Discussion: the ati/radeon open source drivers & radeon repository

ChemBro · 2010-06-07 20:52:33

megawebmaster wrote:

Huh, Xorg server 1.9 RC1 is going to be released today and 1.8 still didn't hit extra... How could it be?

Showstopper bugs for Arch Linux.

roughL · 2010-06-10 13:39:56

regnak90 wrote:

I was wondering if it is possible to maintain video aspect ratio with this driver ? I am trying to play Eschalon Book I fullscreen but the game screen is stretched.

You could try to use the radeontool:

usage: radeontool [options] [command]
         --debug            - show a little debug info
         --skip=1           - use the second radeon card
         dac [on|off]       - power down the external video outputs (on)
         light [on|off]     - power down the backlight (off)
         stretch [on|off|vert|horiz|auto|manual]   - stretching for resolution mismatch 
         regs               - show a listing of some random registers
         regmatch <pattern> - show registers matching wildcard pattern
         regset <pattern> <value> - set registers matching wildcard pattern

Don-DiZzLe · 2010-06-10 18:11:27

Is radeon_ucode still necessary like it says on the first post cuz it kinda conflicts wit linux-firmware?

I have followed the catalyst guide from the wiki to install the fglrx driver and it is working fine altho everything is so damn slow now.
I've recently changed my geforce 7600gt for ai radeon 3650. Altho the card is much faster dan my old one the driver performance is horrible.
Playin games is extremely laggy.

So i'ma swap to the radeon see if that will help the performance.

I am on xorg 1.8 and the latest test kernel 2.6.34 (are the kernels on the first page really necessary?) and I've compiled these packages:

dri2proto-git-20100610-1-i686.pkg.tar.xz
glproto-git-20100610-1-i686.pkg.tar.xz
libdrm-git-20100610-1-i686.pkg.tar.xz
libgl-git-20100610-1-i686.pkg.tar.xz
mesa-demos-git-20100610-1-i686.pkg.tar.xz
mesa-r600-r700-git-20100610-1-i686.pkg.tar.xz
xf86-video-ati-git-20100610-1-i686.pkg.tar.xz

So by installing these pkgs and changing my xorg.conf to:

Section "ServerLayout"
        Identifier     "X.org Configured"
        Screen      0  "aticonfig-Screen[0]-0" 0 0
EndSection

Section "Device"
        Identifier  "aticonfig-Device[0]-0"
        Driver      "radeon"
EndSection


Section "Screen"
        Identifier "aticonfig-Screen[0]-0"
        Device     "aticonfig-Device[0]-0"
        DefaultDepth     24
EndSection

should enable 3D acceleration?

kfgz · 2010-06-10 18:46:25

Don-DiZzLe wrote:

Is radeon_ucode still necessary like it says on the first post cuz it kinda conflicts wit linux-firmware?

It's not necessary if you have linux-firmware installed. Yes, it conflicts with radeon_ucode.

So by installing these pkgs and changing my xorg.conf to:
...
should enable 3D acceleration?

You don't need git packages to enable 3D accel. You also don't need packages from testing.

Don-DiZzLe · 2010-06-10 19:10:16

Hmm, then what's the point of this howto?

kfgz · 2010-06-10 19:29:03

Read last few pages and you'll know

ijanos · 2010-06-10 20:17:14

By the way, the voltage contorl starting to kick in, my card is down to 57°C which is pretty close to the windows driver's 52°C

Perry3D · 2010-06-15 08:59:47

The r300-r500 gallium driver make some good progress: http://tirdc.livejournal.com/27350.html
Is it possible to compile mesa with r300 gallium and r600 classic?

regnak90 · 2010-06-15 09:00:18

Is caster working for anyone ? The game crash as soon as 3D is needed and dmesg is full of : radeon 0000:01:00.0: vbo resource seems too big for the bo.
This game only need OpenGL 1.1 so it should work.

endlessroad1991 · 2010-06-15 10:30:24

@Perry3D:
In mesa-r600-r700-git PKGBUILD, "--with-dri-drivers=swrast,r600 \" tells us which drivers should be built.
you can add r300g to this line.
In the future, you can replace r600 with r600g.

Perry3D · 2010-06-15 10:35:19

Thanks,
i thought there was a configure paramater like --enable-radeon-gallium. But that is a nice solution.
At the moment i build mesa-full with --with-dri-drivers=swrast,radeon,r200,r300,r600. Maybe i will switch r300 to gallium.

lordheavy · 2010-06-15 14:28:08

Perry3D wrote:

Thanks,
i thought there was a configure paramater like --enable-radeon-gallium. But that is a nice solution.
At the moment i build mesa-full with --with-dri-drivers=swrast,radeon,r200,r300,r600. Maybe i will switch r300 to gallium.

