 0x4ec49b8f
Exception code: 0xc0000005
Fault offset: 0x00038dc9
Faulting process id: 0x798
Faulting application start time: 0x01cd46865baf113c
Faulting application path: C:\Program Files (x86)\Bethesda Softworks\Fallout 3\Fallout3.exe
Faulting module path: C:\Windows\SysWOW64\ntdll.dll
Report Id: d03275bd-b27d-11e1-a541-8c89a5c3853d"

Error 09/06/2012 22:05:39 Application Error 1000 (100) "Faulting application name: crysis64.exe, version: 1.1.1.5767, time stamp: 0x471f96dc
Faulting module name: CryRenderD3D9.dll, version: 1.1.1.5767, time stamp: 0x471f9751
Exception code: 0xc0000005
Fault offset: 0x00000000000271e3
Faulting process id: 0xe00
Faulting application start time: 0x01cd4682f3c7df84
Faulting application path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin64\crysis64.exe
Faulting module path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin64\CryRenderD3D9.dll
Report Id: dd6594f1-b276-11e1-a541-8c89a5c3853d"

Error 09/06/2012 21:55:31 Application Error 1000 (100) "Faulting application name: crysis64.exe, version: 1.1.1.5767, time stamp: 0x471f96dc
Faulting module name: KERNELBASE.dll, version: 6.1.7601.17651, time stamp: 0x4e21213c
Exception code: 0x0000087a
Fault offset: 0x000000000000cacd
Faulting process id: 0x15b4
Faulting application start time: 0x01cd4681657a81e0
Faulting application path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin64\crysis64.exe
Faulting module path: C:\Windows\system32\KERNELBASE.dll
Report Id: 73147951-b275-11e1-a541-8c89a5c3853d"

Error 09/06/2012 21:32:37 Application Error 1000 (100) "Faulting application name: Crysis.exe, version: 1.1.1.6115, time stamp: 0x47c58d57
Faulting module name: KERNELBASE.dll, version: 6.1.7601.17651, time stamp: 0x4e211319
Exception code: 0x0000087a
Fault offset: 0x0000b9bc
Faulting process id: 0x894
Faulting application start time: 0x01cd467e884e5ce3
Faulting application path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin32\Crysis.exe
Faulting module path: C:\Windows\syswow64\KERNELBASE.dll
Report Id: 3fa1ccce-b272-11e1-a492-8c89a5c3853d"

Error 09/06/2012 15:43:12 Application Error 1000 (100) "Faulting application name: Crysis.exe, version: 1.1.1.6115, time stamp: 0x47c58d57
Faulting module name: KERNELBASE.dll, version: 6.1.7601.17651, time stamp: 0x4e211319
Exception code: 0x0000087a
Fault offset: 0x0000b9bc
Faulting process id: 0x5c0
Faulting application start time: 0x01cd464df474dcb2
Faulting application path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin32\Crysis.exe
Faulting module path: C:\Windows\syswow64\KERNELBASE.dll
Report Id: 6f84a0e6-b241-11e1-a483-8c89a5c3853d"

Error 09/06/2012 15:24:47 Application Error 1000 (100) "Faulting application name: Crysis.exe, version: 1.1.1.6115, time stamp: 0x47c58d57
Faulting module name: KERNELBASE.dll, version: 6.1.7601.17651, time stamp: 0x4e211319
Exception code: 0x0000087a
Fault offset: 0x0000b9bc
Faulting process id: 0x13d8
Faulting application start time: 0x01cd464b432d9362
Faulting application path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin32\Crysis.exe
Faulting module path: C:\Windows\syswow64\KERNELBASE.dll
Report Id: dce2f1d8-b23e-11e1-a480-8c89a5c3853d"

Error 09/06/2012 15:21:31 Application Error 1000 (100) "Faulting application name: crysis64.exe, version: 1.1.1.6115, time stamp: 0x47c58f4f
Faulting module name: CryRenderD3D9.dll, version: 1.1.1.6115, time stamp: 0x47c58fca
Exception code: 0xc0000005
Fault offset: 0x0000000000027053
Faulting process id: 0xafc
Faulting application start time: 0x01cd464aa41cb5d2
Faulting application path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin64\crysis64.exe
Faulting module path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin64\CryRenderD3D9.dll
Report Id: 684ce093-b23e-11e1-a480-8c89a5c3853d"

