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

venky80 · 2009-12-09 04:03:11

acmps wrote:

venky80 wrote:
Man awesome if someone could host it as a separate repo...I am on x86_64 but this is good
If someone can borrow me a few megabytes on a server I'd do that - I'd even install Arch64 and do x86_64 packages.
Anyone?

Can someone host his files..a repo for x86_64 would be aweosome!
Is it possible you can upload x86_64 packages somewhere!

Last edited by venky80 (2009-12-09 04:05:26)

venky80 · 2009-12-09 04:53:47

Cdh wrote:

I don't built that packages, that are the people from nightly.user-helfen-usern.de
They have x86_64 packages too
[nightly]
Server = http://nightly.user-helfen-usern.de/view.php?file=ix86_64
I think.
Actually Kernel Modesetting does work quite good with late start and then 3D in general works good too, but Warcraft3 in Wine won't start with KMS enabled (dri2 buggy?)
Without KMS it works.

cant seem to install anything from this repository ...
when I do pacman -Ss mesa-git it doesn't find anything

found the issue the repo name is
[nightly]
Server = http://nightly.user-helfen-usern.de/x86_64

Last edited by venky80 (2009-12-09 05:46:20)

Cdh · 2009-12-09 05:53:32

Yea it sucks that there is no documentation for the real entry in the pacman.conf
With try & error I got this working:
Server = http://nightly.user-helfen-usern.de/view.php?file=i686

When you try to download a file through webinterface you see that files are downloaded through

http://nightly.user-helfen-usern.de/view.php?file=x86_64/filename

So the correct entry supposedly has to be

[nightly]
Server http://nightly.user-helfen-usern.de/view.php?file=x86_64

Last edited by Cdh (2009-12-09 05:54:26)

Perry3D · 2009-12-09 09:00:50

thepizzaking wrote:

Just a note, the xf86-video-ati package from testing seems to work fine (ie. no need for -git version) for me. So all it seems that's needed now which isn't in testing are the -git versions of the mesa packages (or if you'd rather build v7.7rc2, that works too).

Did you tested it? Maybe the driver has to be linked against the mesa-package (ati-dri). And this one in testing seems to be linked against mesa from extra.

Perry3D · 2009-12-09 10:35:37

Celeblin wrote:

I have a Dell Studio 1537 with a Mobility Radeon HD3450 (M82 code), and I'm not ready to use experimental packages because I use the laptop for study, but I wanna know if somebody with the same chipset could explain his impressions with opensource-git drivers. That's because with the current video-ati and video-radeonhd my gpu temperature is too high and the fan is on all the time, consuming the battery so I must use the fglrx driver, but I'm really tired of ATI's behaviour with Linux.
Thanks a lot guys!

http://wiki.x.org/wiki/RadeonFeature

There is power management in the radeon driver. But it is not complete and at the moment the catalyst driver will produce less heat as the radeon driver. Maybe we will see the rest of the power management features in kernel 2.6.33.
You can try these 2 options:
- ForceLowPowerMode
- DynamicPM
Use man radeon or google for these parameters.

Cdh · 2009-12-09 11:00:19

Is there any way to control the RPM of the fan on a Radeon 4650? It's really loud at full speed while the card get's not even warm.

venky80 · 2009-12-09 16:12:05

Cdh wrote:

Yea it sucks that there is no documentation for the real entry in the pacman.conf
With try & error I got this working:
Server = http://nightly.user-helfen-usern.de/view.php?file=i686
When you try to download a file through webinterface you see that files are downloaded through
http://nightly.user-helfen-usern.de/view.php?file=x86_64/filename
So the correct entry supposedly has to be
[nightly]
Server http://nightly.user-helfen-usern.de/view.php?file=x86_64

nope it is

[nightly]
Server http://nightly.user-helfen-usern.de/x86_64

Celeblin · 2009-12-09 18:02:39

Perry3D wrote:

http://wiki.x.org/wiki/RadeonFeature
There is power management in the radeon driver. But it is not complete and at the moment the catalyst driver will produce less heat as the radeon driver. Maybe we will see the rest of the power management features in kernel 2.6.33.
You can try these 2 options:
- ForceLowPowerMode
- DynamicPM
Use man radeon or google for these parameters.

Thank you! With that and ClockGating the temp is now near 55º. If somebody wanna use it must disable KMS, because they are not compatible with it for now.

Bye bye catalyst

Cdh · 2009-12-09 21:16:44

You are right, venky. The first time I tried to just add /i686 it didn't work and the repo just showed up empty. I guess I was doing something wrong because now it works. The other way did work too for me.

I'm really amazed. Without KMS the driver is very fast and stable for now. Ok, Prey was very slow and World of Warcraft had some graphical issues and was also very slow on max details, but "simpler" games like nexuiz or warcraft3 in wine work really good and fast.
With catalyst I have seen the weirdest graphical glitches in Warcraft3, but now with radeon it was completely smooth.

But with KMS (and DRI2) it does not work so well with my radeon 4650. I guess it is just not ready...

thepizzaking · 2009-12-09 23:33:39

Perry3D wrote:

thepizzaking wrote:
Just a note, the xf86-video-ati package from testing seems to work fine (ie. no need for -git version) for me. So all it seems that's needed now which isn't in testing are the -git versions of the mesa packages (or if you'd rather build v7.7rc2, that works too).
Did you tested it? Maybe the driver has to be linked against the mesa-package (ati-dri). And this one in testing seems to be linked against mesa from extra.

I'm running libdrm 2.4.16-1 from testing, xf86-video-ati 6.12.99.git20091207-1 from testing, and mesa 7.7rc2 (also running KMS with kernel 2.6.32). glxgears runs, whereas before these packages it didn't. Also, the only thing I needed to build were the mesa packages, I didn't need to rebuild xf86-video-ati.

venky80 · 2009-12-10 06:20:44

Celeblin wrote:

Perry3D wrote:
http://wiki.x.org/wiki/RadeonFeature
There is power management in the radeon driver. But it is not complete and at the moment the catalyst driver will produce less heat as the radeon driver. Maybe we will see the rest of the power management features in kernel 2.6.33.
You can try these 2 options:
- ForceLowPowerMode
- DynamicPM
Use man radeon or google for these parameters.
Thank you! With that and ClockGating the temp is now near 55º. If somebody wanna use it must disable KMS, because they are not compatible with it for now.
Bye bye catalyst

could you post your xorg.conf i am having the same issue

Celeblin · 2009-12-10 10:53:57

venky80 wrote:

could you post your xorg.conf i am having the same issue

Sure, I use this minimal xorg:

Section "Device"
        Identifier  "Card0"
        Driver      "radeon"
        VendorName  "ATI Technologies Inc"
        BoardName   "Mobility Radeon HD 3400 Series"
        BusID       "PCI:1:0:0"
        Option     "ClockGating"                "true"
        Option     "DynamicPM"                  "true"
        Option     "ForceLowPowerMode"          "true"
EndSection

And you need to add "nomodeset" to kernel parameters for deactivating KMS, that is not compatible with power management for now.

I hope it helps you!

mixi · 2009-12-10 15:03:25

Cdh wrote:

Yea it sucks that there is no documentation for the real entry in the pacman.conf

The entry found under http://tracker.kromonos.net/projects/show/nightlyarch should be fixed now. The right link could have been found under http://wiki.archlinux.org/index.php/Uno … positories or https://forum.archlinux.de/?id=20;page= … read=13363.

Perry3D · 2009-12-10 20:08:28

Some good news and some bad news (power managment): A Great Present In The Linux 2.6.33 Kernel

xheyther · 2009-12-10 22:17:44

If you read the reply of linus to the drm-pull for D. Airlie it's... brilliant.

This guy really has a talent when it comes to trolling...

Chosen part :

No, the biggest missing feature is that Fedora is _still_ shipping
Nouveau, and I'm _still_ not seeing Red Hat people actively trying to get
it merged into mainline.
What the _hell_ is going on?
Linus

Last edited by xheyther (2009-12-10 22:17:54)

bcat · 2009-12-10 23:35:08

xheyther wrote:

If you read the reply of linus to the drm-pull for D. Airlie it's... brilliant.
This guy really has a talent when it comes to trolling...
Chosen part :
No, the biggest missing feature is that Fedora is _still_ shipping
Nouveau, and I'm _still_ not seeing Red Hat people actively trying to get
it merged into mainline.
What the _hell_ is going on?
Linus

Perhaps you should double-check your definitions. Linus is indeed very outspoken about how he likes to handle kernel issues, but that's not trolling.

csslayer · 2009-12-11 05:25:19

I finally found that why I cannot enable compiz, maybe many people have the same problem with me. I didn't add my user to video group. And not set the mode in xorg.conf....

xheyther · 2009-12-11 10:16:42

bcat wrote:

xheyther wrote:
[...]
Perhaps you should double-check your definitions. Linus is indeed very outspoken about how he likes to handle kernel issues, but that's not trolling.

