AMD Catalyst 9.8 Delivers New Kernel Support

[phoronix]

Phoronix: AMD Catalyst 9.8 Delivers New Kernel Support

It's taken quite a while, but AMD has finally delivered support for the Linux 2.6.29 kernel within their Catalyst driver. The Linux 2.6.29 kernel was released in March, but it has taken until today for AMD to release any support for this kernel...

http://www.phoronix.com/vr.php?view=NzQ2Mw

[blindfrog]

Great! Finally I can switch to F11 repos and do yum upgrade :)

[RealNC]

Seems to work OK in Gentoo with kernel 2.6.30-r5. Resizing/maximizing problem still there. The mplayer (gl backend) GUI corruption in KDE4 with desktop effects enabled problems seem to have been fixed though; haven't seen any menus and buttons in SMPlayer going missing/replaced with garbage yet.

[rafi]

No change with Xv+Composite, but one thing is cool. I ran Street Fighter 4 benchmark with wine 1.1.26. Before my computer froze when I changed the resolution of the test, now it is ok. Still not playable (image is flickering) but it is a good step forward.

[d2kx]

No change with Xv+Composite, but one thing is cool. I ran Street Fighter 4 benchmark with wine 1.1.26. Before my computer froze when I changed the resolution of the test, now it is ok. Still not playable (image is flickering) but it is a good step forward.

What problems do you experience with Xv+Composite?

[rafi]

Well, like with the 9.7, the image get blurred when the window of the video player is on the left of the screen and the video is black after minimizing the video player (vlc and smplayer tested). To get the image back I just have to resize the window, so there is no real problem with Xv+composite, I just wanted to say that it didn't get better. Anyway the little improvement in wine is cool.

[gsacks]

Ho-hum. No XvBA support. So glad I switched to Nvidia for my HTPC. My 780g board is now relegated to the living room web browser 'net top'. Glad to see AMD making progress, but the lack of accelerated video support on par with VDPAU is really just sad at this point. Just sad. I'll say it again. Sad.

[cruiseoveride]

91.7mb? Don't you just love to hate AMD/ATi

[m4rgin4l]

Phoronix: AMD Catalyst 9.8 Delivers New Kernel Support

It's taken quite a while, but AMD has finally delivered support for the Linux 2.6.29 kernel within their Catalyst driver. The Linux 2.6.29 kernel was released in March, but it has taken until today for AMD to release any support for this kernel...

http://www.phoronix.com/vr.php?view=NzQ2Mw

After a couple of minor adjustments to my system to ditch the default Fedora setup for radeon (blacklist, mkinitrd, etc.) I managed to install the driver and it worked for almost A WHOLE MINUTE before hanging the system. NICE. FANTASTIC. AWESOME.

Great work guys, thanks!

[Tares]

Okay, I've done fast tests comparing 9.7 and 9.8, here are the results :
glxgears :
-------
fglrx 9.7 :
53380 frames in 5.0 seconds = 10675.902 FPS
52212 frames in 5.0 seconds = 10442.344 FPS
54455 frames in 5.0 seconds = 10890.948 FPS
54634 frames in 5.0 seconds = 10926.746 FPS
54642 frames in 5.0 seconds = 10928.370 FPS

fglrx 9.8 :
53251 frames in 5.0 seconds = 10650.068 FPS
53398 frames in 5.0 seconds = 10679.592 FPS
53512 frames in 5.0 seconds = 10702.357 FPS
53515 frames in 5.0 seconds = 10702.912 FPS
52923 frames in 5.0 seconds = 10584.525 FPS
-------

fgl_glxgears :
-------
fglrx 9.7 :
16354 frames in 5.0 seconds = 3270.800 FPS
14017 frames in 5.0 seconds = 2803.400 FPS
16523 frames in 5.0 seconds = 3304.600 FPS
14438 frames in 5.0 seconds = 2887.600 FPS
15332 frames in 5.0 seconds = 3066.400 FPS

fglrx 9.8 :
16568 frames in 5.0 seconds = 3313.600 FPS
16774 frames in 5.0 seconds = 3354.800 FPS
16844 frames in 5.0 seconds = 3368.800 FPS
15977 frames in 5.0 seconds = 3195.400 FPS
16838 frames in 5.0 seconds = 3367.600 FPS
-------

q3a :
-------
fglrx 9.7 :
840.0 fps
772.1 fps
833.3 fps
807.3 fps
802.5 fps

fglrx 9.8 :
660.1 fps
751.4 fps
825.7 fps
705.5 fps
537.5 fps
-------

hon :
-------
fglrx 9.7 :
~60 in game

fglrx 9.8 :
~80 in game :
-------

tropics 1.2 aax2 :
-------
fglrx 9.7 :
30.8 fps
776 score

fglrx 9.8 :
38.0 fps
957 score
-------

Okay, but don't get your hopes too high :/ although tropics improved quite a lot, but that because I've got errors in rendering :/ whole benchmark was missing water and screen was flickering before I've turned aa on :( so something is wrong with this drivers.

As for hon, well i've measured it durning a game play, so it's not really reliable because I wasn't watching the fps all the time, but I think there is a little improvement.

My system and rig :
ubuntu 9.04 amd64
E6550@3.15GHz
HD4850@stock
4GB DDR2-900

Can anyone test unigine tropics ? I'll try sanctuary now.

