Graphics Cards

[lowl1fe]

Ever since we upgraded our GPU to a Radeon 7800xt (coming from a 1070gtx), the system became quite unstable. Random pc restarts happen in games or even while being on desktop or watching videos.

Our Specs:

• Motherboard: MSI MAG B650 Tomahawk WIFI (latest Bios)

• CPU: AMD (AM5) CPU 7600x

• Memory: Corsair Vengeance DDR5 2x 16gb (30CL at 6000hz)

• GPU: Radeon 7800xt - Sapphire Nitro+

• NVME: 1 TB Samsung 980 and a 2 TB Samsung 980 pro

• PSU: Corsair 1000w

• Windows 10 (22h2)

What we tried so far (without success):

• Upgraded our PSU from 850w (Corsair) to a 1000w (Corsair)

• Fresh installation of Windows 10

• Tried the "Silent Bios" on the GPU (without OC)

• Tested older and newer BIOS versions for the mb

• Uninstalled the Adrenalin Software and used DDU to get rid of older drivers and then installed only the plain drivers without the software

• Installed the latest chipset and on-board audio drivers

• Tried to play without EXPO enabled

• bought a new DisplayPort cable

• Connected the GPU with the PSU with two seperate cables

The list is actually a bit longer but I cannot remember everything we tried, however, there is one thing that is quite strange. The restarts happen mostly within one hour after switching on the PC for the first time.

Both temps (CPU/GPU) are actually really good (CPU at 40° idle//GPU at around 60-70° under load). So this is not really a reason for the restarts.

We managed to recreate some restarts by "alt-tabbing" out of certain games.

Now we ran out of ideas and would be glad if some of you had some suggestions.

PS: What I forgot to mention, we had an older PC (AMD 5800/32gb G.Skill) and even there the GPU had these issues.

Best regards and thank you for reading.

[lowl1fe · Accepted solution]

The new drivers seem to have fixed it. No crashes for two days straight.

[zellith]

I have a problem of a restarting computer with a 7800xt also.  So far it only happens when we try to play elder scrolls online, but we havent really been trying other games.  Ive ran cinebench, 3dmark, prime95.  It all works.  But gaming?  REBOOT.

I'm at my wits end.  Gonna reinstall windows 10 and see if the issue persists.

PLEASE respond to my post if you find a solution.

Motherboard: Gigabyte B550 Aorus Elite v2

CPU: 5800X3D

Memory: 2400Mhz Corsair Vengance 32 GB

GPU: Powercolor 7800XT Limited Edition

NVME: Samsung 1TB 980 PRO

PSU: EVGA 1000W SuperNOVA G6

Windows 11

[zellith]

Found another thread and did a few tests.  We are now using 2 display ports and 0 hdmi.  Elder Scrolls Online no longer seems to cause reboots.  Had it on all yesterday and had no problems whatsoever.  If I come back to complain, that means we had a reboot.  Let's hope I dont reply to this message in future!

[lowl1fe]

I am using a Displayport already

[AntoinePL]

Nice workaround, but not really satisfying solution though. A 550€+ gpu should work without crashing the computer no matter chosen video outputs.

Acknowledgment from AMD still expected.

[zellith]

Was running fine until it wasnt.  Just had a reboot.  Updated to the latest graphics drivers and am going to see if it persists.

[vincent294]

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

Worth noting this still happens with single monitor, and mostly in Cyberpunk 2077. I've also seen MCE events in Event Viewer and heap corruption errors, and might be more common after driver updates? Probably a bad pointer or race condition that's hard to reproduce, and seems to only affect people on HDMI?

[Radeonwarrior1993]

• Same Green screen flickering just 5-6 mins of powering on the computer with graphic card.I have radeon 7900XT from Sapphire.Tried all the above step and same result. 
  Now due to necessity of the computer use. Using the computer after plugin out the graphic card power cables. System work smooth without graphic card. 
  
  System :-
  CPU :- Ryzen 9 7900X 12-Core Processor
• GPU:- rx7900xt 20gb from sapphire
• Mobo :- MSI b650P - WIFI
  PSU :- bequiet 800Watt
• RAM :-  gskills 32gb DDR5@ 6000
• SSD :- Samsung 2tb nvem 980 pro 

[Wembley76]

The same problem.
I use ASRock B650M Pro RS with Windows 11 Home Version 22H2 AMD Ryzen 7 7700 GPU + READON RX7800XT+DDR5 32GB.

