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

Celeblin · 2009-12-13 12:49:55

R00KIE wrote:

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.

I read that ForceLowPowerMode is not the lowest possible mode, and I don't know too, if when the system is idle (that is when DynamicPM works) the power mode that activates is the lowest one, so because I have no conflicts or problems I activated that two options.

Maybe you are right and the mode that uses "ForceLowPowerMode" is the same that "DynamicPM" when the system is idle, so DynamicPM is not necessary, but I really don't know.

acmps · 2009-12-13 13:13:42

xheyther wrote:

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

I don't understand what's all the fuss about early start. Just install radeon-initrd from AUR and add radeon in HOOKS (/etc/mkinitcpio.conf). Then rebuild your initrd (sudo mkinitcpio -p kernel26).

acmps · 2009-12-13 13:44:35

Celeblin wrote:

R00KIE wrote:
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.
I read that ForceLowPowerMode is not the lowest possible mode, and I don't know too, if when the system is idle (that is when DynamicPM works) the power mode that activates is the lowest one, so because I have no conflicts or problems I activated that two options.
Maybe you are right and the mode that uses "ForceLowPowerMode" is the same that "DynamicPM" when the system is idle, so DynamicPM is not necessary, but I really don't know.

So, to clarify...

ForceLowPowerMode always enables the low power mode, no matter what happens.
DynamicPM enables low power mode when the GPU is not being used excessively (gaming is using it excessively)

Am I right?

flamelab · 2009-12-13 13:48:07

And none of them works right now with KMS enabled, for your information

acmps · 2009-12-13 13:52:42

flamelab wrote:

And none of them works right now with KMS enabled, for your information

That's really a shame. I love KMS, and I prefer to loose 15min of battery than to loose KMS.

Celeblin · 2009-12-13 14:35:10

acmps wrote:

DynamicPM enables low power mode when the GPU is not being used excessively (gaming is using it excessively)

No exactly, what you say is done by ClockGating.
DynamicPM enables the low power mode when using DPMS, like suspend, standby and, I think, when the screen turns off.

From "man radeon":

Option "ClockGating" "boolean"
              Enable dynamic clock gating.  This  can  help  reduce  heat  and
              increase  battery  life  by  reducing  power  usage.  Some users
              report reduced 3D performance with this enabled.  The default is
              off.

       Option "ForceLowPowerMode" "boolean"
              Enable  a  static low power mode.  This can help reduce heat and
              increase battery life by reducing power usage at the expense  of
              performance. The default is off.

       Option "DynamicPM" "boolean"
              Enable  dynamic power mode switching.  This can help reduce heat
              and increase battery life by reducing power usage when the  sys‐
              tem is idle (DPMS active). The default is off.

Because I don't have any problem or conflicts, I decided to active all of them and now I have 55º of temp, when before I had 70º. Without KMS I have some little artifacts, but still is all well usable, and compiz works smooth.

Last edited by Celeblin (2009-12-13 14:38:57)

badenov · 2009-12-13 14:50:45

Are these options (DynamicPM, ...) really work? I've got Mobility Radeon 4670, kernel-git from aur, other stuff from nightly build repo, fan in my laptop is always on (with ati and radeonhd driver) and temperature is between 68-91oC (45-55oC with catalyst drivers).
xorg.log says:

(WW) RADEON(0): Option "DynamicPM" is not used
(WW) RADEON(0): Option "ClockGating" is not used
(WW) RADEON(0): Option "ForceLowPowerMode" is not used
(WW) RADEON(0): Option "DMAForXv" is not used

Celeblin · 2009-12-13 15:06:37

badenov wrote:

Are these options (DynamicPM, ...) really work? I've got Mobility Radeon 4670, kernel-git from aur, other stuff from nightly build repo, fan in my laptop is always on (with ati and radeonhd driver) and temperature is between 68-91oC (45-55oC with catalyst drivers).
xorg.log says:

Did you disable KMS?

badenov · 2009-12-13 15:21:22

Celeblin wrote:

Did you disable KMS?

Yes, I tried different configurations. No warnings, but temperature is still the same.

