Driver version 530.41.03 causes several issues which makes gaming or working with 3D tools completly impossible. elite dangerous starts normal, but after being ingame for ~10sec. some weird gfx glitches with 3D objects are showing. if i continue the game even the hud elements are glitching.
in csgo (opengl) if my character (or the enemy) goes into a area with shadows the character gets completly black.
the same issues happen in many other games and also tools like blender, davinci resolve etc.
the only way for me to fix this, was downgrading the driver to a previous version.
this will also happen on other pc’s. ryzen3600-desktop with a gtx1080,i7-laptop with a gtx1060, i5-laptop with a gtx1050,
I am also having these issues in Arch Linux under the 530.41.03, specifically in Elite Dangerous. It is unplayable with this driver installed due to graphical glitches.
Yeah, something isn’t right with recent driver releases. On both my GTX 960 and 1080 I get in-game models flickering in and out of existence in Metro: Redux as well. Halo MCC has graphical glitching where model vertices seem to stretch for like a few hundred milliseconds or something.
I thought my 1080 was/is dying since I started experiencing a bunch of weird issues all at once but seeing the reports of people having issues even on Windows makes me think it’s a driver issue. FWIW, I get XID 8 under Linux when running superposition which according to Nvidia’s XID page isn’t hardware related. I get the crashes on Windows as well but applying a negative clock fixes it…?
Not trying to hijack the thread with unrelated BS, just saying that there is something going on with recent drivers and the lack of any communication from Nvidia is frustrating. People might be throwing out perfectly fine hardware thinking it’s dying.
Edit: just going to point out that the Taiga level in Metro Exodus is far worse. It’s so bad you literally cannot play using DX12.
I then switched BIOS to the performance one(forgot I was on stock), increased the power draw to 280, and then applied a +1000Mhz OC to the memory. Still nothing.
Also ran memtestG80:
Final error count after 50 iterations over 7000 MiB of GPU memory: 0 errors
Again, with 280 watt power limit and +1000mhz memory OC.
I don’t want to hijack the thread any further so if there are any doubt left on the GPU being faulty i’ll move the discussion to a different thread. I have screenshots of clocks too. Boosts to 2025 as it always has.
Side note: CUDA iterates the GPUs in reverse. Kinda weird.
Made a Playthough of RE7. Game seems fine at first, but after ~30min into the game, the lightning is slightly off. Which doesnt happens on 525.89.
I’m currently playing RE8 also had weird lightning after some playing, but more annoying was getting constant Image Freeze (all displays where on freeze, not just the game) but audio was still running normal, this happens again after some playing (10-30min).
Went again back to 525.89 and issues are gone
any ETA when we finally get a fix or rollback for this completly broken driver ?
i’m currently on version 525.89 which is not compatible with the new 6.3 kernel my distro dropped.
so downgrading isnt a option for too long, hurry plz up with your decicion.
We have already filed a bug 4127336 internally for the tracking purpose and do have local repro as well.
Team is investigating issue on the priority, shall keep posted on it.
Sorry for the late response, had some stuff to deal with.
I done a lot of game testing (csgo, cs2, elite dangerous, no mans sky, monster hunter world, quake 1 & 2.
All with the same result. regardless which kernel i choose, every game is experiencing microstutter to an extend, which makes it unplayable.
tested opengl & vulkan.
the only options are currently, 1.stick with driver version 525.xx or 2. use other manufacterer.
@Pad.Trick
I played game CSGO and Elite dangerous with driver 535.86.05 for almost an hour but did not observe any stutter.
I would request you to share video highlighting the same so that I can focus on the same point for local repro.