@amd I have sent a bug report several times and have not received any response to date. Thank you AMD

[lowl1fe]

I realized that after having that crash/restart, the PC runs perfectly fine, which is quite strange. It mostly happens when I switch on the PC for the first time of the day. I also tested it in Elders Scrolls Online and I was able to recreate the crash/restart while tabbing out to desktop.

[Raven621]

Same, it follows the same pattern Wich, well it can’t be related to hardware issues, the crash follows logic, a lot of people point AMD drivers (again) 

[spraysn]

Same here. After booting and stressing the GPU with furmark, I get a coil whining and a crash/restart. Could reproduce the error once more. Since then no coil whining and no crashes. It only works when the PC is “cold”.

[Opposing1831]

I have almost the same combo and and been pulling my hair out with the vga led and overall system unstable with exo on. The mobo came with the latest bios release, so I am good there. I do get reliable boots if I go disable the wait for VGA device. I have noticed that if the system boots and then I get the white led, if I restart quickly, it will boot. After I disabled the wait for VGA, I boot everytime.

That leads me to the next problem, the second I try to set the memory to 6000 via the profile or manual, the system gets unstable with memory problems. I am currently using 1 stick and reloading windows and am going to return this Corsair memory for something else.

MSI P 650 -P Wifi

7800 x3d

7800xt Sapphire Nitro+ OC

32GB Corsair DDR5 6000

[zyplex]

Same problem here, reinstalled Windows 11. Tried 3 different BIOS versions, all available driver stacks. PC restarts randomly after undefined time. After first restart, PC is stable.

I "fixed" it with driver installation "driver only" not full install.

[ethanjm123]

Are you saying that you were having the same problems as OP (GPU disabling itself, black screens requiring restarts) and only downloading the drivers is what fixed it? I'm desperate to fix my new pc

[vincent294]

I don't have time to test much, but I can say that installing amd-software-adrenalin-edition-23.20.17.05-win10-win11-oct24-rdna.exe driver only still crashed in Cyberpunk 2077. I was able to get ~1.5 hours of play in, when I usually last 10 minutes to an hour before, so it might be less often, or maybe it's just luck. I'm gonna try rolling drivers back to whql-amd-software-adrenalin-edition-23.9.1-win10-win11-for-radeon-rx7800xt-rx7700xt.exe but Counter Strike 2 is a stutter fest.

[Ionrage]

I get this a lot and this will happen a lot within say 10 minutes of streaming a game on youtube or on twitch.

This isn't even using the AMD hardware encoder, it does crash faster when using it though.

I thought OBS was conflicting but didn't matter as was watching a youtube video and it did the same thing.

I use 4 sticks of Ram and even taken 2 out, still does it.

The crashes I get on the 7800XT is as people describe on these forums, there really seems to be something wrong with this GPU.

[AntoinePL]

Same problem described here.

[lowl1fe]

New Drivers are out as of today (02.11.23).

Here a patch note of that recent driver update:

Fixed Issues:

Intermittent black screen or code 31 error in device manager after reboot on some AMD Graphics Products, such as the Radeon™ RX 6700 XT.

Well well well...seems we might get lucky after all.

https://www.amd.com/de/support/graphics/amd-radeon-rx-7000-series/amd-radeon-rx-7800-series/amd-rade...

https://www.amd.com/en/support/kb/release-notes/rn-rad-win-23-11-1

[ethanjm123]

Didn't fix the problem for me I only got an hour and a half into Overwatch til my 6700xt disabled itself. I was forced to restart and it gave me the same error 22 and 31 codes. That's with this new driver update, a clean install of Windows 11 on a fully wiped SSD, AND a reset Motherboard with every PC component (except my CPU) unplugged and replugged. I'm waiting to get my RMA request hopefully confirmed with XFX. I've spent 10s of hours at this point troubleshooting this problem and I'm so frustrated

[lowl1fe]

Unfortunatelly neither for me. At least AMD knows about it and tries to fix it. Let's see how long it takes AMD to finally solve the problem.

[AntoinePL]

It's not an "intermittent" black screen though, it's a full crash and reboot with black (or green) screen. I'm not sure it's the same problem.

Also, I still had a crash yesterday with 23.11.1 &  7800 XT in Cyberpunk.

[Ketamean55]

I HAVE BEEN HAVING THE SAME ISSUE!!  , hoping it's driver related because I just upgrade from a 3080ti that was slowly dying and cannot be bothered again.