EDIT :
okay, I've tested few things with tropics and found out that first run gives me like 0-4fps and screen corruption (blue and black lanes). But when I change to aax2, everything is looking fine, only water is missing... BUT... when I turn on aax4 or more everything is fine ;o no screen flickering, no corruption, water is perfect... ;x it has to be driver thing :x

[RealNC]

91.7mb? Don't you just love to hate AMD/ATi274MB unpacked :D

[RealNC]

lol, I just got a feed item from their feed about the new release. I try to open it:

"File not found".

Extraordinary support.

[plastikman]

Great! Finally I can switch to F11 repos and do yum upgrade :)

Don't count on it...looks like FC11 + DRI2 + FGLRX = FTL

Each time I use FGLRX on FC11 X86_64 I get hard locks after using Gnome for about 1-10 minutes (nothing in the logs).


Does any one have better experience?

I am using an HP 6930p with RadeonHD 3650

[bridgman]

I probably don't need to say this, but just wanted to make sure you *were* disabling kernel modesetting before running fglrx.

[plastikman]

I probably don't need to say this, but just wanted to make sure you *were* disabling kernel modesetting before running fglrx.

AFAIK KMS is not working. but i do not have it specifically disabled nor am I really sure how one would go about disabling. :/

BTW bridgman I drive by the AMD building in Sunnyvale everyday on the way to work :)

Edit:

I do have the following in my blacklist and a recompiled INITRD

# Radeon and DRM
blacklist drm
blacklist i2c_core
blacklist i2c_algo_bit
blacklist radeon

[bridgman]

Yeah, after I posted that I remembered that fglrx only supports 6xx and up which don't have KMS in F11.

[plastikman]

Yeah, after I posted that I remembered that fglrx only supports 6xx and up which don't have KMS in F11.

hehe no worries.

I realize that FC11 is unsupported but I am curious to know what the issue is. Maybe it is not a DRI2 issue and I just think that because on the same HW in FC10 FGLRX works but in FC11 no love. I am also happy to hear that i am not the only one that has this issue.

[Death Knight]

There is some change with xv on 4850 card. I got cropped XV output before. But now, movies are not cropped. Computer "freezing" for 2-3 seconds some times, except that its good release I think.

[nanonyme]

Yeah, I don't think KMS gets enabled by default yet in F11 for r6xx/r7xx. It does in current Rawhide though so you gotta play pretty careful. :)

[Technojunkie]

Anyone else having trouble getting 9.8 to compile on 64-bit F11 with kernel-2.6.29.6-217.2.7.fc11.x86_64 ? My fglrx-install.log file says:

[Message] Kernel Module : Trying to install a precompiled kernel module.
[Message] Kernel Module : Precompiled kernel module version mismatched.
[Message] Kernel Module : Found kernel module build environment, generating kernel module now.
AMD kernel module generator version 2.1
doing Makefile based build for kernel 2.6.x and higher
rm -rf *.c *.h *.o *.ko *.GCC* .??* *.symvers
make -C /lib/modules/2.6.29.6-217.2.7.fc11.x86_64/build SUBDIRS=/lib/modules/fglrx/build_mod/2.6.x modules
make[1]: Entering directory `/usr/src/kernels/2.6.29.6-217.2.7.fc11.x86_64'
CC [M] /lib/modules/fglrx/build_mod/2.6.x/firegl_public.o
CC [M] /lib/modules/fglrx/build_mod/2.6.x/kcl_acpi.o
/lib/modules/fglrx/build_mod/2.6.x/kcl_acpi.c:431:6: warning: #warning "ACPI notification wrapping won't work."
/lib/modules/fglrx/build_mod/2.6.x/kcl_acpi.c:464:6: warning: #warning "ACPI notification wrapping won't work."
/lib/modules/fglrx/build_mod/2.6.x/kcl_acpi.c:490:6: warning: #warning "ACPI notification wrapping won't work."
/lib/modules/fglrx/build_mod/2.6.x/kcl_acpi.c:514:6: warning: #warning "ACPI notification wrapping won't work."
CC [M] /lib/modules/fglrx/build_mod/2.6.x/kcl_agp.o
CC [M] /lib/modules/fglrx/build_mod/2.6.x/kcl_debug.o
CC [M] /lib/modules/fglrx/build_mod/2.6.x/kcl_ioctl.o
CC [M] /lib/modules/fglrx/build_mod/2.6.x/kcl_io.o
CC [M] /lib/modules/fglrx/build_mod/2.6.x/kcl_pci.o
CC [M] /lib/modules/fglrx/build_mod/2.6.x/kcl_str.o
CC [M] /lib/modules/fglrx/build_mod/2.6.x/kcl_wait.o
LD [M] /lib/modules/fglrx/build_mod/2.6.x/fglrx.o
Building modules, stage 2.
MODPOST 1 modules
CC /lib/modules/fglrx/build_mod/2.6.x/fglrx.mod.o
LD [M] /lib/modules/fglrx/build_mod/2.6.x/fglrx.ko
make[1]: Leaving directory `/usr/src/kernels/2.6.29.6-217.2.7.fc11.x86_64'
build succeeded with return value 0
duplicating results into driver repository...
done.
You must change your working directory to /lib/modules/fglrx
and then call ./make_install.sh in order to install the built module.
- recreating module dependency list
- trying a sample load of the kernel modules
failed.
[Error] Kernel Module : Failed to install compiled kernel module - please consult readme.

Hardware: Phenom II, AMD780G (Radeon HD 3200).

[Erikina]

Thanks for the release.