The way he tell people that he is hungry (and leaving room for people to misunderstand to who he is really shooting at) it's definitely trolling. They finally came to an agreement it seems that someone will port nouveau to the firmware loader infrastructure and that if the microcode legal issue can be solved, we will see nouveau in the mainline kernel.

If the nouveau driver get pulled in the mainstream kernel, we could, us ATI radeon users, see several benefits. The nouveau driver shares some (essential) code with our radeon part (DRM and KMS mainly). so this code wil get more tests and possibly more improvements on performance topic. Other benefits can be forseen for Galium3D and all of it state tracker.

Nezmer · 2009-12-11 13:16:20

HDMI support in KMS

insanemal · 2009-12-11 14:01:26

Hey anybody here running 64bit and use a 32bit chroot for stuff?

I can't get 3D working inside the chroot.

It gives me errors about the kernel not being ready or rejecting the stream.. or something.. I'd run glxgears and get the exact error code.. but my chroot is very broken..

Lets just say trying to build the git versions of the drivers causes odd errors involving it detecting the 64bit kernel.. but trying to compile against the 32bit libraries.

Anyway.. anybody got any ideas?

Napaim · 2009-12-12 21:13:43

I'm using the ati-git driver and getting 3200fps with glxgears as root but 600fps as a normal user, am I missing a step in the configuration? It seems that normal users are using a different version of mesa.. with `glxinfo | grep Mesa` I get
Normal user:

OpenGL vendor string: Mesa Project
OpenGL version string: 2.1 Mesa 7.8-devel

Root:

OpenGL renderer string: Mesa DRI R600 (RV770 9442) 20090101  TCL
OpenGL version string: 1.5 Mesa 7.8-devel

A pointer would be nice, I'm at a loss now.

Perry3D · 2009-12-12 21:27:08

Is your user in the video group?

xheyther · 2009-12-12 21:45:40

For those wanting to activate early KMS, this thread in the testing forum give directions : http://bbs.archlinux.org/viewtopic.php?id=86211&p=1

Napaim · 2009-12-12 22:44:17

Perry3D wrote:

Is your user in the video group?

Indeed not.
That's my problem solved, thank you!

R00KIE · 2009-12-13 00:33:54

Celeblin wrote:

venky80 wrote:
could you post your xorg.conf i am having the same issue
Sure, I use this minimal xorg:
Section "Device"
        Identifier  "Card0"
        Driver      "radeon"
        VendorName  "ATI Technologies Inc"
        BoardName   "Mobility Radeon HD 3400 Series"
        BusID       "PCI:1:0:0"
        Option     "ClockGating"                "true"
        Option     "DynamicPM"                  "true"
        Option     "ForceLowPowerMode"          "true"
EndSection
And you need to add "nomodeset" to kernel parameters for deactivating KMS, that is not compatible with power management for now.
I hope it helps you!

Shouldn't "ForceLowPowerMode" "true" already take care of making the gpu use as little power as possible? I guess it should be somewhat incompatible with dynamicpm because one can either have the lowest power consumption possible (lowest clock speed possible) or have dynamic clocking.

I use "ForceLowPowerMode" too but I don't use any other options .... clock gating makes sense and may be worth a try though.

Arch Linux

#351 2009-12-09 04:03:11

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

#352 2009-12-09 04:53:47

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

#353 2009-12-09 05:53:32

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

#354 2009-12-09 09:00:50

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

#355 2009-12-09 10:35:37

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

#356 2009-12-09 11:00:19

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

#357 2009-12-09 16:12:05

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

#358 2009-12-09 18:02:39

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

#359 2009-12-09 21:16:44

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

#360 2009-12-09 23:33:39

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

#361 2009-12-10 06:20:44

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

#362 2009-12-10 10:53:57

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

#363 2009-12-10 15:03:25

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

#364 2009-12-10 20:08:28

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

#365 2009-12-10 22:17:44

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

#366 2009-12-10 23:35:08

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

#367 2009-12-11 05:25:19

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

#368 2009-12-11 10:16:42

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

#369 2009-12-11 13:16:20

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

#370 2009-12-11 14:01:26

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

#371 2009-12-12 21:13:43

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

#372 2009-12-12 21:27:08

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

#373 2009-12-12 21:45:40

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

#374 2009-12-12 22:44:17

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

#375 2009-12-13 00:33:54

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

Board footer