Aorus 3080 Xtreme Rev 2.0 - DWM.exe Crash

Viper16

[H]ard|Gawd
Joined
May 3, 2005
Messages
1,886
I lucked into the NewEgg Shuffle for a combo overpriced gpu and cpu...Aorus Master Rev 1 Z590 and Zorus Xtreme 3080 Rev 2..to the tune of $1500. I need a upgrade obviously. Well I hadn't purchased a cpu/ram/cooler yet, and installed this GPU on my existing setup. I am currently plagued with the blank screen, monitor disconnect, gpu fans 100%, and the disconnect/reconnect beep of the hardware. ctrl-shift-winkey-b doesnt appears to do anything. I DDU and rolled the latest drivers...still nothing. I have used one rail to power two 8-pins and a separate cable to power the third. The disconnect is random, and typically blips so fast I can't see what is happening. I notice stable 12.1, 12.2, 12.0 voltage on the 3-pins and even stable on the mobo voltage. I can run 3D MARK or play newer games with almost no fault running +2000mhz gpu, pulling +350 watts, and no issues. I go try to load satisfactory or outer worlds and it glitches by the time the first splash screen shows up. It is on a UPS with Sinewave output, I hear not clicking over to battery. I can run Iracing with Oculus, and wheel/pedals with no erroring out. It finally got bad enough that it locked my mbr up and I could not repair to get back into windows....which lead me to believe it may be failing ssd. I have experienced the blank screen and no boot before with the 1080ti in place. Sometimes I can go minutes, sometimes I can an hour or two before it happens.

I reformatted last night, loaded a game / latest nvidia driver on a fresh boot, and got the failure almost immidiately when going into the second splash screen.

I plan to invest in a CPU/RAM, since I already have the MOBO, but was hoping to be more successful with my GPU in its current setup. Is the old Z97 not up to the task? I feel the rails are providing full power and not hiccuping on the PSU.

SOLVED: Gigabyte finally replaced on second time back through warranty. New card as of yet has no issues. Was fairly quick for them to warranty it once explained issues. 🤞
 
Last edited:
Check what voltage the UPS is outputting (mine luckily has a display). I was having a similar issue recently and the cause was, weirdly, a combination of a nearby box fan in a window and an AC unit in a different room causing a lower AC voltage that was *just* enough to cause the same DWM.exe crashing.

Turns out that temperature wasn't the problem, it was the power draw. It wasn't low enough to kick the UPS on (only 105ish volts and it kicks in at 98V) but it was enough to cause something in the GPU to trigger the DWM.exe crash. It just so happened that power draw was lowered when the house was the hottest due to the Air Conditioning unit turning on and more box fans running.

Diagnosing it was an absolute pain in the ass. I did also downgrade my driver's back to 466.(something) and haven't tried updating again. I also set my GPU voltage back to completely normal instead of my usual 6% undervolt.

You can read my recent thread on it here:
https://hardforum.com/threads/1080ti-471-41-drivers-causing-dwm-exe-crashes-solved.2012548/
 
Last edited:
Will give it a try. My UPS will show the voltage input, I will move it around so I can see the display and keep an eye on it. Everytime I have looked at it, it has had a strong 120-121 volts. Thanks!

Went ahead and ordered a my cpu/ram/nvme to finish up my upgrade path.
 
OK, been playing around with this a bit. I finally upgraded the MOBO, RAM, NVMe and Fresh Windows 10 and Drivers. What I noticed, I can play older games with the power turned down below 50% and -150 hz in MSI Afterburner, I am talking hours and hours zero issues. I run 3dmark benchtesting with full power and have not had a crash. I can run it multiple times and zero issues. GPU is +2050mhz and 100% power, pulling +350 watts, no issues.

It is the GPU itself that is resetting, as it is the classic blank screen, gpu fans 100%, disconnect/reconnect windows sound, and wont reset display drivers until reboot, you can still hear the audio going on in the game. ctrl/win/shift/b just beeps.