In light of all the problems people are having, I'll wait till the rpmfusion guys have a go at packaging it.

[RealNC]

Still "not found" in the link to the release notes. IMO just shows how much attention they pay to the Linux release. It's not the first time they don't release properly.

[L33F3R]

91.7mb? Don't you just love to hate AMD/ATi

and? all of a sudden size matters? ;)

[energyman]

Still "not found" in the link to the release notes. IMO just shows how much attention they pay to the Linux release. It's not the first time they don't release properly.

release notes always lag a couple of hours behind because they are done by different people. Now would you calm down? The drivers don't stop working magically without that notes.

[phtpht]

Does it already work well with preemptive kernel? (not voluntary) Preemptive RCU? And big kernel lock?

[AdrenalineJunky]

Still "not found" in the link to the release notes. IMO just shows how much attention they pay to the Linux release. It's not the first time they don't release properly.

wow, your not looking for a reason to hate amd.

the windows link also became active before the release - who cares?

[poofyyoda]

New kernel support is good. Other than that though, I haven't noticed anything else improved..

RandR is still screwed for analogue monitors (since 9.6), spawning new X is crash provoking (since 9.7).

I seriously hope that these issues get fixed sometime, especially the crashing one as it just makes the driver completely useless. If a can't spawn another X server then I can't use my computer for much productivity anymore.


Well, at least the open drivers are progressing quickly anyway

[phtpht]

New kernel support is good. Other than that though, I haven't noticed anything else improved..

RandR is still screwed for analogue monitors (since 9.6), spawning new X is crash provoking (since 9.7).

I seriously hope that these issues get fixed sometime,

It looks like they barely catch with the NEW stuff, and our hopes for getting the OLD stuff working are grim. I dont mean to whine, but seriously, look at the absurd state of it: I bought a pretty expensive piece of hardware and my bug (http://ati.cchtml.com/show_bug.cgi?id=1592) is there, with no reply, rotting at the UNOFFICIAL bugzilla.

[Xipeos]

*sigh* Still doesn't work for me, and it's the same problem I've had for ages - black screen/freeze with no obvious error or warning messages (other than "unknown third party board (..)").
I guess gaming on linux will have to wait some more.

[NeoBrain]

Just installed 9.8, and I'm quite happy about the changes right now:
- minimizing fullscreen OpenGL windows (or games which do support minimizing) works fine again; as long as I have turned off compositing at least
- Nexuiz: bench1 FPS went from min/avg/max 18/33/74 to 16/37/93; Running Nexuiz in Wine runs at 13/33/84
- Wine:
  - slow initialization bug from 9.7 seems to be fixed
  - I can select Shader quality PC3 Low and High in Trackmania Nations Forever now (these settings don't work correctly though because I'm getting quite a few graphical glitches then...); running at normal quality (PC2), I always get > 60 fps

- Compiz alpha blur STILL does not work... C'mon focus blur works just fine, how hard can it be to fix alpha blurring?

Things left to test:
- xinit xterm -- :2 (oh well, would be cool if someone else could test it for me, don't have much time right now. Kept hardlocking my system since about 8.44)

Quite a good release for me, now please AMD, fix Compiz' blur plugin and this driver is perfect for me ;)

[shv1]

How about video flickering under compiz ?

[BlackStar]

- Compiz alpha blur STILL does not work... C'mon focus blur works just fine, how hard can it be to fix alpha blurring?

Actually, this should be trivial to fix if it was a simple shader incompatibility. However, for some reason only the plugin authors know, they decided to assemble the shader piece by piece through C code (strcpy essentially). The code would be trivial to fix if they followed standard shader techniques (e.g. a large shader string modified through conditional compilation), but as things stand the blur plugin is so convoluted it's next to impossible to modify, much less fix.

Edit: no video flickering here with either 9.7 or 9.8 (4850). Video is not vsynced, either.

[NeoBrain]

Actually, this should be trivial to fix if it was a simple shader incompatibility. However, for some reason only the plugin authors know, they decided to assemble the shader piece by piece through C code (strcpy essentially). The code would be trivial to fix if they followed standard shader techniques (e.g. a large shader string modified through conditional compilation), but as things stand the blur plugin is so convoluted it's next to impossible to modify, much less fix.


Oh, okay... Let's hope Compiz 0.9 is doing a better job then :(

[yoshi314]

just looking at issues reported in this whole thread makes me wonder whatare fglrx devs actually doing.

i've stopped using fglrx around 8.35 in 2007 and i'm interested in whether there was at least one solid fglrx/catalyst release since then.

[blindfrog]

Haha great it actually works with Fedora 11 for a 5 SECONDS until it freezes the whole system! :D :D :D Brilliant!

[energyman]

Does it already work well with preemptive kernel? (not voluntary) Preemptive RCU? And big kernel lock?

??? wtf?

emmm... bkl is in linux since it can do smp. So at least 2.0. <Preemptive kernel? of course! for ages.

[cheung]

How about video flickering under compiz ?


I still see video tear even without compiz.

system:debian sid
linux kernel 2.6.30-1-amd64
ATI HD4350.

What can I do?

[energyman]

don't use xv?

[Technojunkie]

ssh'ing into my F11 box and attempting to install 9.8 from the console proved more enlightening:

Installation complete!

