PC Graphics

[constantinevi]

• Issue Description: My computer going black screen while I am gaming, and the music or skype will still work for about 30 seconds but mouse and keyboard are not active anymore, and than PC freeze and restart. After restarting PC, I fine BSOD message with error code 0x00000116 .
• BSOD message:

  A problem has been detected and Windows has been shut down to prevent damage
  to your computer.

  The problem seems to be caused by the following file: dxgkrnl.sys

  If this is the first time you've seen this stop error screen,
  restart your computer. If this screen appears again, follow
  these steps:

  Check to make sure any new hardware or software is properly installed.
  If this is a new installation, ask your hardware or software manufacturer
  for any Windows updates you might need.

  If problems continue, disable or remove any newly installed hardware
  or software. Disable BIOS memory options such as caching or shadowing.
  If you need to use safe mode to remove or disable components, restart
  your computer, press F8 to select Advanced Startup Options, and then
  select Safe Mode.

  Technical Information:

  *** STOP: 0x00000116 (0xffffe4082410d4a0, 0xfffff8054ca60474, 0xffffffffc0000001,
  0x0000000000000003)

  *** dxgkrnl.sys - Address 0xfffff8054c39cf38 base at 0xfffff8054c200000 DateStamp
  0x04e5cf9e

• Hardware:
  • Graphics card: Radeon VII work at default 1802MHz/1000MHz and default voltage setting 1111 mV
  • CPU: Intel i7-4790K
  • Motherboard: ASUS Z97-Pro Wifi/USB3.1
  • RAM: 32GB Kingston DDR3 @1866 MHz
  • PSU: Antec HCG 850W
  • Monitor 1: LG 27UK850-W @4K freesync enabled connect on Radeon VII
  • Monitor 2: Philips 237EQPH connect on Intel Graphic HD 4600
• Software:
  • Windows 10 64 bit 1803
  • Latest graphics driver 19.4.1
  • Latest UEFI BIOS from AMD flashed

I have alos try connect one screen, flash install Win 10, even change my Antec EDGE 650W to this HCG 850W, but don't help. Does I get a bad Radeon VII card? or it's caused by other reason?

I'm considering to have a RMA service.

Update 2019/05/05:

After RMA, I got another card back at 4/24. I try undervoltage and get stable setting at 1800/1200MHz@980mV and 0% power limit. This setting pass Time Spy Extreme stress test at 99.4%, so I choose this as my normal setting.

For next two weeks, I still have same problem: BSOD 0x00000116 and 0x100000ea keep coming to me, but the frequency seems not so high, about once 2 days. These two BDOS error all active in same mode:

screen freeze (sound keep working) -> screen lost signal and turn off (sound stop working) -> wait minutes -> PC restart. And than the BSOD dump record is there.

During these two weeks, I try to turn off Enhanced sync and AMD overlay,  which seems no helping. Now I try  to turn off FreeSync at my LG 27UK850-W.

Temporary conclusion

Before I bought Radeon VII, I use Nvidia GTX1060 (new) and AMD Vega 64 (secondhand), and they all work perfectly. Specially Vega64 have all same radeon setting and works fine, but not Radeon VII. That means this should be VGA card hardware or drivers issue, not caused by other prat of PC, that's why I want to go RMA directly when the new Radeon VII comes issue. Now, The return card also comes with same issue too. It seems be very rare to get another card with same problem, I'm not sure for that it's only caused by VGA. I think I will test following part:

• FreeSync off (testing)
• Change from DP to HDMI
• Change from 4K to 2K

If they all fail, I will try to use this Radeon VII at another PC for more testing. And than would be another RMA maybe...Orz.

[goodplay]

First step would be update to 1809 https://www.tenforums.com/tutorials/16397-repair-install-windows-10-place-upgrade.html

[constantinevi]

I have finished upgrade to 1809 and testing. If there are any error I will report as soon as possible.

