Announcement

Collapse
No announcement yet.

Drivers for AMD (ATI) Radeon HD 5670 in Ubuntu 12.04 LTS (Precise Pangolin).

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Drivers for AMD (ATI) Radeon HD 5670 in Ubuntu 12.04 LTS (Precise Pangolin).

    Hello,

    I've got problems with my AMD (ATI) Radeon HD 5670 in Ubuntu 12.04 LTS (Precise Pangolin).

    I had 2GB of RAM (DDR2, PC6400, 800MHz), I added another 2GB so now I have 4GB (2x2GB) and I'm not able to use fglrx drivers.

    I've applied patch for new kernel http://phoronix.com/forums/showthrea...86-32-bit-arch and installed latest Catalyst 12.4 drivers succesfully without any errors according to WIKI pages http://wiki.cchtml.com/index.php/Ubu...TI.27s_site.29 but unfortunately after system restart I get black or violet screen and I'm not even able to switch to other virtual console with Ctrl+Alt+F1...F6. When I remove 2GB (only 2GB left) system starts normally and everything works fine.

    It's not RAM issue, because I've got Windows 7 installed as well and it works fine, 4GB of memory are visible in the system. Besides when I installed Backtrack 5 which is based on Ubuntu 10.04 LTS, 12.4 catalyst drivers were working correctly with 4GB RAM inserted.


    All installations were made on fresh and updated system. After I install drivers and execute
    Code:
    sudo amdconfig --initial -f
    fglrxinfo output and xorg.conf (before this command there was no /etc/X11/xorg.conf) are as follows (right after installation and before system restart):
    Code:
    przemek@pc:~/catalyst12.4$ sudo amdconfig --initial -f
    Uninitialised file found, configuring.
    Using /etc/X11/xorg.conf
    Saving back-up to /etc/X11/xorg.conf.original-1
    przemek@pc:~/catalyst12.4$ cat /etc/X11/xorg.conf
    Section "ServerLayout"
        Identifier     "aticonfig Layout"
        Screen      0  "aticonfig-Screen[0]-0" 0 0
    EndSection
    
    Section "Module"
    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  "aticonfig-Device[0]-0"
        Driver      "fglrx"
        BusID       "PCI:1:0:0"
    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
    
    przemek@pc:~/catalyst12.4$ fglrxinfo
    display: :0  screen: 0
    OpenGL vendor string: VMware, Inc.
    OpenGL renderer string: Gallium 0.4 on llvmpipe (LLVM 0x300)
    OpenGL version string: 1.4 (2.1 Mesa 8.0.2)
    
    przemek@pc:~/catalyst12.4$
    I moved xorg.conf file to see if it causes problems, but after
    Code:
    sudo mv /etc/X11/xorg.conf /etc/X11/xorg.conf.disabled
    the problem was still present- black or violet screen.

    After xorg.conf restoration
    Code:
    sudo mv /etc/X11/xorg.conf.disabled /etc/X11/xorg.conf
    I tried to tune it according to wiki
    Code:
    sudo amdconfig --tls=0
    
    sudo amdconfig --acpi-services=off
    but without success...

    Blacklisting radeon modules also did not help...
    Code:
    echo "blacklist radeon" | sudo tee -a /etc/modprobe.d/blacklist-radeon.conf
    sudo depmod -a


    Ofcourse when I remove 2GB RAM and leave 2GB- system boots normally and fglrxinfo looks very good:
    Code:
    przemek@pc:~$ fglrxinfo
    display: :0  screen: 0
    OpenGL vendor string: Advanced Micro Devices, Inc.
    OpenGL renderer string: ATI Radeon HD 5670
    OpenGL version string: 4.2.11631 Compatibility Profile Context
    
    przemek@pc:~$ fgl_glxgears
    Using GLX_SGIX_pbuffer
    3178 frames in 5.0 seconds = 635.600 FPS
    4172 frames in 5.0 seconds = 834.400 FPS
    4163 frames in 5.0 seconds = 832.600 FPS
    4153 frames in 5.0 seconds = 830.600 FPS
    4165 frames in 5.0 seconds = 833.000 FPS
    4152 frames in 5.0 seconds = 830.400 FPS
    4141 frames in 5.0 seconds = 828.200 FPS
    4160 frames in 5.0 seconds = 832.000 FPS
    4162 frames in 5.0 seconds = 832.400 FPS
    ^C
    przemek@pc:~$
    Problem is present only when 4GB are inserted... I borrowed 1GB of RAM and tested configuration with 3GB and it was working fine!

