The AMD/ATI Bar & Grill

ChemBro · 2011-02-20 09:12:18

dcc24 wrote:

codeRage wrote:
My laptop has an older x1400 card for which I need an older kernel to support catalyst-old package. Is there anywhere I can find old kernel pkgbuilds?
Even if you do, it will prove extremely hard to maintain such a system. You'll need to downgrade almost everything and that isn't likely to be possible. You should stick with xf86-video-ati.

And/or use it with the gallium-drivers, because they should have about the same performance as the old fglrx-drivers for your card, but... they aren't stable yet.

snack · 2011-02-20 11:28:25

ChemBro wrote:

And/or use it with the gallium-drivers, because they should have about the same performance as the old fglrx-drivers for your card, but... they aren't stable yet.

On a laptop I would say that power consumption could be as important as performance if not more, so catalyst is to be preferred over xf86-video-ati. But I agree that downgrading and maintaining such a system is going to be a nightmare...

dcc24 · 2011-02-20 11:31:55

snack wrote:

ChemBro wrote:
And/or use it with the gallium-drivers, because they should have about the same performance as the old fglrx-drivers for your card, but... they aren't stable yet.
On a laptop I would say that power consumption could be as important as performance if not more, so catalyst is to be preferred over xf86-video-ati. But I agree that downgrading and maintaining such a system is going to be a nightmare...

Actually, on my system (HD4330) battery life differs at most 15 minutes between catalyst and xf86-video-ati. Especially in the past few months, power management improved significantly with the open source drivers.

snack · 2011-02-20 11:35:54

dcc24 wrote:

Actually, on my system (HD4330) battery life differs at most 15 minutes between catalyst and xf86-video-ati. Especially in the past few months, power management improved significantly with the open source drivers.

I gave it a try on my x86_64 laptop with HD 3470. I played a bit with various power management settings, but couldn't find something that has decent performances (ie., not requiring > 1 second in kwin to open a minimized window in the panel) with decent battery life. In my opinion, catalyst is still unmatched when both of this parameters are taken into account.

fzzbt · 2011-02-20 13:59:03

Has anyone with radeon HD68xx gotten the open source drivers working?
xorg crashes for me always with message "KMS required" even tough dmesg says KMS enabled.

codeRage · 2011-02-21 01:49:41

ChemBro wrote:

dcc24 wrote:
codeRage wrote:
My laptop has an older x1400 card for which I need an older kernel to support catalyst-old package. Is there anywhere I can find old kernel pkgbuilds?
Even if you do, it will prove extremely hard to maintain such a system. You'll need to downgrade almost everything and that isn't likely to be possible. You should stick with xf86-video-ati.
And/or use it with the gallium-drivers, because they should have about the same performance as the old fglrx-drivers for your card, but... they aren't stable yet.

Ok, I answered my own question Thanks for help

Last edited by codeRage (2011-02-21 02:02:38)

agapito · 2011-02-21 02:30:41

fzzbt wrote:

Has anyone with radeon HD68xx gotten the open source drivers working?
xorg crashes for me always with message "KMS required" even tough dmesg says KMS enabled.

You need kernel 2.6.38.

Vi0L0 · 2011-02-21 12:13:33

tntexplosivesltd wrote:

Can I delete these directories, or any sub-directories (e.g. src/ or src/archive_files/)?

Ofcourse, in fact you can always remove src and pkg dirs after building package. Well... maybe it's not such a good idea when building -git -svn packages, but still.

tntexplosivesltd wrote:

Also, how can I get the framebuffer when I boot to be my comp's native resolution (1366x768)?

Use vga switch: https://wiki.archlinux.org/index.php/GR … resolution

@GraveDigger & @fzzbt:
i dunno, i tried many times but i'm not able to reproduce your problem. You guys got something in common? Like same gfx?
Btw. we still got archive catalyst repositories, right now with catalyst >= 10.9 : wiki

hopper wrote:

are there any other workarounds available for fixing video stutter except nopat ?
nopat has no effect, I tried (all) other video output drivers they all have the same problem so it's not mplayer's fault. it turns out catalyst 11.1 has a built-in PAT support which is turned on when nopat is used

True, catalyst's built-in PAT is turned on when kernel's nopat is used, and afaik it's working like this from long long time. Does video also stutter with video acceleration turned on?