*** glibc detected *** ./setup.data/bin/x86_64/setup: double free or corruption (fasttop): 0x00000000018314b0 ***

======= Backtrace: =========

/lib64/libc.so.6[0x3683c75a96]

./setup.data/bin/x86_64/setup[0x40a6b5]

./setup.data/bin/x86_64/setup[0x4128fb]

/lib64/libc.so.6(__libc_start_main+0xfd)[0x3683c1ea2d]

./setup.data/bin/x86_64/setup[0x40382a]

======= Memory map: ========

00400000-004bc000 r-xp 00000000 08:02 826582 /root/fglrx-install.Cgo5FZ/setup.data/b

in/x86_64/setup

005bb000-005c3000 rw-p 000bb000 08:02 826582 /root/fglrx-install.Cgo5FZ/setup.data/b

in/x86_64/setup

005c3000-005fc000 rw-p 005c3000 00:00 0

0178b000-01847000 rw-p 0178b000 00:00 0 [heap]

3683800000-368381f000 r-xp 00000000 08:02 1996628 /lib64/ld-2.10.1.so

3683a1e000-3683a1f000 r--p 0001e000 08:02 1996628 /lib64/ld-2.10.1.so

3683a1f000-3683a20000 rw-p 0001f000 08:02 1996628 /lib64/ld-2.10.1.so

3683c00000-3683d64000 r-xp 00000000 08:02 1996749 /lib64/libc-2.10.1.so

3683d64000-3683f64000 ---p 00164000 08:02 1996749 /lib64/libc-2.10.1.so

3683f64000-3683f68000 r--p 00164000 08:02 1996749 /lib64/libc-2.10.1.so

3683f68000-3683f69000 rw-p 00168000 08:02 1996749 /lib64/libc-2.10.1.so

3683f69000-3683f6e000 rw-p 3683f69000 00:00 0

3684000000-3684082000 r-xp 00000000 08:02 1996752 /lib64/libm-2.10.1.so

3684082000-3684282000 ---p 00082000 08:02 1996752 /lib64/libm-2.10.1.so

3684282000-3684283000 r--p 00082000 08:02 1996752 /lib64/libm-2.10.1.so

3684283000-3684284000 rw-p 00083000 08:02 1996752 /lib64/libm-2.10.1.so

3684400000-3684402000 r-xp 00000000 08:02 1996755 /lib64/libdl-2.10.1.so

3684402000-3684602000 ---p 00002000 08:02 1996755 /lib64/libdl-2.10.1.so

3684602000-3684603000 r--p 00002000 08:02 1996755 /lib64/libdl-2.10.1.so

3684603000-3684604000 rw-p 00003000 08:02 1996755 /lib64/libdl-2.10.1.so

3689000000-3689019000 r-xp 00000000 08:02 1996753 /lib64/libgcc_s-4.4.0-20090506.so.1

3689019000-3689219000 ---p 00019000 08:02 1996753 /lib64/libgcc_s-4.4.0-20090506.so.1

3689219000-368921a000 rw-p 00019000 08:02 1996753 /lib64/libgcc_s-4.4.0-20090506.so.1

3693a00000-3693a02000 r-xp 00000000 08:02 1430806 /lib64/libutil-2.10.1.so

3693a02000-3693c01000 ---p 00002000 08:02 1430806 /lib64/libutil-2.10.1.so

3693c01000-3693c02000 r--p 00001000 08:02 1430806 /lib64/libutil-2.10.1.so

3693c02000-3693c03000 rw-p 00002000 08:02 1430806 /lib64/libutil-2.10.1.so

7f41cae08000-7f41cae14000 r-xp 00000000 08:02 1996729 /lib64/libnss_files-2.10.1.so

7f41cae14000-7f41cb013000 ---p 0000c000 08:02 1996729 /lib64/libnss_files-2.10.1.so

7f41cb013000-7f41cb014000 r--p 0000b000 08:02 1996729 /lib64/libnss_files-2.10.1.so

7f41cb014000-7f41cb015000 rw-p 0000c000 08:02 1996729 /lib64/libnss_files-2.10.1.so

7f41cb034000-7f41cb03b000 r--s 00000000 08:02 360019 /usr/lib64/gconv/gconv-modules.cache

7f41cb03b000-7f41cb0bc000 rw-p 7f41cb03b000 00:00 0

7f41cb0bc000-7f41d018f000 r--p 00000000 08:02 2151024 /usr/lib/locale/locale-archive

7f41d018f000-7f41d0192000 rw-p 7f41d018f000 00:00 0

7f41d01b8000-7f41d01ba000 rw-p 7f41d01b8000 00:00 0

7fff445b2000-7fff445cb000 rw-p 7ffffffe6000 00:00 0 [stack]

7fff445ff000-7fff44600000 r-xp 7fff445ff000 00:00 0 [vdso]

ffffffffff600000-ffffffffff601000 r-xp 00000000 00:00 0 [vsyscall]

./ati-installer.sh: line 39: 6724 Aborted ./setup.data/bin/${ARCH_DIR}/setup 2> /dev/null

Removing temporary directory: fglrx-install.Cgo5FZ

[cheung]

don't use xv?


I know that, but there is really no other better option?:confused:

I can compile lastest open source radeon and drm driver, but it does not have 3D support.
I can't use radeon with 3D visual effect of compiz or Kwin.:(

[energyman]

use opengl instead of xv. Nicer picture, better colours and the cpu load is neglible. So what is the problem?

[RealNC]

I know that, but there is really no other better option?:confused:

Nope. However, gl video works quite OK, even for HD content. And it's tear-free except when it's windowed while using compositing (full screen while using compositing is OK.)

