plain hound
I was spectating a modded multiplayer COOP vs Ai game. I joined at the 15 minute mark and caught up by around 18~19 minutes and crashed maybe by 30 minutes?? The framerate was good, my system didn't seem to be lagging, there were no lobby notices of anyone lagging. It was a surprise crash.
sorry if it's a duplicate upload https://log.beyondallreason.dev/logs-20251114T0731-dOzRR99f.zip
somber nebula
AMD PLEASE FIX THIS SHIT ALREADY AAAAAAAAAAAAA
nocturne hound
I have a 7900 GRE. There are a lot of things you can do to contribute to an overall "resolution" of these issues, I will list some here:
The driver rollback and downgrade worked for me. Also, don't use the garbage software that AMD bundles with their products, you just want to install the drivers. Let me know if that helped. -Statistic9166
@somber nebula @plain hound
plain hound
I have a 7900 GRE. There are a lot of things you can do to contribute to an ove...
I've been playing BAR on and off the last two to three weeks and this was the first driver hang I've seen. Last time I really played was late spring and I had full on shut-downs. Other games are just plain making the system shut off. Makes me think it's tripping overcurrent protection within the card. Though a lot of it is just plain heat related.
I think AMDs power profiles are just jacked or compromised components were used by manufacturers. In my case, it's a factory OCd Gigabyte card. I probably need to re-do the thermal paste but it's something I don't have time for and can't afford to screw up atm. So GPU fan override to max and disable AA is my workaround, gets me closer to stable in addition to manually adjusted quality settings. Undervolting seems to help a little too.
I've done the DDU thing countless times. I even jump through other hoops to use Windows own mechanisms to delete the drivers so nothing gets auto-installed. It still crashed.
(I've got an imgur album of the workflow is anyone wants 😂)
Only thing I haven't tried in a while is to disable Resize Bar. though a couple times I had to reset my bios and I had more crashing on stock bios settings than I did with XMP and resize bar.
nocturne hound
Have you run a memtest? Check the Windows event manager application, it gives good information regarding shut downs. Part of the problem I was having was related to the overclocking I was doing, and it was identified in the event manager logs. Another thing that I've heard can cause these issues is a 6+2 power cable being used instead of a full 8. Don't ask me why that could contribute to the issue because I don't know 😛
I'm guessing your power supply is up to snuff? I was about to swapnostics mine out before I checked event manager logs. I'm running on Debian now, much more pleasant gaming experience if I'm being honest. And the AMD GPU drivers are included in the latest kernel, so it's ready to go out of the box.
plain hound
I thought I had run Windows Memtest and Whatever the latest standalone Memtest86 clone there was at the time, when I jumped from 64gb to 128gb and it was fine. Only thing I can think of is maybe a bios update broke compatibility in some unforeseen way? I've seen comments floating around that Memtest is not entirely perfect, in part because the control schemes in DDR have gotten more and more complex since memtest came out, if I'm even remembering that right, so there can still be outliers who test fine but are plagued by infrequent random errors.
What was it in Event Viewer that you were seeing?
A friend of mine suggested BluescreenView and to check Event Viewer. The problem is whatever is causing the black screen shutdown is NOT the result of a bluescreen. Event viewer just shows generic power related messaging and searching the web for what I saw was a waste of time. There was nothing that looked valid enough to be worth the extended dive into the rabbit hole.
I bought a new Seasonic Vertex PX-1200w power supply this past February, which did replace a Seasonic SS-860XP that was older than I thought when I did the swap. Yet I saw ZERO change in overall system stability. The power tables are damn near identical except for the 12v on the newer one being 1200w instead of 852w. There was a minor downgrade on the newer model's -12v rail only being 0.3A instead of 0.5 on the old.
So either my old PSU was getting weaker in a way that the GPU's components couldn't cope with the sag in supply or AMD's recommended PSU minimum is absolute bullshit and these cards need 1500w to handle draw spikes.. 🤦♂️
nocturne hound
Sorry for the late reply.
Memtest isn't the end-all be-all of memory verification, I absolutely agree. But it is at least one tool we have that can be run easily enough to find (hopefully) obvious flaws in the actual performance of the hardware. I assume you're original RAM and the 64GB you added are 100% matched up in specs/manufacturer? Have you tried using 3DMark to load test your system?
In Event Manager (this was well over a year ago that I was actively combing through these logs after a crash) I would see messages related to 'power supply interrupted'. These issues occurred after I started modifying the memory and CPU/GPU overclock values. Removing the XMP profiles in the BIOS and eliminating overclocking fixed them, permanently. These were also normally associated with blackout/shutdown events mostly, not display resets.
Might be able to offer additional insight (from a blind man) if you wanted to provide your system specs and driver info.
hallow idol