I am using an older PSU, with one cable going to 2/3 8-pin plugs and a separate plug for the third power pin. I get a strong 12.1 volt on all plugs, and combined with the PCIE power I can pull +350 watts no problem. It could be a GPU power spike that could be causing a glitch in PSU failing the rail, causing the GPU to reboot, but would it not cause the PSU as a whole to shutdown? my APC UPS is the sinewave, shows a strong 122-123 volt supply. I checked wiring for proper plugging...I watch GPU-z and HWINFO and keep an eye on all temps/rails for proper voltage, and it seems good, but again could be a short enough glitch, trending cannot catch it.

If this is a strong sign of too much for my PSU, I have a new RM1000x at the house, but didnt really want to use it if It was not the problem....

anyone have any experiance in this issue?
 
OK, been playing around with this a bit. I finally upgraded the MOBO, RAM, NVMe and Fresh Windows 10 and Drivers. What I noticed, I can play older games with the power turned down below 50% and -150 hz in MSI Afterburner, I am talking hours and hours zero issues. I run 3dmark benchtesting with full power and have not had a crash. I can run it multiple times and zero issues. GPU is +2050mhz and 100% power, pulling +350 watts, no issues.

It is the GPU itself that is resetting, as it is the classic blank screen, gpu fans 100%, disconnect/reconnect windows sound, and wont reset display drivers until reboot, you can still hear the audio going on in the game. ctrl/win/shift/b just beeps.

I am using an older PSU, with one cable going to 2/3 8-pin plugs and a separate plug for the third power pin. I get a strong 12.1 volt on all plugs, and combined with the PCIE power I can pull +350 watts no problem. It could be a GPU power spike that could be causing a glitch in PSU failing the rail, causing the GPU to reboot, but would it not cause the PSU as a whole to shutdown? my APC UPS is the sinewave, shows a strong 122-123 volt supply. I checked wiring for proper plugging...I watch GPU-z and HWINFO and keep an eye on all temps/rails for proper voltage, and it seems good, but again could be a short enough glitch, trending cannot catch it.

If this is a strong sign of too much for my PSU, I have a new RM1000x at the house, but didnt really want to use it if It was not the problem....

anyone have any experiance in this issue?
Sounds like you've eliminated a lot of other variables. Why not give the new PSU a try? Weren't there actually some Seasonic units a couple of years ago that had a problem with newer video cards and they were tripping OCP really quickly or something? If Seasonic can screw up, I'm sure Thermaltake or whoever makes that PSU can as well. If you have to RMA the card, you will have at least done the troubleshooting.
 
What nvidia driver are you using? I haven't upgraded mine back up to 471 for fear of it crashing again, but I've been running fine on 466.27
What you have going on sounds exactly like what my issue was last week, but since i removed that box fan (and did everything else in that thread) i'm now at 9 days uptime with zero issues.
 
kamikazi - I plan to install the new PSU, just didnt really want to spend another $200.00 if I could resolve it with my 750w PSU from my 2014 build, especially if it was going to do the same failing point.
TheSlySyl - I am running the latest 471.68 - my only other recent weird experience is when both of my monitors shut off, and then both go to sleep, only one will wake back up. I have to turn monitor off and then on again and it will boot up....very odd.
I will try my VR and make sure my wheel/pedals and all that are interfering. I had the GPU on the old 4790k system and it ran VR for over an hour and no issues.
 
if you have a new 1k psu, just hook it up outside the case to test. youll have your answer quickest that way...
edit: tried what thesylsyl said about power/ups?
 
Part of me also wonders if the DWM.exe crash is related to this somehow and multiple monitors (I have 3)
https://hardforum.com/threads/nvrla-exe-may-significantly-increase-cpu-usage-heat.2012758/

Try using DDU and reverting to an old GPU driver. I'm back on 466.27

Also, disconnect down to one monitor and see if it helps stability. It did help me at least to find a way to replicate the crash, which is how i troubleshooted the cause.

FWIW I'm also using an older 750Watt PSU (EVGA GQ) but I think you probably draw more power than mine does (Though my 1080ti does draw 280 watts on the regular.)
 
I’d swap PSU’s no doubt. You’ve eliminated most other stuff.
Yeah I am thinking this is my issue...had thought it from the beginning but wanted to rule out other things. I tried to run Furmark with 50% GPU, the instant I hit 1080P stress the PSU would trip and reset power.
 