The correct command line is (i guess)

./autogen.sh --prefix=/usr \
    --with-dri-driverdir=/usr/lib/xorg/modules/dri \
    --with-dri-drivers=swrast,radeon,r300,r600 \
    --enable-glx-tls \
    --with-driver=dri \
    --enable-xcb \
    --enable-egl \
    --enable-glu \
    --enable-glut \
    --enable-glw \
    --enable-gallium \
    --enable-gallium-r600  \
    --enable-gallium-radeon  || return 1

lordheavy · 2010-06-15 14:33:28

Perry3D wrote:

Thanks,
i thought there was a configure paramater like --enable-radeon-gallium. But that is a nice solution.
At the moment i build mesa-full with --with-dri-drivers=swrast,radeon,r200,r300,r600. Maybe i will switch r300 to gallium.

Seen in configure.ac:

dnl
dnl Gallium Radeon configuration
dnl
AC_ARG_ENABLE([gallium-radeon],
    [AS_HELP_STRING([--enable-gallium-radeon],
        [build gallium radeon @<:@default=disabled@:>@])],
    [enable_gallium_radeon="$enableval"],
    [enable_gallium_radeon=auto])
if test "x$enable_gallium_radeon" = xyes; then
    GALLIUM_DRIVERS_DIRS="$GALLIUM_DRIVERS_DIRS r300"
    gallium_check_st "radeon/drm" "dri-radeong" "egl-radeon" "xorg-radeon"
elif test "x$enable_gallium_radeon" = xauto; then
    GALLIUM_DRIVERS_DIRS="$GALLIUM_DRIVERS_DIRS r300"
fi

dnl
dnl Gallium Radeon r600g configuration
dnl
AC_ARG_ENABLE([gallium-r600],
    [AS_HELP_STRING([--enable-gallium-r600],
        [build gallium radeon @<:@default=disabled@:>@])],
    [enable_gallium_r600="$enableval"],
    [enable_gallium_r600=auto])
if test "x$enable_gallium_r600" = xyes; then
    GALLIUM_DRIVERS_DIRS="$GALLIUM_DRIVERS_DIRS r600"
    gallium_check_st "r600/drm" "dri-r600"
fi

endlessroad1991 · 2010-06-15 15:42:21

Oh, I am definitely wrong.... Really sorry for that.

handy · 2010-06-19 11:46:53

Where An Open ATI Driver Beats The Catalyst Driver
Published on June 19, 2010

http://www.phoronix.com/scan.php?page=a … t_xv&num=1

Perry3D · 2010-06-19 13:23:57

handy wrote:

Where An Open ATI Driver Beats The Catalyst Driver
Published on June 19, 2010
http://www.phoronix.com/scan.php?page=a … t_xv&num=1

That's a weird comparison. They knew that the video-decoding is corrupted in 10.6 but compared it against the OS driver. Without vaapi the catalyst driver really sucks in video rendering.
But i am interested in the 2d-comparison with direct2d :-)

Vi0L0 · 2010-06-19 13:59:22

handy wrote:

Where An Open ATI Driver Beats The Catalyst Driver
Published on June 19, 2010
http://www.phoronix.com/scan.php?page=a … t_xv&num=1

Another Phoronix fail?
For me 1080p videos rendered with vaapi and catalyst 10.6 eat average ~6% (maximum 11%) usage of 1 of my cores (2.4Ghz) when watching video in window and with kwin on. So how could it eat average 95% in Phoronix test?!
Maybe catalyst sux at tested hardware, but i really believe that Michael didn't compiled xvba packages well.
Now my trust to phoronix tests drops significantly...

Ps. Sure xv sux with catalyst and it will take years (maybe aeons) when it will be fixed, but thats well known true.
Pps. Except evergreen cards users i havent seen posts complaining about vaapi support on catalyst 10.6. Achilles' heel of catalyst 10.6 is 2D acceleration which render lots of artifacts for some users...

Perry3D · 2010-06-19 14:13:19

Vi0L0 wrote:

Another Phoronix fail?
For me 1080p videos rendered with vaapi and catalyst 10.6 eat average ~6% (maximum 11%) usage of 1 of my cores (2.4Ghz) when watching video in window and with kwin on.

That is nice. And what's about fullscreen? And how is the quality? Can you see tearing?

