!im dying over here
Ti guess 560 branch should be around mid next month?
H555 is in testing atm
Tbut i gotta say, i havent had any big issue with the 555 branch so far on arch
Ti mean yea, i had to put a few kernel parameters for it to work well, but thats about it
KFirefox is broken atm
Ti think thats due to the explicit sync patch on plasma, isnt it?
Tatleast thats what i heard
T after making sure i had these 3 parameters in my /etc/default/grub, i havent had any crash even on firefox on plasma 6.1 wayland on my 3080 with the 555 branch driver
Tthere were other recommended ones across a few forums, but they are not as important as these 3
KHuh?
KOh Nvidia, nvm.
MI'm curious on when will the modset flag no need inclusion. Nvidia's fbdev is a bit broken. It will occasionally just fail to work completely leaving the screen stuck with artifacts.
There's a lot of weirdness in them, but this is the first time that I actually believe they are testing on sway like they claimed since 515(?)
There's a lot of weirdness in them, but this is the first time that I actually believe they are testing on sway like they claimed since 515(?)
Tcould definitely be
Rlooks like the system76-scheduler is out, and bore+gamemode is in the nvidia-testing images. interesting.....
CGamemode isn't recommended, I think it's only there as a dependency for something
CIt's System76 Scheduler that's being dropped in favor of just running games as a regular program
KCan be used if you like
KJust not something I'm gonna focus on
CI believe you've been saying something to the effect of Gamemode forcing the performance profile leading to worsened performance, no?
CBecause of the CPU being pegged at higher power states instead of boosting opportunistically
KYep
CI believe that's not applicable when pstate is being used though
CAt least going from how ArchWiki describes those drivers
CWhen pstate is set to active it's either a general powersave or performance hint being sent
Hkinda, mozilla seemed to have fucked around and found out
https://github.com/NVIDIA/egl-wayland/issues/110
https://github.com/NVIDIA/egl-wayland/issues/110
GitHub
Reference: #104 (comment) While I don't use KDE I have similar issues with sway although the only application I encountered such crashes where firefox and thunderbird. The issue here is a race ...
LJust wanted to let the devs know, my fiancée is officially fed up with Windows in her ROG Ally. I am commencing the Bazzite installation.
LThis’ll be the 4th Bazzite powered device in this household.
Tah, i see. lets see when that bug gets fixed, but it isnt too much of an issue to begin with tbh
LI know this doesn’t compare to your collection though @HikariKnight
Hcant respond any other way
LI understand, I don’t blame you.
Hyoure catching up though
KDear god man. That's a lot of machines.
KOh wait those are just a lot of displays
Hnvidia test VM (with gpu passthrough), main machine, HTPC (stuck on F39 due to polaris atm but moving to F40 soon once some patches land)
on the go hobby work laptop, legion go, deck (running steamos)
2 penguin plushies my friends sent me as secret santa gifts as jokes (the sunglasses on the small one was a dlc that got sent to me last year, didnt know my secret santa gift came with a season pass)
bunch of controllers i use for testing (plus another 40 in a drawer behind the desk)
on the go hobby work laptop, legion go, deck (running steamos)
2 penguin plushies my friends sent me as secret santa gifts as jokes (the sunglasses on the small one was a dlc that got sent to me last year, didnt know my secret santa gift came with a season pass)
bunch of controllers i use for testing (plus another 40 in a drawer behind the desk)
Tmoved my dots and iso to my ventoy usb drive, wish me luck 
Hgood luck!
K...you have like 50 controllers?
Agnome merged accent colors?
K
Aam i dreaming?
Tdo you have experience in gpu passthrough when only nvidia gpus are in a pc?
Tsince im planning to make use of a second nvidia gpu in a windows vm, so it would be unpractical for me to unload the nvidia module fully
KWe've been blessed
Hthey need to be detected as different GPUs, if you have lets say 2x 3090s you need to add custom scripts tailored for your card into initramfs, this is not something we do
Ti see, i would have 2 4090s, exactly the same manufacturer
Tso i guess thatll be a bit harder?
Hyeah since you need to make a custom script to manually bind the correct card by pci address to vfio-pci at boot before nvidia driver grabs the card as vfio-pci only understands vendorid:deviceid
which would be identical for the 2 cards
which would be identical for the 2 cards