Yeah I am thinking this is my issue...had thought it from the beginning but wanted to rule out other things. I tried to run Furmark with 50% GPU, the instant I hit 1080P stress the PSU would trip and reset power.
I would swap them. Sounds for sure like a PSU issue.
 
Swapped PSU out, can get through Furmark on all tests...no issues. ran three separate 8-pin cables from PSU. Still have reset issue on the lighter graphical consumption games (satisfactory), GPU runs around 65%, I have limited FPS to 120, but there is something in the GPU MHZ spiking that causes the card to shut down. I left the power level @ 100%, but adjusted down -150 mhz on the Afterburner. I played for another couple hours and had no issue.

I think it has to do with the GPU ramping up and hitting the mhz ceiling, that causes the card to shutdown, usually around +2080 mhz.
 
Vbios version? It appears to have a couple of updates with the latest being F71 if I'm looking at the correct GPU. The update change log states support for resizeable bar but no other notes.
 
Swapped PSU out, can get through Furmark on all tests...no issues. ran three separate 8-pin cables from PSU. Still have reset issue on the lighter graphical consumption games (satisfactory), GPU runs around 65%, I have limited FPS to 120, but there is something in the GPU MHZ spiking that causes the card to shut down. I left the power level @ 100%, but adjusted down -150 mhz on the Afterburner. I played for another couple hours and had no issue.

I think it has to do with the GPU ramping up and hitting the mhz ceiling, that causes the card to shutdown, usually around +2080 mhz.
Sounds like a defective card, RMA time.
 
Vbios version? It appears to have a couple of updates with the latest being F71 if I'm looking at the correct GPU. The update change log states support for resizeable bar but no other notes.
94.02.74.40.F5 ---so I assume F5 bios?
IMG_3350.jpg
 
Silly question: Are all the Gigabyte bios' stored within the AORUS Engine (downloaded within its software)? cause I cannot find any information about the bios versions on their awful website.
 
Silly question: Are all the Gigabyte bios' stored within the AORUS Engine (downloaded within its software)? cause I cannot find any information about the bios versions on their awful website.
Don't think there are any for your card. Only Rev. 1.0 has bios updates listed.
 
That's a shame that the card seems to be a lemon. Hopefully, they can take care or you in a reasonable amount of time.
 
All fine?
Doesn't seem to be, I moved to F6 bios with the switch on the card, assuming it is the "silent" bios, it lasted a tad longer on my game (satisfactory), but still did its GPU reset, game still on, audio still playing, GPU fans 100%. The only way I can play that game is with -150 mhz and 50% power on GPU. Of course I am still getting +100 FPS. I cant tell if it is a thermal shutdown or what...I am trying to figure out if I need to register the card on Gigabyte website first, which is partially shutdown due to their attacks, I will most likely start and email and ask them if there is anything in the works, thanks pendragon1 for the recommendation.

I had been exclusively a EVGA user for many generations, but it seems this issue is not uncommon with the other mfg's, just hope I made a good purchase. Afraid of my card being warrantied and getting lost.
 
Doesn't seem to be, I moved to F6 bios with the switch on the card, assuming it is the "silent" bios, it lasted a tad longer on my game (satisfactory), but still did its GPU reset, game still on, audio still playing, GPU fans 100%. The only way I can play that game is with -150 mhz and 50% power on GPU. Of course I am still getting +100 FPS. I cant tell if it is a thermal shutdown or what...I am trying to figure out if I need to register the card on Gigabyte website first, which is partially shutdown due to their attacks, I will most likely start and email and ask them if there is anything in the works, thanks pendragon1 for the recommendation.

I had been exclusively a EVGA user for many generations, but it seems this issue is not uncommon with the other mfg's, just hope I made a good purchase. Afraid of my card being warrantied and getting lost.

Why do you think it will get lost?
 
Why do you think it will get lost?
Lost in Freight or lost in the returns. Been reading some GB stories about sitting in GB receiving for a week or two before they claim receipt, and then another 6-8 weeks for repairs...unsure if backlog or just missing items. I would rather exhaust anything I could do on my end before physically sending it back.
 
Lost in Freight or lost in the returns. Been reading some GB stories about sitting in GB receiving for a week or two before they claim receipt, and then another 6-8 weeks for repairs...unsure if backlog or just missing items. I would rather exhaust anything I could do on my end before physically sending it back.
I've done a few RMAs with gigabyte. Typically takes 2-3 weeks.
 
Ok, finally Gigabyte had there registration services back up and running, got the GPU/MOBO registered. I ran some tests yesterday, played a few other games to see if I could get it to crash, I could not (Cities:Skyline, CyberPunk, Far Cry 5), while I did not hammer the games too long, I was unsuccesful. I did remember on my way to work this morning, that i set the Nvidia software to keep the GPU clock ramped up, so I will go back and turn that off, and play some other games.
 
Figured I would update - I turned off the high performance clock setting in the NVIDIA 3D SETTINGS, have played Iracing 4-5 hours over the last two days and have not had one slip up. It appears my GPU resets are caused by certain games...and not all (Satisfactory, Witcher 3, Outer Worlds). I am running the GPU in F6 bios.

Running Furmark or 3DMark has not caused the card to crash, only when I had the older PSU in there, which fixed Furmark. For those select few games, perhaps I just lower the speed and power of the card when I want to play? I still get +100 FPS in those games at the reduced GPU settings.
 
Figured I would update - I turned off the high performance clock setting in the NVIDIA 3D SETTINGS, have played Iracing 4-5 hours over the last two days and have not had one slip up. It appears my GPU resets are caused by certain games...and not all (Satisfactory, Witcher 3, Outer Worlds). I am running the GPU in F6 bios.

Running Furmark or 3DMark has not caused the card to crash, only when I had the older PSU in there, which fixed Furmark. For those select few games, perhaps I just lower the speed and power of the card when I want to play? I still get +100 FPS in those games at the reduced GPU settings.
Go into the 3d settings and set it to "prefer maximum performance" and see if the crashing continues. If it does I would RMA it. I would not accept a defective/partially working card.
 
Well I sent the card back two weeks ago...had been checking there warranty status page...seems like they could not replicate any of the issues. Bummer. I get the card back on Thursday, so will see if anything was changed in the bios or the card is different.

So I guess I need to get to working on my computer and isolating the problems...never had GPU issues like this in the past, but have always had EVGA cards. Any tips on where to start?
 
Well I sent the card back two weeks ago...had been checking there warranty status page...seems like they could not replicate any of the issues. Bummer. I get the card back on Thursday, so will see if anything was changed in the bios or the card is different.

So I guess I need to get to working on my computer and isolating the problems...never had GPU issues like this in the past, but have always had EVGA cards. Any tips on where to start?
If you need extended help I'd be glad to help you, have spares of everything. Just PM me.
 
Okay, got the card back, nothing was changed on it from the warranty. Warranty states they tested it in many different configurations and did not see any faults. I can run this in 3DMARK all day long and it will not fault. It has to have the dynamic of a video game where the card voltage has the opportunity to spike.



Things I have done:

TEST 1:

PSU – confirmed that I have three different cables going to card, and that they are secure.

MOBO – updated the bios to the current “8FA”, enabled BAR Resizable, left everything else standard (XMP TURNED OFF)

Drivers – DDU removed and installed NVIDIA Current published.

GPU – In OC BIOS (STOCK FROM FACTORY)

TEST: Played Cyberpunk - ~20 minutes in the game will crash randomly.

FPS: 60-65 – GPU Temp >160F – Full Load (all graphics maxed)



TEST 2:

GPU – in QUIET bios (selector switch)

TEST: Played Cyberpunk - ~40 minutes in the game and a crash occurred.

FPS: 45-50 – GPU Temp >160F – full load (all graphics maxed)


My next test, I plan to MSI AFTERBURNER and lower my HZ/Voltage well below 1900 mhz and give it a shot.

After that I plan to turn on XMP and get those setting back (I had this in the past, and it reduced the time between the card shutting off)

Do you think I should just do a fresh install? I have an older 4790K setup that I can get back up and running, it has a 750W Thermaltake in it...we can see how the GPU handles it? I have a EVGA 1080TI card that has zero problems in either of these systems. I did have to upgrade to the 1kw PSU as the 750 was not enough and caused the 3080 to shutdown under load (3dMARK), once I replaced it, it ran 3dmark through the tests.
 