I can compile lastest open source radeon and drm driver, but it does not have 3D support.
I can't use radeon with 3D visual effect of compiz or Kwin.:(

Xv on the radeon driver should work OK. I think.

[d2kx]

Okay, but don't get your hopes too high :/ although tropics improved quite a lot, but that because I've got errors in rendering :/ whole benchmark was missing water and screen was flickering before I've turned aa on :( so something is wrong with this drivers.

The improved performance is for real, the OpenGL code has been optimized. But Michael's statement in the news, that "there is a fix for water rendering in the Unigine Tropics benchmark" does not apply to this release of the AMD Catalyst release but to... another.

[phtpht]

??? wtf?

emmm... bkl is in linux since it can do smp. So at least 2.0. <Preemptive kernel? of course! for ages.

well bkl can be preempted or not. the same goes for rcu, which you did not reply to.

and as for general preemption i recall fglrx DID NOT like it and DID spew shitty messages (http://www.google.com/search?hl=en&q=fglrx+smp_processor_id+preemptible&aq=f&oq=&aqi=) into dmesg (and maybe even worse things).

and by the way was the module_license problem resolved?

[energyman]

Symbol: PREEMPT_VOLUNTARY [=y] │
│ Prompt: Voluntary Kernel Preemption (Desktop) │
│ Defined at kernel/Kconfig.preempt:19 │
│ Depends on: <choice> │
│ Location: │
│ -> Processor type and features │
│ -> Preemption Model (<choice> [=y])

and preempt:rcu:
│ This option reduces the latency of the kernel by making certain │
│ RCU sections preemptible. Normally RCU code is non-preemptible, if │
│ this option is selected then read-only RCU sections become │
│ preemptible. This helps latency, but may expose bugs due to │
│ now-naive assumptions about each RCU read-side critical section │
│ remaining on a given CPU through its execution. │
│

so I stay with no. Because rcu bugs can result in data corrpution. I like the stuff on my harddisks.

[haplo602]

hmm, would be good to include opengl/glsl version in the driver news, so I have all on one page and don't need to dig around :) thanks

[Melcar]

Anyone else getting massive graphical distortions on the Tropics demo?

[RealNC]

Almost two days now and still "not found" for the release notes.

Can it be that nobody in AMD even gives a damn?

[phtpht]

Symbol: PREEMPT_VOLUNTARY [=y] │
│ Prompt: Voluntary Kernel Preemption (Desktop) │
│ Defined at kernel/Kconfig.preempt:19 │
│ Depends on: <choice> │
│ Location: │
│ -> Processor type and features │
│ -> Preemption Model (<choice> [=y])

and preempt:rcu:
│ This option reduces the latency of the kernel by making certain │
│ RCU sections preemptible. Normally RCU code is non-preemptible, if │
│ this option is selected then read-only RCU sections become │
│ preemptible. This helps latency, but may expose bugs due to │
│ now-naive assumptions about each RCU read-side critical section │
│ remaining on a given CPU through its execution. │
│

so I stay with no. Because rcu bugs can result in data corrpution. I like the stuff on my harddisks.

So you more or less confirm what I wrote earlier.

[phtpht]

Almost two days now and still "not found" for the release notes.

Can it be that nobody in AMD even gives a damn?

Let's just assume their team is terribly undermanned.

[kensai]

Good release, at last kernel 2.6.30 support, though 2.6.31 is about to be released. But hopefully 9.9 will have support for it because Ubuntu will be almost close to release by then. Still the leaked 64bit driver I was using was not from 9.8 maybe it was 9.9 or 9.10 beta.

[bridgman]

I think it was a very early 9.9, sort of a pre-alpha.

[PuckPoltergeist]

I'm missing the --xv-pixmap-buffer-type option from the leaked 8.65 version in aticonfig. :D

[Lory]

I have installed the driver, and the TV out is broken. Despite I have the

Option "TVFormat" "PAL-B"

line in my device section of my xorg.conf if I run

xrandr -screen 1 --prop

it gives tv_format:NTSC-JPN. And the image is flickering and there are no colors.

Moreover, after rollback to the 9.7 driver the problem persist. :(

Has anyone any idea?

[nanonyme]

Good release, at last kernel 2.6.30 support, though 2.6.31 is about to be released. But hopefully 9.9 will have support for it because Ubuntu will be almost close to release by then. Still the leaked 64bit driver I was using was not from 9.8 maybe it was 9.9 or 9.10 beta.I'd assume they ship a pre-release and backport the release, as before, if it's not ready in time.

[Tares]

Anyone else getting massive graphical distortions on the Tropics demo?

I've got the same thing. Try to change AA settings, it helps... sometimes.

[bridgman]

I have installed the driver, and the TV out is broken. Despite I have the

Option "TVFormat" "PAL-B"

line in my device section of my xorg.conf if I run

xrandr -screen 1 --prop

it gives tv_format:NTSC-JPN. And the image is flickering and there are no colors.

Moreover, after rollback to the 9.7 driver the problem persist. :(

Has anyone any idea?

I'm guessing that option came from some other driver.

AFAIK you normally set TV standards on fglrx using aticonfig. Run aticonfig --help and look for TV options.

I think you want something like aticonfig --tvf=PAL-B.

[Lory]

Just curious, where did you get that option from, maybe some other driver ?

from the catalyst driver manual (/usr/share/doc/ati/html/ for me)

AFAIK you normally set TV standards on fglrx using aticonfig. Run aticonfig --help and look for TV options.