Last edited by badenov (2009-12-13 15:29:07)

R00KIE · 2009-12-13 15:58:06

As I can understand it goes like this (just trying to make sense of the names):

clockgating: disables the clock (and may also cut the power) to idle parts of the core

dynamicpm: raises and lowers the clock frequency according to load but doesn't disable the clock or turn off completely unused parts of the core (unless clock gating is implied to be on and the clockgating option exists to disable just that if it causes problems)

forcelowpowermode: same as dynamicpm when the core is idle

With KMS this does not work yet, but if you choose not to use KMS when on battery it can make a big difference, in my case with a radeonhd 2400 it is about 30min difference, I guess that for higher end cards it can mean a lot more time than that.

Nezmer · 2009-12-13 16:19:45

Hi guys,
I grabbed a kernel snapshot from drm-radeon-testing head and compiled it. This is the first time I try KMS out and It works great.

One issue though, Xorg.0.log reports that Radeon disabled hardware acceleration. That doesn't happen when I disable KMS. Is this expected behaviour ?

$ lspci|grep -i vga
01:00.0 VGA compatible controller: ATI Technologies Inc M92 [Mobility Radeon HD 4500 Series]

Last edited by Nezmer (2009-12-13 16:20:41)

Nezmer · 2009-12-13 19:36:54

Alright, I needed to grab "R700_rlc.bin" firmware. glxinfo now reports hardware acceleration but I ended up with a corrupted(Is it the right word) display.

acmps · 2009-12-14 13:24:50

No exactly, what you say is done by ClockGating.
DynamicPM enables the low power mode when using DPMS, like suspend, standby and, I think, when the screen turns off.

Thanks! I think I'll disable KMS and use ClockGating!

Celeblin · 2009-12-14 14:28:59

acmps wrote:

No exactly, what you say is done by ClockGating.
DynamicPM enables the low power mode when using DPMS, like suspend, standby and, I think, when the screen turns off.
Thanks! I think I'll disable KMS and use ClockGating!

I must enable ForceLowPowerMode or my laptop gets the 70º temperature, even with the other two options active. DynamicPM is normal, because only works with DPMS is active, but ClockGating seems not to take effect in my power management.

There is more information:
http://www.phoronix.com/forums/showthread.php?t=20413

acmps · 2009-12-14 17:30:08

Celeblin wrote:

acmps wrote:
No exactly, what you say is done by ClockGating.
DynamicPM enables the low power mode when using DPMS, like suspend, standby and, I think, when the screen turns off.
Thanks! I think I'll disable KMS and use ClockGating!
I must enable ForceLowPowerMode or my laptop gets the 70º temperature, even with the other two options active. DynamicPM is normal, because only works with DPMS is active, but ClockGating seems not to take effect in my power management.
There is more information:
http://www.phoronix.com/forums/showthread.php?t=20413

Thanks, that url really gave me the info that I needed!

silenc3r · 2009-12-14 23:22:07

Hi
with kernel .32 and new driver everything seems to work fine (maybe except flash which is kinda catchy) but im getting strange output of glxinfo |grep -i opengl

do_wait: drmWaitVBlank returned -1, IRQs don't seem to be working correctly.
Try adjusting the vblank_mode configuration parameter.
OpenGL vendor string: Advanced Micro Devices, Inc.
OpenGL renderer string: Mesa DRI R600 (RV620 95C4) 20090101  TCL DRI2
OpenGL version string: 1.4 Mesa 7.6
OpenGL extensions:

I guess something's not working properly. How to fix it?

Last edited by silenc3r (2009-12-14 23:22:28)

Perry3D · 2009-12-14 23:26:27

silenc3r wrote:

Hi
with kernel .32 and new driver everything seems to work fine (maybe except flash which is kinda catchy) but im getting strange output of glxinfo |grep -i opengl
do_wait: drmWaitVBlank returned -1, IRQs don't seem to be working correctly.
Try adjusting the vblank_mode configuration parameter.
OpenGL vendor string: Advanced Micro Devices, Inc.
OpenGL renderer string: Mesa DRI R600 (RV620 95C4) 20090101  TCL DRI2
OpenGL version string: 1.4 Mesa 7.6
OpenGL extensions:
I guess something's not working properly. How to fix it?

