Bsod: Graphics_driver_tdr_fault

[Valenciente]

Greetings TPU,

A dear friend of mine have had a longrunning problem now. I've tried a lot of different solutions, but nothing seems to fix the problem.

We've tried formatting his computer with a different OS and much more; he also have the newest drivers for his graphics card and everything else.

His hardware is as follows:
CPU: Intel Core 2 Quad Q6600
MOBO: DFI Lanparty LT X48-T2R
RAM: 2x2GB DDR2 (Can't remember the brand)
GPU: 2xATi HD3870 in CrossFire (512mb each)

I've also dugged out his crash dump file, so if anyone would like to have a look at it, here goes:

VIDEO_TDR_FAILURE (116)
Attempt to reset the display driver and recover from timeout failed.
Arguments:
Arg1: fffffa80067a04e0, Optional pointer to internal TDR recovery context (TDR_RECOVERY_CONTEXT).
Arg2: fffff88004945c80, The pointer into responsible device driver module (e.g. owner tag).
Arg3: 0000000000000000, Optional error code (NTSTATUS) of the last failed operation.
Arg4: 0000000000000002, Optional internal context dependent data.

Debugging Details:
------------------


FAULTING_IP:
atikmpag+6c80
fffff880`04945c80 4883ec28 sub rsp,28h

DEFAULT_BUCKET_ID: GRAPHICS_DRIVER_TDR_FAULT

CUSTOMER_CRASH_COUNT: 1

BUGCHECK_STR: 0x116

PROCESS_NAME: System

CURRENT_IRQL: 0

STACK_TEXT:
fffff880`076119c8 fffff880`05334ef8 : 00000000`00000116 fffffa80`067a04e0 fffff880`04945c80 00000000`00000000 : nt!KeBugCheckEx
fffff880`076119d0 fffff880`05334c02 : fffff880`04945c80 fffffa80`067a04e0 fffffa80`05c69c00 fffffa80`06088010 : dxgkrnl!TdrBugcheckOnTimeout+0xec
fffff880`07611a10 fffff880`0520ff07 : fffffa80`067a04e0 00000000`00000000 fffffa80`05c69c00 fffffa80`06088010 : dxgkrnl!TdrIsRecoveryRequired+0x1a2
fffff880`07611a40 fffff880`05239b75 : 00000000`ffffffff 00000000`000bab72 00000000`00000000 00000000`00000002 : dxgmms1!VidSchiReportHwHang+0x40b
fffff880`07611b20 fffff880`052382bb : 00000000`00000102 00000000`00000005 00000000`000bab72 00000000`00000000 : dxgmms1!VidSchiCheckHwProgress+0x71
fffff880`07611b50 fffff880`0520b2c6 : ffffffff`ff676980 fffffa80`06088010 00000000`00000000 00000000`00000000 : dxgmms1!VidSchiWaitForSchedulerEvents+0x1fb
fffff880`07611bf0 fffff880`05237e7a : 00000000`00000000 fffffa80`04ca9d50 00000000`00000080 fffffa80`06088010 : dxgmms1!VidSchiScheduleCommandToRun+0x1da
fffff880`07611d00 fffff800`02f32a86 : 00000000`fffffc32 fffffa80`05c71b60 fffffa80`03ccf040 fffffa80`05c71b60 : dxgmms1!VidSchiWorkerThread+0xba
fffff880`07611d40 fffff800`02c6bb06 : fffff880`02f63180 fffffa80`05c71b60 fffff880`02f6dfc0 00000000`000cabb3 : nt!PspSystemThreadStartup+0x5a
fffff880`07611d80 00000000`00000000 : 00000000`00000000 00000000`00000000 00000000`00000000 00000000`00000000 : nt!KxStartSystemThread+0x16


STACK_COMMAND: .bugcheck ; kb

FOLLOWUP_IP:
atikmpag+6c80
fffff880`04945c80 4883ec28 sub rsp,28h

SYMBOL_NAME: atikmpag+6c80

FOLLOWUP_NAME: MachineOwner

MODULE_NAME: atikmpag

IMAGE_NAME: atikmpag.sys

DEBUG_FLR_IMAGE_TIMESTAMP: 4c33d53e

FAILURE_BUCKET_ID: X64_0x116_IMAGE_atikmpag.sys

BUCKET_ID: X64_0x116_IMAGE_atikmpag.sys

Followup: MachineOwner

I know it's a display problem, and I'm pretty sure it probably one of his cards that have some sort of problem. Any suggestions before I tell him to get a new card?

Thanks,
Valen

P.S. There is no Overclocking of any kind on the system.
P.P.S. This almost only happens when playing World of Warcraft. I've recently updated his drivers to the newest available, which seemed to make the problem worse.

 

[erixx]

HEAT (inside the VGA cover and cooler because of massive dust) or a dead VGA is my bet...

 

[Kreij]

Have you tried the 3870's card seperately to see which one may be problematic?

 

[Valenciente]

Come to think of it; We did try with one card at one point, but never got to try out the second one.

I live kind of far away from him, and he's not exactly tech savvy, so nothing happens if I ain't visiting him.

I'll try to get his brother to do that for him tomorrow, thanks for reminding me.

 

[Valenciente]

Hi guys. I realize this is an old thread, but I completedly forgot about it. And as he's still having the problem, I thought I'd dig futher into it.

