Question RX 7900 XT Crashes - Help!

Cableman

Member
Dec 6, 2017
78
73
91
I recently replaced my GTX 1070 with a RX 7900 XT. Finally had time to install it today. Used DDU to clear the old drivers. Ran Heaven a few times without any problems. Then started a few games to check the performance. Tried Control. Within 5 mins, I get a black screen and the PC crashes. It started doing the same in Witcher 3. Then I tried Heaven again and was running GPU-Z to monitor the temperature. It never goes beyond 76-77 degrees. After a few minutes, I get the same black screen and crash (at 75 degrees). No other apps are running.

Any ideas? I would appreciate any help.
 

In2Photos

Golden Member
Mar 21, 2007
1,629
1,651
136
What does your CPU utilization look like during these games? GPU? What does the GPU voltage look like? Maybe a screenshot of HWInfo or something similar.
 

fralexandr

Platinum Member
Apr 26, 2007
2,244
188
106
www.flickr.com
I'm pretty sure it's the seasonic psu. Some older seasonics are known not to handle the transient spikes from the nvidia 3xxx/4xxx and amd 6xxxx/7xxxx.

 
Last edited:

Cableman

Member
Dec 6, 2017
78
73
91
I hope it's not the PSU :( By the way, it's Seasonic Prime Titanium 750w purchased in mid-2020.

This whole thing makes no sense to me. I ran Heaven 5 times without any issues. Played some Witcher 3 and Doom on Steam. No issues. Then I decided to download the Epic launcher to try Control (got it for free there). And since I did that, even Heaven crashes after 2-3 mins at the benchmark and I can no longer run anything without crashes. Wouldn't everything be crashing since the start? Could this be some weird software thing? I just don't get how everything can be flawless and then out of a sudden nothing is running...
 
  • Like
Reactions: Kaluan
Jul 27, 2020
16,340
10,352
106
I hope it's not the PSU :( By the way, it's Seasonic Prime Titanium 750w purchased in mid-2020.
Seasonic doesn't make infallible PSUs. I have this: https://www.techpowerup.com/review/seasonic-prime-fanless-700-w/11.html

See all the cons? The PSU is NIB and I don't intend to use it. I would rather sell it to someone who cares more about silence than using this POC to run a gaming rig.

What is your GPU's junction temperature? It's possible that something broke in the PSU from the stress of powering the GPU.
 
  • Like
Reactions: Leeea

Cableman

Member
Dec 6, 2017
78
73
91
GPU Memory Junction temperature was 82 degrees just before crashing. I can still use the computer, Windows is working, standard office tasks are fine. Crashes only when utilizing the GPU. Is there any way to test the PSU?
 
Jul 27, 2020
16,340
10,352
106
Is there any way to test the PSU?
This: https://www.amazon.com/Computer-PC-Tester-Connectors-Enclosure/dp/B076CLNPPK

See if the reported values fall within norms.

But I don't think that can be used to test the PSU's transient spikes handling capability.

I ran a 3090 just fine with a relatively unknown branded 650W PSU so you could try picking up a cheap PSU. Even if your Seasonic doesn't turn out to be the culprit, it's always good to have a spare PSU around.
 

Cableman

Member
Dec 6, 2017
78
73
91
After getting the black screen, I hear the audio for a few seconds before it crashes. I keep trying software solutions, but nothing helps so far.

I can put the old GPU in to run Heaven again, but I doubt it will produce enough power to serve as a test for the PSU.
 
Jul 27, 2020
16,340
10,352
106
I can put the old GPU in to run Heaven again, but I doubt it will produce enough power to serve as a test for the PSU.
It should still be helpful to see if something broke in your PSU. Try running Furmark and Unigine together on the old GPU.

Also, if you can download the UE5 Matrix demo, that's an even better test. It helped me to identify a bad 3090 I got.
 
  • Like
Reactions: Cableman

Leeea

Diamond Member
Apr 3, 2020
3,626
5,368
136
I have a Seasonic Prime 750w. Updated the BIOS before swapping cards. It's an old system, but I've been updating it over time...
in the driver reduce the power usage on the GPU as much as you can.

performance -> tuning -> power consumption

see if the -10% or whatever it lets you do fixes anything.

