When playing The Witcher 3 (from Steam) on a Kabylake Core i7-8650U on a Dell notebook with a UHD620 I get a inconsistent mesh flickering. It occurs in some places only, and is repeatable. But the flickering is very random. I did a video recording of the issue here: https://drive.google.com/file/d/1Y8TBaMtc-dP4N-GRuHcR3OZzx9JaiODg/view This is using DXVK 0.94 bundled with Proton 3.16-6. (and 0.96 too) The Witcher 3 from Steam, 1280x720 all lowest except for texture resolution which is medium. System information: GPU: Intel UHD620 Driver: AMV from mesa git Wine version: Wine 3.16 bundles with Proton 3.16-6 DXVK version: DXVK 0.94 bundled with Proton 3.16-6 and DXVK 0.96 OS: Gentoo Linux, 4.20.4, KDE5 desktop with compositing disabled. Danylo (werman) reproduced it at https://github.com/doitsujin/dxvk/issues/876#issuecomment-458549675 As requested, a bug to track the issue.
The longer the game running the more apparent flickering is and on short traces it is very hard to see it and longer trace, well, they will take much longer to take. So for now I don't have a meaningful trace. But still will investigate the issue. I have: HD 620 Mesa Git Proton 3.16-6 Arch Linux, 4.20.6
hi guys. I also can reproduce this issue on this configuration: i3-8100 GPU - Intel UHD Graphics 630 mesa - tested on latest master and release 18.0.0 (and 18.3.2, system) OS - manjaro Proton 3.16-6 Link to the apitrace (trace was maiden under wine): https://drive.google.com/open?id=1oXmy_WAkJRpKaQAbyFL6J6n6qRsME1M8 What I can say - replaying this apitrace always reproduces flickering (exactly that one, when textures disappear and I see trees and sky (inside the cave) ). I tested this trace on KBL (same configuration with Danylo), and can say that flickering is less, and not stable, as on CFL machine. Hope this information may help to investigate the issue.
Thanks. I built latest GIT and tested. Seems all good :-)
Sorry, wrong issue :-(
I think that this bug and https://bugs.freedesktop.org/show_bug.cgi?id=109630 (vkQuake flickering geometry under Intel) may be connected but still don't know what causes it and how to debug this.
Agreed, it seems to be exactly the same. Some of the geometry disappearing for a frame every so often.
commit bc612536eb2f0f77725103e53077ad5a28036ac3 (HEAD -> master, origin/master, origin/HEAD) Author: Jason Ekstrand <jason@jlekstrand.net> Date: Wed Jul 31 10:42:24 2019 -0500 anv: Emit a dummy MEDIA_VFE_STATE before switching from GPGPU to 3D There is an object-level preemption workaround which requires this. However, even without object-level preemption, we seem to have issues with geometry flickering when 3D and compute are combined in the same batch and this appears to fix it. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=109630 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=111267 Cc: mesa-stable@lists.freedesktop.org Reviewed-by: Kenneth Graunke <kenneth@whitecape.org> Tested mesa-master with the commit on my CFL (checked on attached apitrace). Works without flickers now.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.