Yes, I have tried but not working. Nor with catalyst 9.7.

[kensai]

I think it was a very early 9.9, sort of a pre-alpha.
Then 9.9 won't have kernel 2.6.31 support, 9.10 should. ;)

[Domi09]

Does anyone have a solution for the freeze situation on FC11?
I have installed the 9.8 driver on FC11 64-bit, it works fine until a program of mine tries to connect to the internet, pidgin for instance.
As soon as that happens, the whole system freezes. The open source radeon driver works flawlessly though.
I haven't found any solution for that problem. so I hope anyone of you DOES have a solution.

Looking forward :)

[PuckPoltergeist]

Does anyone have a solution for the freeze situation on FC11?
I have installed the 9.8 driver on FC11 64-bit, it works fine until a program of mine tries to connect to the internet, pidgin for instance.
As soon as that happens, the whole system freezes. The open source radeon driver works flawlessly though.
I haven't found any solution for that problem. so I hope anyone of you DOES have a solution.

Looking forward :)

Really hard freezes? SysRq keys don't work anymore? If you can't get any log message, it will be hard to debug. With gentoo unstable the driver works fine so far.

[Ant P.]

OK so I tried again to get this working and succeeded after turning on a bunch of unwanted IA32 crap in the kernel (which the gentoo ebuild helpfully doesn't mention at all) and downgrading from git Xorg.

So now I have 3D until the memory leaks force me to quit the program or suffer a crash. It feels like IE6-quality software, that's for sure.

[ObiWan]

Still "not found" in the link to the release notes. IMO just shows how much attention they pay to the Linux release. It's not the first time they don't release properly.

Here the typical AMD ReleaseNotes PDF:
http://www2.ati.com/drivers/linux/catalyst_98_linux.pdf

[Silverthorn]

The new randr feature works like a charm. With previous releases I had to disable randr 1.2 to be able to use dual screen.

3D seams like a downgrade.

Error message may appear while playing H264/VC1/mpeg2 media files with
Hardware Acceleration

I wonder what this really means :confused:

[PuckPoltergeist]

Here the typical AMD ReleaseNotes PDF:
http://www2.ati.com/drivers/linux/catalyst_98_linux.pdf

Still mention "Linux kernel 2.6 or above" as requirement. :mad:

[yesterday]

Fedora users shouldn't complain too much

Fedora uses a patched and ahead-of-stream graphics stack. Thus, if you are using Fedora please don't slag off the ATI team, and instead suggest to the Fedora team that be a little less cutting edge. If the Slack, Gentoo, and Arch users aren't encountering problems, it's fair to say that this is a successful release.

I say this as a (very) patient Fedora user

[poofyyoda]

Well its good that the release notes seem to cover a lot more known issues now. Lets hope that they're working on fixing them.

[m4rgin4l]

Fedora users shouldn't complain too much

Fedora uses a patched and ahead-of-stream graphics stack. Thus, if you are using Fedora please don't slag off the ATI team, and instead suggest to the Fedora team that be a little less cutting edge. If the Slack, Gentoo, and Arch users aren't encountering problems, it's fair to say that this is a successful release.

I say this as a (very) patient Fedora user

I'd call this a successful release if the Ubuntu users aren't having problems.

I don't agree with Fedora needing to be less cutting edge. I think it's fair to say that if the OSS driver eventually gets decent support across chipsets, it'll be the Fedora users the ones who'll be testing it first.

What it is absolutely true is that Catalyst and Fedora don't mix. Fedora is constantly pushing the envelope and Catalyst is targeted to a different audience; being very popular amongst retired people living in Florida :)

[energyman]

Still mention "Linux kernel 2.6 or above" as requirement. :mad:

and it works with 2.6.29+30 without patches.

So what is your problem?

[RealNC]

One thing I find funny in the release notes system requirements is "glibc 2.2 or 2.3". 2.2 was released in 2000, 2.3 in 2002.

We're at 2.10 today (which is 2009.)

Edit: here's another "diamond":

"Error message may appear while playing H264/VC1/mpeg2 media files with
Hardware Acceleration"

Uh, wut?

[plastikman]

ssh'ing into my F11 box and attempting to install 9.8 from the console proved more enlightening:

....snip...

./ati-installer.sh: line 39: 6724 Aborted ./setup.data/bin/${ARCH_DIR}/setup 2> /dev/null

Removing temporary directory: fglrx-install.Cgo5FZ


this normally happens when compiling with the radeon and DRM modules loaded..although it is not really worth installing as it will lock your system up within 20 seconds of Gnome loading up.

the joys of FC11

[Kitof]

Error message may appear while playing H264/VC1/mpeg2 media files with Hardware Acceleration

I wonder what this really means :confused:

Xvba support comming soon ?

Gbeauchesne's patch for Xvba support in libva is ready for 5 months (http://www.splitted-desktop.com/~gbeauchesne/libva/patches/) now, so where is xvba-video ?

[Domi09]

I have changed my operating system to ubuntu jaunty now. There was no problem to create debian packages of the new 9.8 driver and to install them. However I have massive problems with Compiz and Xv ...
When I open a window and I play music at the same time, the music hangs for a second. When I maximize the video window in Totem it needs over 2 seconds to react. Smplayer doesn't even play the video in fullscreen mode.

[nanonyme]

Gbeauchesne's patch for Xvba support in libva is ready for 5 months (http://www.splitted-desktop.com/~gbeauchesne/libva/patches/) now, so where is xvba-video ?Did they even say they're going to release it at all?

