RX 5700 XT VIDEO_TDR_FAILURE

[VerTriGO]

Hello everyone,

 

I have a serious issue that happening for almost 3-4 months and I couldn't manage to find any solution until now...

I'm getting blue screen or a freeze at random times when I open games\applications or after shutting them down for example after I finish a game I exit to desktop,

open Discord or chrome fast and then boom freeze, sometimes even when the app or game is running live again it happen randomly. 

 

The error is VIDEO_TDR_FAILURE amdkmdag.sys I will also add DXDIAG file

The fastest way I have found to get the freeze and blue screen for sure is opening cinebench r20 to stress the cpu and the error will appear after first or more tries but eventually it happens. 

I dont why it happenes when I stress the cpu while the graphics card not under stress at all.

error:

*******************************************************************************
*                                                                             *
*                        Bugcheck Analysis                                    *
*                                                                             *
*******************************************************************************

VIDEO_TDR_FAILURE (116)
Attempt to reset the display driver and recover from timeout failed.
Arguments:
Arg1: ffffc30260340010, Optional pointer to internal TDR recovery context (TDR_RECOVERY_CONTEXT).
Arg2: fffff80764c5bc30, The pointer into responsible device driver module (e.g. owner tag).
Arg3: ffffffffc0000001, Optional error code (NTSTATUS) of the last failed operation.
Arg4: 0000000000000003, Optional internal context dependent data.

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

KEY_VALUES_STRING: 1

    Key  : Analysis.CPU.Sec
    Value: 3

    Key  : Analysis.DebugAnalysisProvider.CPP
    Value: Create: 8007007e on DESKTOP-TJ23S8V

    Key  : Analysis.DebugData
    Value: CreateObject

    Key  : Analysis.DebugModel
    Value: CreateObject

    Key  : Analysis.Elapsed.Sec
    Value: 6

    Key  : Analysis.Memory.CommitPeak.Mb
    Value: 75

    Key  : Analysis.System
    Value: CreateObject

BUGCHECK_CODE:  116

BUGCHECK_P1: ffffc30260340010

BUGCHECK_P2: fffff80764c5bc30

BUGCHECK_P3: ffffffffc0000001

BUGCHECK_P4: 3

BLACKBOXBSD: 1 (!blackboxbsd)

BLACKBOXNTFS: 1 (!blackboxntfs)

BLACKBOXPNP: 1 (!blackboxpnp)

BLACKBOXWINLOGON: 1

PROCESS_NAME:  System

STACK_TEXT:  
ffffef8a`4ef4f908 fffff807`5f03ecea : 00000000`00000116 ffffc302`60340010 fffff807`64c5bc30 ffffffff`c0000001 : nt!KeBugCheckEx
ffffef8a`4ef4f910 fffff807`5efee1cb : fffff807`64c5bc30 ffffc302`61af0000 ffffc302`61af00d8 00000000`00000000 : dxgkrnl!TdrBugcheckOnTimeout+0xfe
ffffef8a`4ef4f950 fffff807`5efeecb6 : ffffc302`61af0100 00000000`00000300 ffffc302`61af0000 ffffc302`61af09e0 : dxgkrnl!DXGADAPTER::PrepareToReset+0x1a3
ffffef8a`4ef4f9a0 fffff807`5f03e425 : 00000000`00000100 00000000`00000000 00000000`00000000 ffffc302`60a42040 : dxgkrnl!DXGADAPTER::Reset+0x276
ffffef8a`4ef4fa10 fffff807`5f03e587 : 00000000`00000300 ffffc302`630f4a00 ffffc302`4d49d600 fffff807`45435360 : dxgkrnl!TdrResetFromTimeout+0x15
ffffef8a`4ef4fa40 fffff807`44ef43b5 : ffffc302`60a42040 fffff807`5f03e560 ffffc302`4d49d600 00000000`00000300 : dxgkrnl!TdrResetFromTimeoutWorkItem+0x27
ffffef8a`4ef4fa70 fffff807`44e6bcd5 : ffffc302`60a42040 00000000`00000080 ffffc302`4d4b3040 00000228`4eecab28 : nt!ExpWorkerThread+0x105
ffffef8a`4ef4fb10 fffff807`44fc9998 : ffff9a81`63800180 ffffc302`60a42040 fffff807`44e6bc80 00000228`4eecab78 : nt!PspSystemThreadStartup+0x55
ffffef8a`4ef4fb60 00000000`00000000 : ffffef8a`4ef50000 ffffef8a`4ef49000 00000000`00000000 00000000`00000000 : nt!KiStartSystemThread+0x28