[constantinevi]

Well, it seems not helping.... My PC going dead again. This time is 0x100000ea.

The problem seems to be caused by the following file: atikmdag.sys

THREAD_STUCK_IN_DEVICE_DRIVER

*** STOP: 0x100000ea (0xffff9b0859f42080, 0x0000000000000000, 0x0000000000000000,
0x0000000000000000)

*** atikmdag.sys - Address 0xfffff800541dc17b base at 0xfffff800540e0000 DateStamp
0x5c9af4b0

Final record of Radeon VII is 1800MHz/1000MHz @GPU temp 75 C and Junction temp 101 C.

I don't think this is such extreme temperature condition to overheat crashing.

Everything I have tried told me that this would be the hardware problem, what do you think?

[pokester]

Likely it is thermal throttling like all vegas. Look at the threads that show how to undervolt the vegas for stability. They may offer the solution you need to get that stable. Shame that on round 2 of vega they still can't make a card that works at default settings.

[constantinevi]

That would be very interesting because I have seen more high temperatures like core 80 C and junction goes to 115 C sometimes. And they work in this condition for minutes until the game over, everything seems be fine.

Also, I have try undervolt with a little overclock like 1900/1150 MHz @ 1080mV, I got the same problem as well. That’s why I try driver default clock and volt 1801/1000MHz @ 1112mV to test if the problem caused by the manual adjustment of clock or voltage.

[pokester]

If older drivers don't help and none of those tweaks help then maybe you just have a bad card or an issue with your incoming power?

[constantinevi]

I see. I will use another PC for double check and go RMA service. Thanks for your help.

[blaqq]

Can you go into your Radeon settings for "gaming" and into the "global settings" and set the Frame Rate Target Controll to 50fps, and try gaming. Does it crash still crash while gaming?

[constantinevi]

Sorry for lately reply, I want to go RMA directly thought I have test for many setting. After it fail at another Ryzen 2700X PC, I think it is obviously a bad card. Another thing is that even the 50 FPS work, I don't want to have a video card cost $699 and only let me gaming at 50 FPS....

[pokester]

Glad you figured it out. If you are still in your original return period you could maybe return it to where you bought it instead of RMA even.

[constantinevi]

Sadly, I's out of the period now.

[blaqq]

I have noticed one thing. I can play destiny 2 even if i have over 300fps.. but my BSOD issue is still in other games. I used to BSOD im Final fantasy XIV over 50fps and have noticed that, since my monitor is a 165hz, i set the FPS limiter to 165. However inside the game in video settings i set the FPS limiter to 144, which i think is enough and i dont get BSOD in final fantasy XIV. I am yet to try out this theory in the divison 2. I will try tonight. Because i know i used to be able to play at over 100fps, recent pathes has BSOD me over 80fps. Would you mind trying out my theori and reply if it did work for you? If you crash before you can change settings, maby try set graphic FPS limit to 50, launch the game, set ingame limiter to lets say 120 and then shut down the game and then change graphic fps limiter higher and try launch the game again and play?

[constantinevi]

Thanks for you suggestion, but I have sent my card to MSI. So there are no longer testing.

[blaqq]

Okay  i want ro RMA too but i guess i cannot since my card is 2 years old and has a aftermarket watercooler xD lol

[constantinevi]

Hello, My Radeon VII was back from MSI, Everything seems good now.

By the way, I try undervoltage and get stable setting at 1800/1200MHz@980mV. This setting will pass Time Spy Extreme stress test at 99.4%.

[constantinevi]

Hello, It seems be not lucky, I still have same problem. BSOD 0x00000116 and 0x100000ea keep coming to me. they all active in same mode: screen freeze (sound keep working) -> screen lost signal and turn off (sound stop working) -> wait minutes -> PC restart. And than the BSOD dump record is there.