UPDATE:

I have troublshot as far as I know I could:

(silent bios)
Tried fresh OS install (separate SSD card)
Swamped Cables around on the PSU to the Card
Afterburner Power Input (maxed it out)
Afterburner lowered the max hz to <1800mhz
XMP turned on
NVIDIA Settings (max performance) [ have not seen where to turn debug on to see if that fixed anything ]
GPU temp around 165F

All of the above done still GPU crashes...seems to be at a faster rate even.


What I was semi-succesful and am still testing:

(silent bios)
Power / Temp Limit - Maxed out. (appears to be around 108-110% at high levels)
Core Clock -250
Memory Clock - 300
GPU temp <159F (never exceeds this temp)
No curve edited.

Was able to play Forza Horizon for +2 hours. Previous settings I couldn't get 15 minutes or so. I was playing on my main operating system using the latest Nvidia Drivers. I am trying to build up enough information before I have to make a phone call to Gigabyte Service (manager) and have a discussion.


Again, I can leave it all stock (SILENT BIOS), and run 3dmark and run it all day long, it does not have any issues. When the crash happens, there is typically never a high demand on the card. It is like when it tries to ramp up, it pulls a GPU OCP or OVP and the card shuts down.

I have not tried another PSU...which could be an issue. Just didnt want to drop another $200.00 on a PSU that is brand new....I know i know. :)

I was going to see if I could find a computer repair place in my little town and see if I could pay them something to come and test it on a rig that is known to be functioning, and see if I can reproduce it.

Any other things I should be looking to do? I will double check the ram is on the QVL list for the MOBO...I can remove one stick and see if that make any change.
 
UPDATE:

I have troublshot as far as I know I could:

(silent bios)
Tried fresh OS install (separate SSD card)
Swamped Cables around on the PSU to the Card
Afterburner Power Input (maxed it out)
Afterburner lowered the max hz to <1800mhz
XMP turned on
NVIDIA Settings (max performance) [ have not seen where to turn debug on to see if that fixed anything ]
GPU temp around 165F

All of the above done still GPU crashes...seems to be at a faster rate even.


What I was semi-succesful and am still testing:

(silent bios)
Power / Temp Limit - Maxed out. (appears to be around 108-110% at high levels)
Core Clock -250
Memory Clock - 300
GPU temp <159F (never exceeds this temp)
No curve edited.

Was able to play Forza Horizon for +2 hours. Previous settings I couldn't get 15 minutes or so. I was playing on my main operating system using the latest Nvidia Drivers. I am trying to build up enough information before I have to make a phone call to Gigabyte Service (manager) and have a discussion.


Again, I can leave it all stock (SILENT BIOS), and run 3dmark and run it all day long, it does not have any issues. When the crash happens, there is typically never a high demand on the card. It is like when it tries to ramp up, it pulls a GPU OCP or OVP and the card shuts down.

I have not tried another PSU...which could be an issue. Just didnt want to drop another $200.00 on a PSU that is brand new....I know i know. :)

I was going to see if I could find a computer repair place in my little town and see if I could pay them something to come and test it on a rig that is known to be functioning, and see if I can reproduce it.

Any other things I should be looking to do? I will double check the ram is on the QVL list for the MOBO...I can remove one stick and see if that make any change.
If you can't find any local help again I'd be happy to help, just PM me if need be.
 
UPDATE:

I had taken my whole computer into a local shop, they tested a known working PSU and other MOBO/CPU and the problem carried over with the GPU. For sanity reasons, I had to know and was stoked. I paid them $40.00 to confirm so I would have some sort of information to help go back to Gigabyte.

Finally talked with Gigabyte support and walked them through my issues. I explained that running the card in a benchmark load it wont have issues...and that its random and at times when it feels the GPU is spiking upwards in usage that it will trip. So I sent the card back, as second time just last week and they have reviewed yesterday that a replacement GPU is going to happen. I sent them an almost booklet of the crash dump files and the GPU-Z log of a couple crashes. Apparently it was enough they could reproduce on their end.

So we shall see what I get when it gets here.
 
Back
Top