SYMBOL_NAME:  amdkmdag+abc30

MODULE_NAME: amdkmdag

IMAGE_NAME:  amdkmdag.sys

STACK_COMMAND:  .thread ; .cxr ; kb

FAILURE_BUCKET_ID:  0x116_IMAGE_amdkmdag.sys

OS_VERSION:  10.0.18362.1

BUILDLAB_STR:  19h1_release

OSPLATFORM_TYPE:  x64

OSNAME:  Windows 10

FAILURE_ID_HASH:  {1c94a7e8-453f-c7b9-1484-0a0454a0ee36}

Followup:     MachineOwner
---------

 

 

 

At first I had those parts installed:

sharkoon pure steel case

Rog strix z390-f

RX 5700 XT NITRO+ graphics card

corsair vengence 8x2 3000 mhz RAM

corsair cx750 PSU bronze 

i9 9900k with air cooler

 

After the error appeared I thought maybe its an hardware fault and switched the PSU, the RAM and the cooling for the cpu,

I even went to my friends house because we have the same graphics card (RX 5700 XT NITRO+) and the error still appeared.

 

My current parts are:

Sharkoon pure steel case

Rog strix z390-f

RX 5700 XT NITRO+ graphics card

Corsair dominator 16x2 3200 mhz RAM

Asus rog strix 750W  GOLD PSU 

i9 9900k with asus rog stirx lc 240 water cooling.

 

I have tried almost everything:

 

install and reinstall to newest driver with DDU

Switching to different graphics card but same model

Switching ram

Switching psu

reinstall windows

Updating bios and chipset

disable turbo on cpu to run it on stock clock

switch the power option on windows to maximum performance

disable \ enable RAM XMP

lower the graphics card voltages and clocks

checked temperatures

reinstall the parts from the pc case 

using 2 PCIE cables instead of 1 for the graphics card

on motherboard raising the cpu level to 7 and usage to 140%

 

Any suggestions?

 

Thanks in advance.

 

 

 

 

 

DxDiag.txt

[AMD A10-9600P]

It sounds like a driver issue, have you ever used DDU to completely uninstall the graphics drivers? If not then try that and then try reinstalling the AMD Radeon drivers.

DDU = https://www.guru3d.com/files-details/display-driver-uninstaller-download.html

AMD Radeon Drivers = https://www.amd.com/en/support/kb/faq/gpu-driver-autodetect

[VerTriGO]

15 minutes ago, bengeoghegan11 said:

It sounds like a driver issue, have you ever used DDU to completely uninstall the graphics drivers? If not then try that and then try reinstalling the AMD Radeon drivers.

DDU = https://www.guru3d.com/files-details/display-driver-uninstaller-download.html

AMD Radeon Drivers = https://www.amd.com/en/support/kb/faq/gpu-driver-autodetect

Yes I was using DDU and also clean windows before installing new drivers

[AMD A10-9600P]

3 minutes ago, VerTriGO said:

Yes I was using DDU and also clean windows before installing new drivers

Ok, fair enough, sorry if I'm barking up the wrong tree here but did you make sure you used DDU in Windows safe mode?

[VerTriGO]

2 minutes ago, bengeoghegan11 said:

Ok, fair enough, sorry if I'm barking up the wrong tree here but did you make sure you used DDU in Windows safe mode?

I didn't do it on safe mod but again, I reinstalled windows and after that installing the driver, so it was safe and clean without any old drivers traces.

[AMD A10-9600P]

It is still worth at least trying in safe mode, reinstalling Windows doesn't always do the trick on its own

[VerTriGO]

1 hour ago, bengeoghegan11 said:

It is still worth at least trying in safe mode, reinstalling Windows doesn't always do the trick on its own

I will try when I'll arrive home but meanwhile can you tell me the difference between uninstalling driver on safe mod and do it on clean windows?

[AMD A10-9600P]

45 minutes ago, VerTriGO said:

I will try when I'll arrive home but meanwhile can you tell me the difference between uninstalling driver on safe mod and do it on clean windows?

Safe mode means Windows loads with as few services open as possible, so there shouldn't be many, if any, graphics driver related services open, which should let it uninstall much cleaner. On my laptop I had issues with AMD CPU and GPU drivers, uninstalling using the usual AMD program didn't work, neither did just DDU on it's own, but DDU in safe mode did the trick, hopefully it could help for you too.

