https://bugs.kde.org/show_bug.cgi?id=404966

Sawyer Bergeron <sawyerberge...@gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|NOT A BUG                   |---
     Ever confirmed|0                           |1
             Status|RESOLVED                    |REOPENED

--- Comment #3 from Sawyer Bergeron <sawyerberge...@gmail.com> ---
Sorry, I don't know if I conveyed my experience entirely clearly.

I understand intel graphics are not super powerful, but what's going on here
suggests their raw compute power isn't at fault here:

1: the intel designed monitor for iGPU utilization never peaks above like
20-30%
2: the stutters match my experience with the FPS counter, I'll try other
measurement software to see if they match my experience with the built in
effect.
3: it is a hard drop, the window could be 2/3 of the screen minus one pixel and
everything will run at 60fps with no stuttering, basically never dropping below
55 fps. The moment 2/3 of the screen plus one pixel is using transparency the
FPS drops to 30 fps and never above that, nor below it. It is much more similar
to a cap than something genuinely loading the GPU. From every experience I've
had, if the GPU is genuinely having difficulty keeping up with a DE (GMA
graphics had a bit of trouble a couple years ago) then any drop in FPS is
gradual, so it will go from 60 fps down to 55, 50 and so on with increasing
load. That isn't the behavior here.
4: even when redrawing isn't required for the blurred region (moving a window
around in the region not covered by a window 2/3 screen size) stuttering is
evident, but so long as the transparent region is less than 2/3 screen area
that stuttering does not happen.


If I'm mistaken I apologize for reopening this, the numbers in question just
seem *very* odd if it is simply performance limitations from the iGPU. I'll try
to dig into the source for the transparency effect myself to see if anything
pops out at me in the meantime. It could be something within the Intel drivers
themselves, who knows.

-- 
You are receiving this mail because:
You are watching all bug changes.

Reply via email to