[Kitof]

Did they even say they're going to release it at all?

No, but why did they publish the results of benchs with a beta version of xvba-video ?

[nanonyme]

No, but why did they publish the results of benchs with a beta version of xvba-video ?I haven't heard of AMD/ATi releasing any such benchmarks.

[Nille]

I haven't heard of AMD/ATi releasing any such benchmarks.

Not direkt from AMD but http://www.phoronix.com/scan.php?page=news_item&px=NzM2OA

[Kitof]

I haven't heard of AMD/ATi releasing any such benchmarks.

I don't think Gwenole Beauchesne hacked AMD/ATI's servers to get the beta version of xvba-video ;)

[Nille]

I don't think Gwenole Beauchesne hacked AMD/ATI's servers to get the beta version of xvba-video ;)

Gwenolé Beauchesne is a developer working for a French company on various multimedia APIs who through agreements with AMD has been exposed to the X-Video Bitstream Acceleration API already. Gwenolé previously wrote the VA-API support for MPlayer / FFmpeg and even a VDPAU back-end for VA-API. Recently though he published Linux video decoding acceleration benchmarks.

-----------

[nanonyme]

I wonder if he broke his NDA by putting out benchmarks before the official support for XvBA was announced...

[nanonyme]

Not direkt from AMD but http://www.phoronix.com/scan.php?page=news_item&px=NzM2OADoesn't count. Unless AMD/ATi announced they're going to have XvBA support in the drivers, it's waste of time go around asking for it. As long as they haven't announced it, they can drop XvBA without even commenting on it, really...

[AdrenalineJunky]

Doesn't count. Unless AMD/ATi announced they're going to have XvBA support in the drivers, it's waste of time go around asking for it. As long as they haven't announced it, they can drop XvBA without even commenting on it, really...

truth.

i *think* bridgeman commented a while back that it had only been acknowledged by AMD for embedded systems.

[PuckPoltergeist]

and it works with 2.6.29+30 without patches.

So what is your problem?

It doesn't work with the 2.6.31 prereleases and won't work with comming 2.6.31. The listed requirements are still wrong.

[energyman]

since 2.6.31 has not been released yet, the list is entirely correct. Prerelases are releases. And with 2.6.31 will come 9.9.

[cutterjohn]

truth.

i *think* bridgeman commented a while back that it had only been acknowledged by AMD for embedded systems.I seem to remember that in some post some time ago as well.

Release notes: looks like they just copy & paste the same old "release" notes over and over and over... at least it's better than MSI which just returns NULL (they don't exist). (Look at the supported cards list...)

Well, I'm off to brave 9.8 at some point(may not feel up to the usual disappointment today), but fully expect to return and c&p my own list of bugs yet again...

[Koorac]

This release finally fixes the issue with random desktop freezes which I had with 9.6 and 9.7. Sadly now HoN hangs randomly (but nearly always when the level up animation is played). I guess back to 9.5 unless someone knows a way around the HoN freezes (maybe they only happen at some specific graphics settings?). But other than that a massive improvement over the last drivers. Hopefully 9.9 will finally be the next usable release.

[blabub]

Catalyst 9.8 is not bad at all but same probleme since much version:



Adapter 0 - ATI Radeon HD 4800 Series
Sensor 0: Temperature - 89.50 C

Adapter 1 - ATI Radeon HD 4800 Series
Sensor 0: Temperature - 84.50 C

My system is idle, when playing a game it hits 100 C. :rolleyes:

aticonfig --pplib-cmd "get fanspeed 0"
PPLIB command execution has failed!
ati_pplib_cmd: execute "get" failed!

aticonfig --pplib-cmd "get fanspeed 1"
PPLIB command execution has failed!
ati_pplib_cmd: execute "get" failed!


CrossFire chain for adapter 0, status: enabled
0. 01:00.0 ATI Radeon HD 4800 Series
1. 02:00.0 ATI Radeon HD 4800 Series


Same problem with setspeed. This problem is since I use crossfire.

[WhiteDwarf]

Sadly now HoN hangs randomly (but nearly always when the level up animation is played). I guess back to 9.5 unless someone knows a way around the HoN freezes (maybe they only happen at some specific graphics settings?).

Try disabling dynamic lights in the options (it'll likely be that option if any of them can make a difference).

[RealNC]

Catalyst 9.8 is not bad at all but same probleme since much version:

Same problem with setspeed. This problem is since I use crossfire.

On a 4870, 80 to 85C idle temp is normal. Full load is about 90-95C.

Since you're on Crossfire, higher temps than the above is normal; the cards are close together and heat up more.

[Koorac]

Try disabling dynamic lights in the options (it'll likely be that option if any of them can make a difference).
Sounds like a good idea, I'll try that tomorrow.

[cutterjohn]

Okay, I've done fast tests comparing 9.7 and 9.8, here are the results :
snipped results since they're useless as benchmarks,
but anyways for comparison's sakes mobility radeon 4850 512MB GDDR3
my glxgears FPS stabilize at c. 8200 FPS, and my fgl_glx gears at around 2300 FPS, this is a 500/850 core/mem R770.

I've done ZERO else ATM, and don't really have the heart to test prior bug states ATM to comment on those... no progs run under wine, no games under wine, no video, no sleep, no compiz/beryl, no sleep(power mgmt) testing etc. ad nauseum just plain old 2D unadorned GNOME/Ubuntu variant desktop which still functions OK.