No, this message is fine. IRQ support is not implemented yet. We will get it with kernel 2.6.33. That's just a warning.
We already talked about this in this thread

Kraychik · 2009-12-15 08:57:09

Hi all!
Help me, please.
I did everything as stated in the first post.
Kernel get in AUR.
All compile and install, but nothing works...

# uname -a
Linux arch 2.6.32-rc #1 SMP PREEMPT Mon Dec 14 02:36:36 EET 2009 i686 Intel(R) Core(TM)2 Duo CPU E4500 @ 2.20GHz GenuineIntel GNU/Linux

# dmesg
...
[drm] Initialized drm 1.1.0 20060810                                                   
iTCO_wdt: Intel TCO WatchDog Timer Driver v1.05                                        
iTCO_wdt: Found a ICH9 TCO device (Version=2, TCOBASE=0x0860)                          
iTCO_wdt: initialized. heartbeat=30 sec (nowayout=0)                                   
lp0: using parport0 (interrupt-driven).                                                
[drm] radeon defaulting to kernel modesetting.                                         
[drm] radeon kernel modesetting enabled.                                               
radeon 0000:01:00.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16                        
radeon 0000:01:00.0: setting latency timer to 64                                       
[drm] radeon: Initializing kernel modesetting.                                         
[drm] register mmio base: 0xFE9F0000                                                   
[drm] register mmio size: 65536                                                        
ATOM BIOS: 113                                                                         
[drm] Clocks initialized !                                                             
[drm] Detected VRAM RAM=256M, BAR=256M                                                 
[drm] RAM width 128bits DDR                                                            
[TTM] Zone  kernel: Available graphics memory: 443160 kiB.                             
[TTM] Zone highmem: Available graphics memory: 1036924 kiB.                            
[drm] radeon: 256M of VRAM memory ready                                                
[drm] radeon: 512M of GTT memory ready.                                                
[drm] Loading RV630 CP Microcode                                                       
platform radeon_cp.0: firmware: requesting radeon/RV630_pfp.bin
r600_cp: Failed to load firmware "radeon/RV630_pfp.bin"
[drm:r600_init] *ERROR* Failed to load firmware!
[drm:radeon_driver_load_kms] *ERROR* Fatal error while trying to initialize radeon.
radeon 0000:01:00.0: PCI INT A disabled
radeon: probe of 0000:01:00.0 failed with error -2
...

Xorg log:
...
drmOpenDevice: node name is /dev/dri/card0
drmOpenByBusid: Searching for BusID pci:0000:01:00.0
drmOpenDevice: node name is /dev/dri/card0
drmOpenByBusid: drmOpenMinor returns -1
drmOpenDevice: node name is /dev/dri/card1
drmOpenByBusid: drmOpenMinor returns -1
drmOpenDevice: node name is /dev/dri/card2
drmOpenByBusid: drmOpenMinor returns -1
drmOpenDevice: node name is /dev/dri/card3
drmOpenByBusid: drmOpenMinor returns -1
drmOpenDevice: node name is /dev/dri/card4
drmOpenByBusid: drmOpenMinor returns -1
drmOpenDevice: node name is /dev/dri/card5
drmOpenByBusid: drmOpenMinor returns -1
drmOpenDevice: node name is /dev/dri/card6
drmOpenByBusid: drmOpenMinor returns -1
drmOpenDevice: node name is /dev/dri/card7
drmOpenByBusid: drmOpenMinor returns -1
drmOpenDevice: node name is /dev/dri/card8
drmOpenByBusid: drmOpenMinor returns -1
drmOpenDevice: node name is /dev/dri/card9
drmOpenByBusid: drmOpenMinor returns -1
drmOpenDevice: node name is /dev/dri/card10
drmOpenByBusid: drmOpenMinor returns -1
drmOpenDevice: node name is /dev/dri/card11
drmOpenByBusid: drmOpenMinor returns -1
drmOpenDevice: node name is /dev/dri/card12
drmOpenByBusid: drmOpenMinor returns -1
drmOpenDevice: node name is /dev/dri/card13
drmOpenByBusid: drmOpenMinor returns -1
drmOpenDevice: node name is /dev/dri/card14
drmOpenByBusid: drmOpenMinor returns -1
drmOpenDevice: node name is /dev/dri/card15
drmOpenByBusid: drmOpenMinor returns -1
drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: node name is /dev/dri/card1
drmOpenDevice: node name is /dev/dri/card2
drmOpenDevice: node name is /dev/dri/card3
drmOpenDevice: node name is /dev/dri/card4
drmOpenDevice: node name is /dev/dri/card5
drmOpenDevice: node name is /dev/dri/card6
drmOpenDevice: node name is /dev/dri/card7
drmOpenDevice: node name is /dev/dri/card8
drmOpenDevice: node name is /dev/dri/card9
drmOpenDevice: node name is /dev/dri/card10
drmOpenDevice: node name is /dev/dri/card11
drmOpenDevice: node name is /dev/dri/card12
drmOpenDevice: node name is /dev/dri/card13
drmOpenDevice: node name is /dev/dri/card14
drmOpenDevice: node name is /dev/dri/card15
(EE) RADEON(0): [dri] RADEONDRIGetVersion failed to open the DRM
[dri] Disabling DRI.
(II) RADEON(0): using shadow framebuffer
...

