Graphics Cards

[AlertTubaThrown]

I am experiencing frequent, random crashing to black screen, both during intensive GPU applications (gaming), or just when browsing the internet.

An analysis of the Windows crash .dmp file indicates that it is due to a GPU timeout error, and references this file: AMDKMDAG.sys (see below).

Things I've already tried:

• Clean reinstall Windows 11
• Uninstall drivers using AMD Cleanup Utiltiy in safe mode and reinstall latest drivers
• Installed latest Windows updates (ver 22H2)
• Updated latest motherboard BIOS
• Tested RAM with MemTest and TestMem 5 (Anta777 Absolut) (no errors)
• Ran Intel Processor Diagnostic Tool (passed)
• Ran through this knowledge base article: https://www.amd.com/en/support/kb/faq/gpu-800
• Disabled Intel integrated graphics
• Disabled Freesync
• Ran cmd prompt Dism and SFC commands to verify integrity of windows system files

Any help much appreciated!

System info:

• MSI PRO Z690-A 
• G.Skill Ripjaws 2x16Gb DDR5-6000 PC5-48000 (stock speeds)
• Intel i7 13700k (stock clock)
• EVGA SuperNOVA 750 G3, 80 Plus Gold 750W
• Temperatures are all reasonable under load 

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

VIDEO_ENGINE_TIMEOUT_DETECTED (141)
One of the display engines failed to respond in timely fashion.
(This code can never be used for a real BugCheck; it is used to identify live dumps.)
Arguments:
Arg1: ffffe58d8e91a010, Optional pointer to internal TDR recovery context (TDR_RECOVERY_CONTEXT).
Arg2: fffff806912a8780, The pointer into responsible device driver module (e.g. owner tag).
Arg3: 0000000000000000, The secondary driver specific bucketing key.
Arg4: 00000000000004c4, Optional internal context dependent data.

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

Unable to load image amdkmdag.sys, Win32 error 0n2
*** WARNING: Unable to verify timestamp for amdkmdag.sys

KEY_VALUES_STRING: 1

    Key  : Analysis.CPU.mSec
    Value: 1030

    Key  : Analysis.Elapsed.mSec
    Value: 1836

    Key  : Analysis.IO.Other.Mb
    Value: 0

    Key  : Analysis.IO.Read.Mb
    Value: 0

    Key  : Analysis.IO.Write.Mb
    Value: 0

    Key  : Analysis.Init.CPU.mSec
    Value: 46

    Key  : Analysis.Init.Elapsed.mSec
    Value: 2752

    Key  : Analysis.Memory.CommitPeak.Mb
    Value: 108

    Key  : Bugcheck.Code.LegacyAPI
    Value: 0x141

    Key  : Dump.Attributes.AsUlong
    Value: 18

    Key  : Dump.Attributes.KernelGeneratedTriageDump
    Value: 1

    Key  : Failure.Bucket
    Value: LKD_0x141_IMAGE_amdkmdag.sys

    Key  : Failure.Hash
    Value: {48b738dd-5a92-7ff8-63d0-f075fc680fe0}


BUGCHECK_CODE:  141

BUGCHECK_P1: ffffe58d8e91a010

BUGCHECK_P2: fffff806912a8780

BUGCHECK_P3: 0

BUGCHECK_P4: 4c4

FILE_IN_CAB:  WATCHDOG-20231024-2017.dmp

DUMP_FILE_ATTRIBUTES: 0x18
  Kernel Generated Triage Dump
  Live Generated Dump

TAG_NOT_DEFINED_202b:  *** Unknown TAG in analysis list 202b


VIDEO_TDR_CONTEXT: dt dxgkrnl!_TDR_RECOVERY_CONTEXT ffffe58d8e91a010
Symbol dxgkrnl!_TDR_RECOVERY_CONTEXT not found.

