[Bug 93971] New: video framerate performance regression with UXA

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=93971

            Bug ID: 93971
           Summary: video framerate performance regression with UXA
           Product: DRI
           Version: XOrg git
          Hardware: Other
                OS: All
            Status: NEW
          Keywords: bisected
          Severity: normal
          Priority: medium
         Component: DRM/Intel
          Assignee: intel-gfx-bugs at lists.freedesktop.org
          Reporter: kevmitch at gmail.com
        QA Contact: intel-gfx-bugs at lists.freedesktop.org
                CC: intel-gfx-bugs at lists.freedesktop.org
     i915 platform: IVB

Created attachment 121469
  --> https://bugs.freedesktop.org/attachment.cgi?id=121469&action=edit
Xorg.0.log at 6db906ea8188cd7f731e5ef7391115ee8f5c17c8,

With UXA on my Lenovo Thinkpad T530 Ivy Bridge machine, playing video with mpv
(https://github.com/mpv-player/mpv) worked fine keeping up with my 60hz LVDS or
HDMI TV using the command

    mpv file.mkv --no-config --fs --vo=opengl:scale=spline36
--video-sync=display-vdrop --osd-msg1='${estimated-display-fps}' 

This is the bare minimum opengl output driver with a mid-quality spline36 luma
scaler implemented as an opengl shader. The added "--video-sync=display-vdrop
--osd-msg1='${estimated-display-fps}' " activates display sync and displays the
effective framerate on the screen.

At d909c460743a0cabc5c0ed30e063e99af7b26b17, it seems that vsyncs start to
happen about twice as fast ~120fps on the 60hz displays. Then, three commits
later at 6db906ea8188cd7f731e5ef7391115ee8f5c17c8, it appears that I am no
longer able to render frames fast enough for the refresh rate so that it
bounces around 50 +/- 5 fps on either display. The problem goes away if I
remove the scale=spline36 scaler so that it falls back to cheaper bilinear,
which is why I say this is a performance regression.  Similarly, the hardly
taxing GLXgears appears to render ~60fps just fine.

The problem also does not appear to exist for SNA, but I am unable to use that
mode due to other problems (#91066).

-- 
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the bug.
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20160202/dee9fad1/attachment.html>