----------------------------------------

if that does not work, perhaps flip vsync on, that will reduce power usage also

----------------------------------------

another way to reduce power usage is thermal throttling it.

Setup a custom fan profile in the performance tuning page that provides completely inadequate cooling.

It will get to hot, thermal throttle, and use far less power. This does not violate the warranty, and for testing will help you determine if you have a PSU issue.

It should not crash from overheat.

----------------------------------------

All else fails, RMA IT!

Why it is not functioning is less important then having a functioning GPU. If your system does not have the PSU for it the card is useless to you.
 
Last edited:

Cableman

Member
Dec 6, 2017
78
73
91
It should still be helpful to see if something broke in your PSU. Try running Furmark and Unigine together on the old GPU.

Also, if you can download the UE5 Matrix demo, that's an even better test. It helped me to identify a bad 3090 I got.
I'll try that. You got a bad 3090? Could my issue be with the 7900xt?
 

In2Photos

Golden Member
Mar 21, 2007
1,629
1,651
136
I hope it's not the PSU :( By the way, it's Seasonic Prime Titanium 750w purchased in mid-2020.

This whole thing makes no sense to me. I ran Heaven 5 times without any issues. Played some Witcher 3 and Doom on Steam. No issues. Then I decided to download the Epic launcher to try Control (got it for free there). And since I did that, even Heaven crashes after 2-3 mins at the benchmark and I can no longer run anything without crashes. Wouldn't everything be crashing since the start? Could this be some weird software thing? I just don't get how everything can be flawless and then out of a sudden nothing is running...
This is why I brought up the BIOS earlier. My son's PC ran fine during stress testing but then games started crashing. The BIOS update fixed his PC. Heaven is a constant load on the GPU. Games fluctuate. Even though heaven may appear to be a larger load it doesn't represent everything the system will see.

I'd try another PSU to see if the problem persists.
 

blckgrffn

Diamond Member
May 1, 2003
9,128
3,069
136
www.teamjuchems.com
The post 2020 GPUs have much larger spikes than older GPUs. I have an A tier BitFenix 650w that got exposed by 6800xt and higher GPUs for not having the rail strength.

Don’t keep gambling the GPU and the whole rest of the rig on a PSU that’s probably not cutting.

Control with RT is going to likely pull power as a worst case scenario in a spikey manner. If you pull the power limits way back and it works, even better.

Get a 750W or better - better is good - quality unit that been built with these new generation of GPUs in mind.
 

Cableman

Member
Dec 6, 2017
78
73
91
The post 2020 GPUs have much larger spikes than older GPUs. I have an A tier BitFenix 650w that got exposed by 6800xt and higher GPUs for not having the rail strength.

Don’t keep gambling the GPU and the whole rest of the rig on a PSU that’s probably not cutting.

Control with RT is going to likely pull power as a worst case scenario in a spikey manner. If you pull the power limits way back and it works, even better.

Get a 750W or better - better is good - quality unit that been built with these new generation of GPUs in mind.

It was Control with RT that started this whole mess...

I'll look into newer PSUs. Any recommendations besides the ones above?
 

Cableman

Member
Dec 6, 2017
78
73
91
in the driver reduce the power usage on the GPU as much as you can.

performance -> tuning -> power consumption

see if the -10% or whatever it lets you do fixes anything.

----------------------------------------

if that does not work, perhaps flip vsync on, that will reduce power usage also

----------------------------------------

another way to reduce power usage is thermal throttling it.

Setup a custom fan profile in the performance tuning page that provides completely inadequate cooling.

It will get to hot, thermal throttle, and use far less power. This does not violate the warranty, and for testing will help you determine if you have a PSU issue.

It should not crash from overheat.

----------------------------------------

All else fails, RMA IT!

Why it is not functioning is less important then having a functioning GPU. If your system does not have the PSU for it the card is useless to you.

OK, so I tried all of the above.

Set the power limit to -10%. Crashed.

Lowered the fan speed to the minimum %. It didn't go past 73 degrees before crashing so it didn't even thermal throttle before crashing.