PROCESS_OBJECT: 00000000000004c4

PROCESS_NAME:  System

STACK_TEXT:  
fffffb8b`a3fe7550 fffff806`61b78f24     : ffffe58d`818df030 ffffe58d`818df030 ffffe58d`8b296d10 ffffe58d`8e91a010 : watchdog!WdpDbgCaptureTriageDump+0xb7
fffffb8b`a3fe75c0 fffff806`619e153d     : ffffe58d`8e91a010 fffffb8b`a3fe7780 00000000`00000000 fffff806`a13c5169 : watchdog!WdDbgReportRecreate+0xd4
fffffb8b`a3fe7620 fffff806`a12fbe6b     : ffffe58d`00000000 ffffe58d`8961fbd0 ffffe58d`8e91a010 ffffe58d`89803000 : dxgkrnl!TdrUpdateDbgReport+0x11d
fffffb8b`a3fe7680 fffff806`a13c54d9     : ffffe58d`89803000 00000000`00000000 ffffe58d`8961f000 ffffe58d`89803001 : dxgmms2!VidSchiResetEngine+0x70f
fffffb8b`a3fe7830 fffff806`a1390093     : ffffe58d`8961f000 00000000`00000000 00000000`00000000 00000000`00000000 : dxgmms2!VidSchiResetEngines+0xb1
fffffb8b`a3fe7880 fffff806`a135e33f     : 00000000`00000000 00000000`00000000 00000000`000148a5 00000000`00989680 : dxgmms2!VidSchiCheckHwProgress+0x31d13
fffffb8b`a3fe7900 fffff806`a12c7c29     : 00000000`00000000 ffffe58d`8961f000 fffffb8b`a3fe7a39 00000000`00000000 : dxgmms2!VidSchiWaitForSchedulerEvents+0x37f
fffffb8b`a3fe79d0 fffff806`a1374b25     : ffffe58d`827b7000 ffffe58d`8961f000 ffffe58d`827b7010 ffffe58d`870b7270 : dxgmms2!VidSchiScheduleCommandToRun+0x309
fffffb8b`a3fe7aa0 fffff806`a1374a9a     : 00000000`00000000 fffff806`a13749d0 ffffe58d`8961f000 ffffe58d`7279b040 : dxgmms2!VidSchiRun_PriorityTable+0x35
fffffb8b`a3fe7af0 fffff806`5ad00677     : ffffe58d`85c763c0 fffff806`00000001 ffffe58d`8961f000 006fe47f`b19bbdff : dxgmms2!VidSchiWorkerThread+0xca
fffffb8b`a3fe7b30 fffff806`5ae176a4     : ffffd300`a9c51180 ffffe58d`85c763c0 fffff806`5ad00620 41c61629`a3bc60b5 : nt!PspSystemThreadStartup+0x57
fffffb8b`a3fe7b80 00000000`00000000     : fffffb8b`a3fe8000 fffffb8b`a3fe1000 00000000`00000000 00000000`00000000 : nt!KiStartSystemThread+0x34


SYMBOL_NAME:  amdkmdag+f8780

MODULE_NAME: amdkmdag

IMAGE_NAME:  amdkmdag.sys

STACK_COMMAND:  .cxr; .ecxr ; kb

FAILURE_BUCKET_ID:  LKD_0x141_IMAGE_amdkmdag.sys

OSPLATFORM_TYPE:  x64

OSNAME:  Windows 10

FAILURE_ID_HASH:  {48b738dd-5a92-7ff8-63d0-f075fc680fe0}

Followup:     MachineOwner
---------

[vincent294]

https://community.amd.com/t5/graphics-cards/7800-xt-crashing-with-dual-monitor-setups-dp-hdmi-green-...

I've only had black/green screen shutdown/reboot while gaming, and a BSOD when shutting down. I'm pretty new to poking Event Viewer, and never used a DMP file before. Looking at the DMP file from kernel mode heap corruption in WinDbg:

BLACKBOXWINLOGON: 1

CUSTOMER_CRASH_COUNT:  1

PROCESS_NAME:  AUEPMaster.exe

SYMBOL_NAME:  AMDRyzenMasterDriver+1bb7

MODULE_NAME: AMDRyzenMasterDriver

IMAGE_NAME:  AMDRyzenMasterDriver.sys

IMAGE_VERSION:  2.2.0.0

STACK_COMMAND:  .cxr; .ecxr ; kb

BUCKET_ID_FUNC_OFFSET:  1bb7

FAILURE_BUCKET_ID:  0x13a_11_4gaT_AMDRyzenMasterDriver!unknown_function

OS_VERSION:  10.0.22621.1928

BUILDLAB_STR:  ni_release_svc_prod3

OSPLATFORM_TYPE:  x64

OSNAME:  Windows 10

FAILURE_ID_HASH:  {bffc41d4-0974-7413-9884-14c09d373dc1}

Followup:     MachineOwner
---------

AMD Software: Adrenalin Edition > Settings > Preferences > AMD User Experience Program

To be safe, I'll leave the AMD User Experience Program. I even suspect this BSOD I experienced is unrelated to the other crashes after reading this DMP file. The less I page fault the better.

Edit:

Also will reinstall drivers as minimal install.

[vincent294]

Yeah, still getting green screens on minimal install, but nice to have a few extra MB on my SSD.

[AlertTubaThrown · Accepted solution]

May have found a solution - at least it seems to be holding stable without crashing for the last week since I posted.

Rolled back AMD drivers from latest version to 23.9.1

Simple as that.

Hope this helps someone else out there - was driving me crazy.

[vincent294]

I've been trying that right now, assuming the reason YouTubers didn't notice this issue is day 1 drivers worked. https://www.amd.com/en/support/kb/release-notes/rn-rad-win-23-9-2 changelog sure looked sus to me as well. The reason I haven't suggested it yet is I haven't played much Cyberpunk 2077 so maybe it's just crashing less. Since 23.10.1, crashes were green screen instead of black screen, happened more often, and usually followed by reboot instead of shutdown. The big catch is Counter Strike 2 performance is crap, but me and my friends play Overwatch 2 anyways because the players are less toxic. Alan Wake II driver fixed an issue but I haven't tried that on the old driver yet, so no idea if I'll have to swap drivers per game like I did on my RX 560 before I sold it because of that. Unlike 2019, the Nvidia price gap is much wider, and while I was tempted to support my employer's Arc A770, I wanted a bigger upgrade from my RTX 2060. Reminds me of my old Ryzen 5 2500U, I put up with lots of driver issues but by 2020 drivers were a dream.

[ykyng]

I may have found an approach that I didn't find documented yet:

I have exactly the same card as the OP. I noticed this happens randomly over time independently from load on the card. Mine even crashed when I left the machine on overnight just doing nothing. I read that many people ruled out anything in their hardware (RAM, temperature, voltage e.g.) so I suspected something in the energy options to be the culprit. And I can tell it doesn't happen for me after deactivating Hibernating, Hybrid Standby, selective USB power saving and don't allow the systems to power off my displays (both connected via HDMI) anymore.

Maybe this is of help for some of the tortured 7xxx owners here. This approach is truly more simple than turning your hardware upside looking for faults that were never there before installing your new Radeon card.

Happy gaming,

ykyng

[AlertTubaThrown]

Can confirm in repeated test that rolling back the driver seems to be a solution for now.

Did a clean install of windows with latest AMD drivers: 23.10.2.  Running Furmark for <10 minutes hard crash to black screen.

Ran AMD Cleanup Utility to uninstall the drivers in safe mode.  Installed 23.9.1 drivers and looking stable.  Furmark ran for half an hour with no issues.

As ykyng pointed out, I would get these crashes when under no load too, just browsing the internet it would suddenly crash to black screen out of nowhere.  But running Furmark seems to be a way to get it to crash more consistently in a short space of time.

[vincent294]

Got a black screen shutdown with event ID 6008 in the Oda boss fight in Cyberpunk 2077. I wasn't doing good anyways lol. So it still crashes on 23.9.1, but not as much.

[AlertTubaThrown]

Just wanted to update that the black screen crashes in Furmark still occur with the latest driver, 23.11.1 and do not occur when rolled back to 23.9.1 (amdkmdag.sys is the problem again).

Is AMD ever going to get around to fixing this or am I going to be stuck using outdated drivers for the rest of the life of my graphics card?