    Used kernel supports 4GB of RAM (visible after free -m when fglrx drivers are not installed)
    Code:
    przemek@pc:~$ uname -r
    3.2.0-25-generic-pae
    przemek@pc:~$
    The same situation is with fglrx drivers from repos...


    What kind of logs can I check? Can anyone help me with this issue?

    Looking forward,
    Przemek.

    P.S. I tried also 12.6 Beta version of catalyst drivers, mentioned patch was not need to successfully generate packages, but after installation situation is still the same- black or violet screen when 4GB are inserted...

  • przemnet
    replied
    Problem SOLVED!

    Hello Everyone!

    Problem is SOLVED!

    As per recommendation from fglrx bugtracker http://ati.cchtml.com/show_bug.cgi?id=560 - BIOS upgrade on my motherboard (Gigabyte GA-MA69G-S3H) was the solution! I upgraded to the newest version which is F8E. I was able to install latest Catalyst 12.6 drivers succesfully.

    I had problems with upgradeing BIOS from Windows XP and Windows 7, but upgrade directly from BIOS with Q FLASH worked perfectly! To do it, just download BIOS from Gigabyte site, extract the archive, copy BIOS file to USB stick, plug USB stick into computer, reboot system, hit "End" button during startup and select BIOS upgrade from HDD- you can find lots of videos on YouTube about that.

    Many thanks for your help and support!

    Looking forward,
    Przemek.

    P.S. I wanted to edit my initial post to put solution there (to avoid going through full discussion to search the answer), but I can not do it...

    Leave a comment:


  • bridgman
    replied
    Originally posted by barkas View Post
    By the way, what should I expect when I write a support ticket at the AMD site? Should I get some answer? An automated email or something? A ticket number? Because a ticket without number is completely useless.
    The tracker assigns a number automatically when a new bug is entered. For example, przemnet's ticket was #560.

    Leave a comment:


  • przemnet
    replied
    Problem SOLVED!

    Hello Everyone!

    Problem is SOLVED!

    As per recommendation from fglrx bugtracker- BIOS upgrade on my motherboard (Gigabyte GA-MA69G-S3H) was the solution! I upgraded to the newest version which is F8E. I was able to install latest Catalyst 12.6 drivers succesfully.

    I had problems with upgradeing BIOS from Windows XP and Windows 7, but upgrade directly from BIOS with Q FLASH worked perfectly! To do it, just download BIOS from Gigabyte site, extract the archive, copy BIOS file to USB stick, plug USB stick into computer, reboot system, hit "End" button during startup and select BIOS upgrade from HDD- you can find lots of videos on YouTube about that.

    Many thanks for your help and support!

    Looking forward,
    Przemek.

    P.S. I wanted to edit my initial post to put solution there (to avoid going through full discussion to search the answer), but I can not do it...

    Leave a comment:


  • barkas
    replied
    Originally posted by Kano View Post
    The 12-6 final driver has got a lower version number than the driver packaged in U 12.4. Not optimal to install somehow. Basically that driver should work with hd 5670, but it is a pretty buggy driver overall. The 12.6 legacy (why 12.6 not 12-6?) is better but would have got a even lower version number - correct name would have been 12-5. The 12-6 drivers are both crap when tested with xserver 1.12 - they both need libpciaccess hacks on debian or with u updated to 1.12 via ppa.
    I know from experience that if a driver is newer, that does not have to mean it is better.

    EDIT: And that is exactly what I meant by a bug that is trivial to find. Reminds me of the time I beta tested Windows ME. I survived until the second beta, where the complete DNS resolving functionality was dead. No, I did not post that bug, they should have found it themselves. Never got a beta 3, go figure .

    If those bugs crop up, it simply means you have a huge QA problem.