So, we've determined that he still gets the error with either of his 3870's.
Also, I've currently borrowed my 5850 to him, as I'm not using it at the moment.
He STILL gets the error, however it's much less frequent.

I've read around a bit on the web, and from what I've gathered it can be related to his motherboard's VGA BIOS. Any ideas on that? (I don't know much about BIOS's)

Any thoughts on what might help here?

Thanks.

 

[Steevo]

I would bet there is a PCIe clock or timing issue, try a different motherboard, or bump the voltage to the northbridge a few tenths.

 

[mastrdrver]

What power supply is he running?
Also make sure that the PCIe power saving is off in the power options.
more on AMD Game forum

 

[Valenciente]

@Steevo: Thanks for the suggestion, but we don't have another motherboard to try I'll give the voltage bump a shot though.

@mastrdrver: At the moment he's running an Antec. Not sure which exact model. But it have happened with several different PSU's, including Corsair and OCZ. I'm pretty sure the PCIe power saving option is off, but I'll have a look. Thanks.

And we've already tried about a ton of the suggestions on that site, no luck My own best bet would be his system RAM, something also suggested at that site. I dont have any DDR2 ram anywhere though, so we can't really try that >.<

 

[micropage7]

if there still err on it try to check the bios, try to check the vga part, clock, voltage and so
how bout the voltage? Is that enough?
Does it start periodically or like that
have you checked the heat: vga, the chipset, processor

 

[Valenciente]

The bios is really just using default value's, but I'll have a look at it.

The error ONLY happens when playing World of Warcraft. And it's totally random at that..
(Another friend I play with is also getting this kind of graphics driver crashing while playing WoW, but he's using nVidia)
He can play for hours on end, and the graphics driver just stops, giving him a black screen. Most of the time it will start back up, so he can play again. The amount of time it takes to start up the graphics driver is also quite random. Sometimes it takes 5 seconds, and at times it can take around a minute.
Now and then his computer crashes into a bluescreen as well, but it's much less frequent now that he uses my 5850.

There is no trouble with the heat at all, I've checked a dozen of times.

 

[mastrdrver]

If it is system memory causing the problem have him burn memtest86+ to a cd and boot from the cd. Let it loops for a few hours as it can take time sometimes for errors to show.

Do you know yet if it even happens with the card you loaned to him?

 

[Valenciente]

I'll try the memtest as well when he visits me next time.

It does happen with my card as well. It's much less frequent, but it does happen.

 

[mastrdrver]

Might also have him check the voltages of the psu too.

 

[Valenciente]

He just old me yesterday that as far as he remembers, this problem didn't happen with his old PSU. So it might actually be his newer PSU from Antec. I'll have to check up on the voltages of it, and maybe lend him my spare PSU (which I've never had any such problems with) to test if that works.

Thanks.

 

[mastrdrver]

Is he using molex 4-pin to PCIe power adapters to power the cards? Most (not all) Antec units are have "multiple rails" so it could be that (if he is using adapters) be overloading one of the rails.

 

[Valenciente]

I'm pretty fairly 99% sure that he isn't using any adapters.

 

[lemoen]

Problem solved by memory voltage change

I had TDR faults several times a day for weeks - very frustrating. I swapped out components (RAM, PSU, graphics card) and fiddled with BIOS settings to no avail, until I read some advice to newbies about taking some care with RAM voltages.

I bumped my RAM voltage up from the recommended 1.65 V to 1.8 V and the problem went away instantly - haven't had a single BSOD for over a week.

 

[BiggieShady]

I had TDR faults several times a day for weeks - very frustrating. I swapped out components (RAM, PSU, graphics card) and fiddled with BIOS settings to no avail, until I read some advice to newbies about taking some care with RAM voltages.

I bumped my RAM voltage up from the recommended 1.65 V to 1.8 V and the problem went away instantly - haven't had a single BSOD for over a week.

Good for you, but I doubt it is relevant in this 3 year old thread about 3000 Ati Radeon series (finally I get to say Ati )

 

[lemoen]

Well it's one of the top pages that comes up when you Google TDR faults. And the subject is "Bsod: Graphics driver tdr fault". But hey, the road to hell is paved with good intentions.

 

[Law-II]

I bumped my RAM voltage up from the recommended 1.65 V to 1.8 V

Hi

Depending on the hardware installed in a system it is not recommended in all cases to up voltage to 1.8 as it may damage CPU memory controller *if on cpu die

atb (all the best)

Law-II

 

[claylomax]

the road to hell is paved with good intentions

Ha! Well said!

 

[Aquinus]

bumped my RAM voltage up from the recommended 1.65 V to 1.8 V and the problem went away instantly - haven't had a single BSOD for over a week.

That will kill the IMC on an Intel CPU very quickly. I strongly recommend against this. If you're not running at stock, you should revert everything now and try again.

 

[Valenciente]

A think a mod should lock this or something... o.o

 

[claylomax]

A think a mod should lock this or something... o.o

Why? There's nothing wrong with this thread; I like it.

 

[Valenciente]

Why? There's nothing wrong with this thread; I like it.

Well it's 3 years old, no longer relevant whatsoever, and getting highly off-topic. But maybe I'm just being grumpy