[nomana]

I am running into the same issue of 7800xt crashing the PC in games (gray screen on display port, with green lines) but only after a PC restart (and within a game randomly regardless of the load). Once it crashes and the PC comes back up, it runs fine for hours. I returned an ASRock 7800XT earlier thinking it was faulty hardware, but then I got a new Sapphire 7800XT after a month or so, and it has the same crash.

Before I returned the earlier card, I tried everything - changing power supply, XMP profiles, SAM on and off, reseating graphic card, changing power cables, DDU, went back to older chipset drivers etc but nothing helped. I thought it was surely a hardware issue.

Now though it seems like too big of a coincidence to get two faulty cards in a row. This has to be a driver or a BIOS problem, where something isn't set correctly on the GPU after it comes up from a long power-off time.

I'll open a trouble ticket on AMD portal and will continue experimenting with other things on the PC. The card otherwise runs so well, that I don't want to return it for something that can be fixed in the software, especially when not everyone is running into this issue (I'd assume so, otherwise there'd be more noise in the support forums, or may be people just leave their PC on 24/7)

[swagboss]

Just tossing my hat into the ring, exact same issue. At this point I'm 90% sure it's related to the BIO's on the cards themselves. 

[zyplex]

I had actually thought that I had solved the problem with a "driver only" installation. Unfortunately not.

I have now sent the card back to the retailer. I honestly hope that he will confirm this and I will get my money back.

[Erdg]

It is ridiculous that AMD has still not addressed this issue. I had spent so much time and tried everything under the sun before realising how widespread this issue was. There are so many threads here and on all the other tech forums. After the cards have been out for so long with no resolution I have submitted a claim with the retailer. Hopefully they accept the claim and I can go back to NVIDIA.

[nomana]

If you keep the PC on, this problem may not happen. I am not calling that a solution, instead just trying this to understand the situation better.

I mentioned already in this thread that I returned my first 7800XT thinking it was a faulty card, but the second one (a different brand) is showing the same issue. Now I am narrowing it down to whether the problem only happens after a PC restart (or waking up from S3 sleep) and the PC works perfectly fine after the crash for hours. So far that seems to be the case. I am keeping the PC on for few days to test this theory. If the problem is indeed something to do with GPU/board initialization after a cold start, then it's definitely fixable through drivers or GPU BIOS.

The card performs so well otherwise (frame rates, thermal, noise) that I really don't want to send it back.

I'll keep this thread updated with what I find and observe.

[SuizoX]

I have the exact same issue. I was thinking that it could be caused by my Ram ,but reading this threat I can see that it is caused by my 7800xt nitro+. 

I hope they find soon a solution, otherwise I love my card. 

[nomana]

The PC has seen no crashes, and it has been on for three days now. I just have a screen-blankout timer running, otherwise the PC is on. I have tried all sort of games which would earlier randomly crash the PC, but there are absolutely no issues.

I can conclude with enough confidence that this 7800XT crash happens only when the PC has started recently. It doesn't always happen on restarting, but whenever it crashes it's after the restart and within few minutes of trying out a game for the first time.

I'll now see whether there's any difference in complete shutdown (and if so, the duration), sleep and hibernation cycles. I'll update later.

For those of you running into this issue, either keep the PC on, or after a restart be ready for "one" crash (so don't play Diablo4 in hardcore mode, for example - true story)

[necramonium]

Same pattern for me, it will usually crash, and strangely, mostly when editing in Adobe Premiere Pro, or when upscaling a video in Topaz Video AI. RIGHT after a cold boot, it will crash. I tried today to see how long i can play a videogame until it crashes, i played for like 3 hours and nothing! And the videogame is allot more strain on my system.

[0rGaan1c]

I think I will write my experience with this card, some of the issues were because of my own incompetency, but it was my first PC build so...
So, I built this PC around 2 months ago, and I started having issues with the GPU after around a week tried several things (windows reinstall, bios update, etc) but the issues increased with time (like pc completely shutting down when gaming, sometimes, the VGA light on motherboard staying ON and monitor not getting any signal, OCCT PSU test failing) found out that it was because of loose PSU connections so I fix those (on 19-Nov-2023, it was so hard to get the PCIe wire flushed in, lol https://imgur.com/a/H0FMr1e) and it solved all of my issues except this one and it's literally the same as yours.