What causes failure?

guzz46 · 2009-12-15 09:33:27

Do you have radeon firmware in /lib/firmware if not download this

if you are x86_64
http://www.schlunix.org/archlinux/testi … pkg.tar.gz

or this if you are i686
http://www.schlunix.org/archlinux/testi … pkg.tar.gz

I have a custom kernel so i didn't install the package i just extracted the contents and copied the radeon folder to /lib/firmware

Last edited by guzz46 (2009-12-15 09:36:17)

Perry3D · 2009-12-15 11:11:25

Or install the kernel from testing instead from AUR.

venky80 · 2009-12-15 21:19:24

can anyone confirm if KWIN composting works with kde 4.4 for their r6xx ati graphics card?

Nezmer · 2009-12-16 01:55:55

Nezmer wrote:

Alright, I needed to grab "R700_rlc.bin" firmware. glxinfo now reports hardware acceleration but I ended up with a corrupted(Is it the right word) display.

If anyone cares, Alex fixed this with this patch:
http://comments.gmane.org/gmane.comp.vi … evel/41192

Intrepid · 2009-12-18 05:24:03

Anyone here run Arch64 and have trouble with 32-bit opengl programs also? I can't run Zsnes or pSX in OpenGL. I think it's because lib32-libgl is not in git.

Last edited by Intrepid (2009-12-18 05:25:43)

Dinth · 2009-12-18 16:57:11

Im also asking for lib32-libgl-git package. Now 3d acceleration in wine application work terribly slow (about 0.3 fps, propably software mode) for me

[edit]
Ive tried to make lib32-libgl-git, lib32-libdrm-git and lib32-mesa-git packages myself, by modifing orginal PKGBUILD to use nightly repository libgl, libdrm and mesa packages. But unfortunately still games run terribly slow or crashes with error that direct rendering is disabled

Last edited by Dinth (2009-12-18 18:37:28)

Intrepid · 2009-12-18 19:39:03

Glad to hear I'm not alone (or crazy!)

Arch Linux

#376 2009-12-13 12:49:55

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

#377 2009-12-13 13:13:42

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

#378 2009-12-13 13:44:35

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

#379 2009-12-13 13:48:07

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

#380 2009-12-13 13:52:42

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

#381 2009-12-13 14:35:10

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

#382 2009-12-13 14:50:45

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

#383 2009-12-13 15:06:37

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

#384 2009-12-13 15:21:22

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

#385 2009-12-13 15:58:06

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

#386 2009-12-13 16:19:45

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

#387 2009-12-13 19:36:54

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

#388 2009-12-14 13:24:50

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

#389 2009-12-14 14:28:59

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

#390 2009-12-14 17:30:08

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

#391 2009-12-14 23:22:07

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

#392 2009-12-14 23:26:27

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

#393 2009-12-15 08:57:09

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

#394 2009-12-15 09:33:27

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

#395 2009-12-15 11:11:25

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

#396 2009-12-15 21:19:24

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

#397 2009-12-16 01:55:55

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

#398 2009-12-18 05:24:03

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

#399 2009-12-18 16:57:11

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

#400 2009-12-18 19:39:03

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

Board footer