Error 09/06/2012 15:14:03 Application Error 1000 (100) "Faulting application name: LANConfigSvc.exe, version: 6.1.8800.16385, time stamp: 0x4d6ed265
Faulting module name: KERNELBASE.dll, version: 6.1.7601.17651, time stamp: 0x4e21213c
Exception code: 0xe0434352
Fault offset: 0x000000000000cacd
Faulting process id: 0x4f0
Faulting application start time: 0x01cd4648effeb82c
Faulting application path: C:\Program Files\Windows Server\Bin\LANConfigSvc.exe
Faulting module path: C:\Windows\system32\KERNELBASE.dll
Report Id: 5d072cf6-b23d-11e1-af43-8c89a5c3853d"

Error 09/06/2012 15:03:20 Application Error 1000 (100) "Faulting application name: Crysis.exe, version: 1.1.1.6115, time stamp: 0x47c58d57
Faulting module name: atidxx32.dll, version: 8.17.10.432, time stamp: 0x4f7e5155
Exception code: 0xc0000005
Fault offset: 0x00023b29
Faulting process id: 0x1a30
Faulting application start time: 0x01cd46486096b9da
Faulting application path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin32\Crysis.exe
Faulting module path: C:\Windows\system32\atidxx32.dll
Report Id: de1c1b26-b23b-11e1-b989-8c89a5c3853d"

Error 09/06/2012 15:03:13 Application Error 1000 (100) "Faulting application name: Crysis.exe, version: 1.1.1.6115, time stamp: 0x47c58d57
Faulting module name: KERNELBASE.dll, version: 6.1.7601.17651, time stamp: 0x4e211319
Exception code: 0x0000087a
Fault offset: 0x0000b9bc
Faulting process id: 0x1a30
Faulting application start time: 0x01cd46486096b9da
Faulting application path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin32\Crysis.exe
Faulting module path: C:\Windows\syswow64\KERNELBASE.dll
Report Id: d9d4eb8d-b23b-11e1-b989-8c89a5c3853d"

Error 09/06/2012 14:59:45 Application Error 1000 (100) "Faulting application name: crysis64.exe, version: 1.1.1.6115, time stamp: 0x47c58f4f
Faulting module name: CryRenderD3D9.dll, version: 1.1.1.6115, time stamp: 0x47c58fca
Exception code: 0xc0000005
Fault offset: 0x000000000001db63
Faulting process id: 0x142c
Faulting application start time: 0x01cd4646cc70150a
Faulting application path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin64\crysis64.exe
Faulting module path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin64\CryRenderD3D9.dll
Report Id: 5dc2d97b-b23b-11e1-b989-8c89a5c3853d"

Error 09/06/2012 14:48:43 Application Error 1000 (100) "Faulting application name: crysis64.exe, version: 1.1.1.6115, time stamp: 0x47c58f4f
Faulting module name: KERNELBASE.dll, version: 6.1.7601.17651, time stamp: 0x4e21213c
Exception code: 0x0000087a
Fault offset: 0x000000000000cacd
Faulting process id: 0x434
Faulting application start time: 0x01cd4645ca4914ae
Faulting application path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin64\crysis64.exe
Faulting module path: C:\Windows\system32\KERNELBASE.dll
Report Id: d344455d-b239-11e1-b989-8c89a5c3853d"

Error 09/06/2012 14:41:33 Application Error 1000 (100) "Faulting application name: crysis64.exe, version: 1.1.1.6115, time stamp: 0x47c58f4f
Faulting module name: KERNELBASE.dll, version: 6.1.7601.17651, time stamp: 0x4e21213c
Exception code: 0x0000087a
Fault offset: 0x000000000000cacd
Faulting process id: 0x17dc
Faulting application start time: 0x01cd4644b388f35d
Faulting application path: C:\Program Files (x86)\Electronic Arts\Crytek\Crysis\Bin64\crysis64.exe
Faulting module path: C:\Windows\system32\KERNELBASE.dll
Report Id: d2f937ea-b238-11e1-b989-8c89a5c3853d"
 
Ok. Forgot to say I've also run Intel Burn In test on maximum 20 cycles no issues.

Since I've been home this evening I've run OCCT again four times with and without AA and shaders set to 1 and 8. The disc image thing was completely static from the start and show errors almost immediately. I ran FurMark 15min benchmark with and without AA, both completed perfectly.

I've been looking at the differences between FurMark and OCCT. It seems, if I understand correctly, FurMark uses OpenGL, OCCT uses DirectX (as do both Crysis and Fallout 3)?

