On Sun, 29 Nov 2020 23:15:53 +0000 juanma1980--- via enlightenment-users <enlightenment-users@lists.sourceforge.net> said:
> Hi. Certainly I didn't give much detail, I only want to know if anyone else > has been having issues (As I compile from git I assume that there can be > issues sometimes so I think that this mail list could be a good place for ask > about). no issues here - but also no nvidia. no nouveau. no hybrid graphics. my plain intel systems are fine - my amd gpu desktop system is fine. > I'm using opengl, in a i7 laptop with intel+nvidia (nouveau). I compile E > directly from git, this issue began 2-3 days ago, when I update the system. > Currently my E version is 0.24.99.24339, Kernel 5.9.11, Efl also from git, > commit b59b60502122b505d20e91ed4dbddcee3abc9ce5, and Xorg 1.20.8 perhaps... the issue is in the os update, not efl or e? i'm instantly suspicious of optimus/hybrid systems. while in theory they can work, people have had no end of trouble with them. i wouldn't buy or touch such a machine at all - not worth the pain. but nothing changed with respect to the pager or even rendering of miniature windows. i ADDED a new kind of miniature window that includes frame etc. for the new winlist thing - pager doesn't use it. nothing changed with the background. the only thing it might be is last friday i dropped a short shader addition that adds dithering. in theory there could be some horrible side-effects if your shader compiler is horribly broken. i have my doubts it is - at least it's working fine without issues on my intel gfx boxes, amd, raspberry pi, and even my panfrost devices. dithering isnt applied to just pager - it applies to everything. b59b60502122b505d20e91ed4dbddcee3abc9ce5 is the commit hash in efl. it should not be expensive unless again - pathological case in shader compiler generating really bad code. if it's this - simply revert that patch locally in efl and re-build (you can git reset --hard origin to reset your revert later if it makes no difference). > The first thing that caught my attention was that pager wasn't rendering my > desktop background, only the windows (i can attach a image showing the lack > of the desktop background image at pager). Perhaps an evas related problem? background is a thumbnail of the wallpaper - it's actually pre-rendered at lower resolution and even saved on disk. nothing has changed here for a very long time. > Also I can attach meson_options from both enlightenment and efl if that helps > or any other thing/test/whatever > > p.s. Wow, rasterman? Let me say I'm amazed, I admire your work since my > first contact with E16, 20 years ago aprox. Great job. thanks :) > === In soft we trust === > > 29 de noviembre de 2020 13:44, "Carsten Haitzler" <ras...@rasterman.com> > escribió: > > > On Sun, 29 Nov 2020 11:09:30 +0000 juanma1980--- via enlightenment-users > > <enlightenment-users@lists.sourceforge.net> said: > > > >> Hi all. I'm from Valencia (Spain) and a long time E and slackware user. > >> Recently I've faced a problem with pager module that is spending cpu as > >> there's no morning on earth. I've two options: buy a nuclear reactor to > >> power my computer or subscribe to this list and ask if anyone has this > >> problem too. p.s. english it's not my native language, apologies for the > >> mistakes. > > > > that isn't very detailed or useful.... the pager just renders a scaled down > > copy of all your windows - the way it works (well it plus evas) is just > > windows that update will be re-drawn plus some small space around (rounded > > up to tiles). pager will not be updating anything unlesss something on > > screen is. rendering updates does involve sampling quite a reasonable > > amount of memory for the pixmap sources. if you are using gpu (gl) then it > > should manage this without any trouble at all. But given hos GPUs work it's > > not much of a cost to keep "smooth scaling of window content" on and it > > looks far better. If you are software compositing, then "smooth scaling of > > window content" under composite settings -> rendering will be a lot more > > expensive. it makes the cpu sample the entire source window content and > > calculate a full super-sampled downscaled version. if you start e with > > software rendering in the wizard, smooth scaling is off by default. > > > > first - are you using software or gl? believe it or not i and man other > > users i have heard from have accidentally switched to software and didn't > > notice for days, weeks or even months, then they have some performance > > bottlenecks in some places, notice a lot of cpu being used by enlightenment > > and find they are not using GL. so it's happened even to me. second - if > > you can't use gl... try turning off smooth scaling under composite settings > > -> rendering > > > > -- > > ------------- Codito, ergo sum - "I code, therefore I am" -------------- > > Carsten Haitzler - ras...@rasterman.com > > > _______________________________________________ > enlightenment-users mailing list > enlightenment-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/enlightenment-users -- ------------- Codito, ergo sum - "I code, therefore I am" -------------- Carsten Haitzler - ras...@rasterman.com _______________________________________________ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
