New (open box) 7900 XTX issues - spontaneous reboots - need help

W

Wag

[H]ard|Gawd
Joined
Aug 29, 2006
Messages
1,354
I finally got around to installing my open-box Red Devil 7900xtx. I had to do a fresh Win11 install, but even now I am still getting spontaneous reboots- trying to narrow it down. I have a month to return to the card to Microcenter, so I want to try and figure out if it's just an instability issue with AMD in general or something else, possibly a bad card (it was a return).

I have an Asus X670e Creator wifi mobo, 7900x and the Red Devil is the new edition. My previous 2080Ti was rock solid so I doubt it's my EVGA 1300 G2, thing is a beast (albeit a 7-8yr old one). It is old and has powered some pretty hefty setups, including dual 1080Ti SLIs a while back but should still be good.

In any event, I have one reproduceable reboot- if I turn Auto-HDR on. Eventually I will just reboot- straight to post. I am undervolting my ram somewhat, I'm wondering if that might be causing the instability. Do people suggest I just go into the bios and set everything to defaults and see how it goes?

Everything is current, latest BIOS, Firmware for everything, including my Gigabyte F43U display, with Win11 driver.

2080Ti was fine with Auto-HDR and everything else.

I haven't even tried gaming yet. I ran a short stability test in GPU-Z, seemed ok, temps were in the 66C range.
 
You say you had to do a fresh Win11 install. Did you do so because of initial issues and if so were they different then now?
Ya set your bios to default as a starter, you can tune later.
 
funkydmunky said:
You say you had to do a fresh Win11 install. Did you do so because of initial issues and if so were they different then now?
Ya set your bios to default as a starter, you can tune later.
Click to expand...
Similar, yes. Even after DDU I was getting spontaneous reboots.

I'm turned off my undervolt and still spontaneous reboots. I turned off HDR. Next I will totally default the mobo BIOS.

I'm beginning to suspect I have a bad card- either that or PSU. New PSU? I don't know, rock solid with the 2080ti. I guess I can try different outputs on my PSU and see what happens. I suppose if it comes down to it I can repaste the card (doesn't seem to be running hot) but I don't want to mess with it too much if I'm going to return it.

Oh well, worth a shot, I saved a lot of money. Not time and grief and gas.
 
Last edited:
Older PSUs aren't capable of switching fast enough to deal with these newer high-wattage cards, it's been a problem here for over a year. Whenever someone gets a flagship or high-end GPU and pairs it with an old PSU and they get unexpected shut-downs, it turns out to be the PSU.

I haven't been keeping up with PSUs lately so I can't point you in any direction there. What I can tell you is find a 2021 or 2022 model PSU, and if you think it might be the video card, just keep the receipt and put it on a credit card in case I'm wrong and you want to return it.

You could take a completely new PSU from 2015 that was never opened and it would have the same problems. They never thought they'd need to switch as fast as these cards require back then.
 
Wag said:
I finally got around to installing my open-box Red Devil 7900xtx. I had to do a fresh Win11 install, but even now I am still getting spontaneous reboots- trying to narrow it down. I have a month to return to the card to Microcenter, so I want to try and figure out if it's just an instability issue with AMD in general or something else, possibly a bad card (it was a return).

I have an Asus X670e Creator wifi mobo, 7900x and the Red Devil is the new edition. My previous 2080Ti was rock solid so I doubt it's my EVGA 1300 G2, thing is a beast (albeit a 7-8yr old one). It is old and has powered some pretty hefty setups, including dual 1080Ti SLIs a while back but should still be good.

In any event, I have one reproduceable reboot- if I turn Auto-HDR on. Eventually I will just reboot- straight to post. I am undervolting my ram somewhat, I'm wondering if that might be causing the instability. Do people suggest I just go into the bios and set everything to defaults and see how it goes?

Everything is current, latest BIOS, Firmware for everything, including my Gigabyte F43U display, with Win11 driver.

2080Ti was fine with Auto-HDR and everything else.

I haven't even tried gaming yet. I ran a short stability test in GPU-Z, seemed ok, temps were in the 66C range.
Click to expand...
Try another set of GPU PSU cables. see my post on kernel error 41.
 
drutman said:
Try another set of GPU PSU cables. see my post on kernel error 41.
Click to expand...
Ok, I will play with it a bit. As it stands now I’m just getting white/grey screen spontaneous reboots quite frequently in Win11. I stated earlier I have a EVGA 1300 G2 PSU

Maybe I should run Linux and see what happens?
 
Wag said:
Ok, I will play with it a bit.
Click to expand...
drutman said:
I would pop in a freshly formatted SSD with just win 11 installed and load a game to se if it crashes. This will tell you if it is driver /software conflict issue related. HW swapping is a PIA. I found out that my Cablemod 6 and 8 pin custom harness not making contact was the issue. I swapped out every piece of HW except the MB.

https://cdn.wccftech.com/wp-content/uploads/2023/07/youtube-video-gif.gif
Click to expand...
This is a fresh install. I had difficulty even making it through the install- Windows 11 was having problems installing.
 
