[Mplayer-dev-eng] x11 vs. dga

Andreas Ackermann (Acki) asackerm at stud.informatik.uni-erlangen.de
Mon Jun 18 00:18:11 CEST 2001


Hi Balazs,

now I've made the doublebuffering feature switchable from the commandline.
By default it is off (this unfortunately also disables OSD in the DGA
driver, sorry, but it would flicker like hell if I did it anyway !!!).

The X Server does a BUSY wait if I use the XDGASetViewport() call which
does the trick for doublebuffering. I always thought this would be handled
within a interrupt ... I was so sure, I never bothered to check this ...
:-(

Here are some speed values for ChickenRun, I've always been using this one
as a reference:

-bpp 24
DGA + Doublebuffering + OSD      99%
DGA - Doublebuffering - OSD      72%
X11 - OSD                        97%

The percentage values are the sum of CPU usages of X11 and mplayer summed
up during the same section of the movie.

The X-Server was in all cases running at 32 bits per pixel. For DGA ouput
I used -bpp 24 which effectively means a conversion from 24 bpp to 32
bpp. For convenience the above values with -bpp 32: (the X11 driver
ignores the -bpp switch, so its output is listed only once ...)
Graphics card: riva128 (nv.o)
Processor:     AMD K6-II+ @525MHz

-bpp 32
DGA + Doublebuffering + OSD      99%
DGA - Doublebuffering - OSD      78%

Now for -bpp 15
DGA + Doublebuffering + OSD      78%
DGA - Doublebuffering - OSD      52%


I don't quite understand the *BIG* differnce between -bpp 24 and -bpp 15
but the result reproduces (at least with top).

I *believe* speed penalty for doublebuffering depends on the selected
refresh rate - the higher the vertical refresh rate is, the less time is
needed on the average to wait for the vertical blank.
However higher refresh rates increase transfer volume on graphics memory
and thus slow down the copying ...

-Andreas
-------------------------------------------------------------------
        http://www.acki-netz.de  email: acki at acki-netz.de
   //   (+49)[0]9131/409500 or (+49)[0]9286/6399
 \X/    acki or acki2 on #rommelwood

   No trees were killed in the sending of this message. However
     a large number of electrons were terribly inconvenienced.

On Fri, 8 Jun 2001, Balazs Tibor wrote:

> Hi!
>
> I'm confused
>
> -----------
> CommandLine: '-v' '-osdlevel' '0' '-vo' 'dga' '/mnt/hdc4/home/tibcu/MPLAYER/405_DivX.avi'
>
> A: 150.0 ( 150.0)  V:  89.1  A-V: 60.915 ct:  8.772  2674  49% 46% 39.5%
> 2625acm
>  converted 372 -> 8192
> A: 150.0 ( 150.1)  V:  89.1  A-V: 60.922 ct:  8.776  2675  49% 46% 39.5%
> 2626acm
>  converted 372 -> 8192
> A: 150.1 ( 150.1)  V:  89.1  A-V: 60.962 ct:  8.779  2676  49% 46% 39.5%
> 2627acm
>  converted 372 -> 8192
> A: 150.1 ( 150.1)  V:  89.2  A-V: 60.969 ct:  8.782  2677  49% 46% 39.5%
> 2628acm
>  converted 372 -> 8192
> ------------
>
> ------------
> CommandLine: '-v' '-osdlevel' '0' '-vo' 'x11' '/mnt/hdc4/home/tibcu/MPLAYER/405_DivX.avi'
>
> A: 150.0 ( 150.0)  V: 134.3  A-V: 15.699 ct: 12.322  4032  49%  9% 26.0%
> 3916A:
> 150.0 ( 150.0)  V: 134.4  A-V: 15.689 ct: 12.325  4033  49%  9% 26.0%
> 3917acm co
> nverted 372 -> 8192
> A: 150.0 ( 150.1)  V: 134.4  A-V: 15.677 ct: 12.328  4034  49%  9% 26.0%
> 3918acm
>  converted 372 -> 8192
> A: 150.1 ( 150.1)  V: 134.4  A-V: 15.700 ct: 12.332  4035  49%  9% 26.0%
> 3919A:
> 150.1 ( 150.1)  V: 134.5  A-V: 15.690 ct: 12.335  4036  49%  9% 26.0%
> 3920acm co
> nverted 372 -> 8192
> A: 150.1 ( 150.2)  V: 134.5  A-V: 15.681 ct: 12.338  4037  49%  9% 26.0%
> 3921acm
>  converted 372 -> 8192
> ------------
>
> WHY is dga output sow slow?!? and why does it influence audio out? (39.5%
> compared to 26%)
>
> If someone is interested (maybe Acki? Acki, have you come back from
> vacation?), I can upload all outputs to thot (5 Mb without
> gzipping them), including -framedrop, -hardframedrop and
> -nodshow -hardframedrop outputs too
>
> bye
>                                 -=+=- Tibcu -=+=-
>
>
> -framedrop (at A: 150.0) last number:
> 827 with dga
> 495 with x11
>
> -hardframedrop (at A: 150.0) last number:
> 1206 with dga
> 617  with x11
>
> -nodshow -hardframedrop (at A: 150.0) last number:
> 709 with dga
> 529 with x11
>
>
> MPlayer 0.17cvs-010605-15:48  (C) 2000-2001 Arpad Gereoffy <arpi at thot.banki.hu>
>
> was used
>
>
> _______________________________________________
> Mplayer-dev-eng mailing list
> Mplayer-dev-eng at lists.sourceforge.net
> http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng
>



_______________________________________________
Mplayer-dev-eng mailing list
Mplayer-dev-eng at lists.sourceforge.net
http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng



More information about the MPlayer-dev-eng mailing list