Vi0L0 wrote:

Ps. Sure xv sux with catalyst and it will take years (maybe aeons) when it will be fixed, but thats well known true.
Pps. Except evergreen cards users i havent seen posts complaining about vaapi support on catalyst 10.6. Achilles' heel of catalyst 10.6 is 2D acceleration which render lots of artifacts for some users...

But i heard that there were also problems with the old version and direct2d?

Vi0L0 · 2010-06-19 14:28:12

Perry3D wrote:

Vi0L0 wrote:
Another Phoronix fail?
For me 1080p videos rendered with vaapi and catalyst 10.6 eat average ~6% (maximum 11%) usage of 1 of my cores (2.4Ghz) when watching video in window and with kwin on.
That is nice. And what's about fullscreen? And how is the quality? Can you see tearing?

For me (hd4850) at fullscreen mode it's average ~4% (maximum 10%). Since catalyst 10.5 quality is very good - same as with xv/gl/gl2 video output (before 10.5 i was experiencing bad colours). When compositing is on theres no tearing at all in fullscreen mode (no tearing at all in any mode when compositing is off).

Perry3D wrote:

Vi0L0 wrote:
Ps. Sure xv sux with catalyst and it will take years (maybe aeons) when it will be fixed, but thats well known true.
Pps. Except evergreen cards users i havent seen posts complaining about vaapi support on catalyst 10.6. Achilles' heel of catalyst 10.6 is 2D acceleration which render lots of artifacts for some users...
But i heard that there were also problems with the old version and direct2d?

Yes, i also had problems with direct2d before catalyst 10.6 - but that funcionality was in developement and was off by default. With 10.6 it's always on (its now named ATi 2D Acceleration) and now i have no problems with it.

Perry3D · 2010-06-19 14:34:31

Sounds very good . The catalyst driver is making some progress.
But if i understand right you only get tearing if compositing is enabled and you watch in window mode?

I hope phoronix will do a good job when they benchmark 2d-performance .

Vi0L0 · 2010-06-19 14:43:58

Perry3D wrote:

But if i understand right you only get tearing if compositing is enabled and you watch in window mode?

Yes. Plus offcourse flash was and is always tearing - so i kinda miss oss driver and hes 'v-sync' ;P.
Hope that Adobe will do video acceleration in next linux flashplugin release (yes, Adobe sux very very VERY much cuz of no flash x86_64 linux support in flashplugin 10.1).

Perry3D · 2010-06-19 15:17:33

I would prefer a wide spreading of the html5 video tag in combination with flash-blocker for some flash-sites.

Cdh · 2010-06-19 22:50:15

Perry3D wrote:

They knew that the video-decoding is corrupted in 10.6

How is it actually corrupted?
It seems to be working fine for me...

Actually CPU usage is a bit higher than with 10.5 but still runs fine.

Perry3D · 2010-06-20 16:46:18

And there it is, the 2d- comparison: http://www.phoronix.com/vr.php?view=15040
But i don't like the article? Why did they use the old mesa 7.7 stack? New Catalyst vs. old Mesa, very funny

And in gtkperf they only tested the radio buttons and combo box. And they didn't write how many rounds they used.
And in x11perf -aatrapezoid300 they get 500? I get 2700. Weird test.

@Cdh: I mean it is corrupted on their system.

gtklocker · 2010-06-20 17:05:59

Scrolling is laggy (very laggy) after the last update.

Edit: not only scrolling.

Last edited by gtklocker (2010-06-20 17:12:28)

Arch Linux

#1401 2010-06-07 20:52:33

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1402 2010-06-10 13:39:56

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1403 2010-06-10 18:11:27

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1404 2010-06-10 18:46:25

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1405 2010-06-10 19:10:16

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1406 2010-06-10 19:29:03

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1407 2010-06-10 20:17:14

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1408 2010-06-15 08:59:47

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1409 2010-06-15 09:00:18

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1410 2010-06-15 10:30:24

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1411 2010-06-15 10:35:19

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1412 2010-06-15 14:28:08

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1413 2010-06-15 14:33:28

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1414 2010-06-15 15:42:21

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1415 2010-06-19 11:46:53

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1416 2010-06-19 13:23:57

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1417 2010-06-19 13:59:22

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1418 2010-06-19 14:13:19

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1419 2010-06-19 14:28:12

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1420 2010-06-19 14:34:31

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1421 2010-06-19 14:43:58

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1422 2010-06-19 15:17:33

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1423 2010-06-19 22:50:15

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1424 2010-06-20 16:46:18

Re: Discussion: the ati/radeon open source drivers & radeon repository

#1425 2010-06-20 17:05:59

Re: Discussion: the ati/radeon open source drivers & radeon repository

Board footer