VGA arbiter: removing RAC
Tiago Vignatti
vignatti at c3sl.ufpr.br
Wed Dec 19 14:45:53 PST 2007
Hi guys,
The Resource Access Control inside Xorg is the guy responsible to take
care of the various resources of memory and to share them in a wise
manner when two or more graphics devices are active (think multi-head).
As an alternative from RAC we can rely on VGA arbiter. So me and Paulo
Zanoni spent the last days implementing "the glue" of Xorg to use the
arbiter code with this purpose.
Now we're trying to do some experiments to see what's the performance
difference using the RAC and the arbiter. We thought in two tests: (a) a
multi-head X with RAC and other with the arbiter. And (b), test a
single-head X using the arbiter and other not using. In the (a), we can
evaluate if the usage of the arbiter overloads the X server and (b) if
there's some difference in use RAC or arbiter in a multi-head
environment. So we ran the x11perf tool and reported a very tiny
worthless gap in both tests (used with -comppixwin500, -move and
-rect500 options). All the experiments you can see here:
http://people.freedesktop.org/~vignatti/logs-between-rac-arb_19Dez2007.tar.gz
Am I missing something about how I did these results or everything seems
fine?
(I made the mistake to test this things using CFLAGS='-g3 -O0'. Anyone
knows if this will result in different conclusions between the two test
cases? I'm using the git upstream code of server-1.4-branch and nv driver)
Anyway, we will keep our working and the next challenge will be to think
about DRI, xv and GL.
my regards,
--
Tiago Vignatti
C3SL - Centro de Computação Científica e Software Livre
www.c3sl.ufpr.br
More information about the xorg
mailing list