Ever since I fixed the loose connections(19-Nov-2023) I have only faced this "PC crashing and restarting" issue about 5 times in total (every time while playing cyberpunk) and it usually happens after 10 minutes or so, and after the crash the PC works pretty fine, I can game for hours.
Initially, I thought the crash happened when I started my PC after a long time (20 hours or so), but that doesn't seem to be the case since today it crashed in the afternoon and I did use the PC in the morning.

Not sure if it has something to do with hardware or if it's just a driver-related issue. But since you mentioned you got a different card and it is having the same issue, then probably it's not hardware-related.

[Tredans]

Same things is happening to me. Current setup is 7800xt and ryzen 7. When I installed the GPU and CPU i put probably 100 hours into Starfield with ZERO issues. then even more hours into some indie games. Then suddenly this started happening. It has to be a Driver issue with the GPU. I have updated everything to the most recent Driver and same issue. From fresh start everything runs fine until i start any game even the lowest resource games and within 10 min force restart. 

I am betting if the gpu driver is rolled back to an older version from lets say 3 or 4 months ago this issue would go away, but i cant find them. 

[randomdude]

Hi,

I think I have the same problem with my new setup. Screen just goes black, one time I think it also went green. Sometimes the audio still continues for some seconds, sometimes it just drops out with a weird repeating sound. The thing that I found interesting is that it gets stuck with the black screen and doesn't respond but after some time it reboots itself back into windows. The event viewer only shows a critical event 41 due to improper power down.
It happens via with a single HDMI projectort connected as well as as a single displayport screen. Furmark, AIDA stress test, Prime95, and combinations thereof all worked without stability issues. Reported voltages of the PSU look unsuspicious.
Most of the time my system froze right at the moment when I wanted to leave full screen of youtube or netflix. First I thought it is related to EXPO so I disabled it. It still happens. Unfortunately I didn't have so much time for thorough testing and gaming of the system. In Star Wars BFII, and CoD Cold War I didn't see this problem and gaming was stable for hours straight.

Maybe interesting for more reproducible testing: just yesterday I saw that Assasins Creed Syndicate is given away for free from Ubisoft, so I thought I'll give it a try. This game triggers the problem *very* reliable for me. Most of the time it took just 5 to 10 minutes to run into a freeze.
Maybe one of you guys can try if this game also triggers this problem for you?

The setup:
CPU: Ryzen 7 7800X3D
GPU: RX 7800 XT Sapphire NITRO+
MB: MAG MSI X670E MSI Tomahawk WIFI
RAM: 64GB (2x32) G.Skill Trident Z5 NEO DDR5-6000
Disk: Samsung 990 Pro 4TB
PSU: Seasonic Platinum-660
(all drivers and BIOS up to date)

[Deimos2202]

Got 7800XT Green screen also. Will try to download AC and test it

[randomdude]

Well... forget about that better reproducability. I tried to do more tests with AC on the same savegame that crashed several times in a row. And of course now I cannot reproduce it anymore

Maybe this is related to the behavior described also by @nomana  above: "Once it crashes and the PC comes back up, it runs fine for hours."
In my case it definitely crashed multiple times within a few minutes. But yeah maybe it gets more stable after some crashes.

[Deimos2202]

So... After the Test's with Asassin Creed Syndicate I can confirm that this game Can crush card really fast 5-7 minutes, BUT in my case 7800XT Crashes and after that I receive message in AMD Adrenalin about Unknow Error and Resseting Card to Default settings. The interesting thing that it My Card Crashes only One time per day! So it's crashing when PC was long time OFF ans system is "cold"

[randomdude]

Thanks for testing it with ACS. "Good" to hear that you can reproduce it quickly too. I didn't get the Adrenalin error, though. It just got stuck and it automatically rebooted after 30 seconds or so.

To add another observation: I just ran furmark with GPU and CPU stress test to first warm up my system before giving ACS another try. I left the test running while browsing, everything was stable. After temperatures settled for some time I stopped Furmark and started ACS with my quickly crashing savegame and not even a minute into the game, it crashed again. This tells me that it is probably not related to the temperature / warmup period itself but some other thing that goes away after the crash (or multiple crashes).

[lowl1fe]

The crashes have to be driver or OS related. On reddit someone tested it on Linux and he had no crashes at all.

[vincent294]

Not sure it's just Windows:

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

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

Also good to note what distribution they are using (Ubuntu Desktop 22.04.3 HWE kernel, Fedora Workstation 39, etc).