I think the DirectX 12 (beta) option on the launcher seems to be the issue! I have no recollection of ever toggling that on, but with DirectX 11 toggled the game hasn’t crashed. Thank you very much for pointing that out.
I do now get flashing of weird past frozen frames on the screen. That is probably some weirdness with VRR/G-Sync/refresh rates or something else I can tune out. But the game isn’t crashing resulting in a loss of progress, so that is a major win.
Zen vs Linux kernel didn’t appear to make a difference. Also, it didn’t matter where in the game I was, it would hang and crash anywhere. Thanks!
EDIT: I played about an hour tonight. It runs like crap, but doesn’t crash. Relays, open maps are very hitch and go, and there is lots of screen flashes everywhere. Dropping resolution makes it worse, modifying refresh rates makes it worse. Disabling DLSS doesn’t make it better.
Warframe is working nicely with DirectX 11 toggled and module 550.67 on kernel 6.8.2, using gamescope and native resolution options launched through steam. Sometimes it freezes on opening after the launcher but it’s rare and could very well be a wayland/hyprland thing. Running it without gamescope works, but it flickers like crazy and it’s too hitch and go for me.
Without gamescope Warframe flickers like that, though for me it doesn’t seem to flicker to black as in your video- it flickers like it is sending frames in the wrong order or something. But definitely with the same regularity and intensity as your video. Also it could be that your video doesn’t capture the same as what you see in real time, of course.
Gamescope with these launch options through Steam fixes the flickering for me “gamescope -W 2560 -H 1440 -r 165 -f – %command%” (1440p resolution with 165 Hz screen, full screen).
Curiously, Deep Rock Galactic, No Man’s Sky, and other games do not flicker without using Gamescope.
Here is my example of Warframe flickering in wayland session:
This is without gamescope or anything like that, running under DirectX 11, launching Warframe directly from Steam. Toggled to DirectX 12 in the launcher the game runs smoother generally with less hitching but it still exhibits the major flicker every few seconds and then crashes with Xid 109 CTX Switch Timeout errors after playing 10+ minutes.
Thanks for the link! I will follow that post pertaining to debugging the glitches. The video is off-topic, but the same game (Warframe) does crash with “Xid 109 CTX Switch Timeout” when DirectX 11 is chosen in the launcher.
Wasteland 3 still crashing for me consistently. Other games in my catalog (to be fair, none of them listed in this thread) are working without issues.
In one weekend it crashed about 11 times… Less than a 48 hour window.
Edit: To clarify, as someone asked previously from the Nvidia staff, yes, it is running in Proton (a GE variant) via Lutris. However, I have tried multiple versions of dxvk, proton-ge, etc… Different configurations, to no avail.
I have another machine that I run WL3 on via Proton and there’s no issues.
The machine I am running it on that DOES have issues, however, is only having problems with WL3. Tried a re-install as well, but the issue persists.
I ran Everspace (natively and with proton, both) for several hours. No issues. No crashes.
Some further details… I think this is specific to Nvidia/Optimus, as that is what my ASUS laptop uses (1650, Intel). I ran WL3 on proton (GE variant) on my buddy’s AMD laptop and zero issues. No crashes.
Just happened again… Uploaded as much output as I was able to scrape together.
The issue happens completely at random. It’s not a specific save or event that is triggering this. I’m, so far, at about 40+ crashes since I’ve gotten this laptop and it only really seems to happen with Wasteland 3.
I’ve been facing this issue for months now with my RTX 2060 Mobile, mostly with newer or more GPU-intensive games.
Just upgraded to 555.42 and confirmed the issue still persists (at least with Diablo IV), in case anyone was wondering. Logs are attached to this post. It did last longer before it crashed - instead of crashing every 20 or so minutes, it had really long stutters every once in a while, but eventually crashed anyway.
I had compiled many bug reports in the past weeks on 550 drivers, but I’m not sure they’ll be of much use now. I can upload them if the developers deem relevant. I’ll collect reports on the newest driver and upload them here once I have some time. I can, however, confirm that I’ve faced Xid 109 errors on Baldur’s Gate 3, Persona 5 Royal, Elden Ring, Total War: Warhammer III, Diablo IV and, funny enough, The Last Spell, which is not GPU-intensive at all. All of them running on Proton, except Total War: Warhammer III.
I really hope this gets fixed for good soon. It has been really frustrating to be unable to use my GPU for the purpose I intended it for for months.
Is anyone from Nvidia still looking into this issue or aware of this thread? Don’t mean to be pedantic… Just seems like a handful of games were accounted for and the rest of the reports in here fell to the wayside.
Did a few tests on the newest beta drivers, 555.52.04, and the problem persists. I have attached the journalctl and nvidia-bug-report outputs after a crash in less than 5 minutes of running Elden Ring.
Current system is Arch LInux on kernel 6.9.3, and an RTX 2060 Mobile card with the latest 555.52.04 beta drivers. Can confirm Diablo IV still presents the same issue, but I’ve seen reports of memory leaking due to the game itself on Linux, so I haven’t generated any logs for that one. Also, curiously enough, the game The Last Spell has stopped giving me Xid109 errors, and runs smoothly now (at least for up to an hour of gameplay).
I hope this gets fixed before the very expected release of Elden Ring’s DLC in 15 days - or, at the very least, we get a development update on this issue -, but I’m not optimistic.
It looks like the nvidia library did not report the work done and went into waiting (__futex_abstimed_wait_common64). All threads are blocked, and some processor cores and GPUs are loaded at 100%.
