Version: unspecified (using KDE 4.7.0) OS: Linux Please implement GL_EXT_x11_sync_object support to reduce tearing with nVidia proprietary driver. http://www.phoronix.com/vr.php?view=OTU0NQ Reproducible: Didn't try Expected Results: No tearing with nVidia proprietary driver.
Patches are welcome
Would this solve the always-present jerking windows mode in KDE with NVIDIA? Tearing, slow window scrolling, *very* slow window switching, lagging when disk is being accessed - why the window manager slows down to a crawl when the disk is being used? - etc...
(In reply to comment #2) > Would this solve the always-present jerking windows mode in KDE with NVIDIA? No, and it will unlikely aid on avoiding tearing either. - this object is to glXWaitX() cross dpy's (what's probably necessary because the compositor operates on another dpy than the composited client - will test that) but will NO WAY cover: - tearing while moving windows - pixmap allocation timeouts UNLIKELY cover: - tearing while playing videos POSSIBLY cover: - tearing while scrolling > Tearing try to turn off twinview (in /etc/X11/xorg.conf!) > slow window scrolling with v'sync? try: https://git.reviewboard.kde.org/r/104247/ (will not cleanly apply to current master, gonna push update tonight or tomorrow) > *very* slow window switching likely oxygen+nvidia issue: "oxygen-setting"/"window decorations"/animations/ uncheck "window active state change transitions" > lagging when disk is being accessed - why the window manager slows down to a crawl when the disk is being used? no idea, but it's related to the GL compositor, probably GL_texture_from_pixmap driven and also likely rather I/O than HDD related Last note: Please do NOT turn this bug into some OT rant where you express and investigate on various loosely related issues If you want to say something, approach forum.kde.org, the kwin mailing list or -if you must use BKO- open a new "bug" Hijacking other ppls bugs is not very nice. Thanks.
Hi there, I'm really sorry if I hijacked the thread and I apologize for the rant. It was really not my intention. I was just wondering if this feature would help to solve these problems, as something similar about GL_EXT_x11_sync_object was mentioned on this bug report: https://bugs.kde.org/show_bug.cgi?id=258971 Thanks for your patience.
don't worry - i was just telling (to prevent what happens right now - we spam this bug with our communication ;-)
Feature request renamed because GLX_EXT_buffer_age is more likely help here. Looks like work on this is already going on :) https://git.reviewboard.kde.org/r/107198/
Sorry for up-duping. *** This bug has been marked as a duplicate of bug 319995 ***