Graphics Cards

[heartless_ca]

I'm stuck.

My problem is my pc hard freezes or BSOD that requires a complete computer hard reboot. I've had this card for roughly a month and I have tried and replaced everything I can think of but I've had nothing but issues. 
The BSOD is THREAD_STUCK_IN_DEVICE_DRIVER atikmdag.sys/dxgkrnl.sys (according to BlueScreenView), however, I rarely see this bsod. Most of the time it is a hard freeze and I have to hard reboot. This issue is so intermittent and I cannot recreate it reliably as I can go 
1-2 days with no issues gaming (Division 2) for roughly 5 hours and the next day I might have 4 hard freezes once every 20-30 minutes. To date I have not had the issue while not under load ( surfing web).
Never had any of these issues when I had my GTX 1060 6GB prior to this gpu upgrade.

My setup:

CPU: Intel i5 8600k, stock (no oc)
GPU: MSI OC Airflow Vega 56, stock
Mobo: Asus Z370-A (latest firmware)
RAM: Corsair Vengeance LPX 16GB (3200mhz rated, but stock mhz)
SSD (Primary): Samsung 860 EVO 500GB
HDD (Secondary): Western Digital 1TB 7200RPM HDD
PSU: Corsair RM750x 750W 80 PLUS Gold

Monitor:Samsung 24" c24fg70 1080p 144Hz (not using freesync)

OS: W10, latest build

What I've tried:

I have an MSI Card but do NOT have MSI afterburner or any other software besides the stock AMD suite.

Originally had OC on CPU and Ram. I quit that immediately when I first had a crash. Haven't been OCing for a few weeks but have the crashes frequently still so I don't think it was related.

Reinstalling/ downgrading video drivers normally (19.1.1 -> 19.3.x -> 19.4.1) On 19.1.1 I get the black screen lock up issue and I do not get that with 19.3.x -> 19.4.1 so I've been sticking to the recent.

Using DDU to uninstall the video drivers in Safe Mode, then reinstalling the latest driver (19.4.1)

Setting Windows power options to High Performance(No difference)

Purchasing a brand new PSU (had a 600 watt corsair with seperate 8 pins, now have a 750 watt gold rated with seperate 8 pins).

Purchasing additional case fans for cooling. This does not seem to be a temperature issue as my cpu never goes above 60c under load and GPU 75c under load for hours but I thought I would mention it.

Setting power to +50% in wattman. No difference.

Wiping Windows and starting over. (No difference)

When this thing works, it's great but it seems so unstable and I can't rely on it to just work.
Does anyone have any idea on what I can try next? Should I RMA the card?

[goodplay]

Have you tried,

[heartless_ca]

Not sure what you want from the dxdiag but the reliability report looks bad. Here are a few things:

Please let me know what other info I can provide to help.

[goodplay]

Did you use DDU to uninstall the previous nv drivers, maybe check again that they have all been removed.

Have you been installing the amd drivers over the top of old, or using the custom clean install option(not ddu).

You can search the web for those livekernel errors and try the various troubleshooting steps.

If the freezes are only in game/s lower some graphics settings, read some (name of game) graphics performance tweak guides.

You could post the dxdiag file if you want, might see something. Save it and attach as a file (use advanced editor) for reply.             

[heartless_ca]

Appreciate the help btw, people like you deserve a shout out.

Dxdiag attached.

I install the drivers on top of the old ones normally, especially after I reformatted windows (to make sure nv stuff was completely gone). 

New version released yesterday I think.

I will try a clean install option this time around. 

I'll keep you posted.

[goodplay]

The only AMD problem shown is AUEPMaster , although possibly not responsible, you don't want it.

Other issues, as an example,

Corsair Link ? (disable/uninstall/update/reinstall ?).

There are others that you can scroll down to the WER section of your dxdiag and web search the event and/or prob sig. for possible fixes.

Or you could try posting at https://www.tenforums.com/bsod-crashes-debugging/ 

Their debug of software/hardware is usually good. 

[heartless_ca]

Steps taken:

Opt-out of the recommended AUEPMaster.

Uninstall all corsair utilities - using the generic driver.

Reboot.

Played PUBG for roughly 40 minutes... BSOD. - Thread stuck in device driver.

Not sure if you want me to post the mini dump or how to do so. Also uploaded new dxdiag.

I see this:

+++ WER9 +++:
Fault bucket 1883009619513227646, type 5
Event Name: RADAR_PRE_LEAK_64
Response: Not available
Cab Id: 0

Problem signature:
P1: TslGame.exe
P2: 5.4.15.6
P3: 10.0.17763.2.0.0
P4:
P5:
P6:
P7:
P8:
P9:
P10:

TslGame.exe is Pubg, happened after ~40 minutes.

[heartless_ca]

Dxdiag2

[goodplay]

I think you need to go to tenforums(link I posted), they look at a lot more than just dxdiag for troubleshooting as there may be issues a lot deeper than what dxdiag reports.

[heartless_ca]

Man I’m super bummed out. I never had a single issue with my 1060 whatsoever  for over a year. Thanks for your help though.

[muon]

I get this too on linux with various distro's with my Sapphire Vega 56 Pulse. It happens intermittently usually a few times a day if I am playing games, the screen will just freeze and I have to hard reboot. I have used various mesa drivers and linux kernels all with same result. I have a corsair RM850x psu so it is similar to ops which might be the issue? Strangely when I have gamed in windows it doesn't seem to be happen though I haven't tested it as much so it might just be because I dont use it enough to see it. I bought it as amd are supposed to be good with linux which I have not found the case to be at all, very disappointed in the card. I'm not sure where to go to investigate further.

[higsta]

I too have this issue, with a MSI OC Airboost Vega 56.

I have tracked it down to Driver issue, had the minidumps analysed over at Microsoft community 3 response all with same outcome that it was a driver issue.

Analysis Results 

File Name: 050619-5125-01.dmp 
THREAD_STUCK_IN_DEVICE_DRIVER_M (100000ea) 
*** WARNING: Unable to verify timestamp for atikmdag.sys 
DEFAULT_BUCKET_ID: GRAPHICS_DRIVER_FAULT 
IMAGE_NAME: dxgkrnl.sys 

and searching for vega 56 driver issues i have found numerous post regarding the same issue across many different brands but Gigabyte and MSI seem to be the majority.

I have had this card 2 months and it has done this with every set of drivers since i've had it from 19.3.3 to 19.4.3 the recommendation was to ask MSI for the recommended driver, and their response was the one that they post on their support page which is 19.1.1 so im now running that.

Thing is my system can run anywhere from 5 days with no issues to 5 crashes in as many hours. so only time will tell I guess.

[heartless_ca]

Does 19.1.1 work for you? I rma my card I’m awaiting a return. Using my old gpu without issue for 3 weeks now.

[higsta]

Yes 19.1.1 is now 16 days no issues at all !

I too rma'd the 1st Vega 56 I had and my old GPU ran with no issues when the new one arrived I was getting all the same problems again but now it looks like 19.1.1 downloaded from the msi site is working fingers crossed.

[higsta]

UPDATE - Day 21 an system crashed exactly the same as it had been doing!!! 

Day 22 no crashes

Day 23 no crashes 

Day 24 CRASHED again! 

Can this really be a driver issue?

21 days with no issues

2 crashes in 3 days.

And I have not installed any other software during this time.