Sounds like a connection issue, mine was totally random (my SSD swap did what yours did, bombed on install).
Two days later I accidently touched the cable while powered on and the PC shutdown with the same kernel error. This is how I diagnosed it.
The Cablemod harness is lightweight and the connector is slimmer than the stock EVGA connector.
Stock EVGA cable is harder to snap in and more robust, not sure if it is a flaw or just the way Zotac spaces the connections.
 
Typically if it's a PSU problem it will not have issues in low load scenarios like installing windows or desktop use. I know what Axman is talking about though, had issues with that with the 30 series. My best guess here is the gpu is the problem. Might be able to get the MC guys to test the card for you.
 
I suspect one of the
Furious_Styles said:
Typically if it's a PSU problem it will not have issues in low load scenarios like installing windows or desktop use. I know what Axman is talking about though, had issues with that with the 30 series. My best guess here is the gpu is the problem. Might be able to get the MC guys to test the card for you.
Click to expand...
They claimed they did, or they do before they put it out for resale, although I tend to wonder about that. They probably just test for a minute or two to see if it posts.
 
Wag said:
I suspect one of the

They claimed they did, or they do before they put it out for resale, although I tend to wonder about that. They probably just test for a minute or two to see if it posts.
Click to expand...
Agreed, I highly doubt they even check all the outputs.
 
Test the card in another system, and test with your system set to totally stock settings, no overclocking of anything. That includes undervolting the memory.

Also test the voltage on the 12V rail (preferably with a hardware voltmeter, not the motherboard's onboard monitoring). While you run a game-like application, such as 3DMark. If you're seeing substantially less than 12V, suspect the power supply.

A bad card is a possibility (maybe even an elevated one with a card that you know was returned for some reason), but generally, I would expect a bad card to produce different symptoms than this, such as hard shutdowns, black screen with full blast fans, maybe blue screen, et cetera. Hard reboot suggests either motherboard or power supply, to me.
 
RazorWind said:
Test the card in another system, and test with your system set to totally stock settings, no overclocking of anything. That includes undervolting the memory.

Also test the voltage on the 12V rail (preferably with a hardware voltmeter, not the motherboard's onboard monitoring). While you run a game-like application, such as 3DMark. If you're seeing substantially less than 12V, suspect the power supply.

A bad card is a possibility (maybe even an elevated one with a card that you know was returned for some reason), but generally, I would expect a bad card to produce different symptoms than this, such as hard shutdowns, black screen with full blast fans, maybe blue screen, et cetera. Hard reboot suggests either motherboard or power supply, to me.
Click to expand...

Unfortunately the only way to test in another system is to yank the PSU and use my 5820k which I don’t want to do. I had a lower lumbar fusion which makes it difficult for me to bend over for an extended period of time. It took me 3x-4x longer to build my new setup than it should have.

I will try to swap out the cables, use other outputs on the PSU. EVGA cables are sleeved and very difficult to bend, that might have something to do with it. I think 2 of the 8pin cables I’m using have never been used- I’ll look and see if I have some others and make sure the card ram is seated properly. It’s difficult for me to even do that.
 
I don't think the cables are your problem, but you could try swapping them out if you have spares.

You should check the actual voltage of the 12V rail while the system is working, though. A healthy power supply will keep the 12V rail at 12V. A flaky one may let it droop down to 11V, and that could definitely be a cause of your symptoms.

You can use HWInfo if measuring with a multimeter isn't an option for you. I'm pretty sure most motherboards report it. Worst case is that the problem isn't your power supply, and this allows you to rule that out. Next suspect would be the motherboard, IMHO.
 
RazorWind said:
I don't think the cables are your problem, but you could try swapping them out if you have spares.

You should check the actual voltage of the 12V rail while the system is working, though. A healthy power supply will keep the 12V rail at 12V. A flaky one may let it droop down to 11V, and that could definitely be a cause of your symptoms.

You can use HWInfo if measuring with a multimeter isn't an option for you. I'm pretty sure most motherboards report it. Worst case is that the problem isn't your power supply, and this allows you to rule that out. Next suspect would be the motherboard, IMHO.
Click to expand...
This is probably the correct answer. I bought high end braided cables 10 or so years ago and have never had an issue. I did have an issue with cheap cables that came with a PSU before however.
 
RazorWind said:
I don't think the cables are your problem, but you could try swapping them out if you have spares.

You should check the actual voltage of the 12V rail while the system is working, though. A healthy power supply will keep the 12V rail at 12V. A flaky one may let it droop down to 11V, and that could definitely be a cause of your symptoms.