If this is the case that may help with a resolution? What should I do next? I've already tried web installer/update... Could it still be something else...?
 
It looks like you have covered just about everything. Your temps are in the norms even at 85c under load is high but not so high that it would cause crashs. Now if that were on idle then you would have a problem.

You have tried older drivers for your card? Are any of your games Steam related? What I am wondering about is your 8v when it should be 12v. Even though that one OCZ you have is suppose to be on a good PSU list.

So when you use the intel thing to boost your CPU and run everything else. I wonder if it isn't your PSU that is faulty? It would explain the fact how your system crashed after 1.5 hours and then it went 23 hours.

Anyway the only way to see for sure is to swap out your PSU and then run the programs. I could be wrong but after seeing eveything you did and then the 8v on the 12v rail. I just don't know.

It has something to do with your drivers and not enough power will cause the same problems. Anyway get back to us with answers to the questions and we will se what else we can do. Good luck.
 
^Thanks for the pointers.

Ok. I borrowed a multimeter and followed the guide I found here (really easy to follow):

http://www.hardwareheaven.com/guides/testingPSU/

I then adapted the basic principles of the above guide and applied that to testing the graphics card supply as well.

Readings from 20/24 pin connection whilst CPU under load (running Prime95) as follows:

3.3V rail: Actual 3.42V
5V rail: Actual 5.04V
12V rail: Actual 12.26V

Readings from 6 pin connector whilst GPU/CPU under load (OCCT) as follows:

12V rail: Actual 12.09V

So, 5V and 12V appear comfortably within acceptable parameters 3.3V maybe a touch borderline, but ok. 12V was only rail to show any fluctuation and only by 0.01V between 12.25V and 12.26V on 20/24 pin connector.

I think I'm fairly satisfied that the PSU is not to blame.

I'm leaning towards thinking it's a graphics isolated issue (possibly system memory - on basis of previous Prime95 fails). So either the card is duff (seems odd seeing as FurMark works fine) or it's a driver/DirectX/OS issue.

I have only installed 12.4 drivers/CCC thus far, so no old drivers to remove (I have previously uninstalled 12.4, run Driver Sweeper then re-installed 12.4).

I think I'll try uninstalling 12.4 and install an older version tomorrow evening (had enough for now). Driver only at first, then with CCC as well. Failing that I think my weekend may soon be consumed by Windows re-installation!!!

Any other ideas greatly appreciated...
 
Forgot to pick up on your Steam comment ^DM186

Neither game is on Steam, but I do have some Steam games (none installed ATM).

What were you leading to in relation to Steam?

Could this have been a potential cause? Or is Steam a resolution? I could try HL2?
 
Don't worry about Steam. If your games didn't come from Steam then then we can rule that one out. Yes you can rule out the PSU as well. One thing us a driver sweeper to get rid of all the drivers.

Then when installing the new drivers, when you come to choose between express install (Recommended) or Custom. Choose custom. Then you will have another choice and make sure you choose Clean install.

Here is a few driver sweepers incase you don't have any. Have you swap out your GPU with another one to rule out it is your GPU? I know you did all the stress test and stuff but that works different than a game.

When you were using the volt meter did you try running the game and check the voltage at the same time? One more question, on your SSD is your game loaded on it with your OS?

http://download.cnet.com/Driver-Sweeper/3000-2094_4-75219868.html

http://performersoft.com/dr.php?brand=Updated&gclid=CKmppoaHna8CFUZgTAodTzOD4Q

http://phyxion.net/item/driver-sweeper.html
 
Crysis is patched to 1.2. Latest is 1.21 but includes online multiplayer only changes and I have read numerous reports of his causing issues.

I haven't tried patching Fallout 3 yet. I'll do it tonight.

I used Driver Sweeper before when uninstalling. I'll do it again and install an older driver version custom clean as you suggest tonight.

Steam is installed but no games ATM if that is relevant.

I didn't volt test during gameplay, only during OCCT, but both load PSU and present with same error so...

Games and OS are on same SSD...
 
That is good you are on top of things so now I need to think of what you haven't thought of and I don't think I can beat that because you have though of everything. I know I have a dry sence of humor.

I have herd depending on how big your SSD is that you should only have one thing on it at a time. So if you have a 60gigs of one then you should only put the OS on it. But if bigger then should not matter.