GPU temps: For me they DO run HIGHER than under windows even at the basic windows(vista) performance setting.

e.g. typically unde rlinux even w/ clocks at 250/200 (core/mem) aticonfig reports 56.5C, while under windows it's more like 50C power saving, 52C balanced, 54(ish)C high performance. MSI's turbo battery mode(Vista, I suspect that they underclock but I've not investigated so long as things are working well as it's a pointless triviality) get's me down to 44-46C.

Never really recorded temps under linux while running games or other GPU intensive apps, so I won't bother to mention various peaks here...

[storma]

For the HoN freezes, try nopat at boot.

[blabub]

On a 4870, 80 to 85C idle temp is normal. Full load is about 90-95C.

Since you're on Crossfire, higher temps than the above is normal; the cards are close together and heat up more.

I´ve installed for testing shortly Windows 7. The temps are 65 to 70C and I can set the fan speed.

My problem is, that my case is nearly burning with the two cards (an usb cable was melting cause of this!) in it.

I would be so much happy when I can set the fan speed, I missing the noise of the fan. :rolleyes:


At all, this is still a bug:
aticonfig --pplib-cmd "get fanspeed 0"
PPLIB command execution has failed!
ati_pplib_cmd: execute "get" failed!

aticonfig --pplib-cmd "get fanspeed 1"
PPLIB command execution has failed!
ati_pplib_cmd: execute "get" failed!

Is there another way to set the fan speed?

[storma]

Is there another way to set the fan speed?

http://www.phoronix.com/forums/showthread.php?t=17331&highlight=overdrive

[Tares]

For the HoN freezes, try nopat at boot.

I've tried that, but I don't think it works here. I'm on ubuntu amd64 and dmesg didn't said anything about pat ;x

[Domi09]

Hi, I changed from fedora 11 to ubuntu, only to get the new ATI driver work, well it does, but it works very badly.
I do not know how so many of you can say with their cards, the driver works perfectly well, because with my card HD 4670 it definately doesn't.
When I idle for a while the X-Server freezes ...
When I open windows in full size, it freezes for a short time and videos are playbacked in a miserable way.
I tried to improve with adding "nopat" to my boot parameters and through Xorg.conf options, but no success.
Here's my current Xorg.conf, maybe someone knows what to do to get this problems solved.



Section "ServerLayout"
Identifier "aticonfig Layout"
Screen 0 "aticonfig-Screen[0]-0" 0 0
Option "AIGLX" "on"
EndSection

Section "Files"
EndSection

Section "Module"
EndSection

Section "Monitor"
Identifier "Configured Monitor"
EndSection

Section "Monitor"
Identifier "aticonfig-Monitor[0]-0"
Option "VendorName" "ATI Proprietary Driver"
Option "ModelName" "Generic Autodetecting Monitor"
Option "DPMS" "true"
EndSection

Section "Device"
Identifier "Configured Video Device"
EndSection

Section "Device"
Identifier "aticonfig-Device[0]-0"
Driver "fglrx"
BusID "PCI:1:0:0"
Option "TexturedVideo" "on"
Option "TexturedVideoSync" "True"
Option "OpenGLOverlay" "off"
Option "Textured2D" "on"
Option "TexturedXrender" "on"
Option "UseFastTLS" "1
EndSection

Section "Screen"
Identifier "Default Screen"
Device "Configured Video Device"
Monitor "Configured Monitor"
EndSection

Section "Screen"
Identifier "aticonfig-Screen[0]-0"
Device "aticonfig-Device[0]-0"
Monitor "aticonfig-Monitor[0]-0"
DefaultDepth 24
SubSection "Display"
Viewport 0 0
Depth 24
EndSubSection
EndSection

Section "DRI"
Mode 0666
EndSection

Section "Extensions"
Option "Composite" "Enable"
Option "RENDER" "Enable"
Option "DAMAGE" "Enable"
EndSection

[blabub]

http://www.phoronix.com/forums/showthread.php?t=17331&highlight=overdrive

Thanks for the link but:


Error reading fan speed value
Error reading fan speed value
Error reading fan speed value
Error reading fan speed value
Error reading fan speed value
Error reading fan speed value
Error reading fan speed value
Error reading fan speed value
Error reading fan speed value
Error reading fan speed value
Error reading fan speed value
Error reading fan speed value

[Lory]

I have installed the driver, and the TV out is broken. Despite I have the

Option "TVFormat" "PAL-B"

line in my device section of my xorg.conf if I run

xrandr -screen 1 --prop

it gives tv_format:NTSC-JPN. And the image is flickering and there are no colors.

Moreover, after rollback to the 9.7 driver the problem persist. :(

Has anyone any idea?

The solution was:

Remove the /etc/ati

reinstall the driver
then running aticonfig --tvf PAL-B

Now all is ok.

[Qaridarium]

Hi, I changed from fedora 11 to ubuntu, only to get the new ATI driver work, well it does, but it works very badly.
I do not know how so many of you can say with their cards, the driver works perfectly well, because with my card HD 4670 it definately doesn't.
[/CODE]

i have an 4670 to and it works!

Kubuntu 9.04 with kernel 2.6.30+KDE4.3

[RealNC]

The solution was:

Remove the /etc/ati

reinstall the driver
then running aticonfig --tvf PAL-B

Now all is ok.

This "delete /etc/ati" problem has been mentioned so many times already, but still the catalyst installer doesn't seem to care about it.