You can use HWInfo if measuring with a multimeter isn't an option for you. I'm pretty sure most motherboards report it. Worst case is that the problem isn't your power supply, and this allows you to rule that out. Next suspect would be the motherboard, IMHO.
Click to expand...
The only problem with that is it won't show if transient spikes are causing OCP to trip which is one of the common issues using older PSUs on newer cards.
 
RazorWind said:
I don't think the cables are your problem, but you could try swapping them out if you have spares.

You should check the actual voltage of the 12V rail while the system is working, though. A healthy power supply will keep the 12V rail at 12V. A flaky one may let it droop down to 11V, and that could definitely be a cause of your symptoms.

You can use HWInfo if measuring with a multimeter isn't an option for you. I'm pretty sure most motherboards report it. Worst case is that the problem isn't your power supply, and this allows you to rule that out. Next suspect would be the motherboard, IMHO.
Click to expand...

I don’t think it’s the motherboard, I’ve been using it for almost a month with no problems. HWinfo shows the +12V rail as all 12.192V, no fluctuations.
 
Ebernanut said:
The only problem with that is it won't show if transient spikes are causing OCP to trip which is one of the common issues using older PSUs on newer cards.
Click to expand...
True, but I'm assuming that, if he's stumped on this, he probably doesn't have the oscilloscope that he'd need in order to do that test rigorously.

Wag said:
HWinfo shows the +12V rail as all 12.192V, no fluctuations.
Click to expand...
Interesting that it's showing that high. Is that with a game or benchmark running?
 
RazorWind said:
True, but I'm assuming that, if he's stumped on this, he probably doesn't have the oscilloscope that he'd need in order to do that test rigorously.


Interesting that it's showing that high. Is that with a game or benchmark running?
Click to expand...

No, unfortunately I don't have an oscilloscope, I guess I could buy one.

And no, no game running, nothing. I really haven't installed anything. I just switched around some of the cables and I'm getting the same 12.192V reading.

I also disabled MPO in the registry, no difference.

I might try Linux. At this point I can only think of buying a new PSU or returning the card.
 
Try checking the voltage with a game or 3DMark running. Start the benchmark, alt-tab out, and look at what HWInfo says. It's not that useful to check it when the system is idling, since that's not apparently when your problems are happening.
 
RazorWind said:
Try checking the voltage with a game or 3DMark running. Start the benchmark, alt-tab out, and look at what HWInfo says. It's not that useful to check it when the system is idling, since that's not apparently when your problems are happening.
Click to expand...
It is happening when I'm on the desktop.
 
Hmm, maybe it's is the protection circuitry of 1300 G2. When I first installed my Power Devil 7900XTX with 1300 G2 (also around 7yrs old, powering various systems), the first time I powered on it sounded like a relay or CAP popped (no smell though). Thought it was a loose cable, but connections seemed fine (PSU and PCI-E). Switching to a Seasonic Focus 850w Platinum and no issues. In my case, the G2 never powered on after initial pop so RMA'd the G2 and got a Supernova 1300GT as replacement.
 
Eymar said:
Hmm, maybe it's is the protection circuitry of 1300 G2. When I first installed my Power Devil 7900XTX with 1300 G2 (also around 7yrs old, powering various systems), the first time I powered on it sounded like a relay or CAP popped (no smell though). Thought it was a loose cable, but connections seemed fine (PSU and PCI-E). Switching to a Seasonic Focus 850w Platinum and no issues. RMA'd the G2 and got a Supernova 1300GT as replacement.
Click to expand...

It's still under warranty. I could probably send it back but technically there isn't anything wrong with it- I suppose I can say it's not working with my new GPU, and of course they would test it- probably not worth the shipping cost. Well, at least I see someone else had a problem with the same GPU/PSU combo. I hate wasting money if it's not the issue.

Decko87, I said in the OP I've been using a 2080Ti for weeks now with no issues. I'll probably reinstall it because as it is now it's unusable.
 
Wag said:
It's still under warranty. I could probably send it back but technically there isn't anything wrong with it- I suppose I can say it's not working with my new GPU, and of course they would test it- probably not worth the shipping cost. Well, at least I see someone else had a problem with the same GPU/PSU combo. I hate wasting money if it's not the issue.

Decko87, I said in the OP I've been using a 2080Ti for weeks now with no issues. I'll probably reinstall it because as it is now it's unusable.
Click to expand...
Yeah pop it back in and see if its stable again, only takes a few minutes. I think the card is a dud.
 
Decko87 said:
Yeah pop it back in and see if its stable again, only takes a few minutes. I think the card is a dud.
Click to expand...

Nope, card is fine, PSU fine, AMD is the problem.

Installed Ubuntu on a spare drive, installed latest AMD Ubuntu drivers, Steam, then Shadow of the Tombraider.