[VerTriGO]

4 minutes ago, bengeoghegan11 said:

Safe mode means Windows loads with as few services open as possible, so there shouldn't be many, if any, graphics driver related services open, which should let it uninstall much cleaner. On my laptop I had issues with AMD CPU and GPU drivers, uninstalling using the usual AMD program didn't work, neither did just DDU on it's own, but DDU in safe mode did the trick, hopefully it could help for you too.

To be honest I doubt it will work but I have nothing to lose until someone hopefully will give another solution 

[modestboy92]

Good day to all! I'm from Russia, my English may not be very good, but I have the same problem. My radeon rx580 video card. The same problem appeared a few days ago. The drivers were installed old, I haven't updated them since February. The problem appeared on its own. Updated the drivers to the latest version, nothing has changed. Sometimes artifacts and errors appear immediately after starting the computer, sometimes during games, sometimes while watching videos from YouTube. I hope that someone will help solve this problem.

[JetBlack]

I was having issues similar to this, and the crashes eventually got tied to the Gigabyte lighting software. 
sometimes its the smallest things that make such an impact. 

[VerTriGO]

14 hours ago, JetBlack said:

I was having issues similar to this, and the crashes eventually got tied to the Gigabyte lighting software. 
sometimes its the smallest things that make such an impact. 

I got those errors before I downloaded the aura sync and icue

[VerTriGO]

anyone else?

[VerTriGO]

Still haven't found a solution 

[-rascal-]

2 minutes ago, VerTriGO said:

Still haven't found a solution 

 

I've encountered the video_trd_failure BSoD a couple times, but that is because my undervolt was too much, and then again when my overclock(s) was unstable.

On a NTRO+ 5700 XT as well.

 

What drivers are you using now?

A new set of drivers came out on in June-10-2020, 20.5.1.

I'm currently using the May-26-2020, 20.4.2. WHQL drivers, though.

 

If your GPU is not already reverted to stock / defaults through Radeon Software, do that first.

Next, go into GPU tuning, Enable GPU Tuning, and Advanced Control, and double check your Frequency / Voltage for the 3rd power state (Max Boost state).

[chris1981]

Have you tried to change the bios switch on the graphics card to see if that makes any difference?

[VerTriGO]

18 hours ago, -rascal- said:

 

I've encountered the video_trd_failure BSoD a couple times, but that is because my undervolt was too much, and then again when my overclock(s) was unstable.

On a NTRO+ 5700 XT as well.

 

What drivers are you using now?

A new set of drivers came out on in June-10-2020, 20.5.1.

I'm currently using the May-26-2020, 20.4.2. WHQL drivers, though.

 

If your GPU is not already reverted to stock / defaults through Radeon Software, do that first.

Next, go into GPU tuning, Enable GPU Tuning, and Advanced Control, and double check your Frequency / Voltage for the 3rd power state (Max Boost state).

 

17 hours ago, chris1981 said:

Have you tried to change the bios switch on the graphics card to see if that makes any difference?

I have tried updating to the latest version that came out right now and still getting blue screens or just pc freeze without auto restart until I force it..

I also tried to change the bios switch but no luck..I really dont know what this is anymore and im starting to give up, I checked on other forums and people are saying that they bought a bigger psu and the problem have gone but 750w gold isnt enough for 1 card?! thats sound very strange I have to say..

please anyone have more suggestions? 

[-rascal-]

38 minutes ago, VerTriGO said:

 

I have tried updating to the latest version that came out right now and still getting blue screens or just pc freeze without auto restart until I force it..

I also tried to change the bios switch but no luck..I really dont know what this is anymore and im starting to give up, I checked on other forums and people are saying that they bought a bigger psu and the problem have gone but 750w gold isnt enough for 1 card?! thats sound very strange I have to say..

please anyone have more suggestions? 

 

I'm not expecting toggling the BIOS switch will fix the issue.

Switch has 3 positions; left to right...

Normal/Default (220W) BIOS -- Quiet BIOS (180W / 190W?) -- Remote BIOS (e.g. control which BIOS to use through TriXX software)

 

How does your settings look like in Radeon Software >> Tuning?

Are you able to share a screenshot...using my for an example from another thread I responded to

NOTE: Settings in the screenshot is NOT default/stock, it has been tweaked

[VerTriGO]

On 6/11/2020 at 8:50 PM, -rascal- said:

 

I'm not expecting toggling the BIOS switch will fix the issue.

Switch has 3 positions; left to right...