Vi0L0 · 2011-02-22 20:27:56

silent aur and noisy repo update:
i've been testing no_bkl.patch for a week on 2.6.38 rc5 kernel without problems, also with fresh rc6 it's fine - so i've updated all aur and repo packages with this patch.
It could be useful for those individuals who wants to run kernel without Big Kernel Lock now, or maybe in near future if only arch devs will pull out outcoming 2.6.38 without BKL.

yimm · 2011-03-12 09:05:15

dcc24 wrote:

Actually, on my system (HD4330) battery life differs at most 15 minutes between catalyst and xf86-video-ati. Especially in the past few months, power management improved significantly with the open source drivers.

https://bugs.freedesktop.org/show_bug.cgi?id=31000

dcc24 · 2011-03-12 09:11:57

yimm wrote:

dcc24 wrote:
Actually, on my system (HD4330) battery life differs at most 15 minutes between catalyst and xf86-video-ati. Especially in the past few months, power management improved significantly with the open source drivers.
https://bugs.freedesktop.org/show_bug.cgi?id=31000

That's weird. Maybe catalyst is soo underperforming for me, that the os drivers are relatively performing better.

snack · 2011-03-12 11:37:21

yimm wrote:

dcc24 wrote:
Actually, on my system (HD4330) battery life differs at most 15 minutes between catalyst and xf86-video-ati. Especially in the past few months, power management improved significantly with the open source drivers.
https://bugs.freedesktop.org/show_bug.cgi?id=31000

I tested the open driver recently and I got similar results on my Mobility HD 3470. Power menagement has been greatly improved, and even if temperatures are ok battery life is still greatly reduced wrt Catalyst. I hope some more work will be done in this field even if the RadeonWiki page marks power management as DONE.

DarksideEE7 · 2011-03-13 11:26:09

Is anyone having graphical glitches with flash with KDE/kwin with compositing enabled? This began a few weeks ago. Basically when I go to view a youtube video (only happens on youtube so far, other flash video players are fine) and scroll down the page, the video panel 'stretches' along the scroll path.

Also if I open another tab the video will 'bleed through' and cover portions of that page. I've been able to recreate it in other browsers (Firefox and Opera). I can disable compositing with a hotkey and this fixes it.

I do not experience this on my laptop with Intel graphics, so I'm inclined to believe that it's ATI related.....

This is quite annoying, so if anyone has any idea how to fix it that would be great.

Last edited by DarksideEE7 (2011-03-13 11:26:59)

Vi0L0 · 2011-03-13 14:26:19

DarksideEE7 wrote:

Is anyone having graphical glitches with flash with KDE/kwin with compositing enabled? This began a few weeks ago. Basically when I go to view a youtube video (only happens on youtube so far, other flash video players are fine) and scroll down the page, the video panel 'stretches' along the scroll path.
Also if I open another tab the video will 'bleed through' and cover portions of that page. I've been able to recreate it in other browsers (Firefox and Opera). I can disable compositing with a hotkey and this fixes it.

Same here, i noticed it happens only on youtube, and only on direct video pages (those with /watch?v=). When yt video is embedded on other external not-youtube page it's working fine, also youtube's users channels are ok. The bug exist no matter what browser i will use if kwin is enabled.

Looks like kwin in kde 4.6.1 was changed in a way we didn't want.
Btw kwin dev (only one man is working on it) is using nvidia card - maybe he change something to fit vdpau hardware acceleration better - but it's my shooting in the dark only.

I guess i tried everything and only disabling kwin helps - just like you said.
Another way is to use youtube's html5+webm (through youtube.com/html5) - but sadly not all videos were/are converted to webm format.

When i'll find more time i will search web (or ask on irc) to find similar problems, and eventually i will fill bug report. Anyhow i will wrote more about this here.

Edit: i just spoke with guys from #kde and they told me it's flash fault... and it's touching also other drivers, not only catalyst, but it could manifest in other way - like complete kwin crash after leaving fullscreen flash.

Last edited by Vi0L0 (2011-03-13 16:04:36)

DarksideEE7 · 2011-03-13 20:19:39

Vi0L0 wrote:

DarksideEE7 wrote:
Is anyone having graphical glitches with flash with KDE/kwin with compositing enabled? This began a few weeks ago. Basically when I go to view a youtube video (only happens on youtube so far, other flash video players are fine) and scroll down the page, the video panel 'stretches' along the scroll path.
Also if I open another tab the video will 'bleed through' and cover portions of that page. I've been able to recreate it in other browsers (Firefox and Opera). I can disable compositing with a hotkey and this fixes it.
Same here, i noticed it happens only on youtube, and only on direct video pages (those with /watch?v=). When yt video is embedded on other external not-youtube page it's working fine, also youtube's users channels are ok. The bug exist no matter what browser i will use if kwin is enabled.
Looks like kwin in kde 4.6.1 was changed in a way we didn't want.
Btw kwin dev (only one man is working on it) is using nvidia card - maybe he change something to fit vdpau hardware acceleration better - but it's my shooting in the dark only.
I guess i tried everything and only disabling kwin helps - just like you said.
Another way is to use youtube's html5+webm (through youtube.com/html5) - but sadly not all videos were/are converted to webm format.
When i'll find more time i will search web (or ask on irc) to find similar problems, and eventually i will fill bug report. Anyhow i will wrote more about this here.
Edit: i just spoke with guys from #kde and they told me it's flash fault... and it's touching also other drivers, not only catalyst, but it could manifest in other way - like complete kwin crash after leaving fullscreen flash.

Glad to see I'm not the only one. Countless google searches have been fruitless so far.

I've definitely experienced entire kwin crashes and/or lockups with a grey screen after attempting to leave full screen with esc key instead of clicking on the GUI control. To fix this I can press ctrl+alt+F{1-6}, then back to X using ctrl+alt+F7.

The funny thing is th the more complicated flash video interfaces (such as huludesktop) work nearly perfect, yet the most common ones (like youtube) experience this.

If you need some input and/or another source for the bug tracker let me know and I'll submit some info.

marcos · 2011-03-14 09:53:56

Hi,

I'm having some troubles installing the ati driver from the AUR package. My card is a Radeon HD3600 and I'm running a 2.6.37-ARCH kernel.

I'm using catalyst-utils and catalyst packages. The first installs without problem, but when I try to install catalyst I get this:

marcos@duvel:10:45:26:~/AUR/ati/catalyst 
$ makepkg -sci
==> Making package: catalyst 11.2-1 (Mon Mar 14 10:45:30 CET 2011)
==> Checking Runtime Dependencies...
==> Checking Buildtime Dependencies...
==> Retrieving Sources...
  -> Found ati-driver-installer-11-2-x86.x86_64.run
  -> Found ati_make.sh
  -> Found makefile_compat.patch
  -> Found 2.6.38_console.patch
  -> Found no_bkl.patch
==> Validating source files with md5sums...
    ati-driver-installer-11-2-x86.x86_64.run ... Passed
    ati_make.sh ... Passed
    makefile_compat.patch ... Passed
    2.6.38_console.patch ... Passed
    no_bkl.patch ... Passed
==> Extracting Sources...
==> Starting build()...
==> Setting compilation variables ...
OsVersion says: SMP=1
file /proc/kallsyms says: SMP=1
file /lib/modules/2.6.37-ARCH/build/include/generated/autoconf.h says: SMP=1
file /lib/modules/2.6.37-ARCH/build/include/generated/autoconf.h says: MODVERSIONS=
file /lib/modules/2.6.37-ARCH/build/arch/x86/include/asm/compat.h says: COMPAT_ALLOC_USER_SPACE=arch_compat_alloc_user_space
==> Unpacking archive ...
Creating directory archive_files
Verifying archive integrity... All good.
Uncompressing ATI Catalyst(TM) Proprietary Driver-8.821..................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................
==> Applying patch(es) ...
patching file common/lib/modules/fglrx/build_mod/2.6.x/Makefile
patching file common/lib/modules/fglrx/build_mod/firegl_public.c
==> Compiling fglrx module ...
make: Entering directory `/usr/src/linux-2.6.37-ARCH'
  CC [M]  /home/marcos/AUR/ati/catalyst/src/archive_files/common/lib/modules/fglrx/build_mod/firegl_public.o
  CC [M]  /home/marcos/AUR/ati/catalyst/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_acpi.o
  CC [M]  /home/marcos/AUR/ati/catalyst/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_agp.o
  CC [M]  /home/marcos/AUR/ati/catalyst/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_debug.o
  CC [M]  /home/marcos/AUR/ati/catalyst/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_ioctl.o
  CC [M]  /home/marcos/AUR/ati/catalyst/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_io.o
  CC [M]  /home/marcos/AUR/ati/catalyst/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_pci.o
  CC [M]  /home/marcos/AUR/ati/catalyst/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_str.o
  CC [M]  /home/marcos/AUR/ati/catalyst/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_wait.o
  LD [M]  /home/marcos/AUR/ati/catalyst/src/archive_files/common/lib/modules/fglrx/build_mod/fglrx.o
  Building modules, stage 2.
  MODPOST 0 modules
make: Leaving directory `/usr/src/linux-2.6.37-ARCH'
==> Entering fakeroot environment...
==> Starting package()...
install: cannot stat `fglrx.ko': No such file or directory
    Aborting...

Have already discussed it with Vi0L0 a time ago but I had no time to look over it again till now.

Hope somebody can shed some light over it

Thanks!

agapito · 2011-03-17 09:39:36

Is catalyst working fine with 2.6.38? In my case desktop effects are a bit slower.

hd5770

Vi0L0 · 2011-03-17 14:10:28

@agapito:
it is working fine for me - although from [testing] i took only new kernel and syslog-ng.

Sidenote: i noticed (no matter what kernel im using) that kwin's Blur effect is making desktop slower, also if it's enabled i can see some bugs on shadows - like no shadows or only partial shadows.

xephinx · 2011-03-17 18:31:45

Is there a workaround for this:

sudo pacman -Syu                                                                                                            
:: Synchronizing package databases...                                                                                                   
 testing is up to date                                                                                                                  
 catalyst is up to date                                                                                                                 
 core is up to date                                                                                                                     
 extra is up to date                                                                                                                    
 community-testing is up to date                                                                                                        
 community is up to date                                                                                                                                                                                                                   
:: The following packages should be upgraded first :                                                                                    
    kernel26                                                                                                                    
:: Do you want to cancel the current operation                                                                                          
:: and upgrade these packages now? [Y/n] y                                                                                              
                                                                                                                                        
resolving dependencies...                                                                                                               
looking for inter-conflicts...                                                                                                          
error: failed to prepare transaction (could not satisfy dependencies)                                                                   
:: catalyst: requires kernel26<2.6.38

thanks

Vi0L0 · 2011-03-17 18:56:20

@xephinx:
until kernel26 won't be updated to 2.6.38 in [core] i won't update catalyst package - because ofcourse it will cause damage to non-[testing] users.
catalyst package was and is build always for kernel from [core] repo.

A workaround for [testing] users is to remove catalyst and use catalyst-hook or catalyst-daemon or catalyst-generator package.
All those packages are able to build fglrx kernel module(s) for kernel(s) installed on your system.
If you decide to use catalyst-generator please remember to type

catalyst_build_module all

after installing this package (or after installing new kernel).
catalyst-hook or catalyst-daemon will build fglrx module automatically for booted kernel.
+ catalyst-hook will build this module while installing/updating/downgrading kernel
+ catalyst-daemon will build this module while system start

For more informations please visit wiki: https://wiki.archlinux.org/index.php/ATI_Catalyst - chapter 2.3 and 2.4

Last edited by Vi0L0 (2011-03-17 19:02:27)

devadittya · 2011-03-18 18:45:33

DarksideEE7 wrote:

Glad to see I'm not the only one. Countless google searches have been fruitless so far.
I've definitely experienced entire kwin crashes and/or lockups with a grey screen after attempting to leave full screen with esc key instead of clicking on the GUI control. To fix this I can press ctrl+alt+F{1-6}, then back to X using ctrl+alt+F7.
The funny thing is th the more complicated flash video interfaces (such as huludesktop) work nearly perfect, yet the most common ones (like youtube) experience this.
If you need some input and/or another source for the bug tracker let me know and I'll submit some info.

Same here. I recently reinstalled Arch, moving to 32 bit. I spend a lot of time on youtube, and the 64 bit flash plugin (and 32bit multilib flash) are completely unusable for me. Thus, I moved to 32 bit. To my horror, absolutely nothing would work (videos, youtube..) I started a thread as well (https://bbs.archlinux.org/viewtopic.php?id=115218), but didn't get any replies.

I was randomly lurking on this thread, and read your post. Switched to Compiz, and EVERYTHING works! I have always had issues with KWin. I am still waiting for a fully functional KWin

Vi0L0 · 2011-03-19 13:34:22

MOTD: Unsupported (by catalyst) xserver 1.10 is in [testing]!
I have updated [xorg19] repo with newer packages from [extra], and probably in the days i will update it with xserver 1.9.5 (released 2 days ago) which seems to work fine.
Unfortunatelly that's all i can do atm.

@DarksideEE7 & @devadittya:
today i got more time so i could prepare some bug report... but first the story
A long time ago in a galaxy far, far away...
i tested how youtube flash works with oss radeon driver and there it's working fine - so it seems that this particular bug is related only to catalyst and flash (well and ofcourse kde's 4.6.1 kwin).
I would say that the most appropriate place where i should place bug report is kde's bugtracker - we all know that bugs in open sorce soft are fixed in fastest way - although a couple of days ago when i was speaking with kde devs on #kde they told me that they don't care about buggy flash...
That's too bad because now i have to place bug report on ati or flash bugtracker - and we know how much time we would have to wait for fix: months...
Now i have to pick between ati or flash bugreport... i guess i will pick ati one.
So as soon as i will wrote bug report i will place url to it here.

Edit: Bug report is here: http://ati.cchtml.com/show_bug.cgi?id=85

Last edited by Vi0L0 (2011-03-19 18:55:40)

Vi0L0 · 2011-03-22 13:50:33

@DarksideEE7 & @devadittya:
is youtube working good for you right now?
Cuz it's fine here...

I just dunno how this happened - in recent pacman's updates i can't see anything that could be related, also im not using [testing]... Strange.

Last edited by Vi0L0 (2011-03-22 13:50:55)

trapanator · 2011-03-22 14:02:15

I have a radeon hd3200 and I use opensource drivers (I hate to download 135MB official drivers...).

I have KDE 4.6.1 with kwin effects disabled (I have an Asus 1201T netbook), but when I upgraded to X.org 1.10 I see graphic glitches when I open a menu: before the menu is drawn, in its area there are graphic glitches, but after a 1 sec I see the menu...

Perry3D · 2011-03-22 20:34:23

Hi trapanator,
we need some input:

dmesg |grep -i drm

and

cat /var/log/Xorg.0.log

Arch Linux

#2826 2011-02-20 09:12:18

Re: The AMD/ATI Bar & Grill

#2827 2011-02-20 11:28:25

Re: The AMD/ATI Bar & Grill

#2828 2011-02-20 11:31:55

Re: The AMD/ATI Bar & Grill

#2829 2011-02-20 11:35:54

Re: The AMD/ATI Bar & Grill

#2830 2011-02-20 13:59:03

Re: The AMD/ATI Bar & Grill

#2831 2011-02-21 01:49:41

Re: The AMD/ATI Bar & Grill

#2832 2011-02-21 02:30:41

Re: The AMD/ATI Bar & Grill

#2833 2011-02-21 12:13:33

Re: The AMD/ATI Bar & Grill

#2834 2011-02-22 20:27:56

Re: The AMD/ATI Bar & Grill

#2835 2011-03-12 09:05:15

Re: The AMD/ATI Bar & Grill

#2836 2011-03-12 09:11:57

Re: The AMD/ATI Bar & Grill

#2837 2011-03-12 11:37:21

Re: The AMD/ATI Bar & Grill

#2838 2011-03-13 11:26:09

Re: The AMD/ATI Bar & Grill

#2839 2011-03-13 14:26:19

Re: The AMD/ATI Bar & Grill

#2840 2011-03-13 20:19:39

Re: The AMD/ATI Bar & Grill

#2841 2011-03-14 09:53:56

Re: The AMD/ATI Bar & Grill

#2842 2011-03-17 09:39:36

Re: The AMD/ATI Bar & Grill

#2843 2011-03-17 14:10:28

Re: The AMD/ATI Bar & Grill

#2844 2011-03-17 18:31:45

Re: The AMD/ATI Bar & Grill

#2845 2011-03-17 18:56:20

Re: The AMD/ATI Bar & Grill

#2846 2011-03-18 18:45:33

Re: The AMD/ATI Bar & Grill

#2847 2011-03-19 13:34:22

Re: The AMD/ATI Bar & Grill

#2848 2011-03-22 13:50:33

Re: The AMD/ATI Bar & Grill

#2849 2011-03-22 14:02:15

Re: The AMD/ATI Bar & Grill

#2850 2011-03-22 20:34:23

Re: The AMD/ATI Bar & Grill

Board footer