4k, Vulkan, all settings high, 136fps average on benchmark. No artifacts, everything rock solid. Been using Ubuntu the last few hours, absolutely no problems, no reboots, etc.

First time in forever I purchased an AMD card and it doesn't even work in Windows. I don't know what else there is to say.

I could probably bang my head against a wall forever and still not get it to work. You would think that since I at least that I have an entire AMD build they would work OK together.

Unless anyone has any ideas, I guess I'll return it- probably skip this gen and wait for the next gen of Nvidia cards.
 
Last edited:
Wag said:
Nope, card is fine, PSU fine, AMD is the problem.

Installed Ubuntu on a spare drive, installed latest AMD Ubuntu drivers, Steam, then Shadow of the Tombraider.

1440p (couldn't get it to 4k for some reason), Vulkan, all settings high, 240fps average on benchmark. No artifacts, everything rock solid. Been using Ubuntu the last few hours, absolutely no problems, no reboots, etc.

First time in forever I purchased an AMD card and it doesn't even work in Windows. I don't know what else there is to say.

I could probably bang my head against a wall forever and still not get it to work. You would think that since I at least that I have an entire AMD build they would work OK together.

Unless anyone has any ideas, I guess I'll return it- probably skip this gen and wait for the next gen of Nvidia cards.
Click to expand...
AMD isn't the problem or it would be a widely known issue. Most people with 7900 XTXs have a rock solid experience , myself included. Does the card crash @ 1440p on windows? Is the windows install old?
 
Decko87 said:
AMD isn't the problem or it would be a widely known issue. Most people with 7900 XTXs have a rock solid experience , myself included. Does the card crash @ 1440p on windows? Is the windows install old?
Click to expand...
Fresh install, totally unusable (spontaneous reboots at stock) in windows regardless of the resolution.

And I was able to benchmark Tomb Raider in Ubuntu @ 4k, worked fine. Ran for hours, no reboots.

The only other thing is the PSU. But it work’s absolutely fine in Linux.
 
Wag said:
Fresh install, totally unusable (spontaneous reboots at stock) in windows regardless of the resolution.

And I was able to benchmark Tomb Raider in Ubuntu @ 4k, worked fine. Ran for hours, no reboots.

The only other thing is the PSU. But it work’s absolutely fine in Linux.
Click to expand...
You can try different driver versions and see if it matters. Just make sure you use DDU
 
Troubleshooting is on point. Maybe HDMI cable (Can you check on linux/monitor that refresh rate is at 144hz) or window setting wrong monitor settings?
Some screenshots for comparison (running on Intel z690). I don't know if AMD ever fixed this or just my setup, but always had issues with performance overlays with AMD (rivaturner or even AMD's) so always have turned off:
1689036092246.png
1689036148762.png
1689036191893.png
1689037423627.png
 
Last edited:
Eymar said:
Troubleshooting is on point. Maybe HDMI cable (Can you check on linux/monitor that refresh rate is at 144hz) or window setting wrong monitor settings?
Some screenshots for comparison (running on Intel z690). I don't know if AMD ever fixed this or just my setup, but always had issues with performance overlays with AMD (rivaturner or even AMD's) so always have turned off:
View attachment 582572View attachment 582575View attachment 582577View attachment 582578
Click to expand...

I definitely was at 4k/144Hz in Ubuntu, I set it myself. I couldn’t set that if I didn’t install the newest Ryzen driver first.

I was running Ubuntu off an external drive- I’m wondering if it might be a PCIe conflict of some kind with my nvme drive.

I can do a fresh install of windows one more time and see what happens. Maybe I should install to an external drive.
 
NightReaver said:
How can it be AMD when other people run it just fine? That obviously points to something with your setup.
Click to expand...
Ram issue? I’ll yank a stick. It’s cheap ram that came with my board from MicroCenter.

Honestly I’m stumped.
 
I think I figured it out-

I reinstalled Win11 again and started installing one driver at a time running Shadow of the Tomb Raider benchmarks inbetween- upated windows to the latest version, everything good- and then my last two drivers in the device mgr-

Thunderbolt 4 and Realtek Bluetooth-

Installed the Intel Thunderbolt 4 driver set, within a minute- spontaneous reboot. Uninstalled drivers seems to be working ok at the moment. (I'll probably turn it off in the BIOS for now) I suspected Ubuntu didn't load the TB4 drivers for this board because it's on an AM5 setup.

I specifically purchased this board for the TB4 support for some music production work I was planning on doing. If this turns out to be the problem I guess I'll shoot a message over to Asus- probably needs another firmware update.

Well that was pretty difficult to narrow down. Thanks for all the help.

Edit- Damnit- still doing it even after I uninstalled the drivers. I guess I will disable in BIOS and then try another fresh install- pretty sure this has something to do with it or one of the Win11 22H2 updates.
 
Last edited:
You must log in or register to reply here.
Back
Top