I try to turn off Enhanced sync and AMD overlay. It seems no help. By the way, this problem always happen when I play PUBG and World of Warships. It doesn't happen when I play The Division 1 and 2. Does any one have an idea about this situation? I have no idea what's going on, everything works fine when I use Vega 64 before.

There is the report of DxDiag:

[pokester]

Have you tried setting the power limit at maximum?

[constantinevi]

Yes, I have, but no help.

[pokester]

I wish I had advice to help but I have seen quite a few of these complaints but no resolution. They should work at default settings. They should definitely work with know tweaks for vegas. If not I would return or exchange the card if it is still possible. If not RMA it for another. 

[constantinevi]

Thanks for you reply. First, this card have default setting 1801/1000MHz@1121mV, I have try this setting when the first crash caused after RMA, so I don't think the default setting does any help.

By the way, I undervoltage to 1800/1200MHz@980mV and 0% power limit. This setting pass Time Spy Extreme stress test at 99.4% though result of stress test does means everything.

[qwixt]

I googled your PSU, and it has 2 12V rails. What I suggest trying is using two separate 8 pin GPU power connectors. Plug one into the +12V1 side, and the other into the +12V2 side, then plug one 8-pin connector from each into the GPU.

It's a long shot, but if I googled that PSU correctly, you only have 40A per rail.

[constantinevi]

Thanks for your reply. I know this issue, so I had changed the PSU to Antec HCG 850W, which is only one 12V rail. A lot of testing had done under Antec HCG 850W, conclusion is it seems not caused by PSU.

[blaqq]

Hey.. u know i had pretty horrible BSOD issues too right...

Now i have installed windows 10 with a release from 2017.. Now i have no issues.. Wanna try that copy? if you want i can provide it for you from my googledrive,

email me at derya.y@gmail.com

[constantinevi]

Thanks, I have Win 10 license, so it's not so hard to get Win 10 version like 1703. I would have tried if all test fail.

[blaqq]

Ofcourse you still need a licence for the one i have haha its downladed

from microsoft 2 years ago. Ill check which version i have tonight and tell

you which i got. If you install old version of windows it might be

important to inactivate windows update from services in administrative

tools in the controll panel.

You can then activate windows update from services and go into activate

windows, let it activate with your genuine windows key and then inactivate

windows update from services.

All just so that windows dont download updates and start giving you BSOD

again.

[blaqq]

Yes the Windows version that fixed the BSOD issues is

Version 1709 (OS build 16299.15)

i would also like to note that this windows verision also improved FPS by atleat 40fps up to 100fps more all across the board.

I can also note that in your original post a dxgkrnl.sys error

I could also at times pinpoint the BSOD issues to a KERNEL error.

During all my different tests i could not find any issue with any of my hardware. I do also have a 1000watt PSU from corsair. 

So the issue is as it seems 100% software based. Our type of errors usually come up as "driver error" while searching the google and forums. However, different driver versions never fixed the issue.

The only thing that fixed the issue was an older version of windows, version 1709 in my case. 


If you reinstall windows and see that your BSOD issue is solved. Please share it under these comments as that piece of information could be proven very usefull and i will make sure to contact both microsoft and AMD it.

[constantinevi]

Congratulation. I read the dxdiag report of yours, I think we are facing different problem but all caused by drivers/software. So I don't consider to reinstall Windows for now. As I mention, my Vega64 work fine with the current version of Windows 10. I will have a report later to talk about what I found this week.

[constantinevi]

OK, after FreeSync disable for these three days, BSOD doesn't happened.

For BSOD I hit, Microsoft have docs which describe :

The THREAD_STUCK_IN_DEVICE_DRIVER bug check has a value of 0x000000EA. This indicates that a thread in a device driver is endlessly spinning.

The VIDEO_TDR_FAILURE bug check has a value of 0x00000116. This indicates that an attempt to reset the display driver and recover from a timeout failed.

I hit these BSOD while gaming, but not all of them and every time.