Normal/Default (220W) BIOS -- Quiet BIOS (180W / 190W?) -- Remote BIOS (e.g. control which BIOS to use through TriXX software)

 

How does your settings look like in Radeon Software >> Tuning?

Are you able to share a screenshot...using my for an example from another thread I responded to

NOTE: Settings in the screenshot is NOT default/stock, it has been tweaked

the swtich is on the left side and this is the default settings of my card:

 

[Dr0y]

On 5/29/2020 at 9:46 AM, VerTriGO said:

The fastest way I have found to get the freeze and blue screen for sure is opening cinebench r20 to stress the cpu and the error will appear after first or more tries but eventually it happens. 

Ok so I haven't re-read all the replies and I'm probably just throwing some stupid idea there but could it be related to the integrated gpu that for some strange reasons would just interfere or crash when stressing the cpu ... as I find it really strange that it happens really fast when you stress the cpu in cinebench.

 

Maybe the old psu was not enough and the new psu is faulty ... do you have a third one to try ?

[-rascal-]

5 hours ago, VerTriGO said:

the swtich is on the left side and this is the default settings of my card:

 

 

Damn your state 3 boost clock is on the higher side for default settings. Sapphire lists the GPU Core frequency for the NITRO+ 5700XT as:

- Boost Clock, up to 2010 MHz (can be +/-)

- Game Clock, ~1905 MHz

- Base Clock, ~1770 MHz

 

Even my NITRO+ 5700 XT S.E. is around ~2030 MHz at default settings.

 

Try this, leave the voltage at the default 1193mV, but drop the GPU frequency to say... 2030, 2050, or 2060 MHz.

 

Given the VIDEO_TRD_ERROR that you get, and the 2100 MHz State 3 boost target, looks like an unstable overclock / boost target.

 

Some people have to put the 5700 XT under a water cooling loop to get 2.1 GHz, or higher. In addition to that, modify the power tables, too.

[VerTriGO]

On 6/13/2020 at 8:51 PM, -rascal- said:

 

Damn your state 3 boost clock is on the higher side for default settings. Sapphire lists the GPU Core frequency for the NITRO+ 5700XT as:

- Boost Clock, up to 2010 MHz (can be +/-)

- Game Clock, ~1905 MHz

- Base Clock, ~1770 MHz

 

Even my NITRO+ 5700 XT S.E. is around ~2030 MHz at default settings.

 

Try this, leave the voltage at the default 1193mV, but drop the GPU frequency to say... 2030, 2050, or 2060 MHz.

 

Given the VIDEO_TRD_ERROR that you get, and the 2100 MHz State 3 boost target, looks like an unstable overclock / boost target.

 

Some people have to put the 5700 XT under a water cooling loop to get 2.1 GHz, or higher. In addition to that, modify the power tables, too.

tried it.. still the same, also having pc freeze without any blue screens and if I do get a blue screen its the same error VIDEO_TDR failure

im starting to think maybe intel compatibility with amd graphics card maybe have something to do with it 

[-rascal-]

1 hour ago, VerTriGO said:

tried it.. still the same, also having pc freeze without any blue screens and if I do get a blue screen its the same error VIDEO_TDR failure

im starting to think maybe intel compatibility with amd graphics card maybe have something to do with it 

 

Mixing Intel / AMD CPU with nVidia / AMD GPU should not be problem.

I myself is running an Intel i7-8086K with my 5700 XT...

 

VIDEO_TRD_FAILURE usually comes up when the video card is unstable, or faulty.

I would maybe considering reaching out to Sapphire's support team, and see if they have any ideas.

 

Actually...what vBIOS is your 5700 XT using?

That's the only remaining things I can see that can be different...is the BIOS on the actual GPU.

 

[VerTriGO]

9 hours ago, -rascal- said:

 

Mixing Intel / AMD CPU with nVidia / AMD GPU should not be problem.

I myself is running an Intel i7-8086K with my 5700 XT...

 

VIDEO_TRD_FAILURE usually comes up when the video card is unstable, or faulty.

I would maybe considering reaching out to Sapphire's support team, and see if they have any ideas.

 

Actually...what vBIOS is your 5700 XT using?

That's the only remaining things I can see that can be different...is the BIOS on the actual GPU.

 

I went to my friend's house and used his 5700 xt same model, still crashed, how do I check vBIOS?

[-rascal-]

On 6/21/2020 at 1:26 AM, VerTriGO said:

I went to my friend's house and used his 5700 xt same model, still crashed, how do I check vBIOS?

 

GPU-Z should tell you the vBIOS version.

'BIOS Version' field: