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

Perry3D · 2010-03-23 17:52:44

What have you done? Some hints in dmesg or Xorg.0.log?

gtklocker · 2010-03-23 17:59:38

No, everything's clear.

Perry3D · 2010-03-23 19:21:04

Are you using my repo?

perry64 ~ % glxinfo |grep -i opengl
OpenGL vendor string: Advanced Micro Devices, Inc.
OpenGL renderer string: Mesa DRI R600 (RV770 9442) 20090101 TCL DRI2
OpenGL version string: 2.0 Mesa 7.9-devel
OpenGL shading language version string: 1.10
OpenGL extensions:

/edit: oh, i updated glproto, libdrm,mesa and the driver and 2d-acceleration is very slow. I hope nobody get the packages from today from the repo. But i revert the repo.
No updates for my repo until this bug is resolved: http://bugs.freedesktop.org/show_bug.cgi?id=27284

Last edited by Perry3D (2010-03-25 19:51:47)

Perry3D · 2010-03-24 17:08:55

Michael announced a fglrx vs. radeon comparison for this monday on phoronix: http://www.phoronix.com/forums/showpost … stcount=46

That could be interesting.

gtklocker · 2010-03-26 11:58:31

Hey! I made my own packages and everything is working like a charm!

I'll upload these somewhere...

handy · 2010-03-28 10:43:28

Is anyone aware of a scripting solution that would allow a user to choose from the Grub menu the open-source or the Catalyst solution & so boot their system with whatever they have chosen?

I think that there are many people who would like to use such an option, it would be very convenient for introducing the open-source solution to users & for easily allowing those who use the open-source packages to have the option for Catalyst when desired/required.

Could this script be modified to do the job?

http://ubuntuforums.org/showpost.php?p= … ostcount=9

Last edited by handy (2010-03-28 10:47:09)

gtklocker · 2010-03-28 10:48:58

The packages confict each other, so it can't be done.

Anyway, these are the options:

r300-r600 --> Radeon
r700- ~ --> Catalyst-test (w/ VAAPI etc.)

ijanos · 2010-03-28 10:50:41

gtklocker wrote:

The packages confict each other, so it can't be done.
Anyway, these are the options:
r300-r600 --> Radeon
r700- ~ --> Catalyst-test (w/ VAAPI etc.)

radeon supports r700

gtklocker · 2010-03-28 10:52:46

I know, but you have the VAAPI profit (with UVDv2) using catalyst-test.

Perry3D · 2010-03-28 10:53:08

ijanos wrote:

radeon supports r700

It also supports r800 (yet without acceleration).

dejavu · 2010-03-28 21:05:09

Wow...

The new kernel26-pm is amazingly fast on a hd3450...
And also clocks down, when idle.

Really nice... ;-)

gtklocker · 2010-03-28 21:20:37

New? New new? Really new?

venky80 · 2010-03-28 23:32:02

dejavu wrote:

Wow...
The new kernel26-pm is amazingly fast on a hd3450...
And also clocks down, when idle.
Really nice... ;-)

is it from perry3d repo? Also is the sensors not getting detected problem patched in pm kernel?

dejavu · 2010-03-28 23:54:00

Yes it's from perrys3d repo...

Regarding the sensor problem, I can't say anything about it, because I am pretty sure, that
I have no such sensors in my laptop.

venky80 · 2010-03-29 01:17:07

dejavu wrote:

Yes it's from perrys3d repo...
Regarding the sensor problem, I can't say anything about it, because I am pretty sure, that
I have no such sensors in my laptop.

my sensors stop working after resuming from suspend and laptop gets progressively hot.

Perry3D · 2010-03-29 04:52:13

S3TC Support For Mesa Brought Up Again
I think i will build mesa with the txc_dxtn lib as dependency. I hope you don't live in a country where software patents are a problem .

Phoronix: Catalyst vs. Mesa Performance With Ubuntu 10.04
Conclusion for catalyst: 2D and Video suck, 3D is a lot faster (but they tested the 7.7 version of mesa )

@venky: yes, The patch should be included in kernel26-pm.

Last edited by Perry3D (2010-03-29 06:15:46)

gtklocker · 2010-03-29 06:20:05

Patch it up !

@Perry3D: Could you do a

makepkg --source

on the build directory (of kernel26-pm) and upload the source file somewhere (eg. omploader) ?