I herd that in the past so I really don't know how true it is but it is worth maybe looking at if all else fails. But so far you have been doing a good job troubleshooting. Let us know after the trys you are going to do. Good luck.
 
@DM186: It's true that you do not want to fill up your SSD

@Xeon84: You'd want your SSD at around 60-70% full, maybe 75% is pushing it. Otherwise, TRIM/garbage won't be optimizing the SSD properly. You also do not want to benchmark an SSD too many times, it will decrease the performance permanently.
 
Ok I've uninstalled drivers/CCC again using driver sweeper and manual approach and then installed 12.1 drivers only. Upgraded Fallout 3 to latest version. And...

Still doesn't work. :-(

SSD is 128gb got about 77GB spare ATM so not even half full.

Soon to be wiped as I think I'm going to have to reinstall Windows :-(

I could try graphics card in my other machine, but it's XP 32 bit with AMD CPU and an existing ATI card so could introduce too many other factors...
 
I've got an Nvidia 7600gt 256Mb lying around. Otherwise my old ATI stuff is AGP! (I haven't upgraded for some time!) ATI would be ideal really I guess, as it's a direct swap out. Of course Intel chip does have onboard graphics capability. Could try that out and see what happens. Not really up to gaming but would be interesting to see what happens. Even if I have to run non-native low res and low settings. If it works and dedicated card doesn't it can't be O/S, DirectX or game. Well, less likely anyway.
 
If it is your GPU it just doesn't make any sence that it could be that because of the stress testing you did. I am wondering if something has come loose or a wire is not all the way seated?

The reason I thought is was the PSU is because of the stress test running then freezing then running for 23 hours plus you telling us about the 8v. All of your problems sound like a PSU problem.

Or something is loose or not getting the right voltage. It's not temps. It's not the had drive. The two things you cannot test is the MOBO and the CPU other than temps and usage.

MOBO you have to swap things out. So what else is there that you havn't tried? This has dumbfounded me. Let us know what happens with what you are going to do. Good luck.
 
Ok. I've swapped card out with the Nvidia 7600gt I've got. Had to stick it in 1024x768 on low, but it ran Crysis and played for 20 mins no issues. Swapped back to ATI 6850, kept settings the same, kept freezing up and recovering. It was kicking out around 250fps when running!

I decided to fire up my old xp machine installed Crysis and Nvidia 7600gt and got it running fine. Installed ATI 6850 and it ran for much longer then froze up then ran then froze up. However event log shows hard disk error, not graphics related error. The xp machine is pretty crappy and old now, drive is very noisy, so this is plausible I suppose. Higher load with ATI 6850 probably brought out error quicker than with 7600gt. Plus it was later in gameplay, perhaps error, if disk related, is with a part of the later game.

Anyway results seem somewhat inconclusive.

I think I'll try borrow a newer dx11 ATI card to test in my system or find another more reliable newer system to instal my card in.

@fantastik250
System is complete new build. All components are new. Crysis, Fallout 3 and OCCT benchmark all present same symptoms.
 
Just thought I'd give a quick update.

I found an IT guy locally who helped out with some additional diagnostics using some of his kit.

He reports to have done the following:

- Run system with 750W Antec High-Power Gaming PSU - no improvement
- Swapped out system memory - no improvement
- Swapped out graphics card with 6850 (Factory overclocked) - worked fine!
- Put my 6870 in his system - crashed after 30 mins of Skyrim

Fairly conclusive that it is likely the card at fault.

I'll be getting it RMA'd ASAP.

Let you know what happens.
 
I would have to agree on that one. Something must short out on you card after it heats up I am glad you found someone to help and do let us know what happens and good luck to you.
 
Item has been returned to reseller and they are in agreement that card is at fault. Now needs to be returned to manufacturer for them to test AGAIN and confirm whether they will repair, replace or credit. Another up to 28 days wait! Whoopee!

I'll let you know when I get final resolution...
 
After waiting and chasing for 6 weeks, the reseller finally agreed to send out an MSI 6870 as a replacement for my duff XFX 6870, plus a voucher for some free games. First thing I noticed is how much quieter the MSI is over the XFX (although it runs a little hotter)! Much more pleasant on the ears. But most importantly, 3 gaming sessions (30-45 mins each) on Crysis... no issues thus far!
 
That is good news and again you had to wait so long. I will keep my fingers crossed and hope your new card does the job and I am pretty sure it will. Good luck to you and your new card and have fun gaming with it and your new games. Cheers.