Following games hit BSOD randomly, sometimes is fine all the day, sometimes fail many times per day:

World of Warships, PUBG

Following never hit BSOD, though these games'  play time are not as long as above:

The Division, The Division 2, Assassin's Creed Origins

All game above play at 4K in windowed borderless mode.

Because I hit BSOD average once two days, so I would like to test every option in following way:

Test every option, if BSOD happened, I will go to test next option. If anyone can keep stable with non BSOD for a week, I will assume the BSOD is caused by that setting.

After three days with non BSOD, I try to search about FreeSync, the information shows many problem. I collected some issues which have been posted:

Vega 56/64 crashing, black screen/no video under high load 

Vega 56 Hard Freeze PC and BSOD 

This would be more interesting:

Vega FE LC (BSOD on Monitor Standby or Connection) says

**If freesync is disabled or a monitor is connected without freesync, there is no bluescreen**

If the FreeSync disable setting pass my test, consider I have Vega64 and no issue with FreeSync on, I could assume the problem would be:

• FreeSync driver have bug with some games, but FreeSync not actually work under Vega64 though the setting is on.
• FreeSync driver have bug under Radeon VII, but be stable under Vega64.
• FreeSync driver have bug for some FreeSync monitor connect to Radeon VII, but be stable under Vega64.(Assume there are still lots of Radeon VII's customer work fine with there FreeSync monitor, I can't test it because I don't have another FreeSync monitor)

I will keep my eye on this testing. If there are no more BSOD, I would make a conclusion next Sunday.

[constantinevi · Accepted solution]

Finally, I can confirm that the BSOD is caused by FreeSync. I have test with monitor's FreeSync off for a week, no more BSOD happened after that. So it seems the driver for Radeon VII is still some problems.

If you hit BSOD issue randomly with code 0x000000EA / 0x00000116 when gaming, I hope this will help.

[blaqq]

Glad you figured it out! This BSOD mayhem seems to be caused by so many

different factors for different users. Maby i will make a compliation of

all these solutions.

[meisteradler]

Hello constantinevi. I am using also a radeon VII in a system with i7 4790K and Z97 chipset with 4k monitor samsung LU32H850, connected with displayport. I have the same problem randomly from time to time (weekly), but my problem appears in idle mode (desktop).

BDOD with restarts. I have also now unchecked "ultimate engine" of freesync. Lets se, what happend next time.

btw: my radeon VII is not so good,  1801 Mhz/ 1129mV. May be, this could have anything to do with. (excuse my english, I am german)

[constantinevi]

You are not alone. Vega FE LC (BSOD on Monitor Standby or Connection) this case also happen in idle mode. Otherwise, I don't think it's a problem caused by chip after I have RMA service once. Uh....get another chip with same problem....

BTW, I don't know about "unchecked ultimate engine of freesync", though I disable FreeSync directly from my LG 27UK850-W OSD, not in Radeon Setting.

Cheers.

[meisteradler]

Yes, this is also from the menu of the monitor: 1.Freesync off, 2. Freesync standard engine, 3. freesync ultimate engine. Now I have choosen 1. and wait what will happend next days.

[meisteradler]

BSOD is always there. PC was  for 5min in idle mode after starting, firefox was opened with 2 windows, one window shows tv livestream, then it happend.

 Conclusion: freesync is not the reason in my case.

[constantinevi]

Does the BSOD have same error code? Are you having this problem since buy this card?

[meisteradler]

Sorry I dont have the error code. Before the radeon7 there was a Asus Vega 64 without problems in this PC.

[constantinevi]

I suggest you to provide your hardware list, and describe the phenomenon happened at your PC. Furthermore, could you check the DxDiag to see if there are some error record there. Check if there are any record in BlueScreenViewer, too.

[blaqq]

Have you tried running memtest?

[meisteradler]

No, I don't. This Pc  works flawless since 4 years with different hardware and gpus (nNvidia, AMD)