Last edited by gtklocker (2010-03-29 06:20:24)

Cdh · 2010-03-29 07:10:42

Perry3D wrote:

Conclusion for catalyst: 2D and Video suck

AMD's X-Video Bitstream Acceleration (XvBA) was not used since the open-source ATI driver does not support it

Yes, of course it sucks when you don't use the techniques that make it far superior.

yimm · 2010-03-29 20:07:29

@ Cdh : yes, it's not objective.
And for 2D, by using Direct2D it's fast.

Last edited by yimm (2010-03-29 20:08:05)

Perry3D · 2010-03-29 20:27:49

I added the libtxc_dxtn package to my repo. If i understand right you just need to install the lib and s3 compression is enabled by default.

@gtklocker: http://dl.dropbox.com/u/4832298/x86_64/ … src.tar.gz
@Cdh: You're right. Even direct2d wasn't enabled. He also should compare the catalyst to a newer mesa stack (git or at least 7.8).

Cdh · 2010-03-29 20:47:44

I don't want to pick on the free driver because it is really great already, but there needs to be video acceleration implemented and gallium support needs to be completed so newer opengl features are fast as well.
At the moment they are really slow, I think the CPU does them. For example shadows in many games make the fps drop really low, also you cannot run for example half life 2.

They didn't write something like "video performance with vaapi is better then viedeo performance with the free driver because video acceleration is not implemented in the free driver" or "advanced opengl features do not work or are terribly slow because they are just not ready".
catalyst has some heavy issues too, like in half life 2 with HDR the whole world often goes just dark like there was no lighting at all or sometimes severe flickering and artifacts while gameplay. Also putting 3D applications in the background and using other windows above them does often not work and the 3D application is always in the foreground and ON ALL WORKSPACES! The ATI "AI" function is buggy like hell and results in major graphic errors.

With the free driver I never encountered such problems and I am sure they will be much better then (today's) catalyst when they are finally ready.

Vi0L0 · 2010-03-30 13:37:35

Perry3D wrote:

He also should compare the catalyst to a newer mesa stack (git or at least 7.8).

But according to this http://www.phoronix.com/scan.php?page=a … ance&num=1 theres no difference in performance between mesa 7.7/7.8/7.9-dev

Perry3D · 2010-03-30 13:56:31

Vi0L0 wrote:

But according to this http://www.phoronix.com/scan.php?page=a … ance&num=1 theres no difference in performance between mesa 7.7/7.8/7.9-dev

This benchmark is based on a R5xx card. If you use a R6xx/7xx card there is a big difference between 7.7 and 7.8. For example blit support is introduced with 7.8.

Vi0L0 · 2010-03-30 14:10:49

Perry3D wrote:

Vi0L0 wrote:
But according to this http://www.phoronix.com/scan.php?page=a … ance&num=1 theres no difference in performance between mesa 7.7/7.8/7.9-dev
This benchmark is based on a R5xx card. If you use a R6xx/7xx card there is a big difference between 7.7 and 7.8. For example blit support is introduced with 7.8.

Awww... then sorry
Now all the signs on earth and on heaven shows that we shall not rely on phoronix...

ChemBro · 2010-03-30 15:11:32

Vi0L0 wrote:

Now all the signs on earth and on heaven shows that we shall not rely on phoronix...

Don't rely on their benchmark-results (except the result is too obvious like the difference between catalyst and radeon power). But they are a good source for news. IMO.

Arch Linux

#1076 2010-03-23 17:52:44

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

#1077 2010-03-23 17:59:38

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

#1078 2010-03-23 19:21:04

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

#1079 2010-03-24 17:08:55

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

#1080 2010-03-26 11:58:31

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

#1081 2010-03-28 10:43:28

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

#1082 2010-03-28 10:48:58

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

#1083 2010-03-28 10:50:41

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

#1084 2010-03-28 10:52:46

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

#1085 2010-03-28 10:53:08

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

#1086 2010-03-28 21:05:09

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

#1087 2010-03-28 21:20:37

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

#1088 2010-03-28 23:32:02

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

#1089 2010-03-28 23:54:00

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

#1090 2010-03-29 01:17:07

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

#1091 2010-03-29 04:52:13

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

#1092 2010-03-29 06:20:05

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

#1093 2010-03-29 07:10:42

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

#1094 2010-03-29 20:07:29

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

#1095 2010-03-29 20:27:49

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

#1096 2010-03-29 20:47:44

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

#1097 2010-03-30 13:37:35

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

#1098 2010-03-30 13:56:31

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

#1099 2010-03-30 14:10:49

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

#1100 2010-03-30 15:11:32

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

Board footer