    I realize AMD employees are not allowed to state that publicly though , but I expect them to know and to work at it.
    Last edited by barkas; 07-08-2012, 04:13 PM.

    Leave a comment:


  • barkas
    replied
    Originally posted by bridgman View Post
    It's perfectly true to say "the tracker is... not necessarily monitored" because it's a meaningless statement, but in fact the tracker *is* monitored and bugs *are* fixed.
    By the way, what should I expect when I write a support ticket at the AMD site? Should I get some answer? An automated email or something? A ticket number?

    Because a ticket without number is completely useless.
    Last edited by barkas; 07-08-2012, 04:16 PM.

    Leave a comment:


  • Kano
    replied
    The 12-6 final driver has got a lower version number than the driver packaged in U 12.4. Not optimal to install somehow. Basically that driver should work with hd 5670, but it is a pretty buggy driver overall. The 12.6 legacy (why 12.6 not 12-6?) is better but would have got a even lower version number - correct name would have been 12-5. The 12-6 drivers are both crap when tested with xserver 1.12 - they both need libpciaccess hacks on debian or with u updated to 1.12 via ppa.

    Leave a comment:


  • bridgman
    replied
    Originally posted by barkas View Post
    EDIT: I hope you know that the cchtml bugtracker is unofficial and not necessarily monitored by ati staff,
    It's perfectly true to say "the tracker is... not necessarily monitored" because it's a meaningless statement, but in fact the tracker *is* monitored and bugs *are* fixed.

    Leave a comment:


  • RussianNeuroMancer
    replied
    Originally posted by barkas View Post
    Next month you'll tell me to update to 12.7 because 12.6 is crap.
    I just tell you what I actually read and hear from people about 12.6 release.
    Originally posted by barkas View Post
    I'm not ATIs unpaid beta tester. The GPU in question has been supported for years, if they haven't managed to produce a working driver yet
    Doesn't work in your setup != doesn't work anywhere.
    Originally posted by barkas View Post
    EDIT: I hope you know that the cchtml bugtracker is unofficial and not necessarily monitored by ati staff
    Because of that I mention tech. support.
    Originally posted by barkas View Post
    as was made clear when the fglrx + gnome3 bugs were posted there.
    Gnome Shell topic was already discussed.
    Originally posted by barkas View Post
    Additionally most bugs with that driver are of the trivial to find kind and should have been caught by inhouse beta testing, if there is such a thing.
    For example you talk about some "trivial bugs" in this thread: screen is black on boot, kernel panic at shutdown, graphic corruption after VT switch, but I doesn't have this bugs (for obvious reason - different hardware). What make this bugs "trivial" if nobody properly report it?
    Or topic starter issue, that probably related to motherboard BIOS bug (or even maybe hardware failure as Kep P. assume) but not fglrx bug.
    Originally posted by barkas View Post
    Excuse me if I'm pissed off but I've been having those problems and the selfsame argument for going on 15 years now, first with the windows driver, now with windows and, worse, linux.
    And you doesn't report bugs to vendor for all 15 years? If you doesn't report bugs, don't be surprised is your bugs doesn't get fixed. There is no AMD telepathists department.
    Last edited by RussianNeuroMancer; 07-08-2012, 03:20 PM.

    Leave a comment:


  • barkas
    replied
    Next month you'll tell me to update to 12.7 because 12.6 is crap.

    I'm not ATIs unpaid beta tester. The GPU in question has been supported for years, if they haven't managed to produce a working driver yet, I don't believe one additional update will substantially change things. Next driver update for me may be with ubuntu 12.10. Certainly not sooner.

    EDIT: I hope you know that the cchtml bugtracker is unofficial and not necessarily monitored by ati staff, as was made clear when the fglrx + gnome3 bugs were posted there. Additionally most bugs with that driver are of the trivial to find kind and should have been caught by inhouse beta testing, if there is such a thing.
    Excuse me if I'm pissed off but I've been having those problems and the selfsame argument for going on 15 years now, first with the windows driver, now with windows and, worse, linux.
    Last edited by barkas; 07-08-2012, 02:45 PM.

    Leave a comment:

Working...
X