Announcement

Collapse
No announcement yet.

AMD Catalyst 13.1 For Linux Officially Does X.Org 1.13

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

  • AMD Catalyst 13.1 For Linux Officially Does X.Org 1.13

    Phoronix: AMD Catalyst 13.1 For Linux Officially Does X.Org 1.13

    Catalyst 13.1 for Linux was released on Thursday as the first AMD Linux binary blob of 2013. This driver is notable since it officially supports X.Org Server 1.13...

    http://www.phoronix.com/vr.php?view=MTI3NzU

  • #2

    Comment


    • #3
      This must be the end of the world or somethin' AMD published the release note! Unfortunately the number of bugs that are fixed is too small. With the previous version of Catalyst (9.00), X crashes unexpectedly on KDE (both 4.8.x and 4.9.x) so I downgraded Catalyst to 8.98. I hope they have fixed this annoying bug so far.

      Comment


      • #4
        Still no news about legacy drivers. -.-

        Comment


        • #5
          Originally posted by asim75 View Post
          Still no news about legacy drivers. -.-
          Was there supposed to be a new release for those too?

          Anyway, release notes! WOOHOOO!! Off to test this version.

          Comment


          • #6
            [367245]: X crash for AMD PowerXpress™*A+I High-Performance mode on Ubuntu 12.10
            Not helpful. What crash? With what software versions is it supposed to work?

            I tried it with an AMD Enduro notebook. "Downgraded" to all the stable versions, X.Org X Server 1.13.1, mesa 9.0.1 and xf86-video-intel 2.20.18.

            Didn't work. X start looks ok, until…
            Code:
            [   749.149] (EE) fglrx(0): Map intel primary surface to user space failed!
            [   749.149] 
            Fatal server error:
            [   749.149] AddScreen/ScreenInit failed for driver 0
            Any plans to make it work on current stable versions?

            Edit:
            1. Downgraded to xf86-video-intel 2.20.2, and X.org 1.12, no difference.
            2. Added
            Code:
            Option  "Ignore"  "true"
            to the Monitor section in the pxp amd xorg.conf. Difference!
            Code:
            [  2496.156] (II) fglrx(0): Output LVDS1 using monitor section aticonfig-Monitor[0]-0
            [  2496.156] (**) fglrx(0): Option "Ignore" "true"
            [  2496.228] (II) fglrx(0): Output VGA1 using monitor section aticonfig-Monitor[0]-0
            [  2496.300] (II) fglrx(0): Output HDMI1 using monitor section aticonfig-Monitor[0]-0
            [  2496.343] (II) fglrx(0): Output DP1 using monitor section aticonfig-Monitor[0]-0
            [  2496.344] (II) fglrx(0): Output HDMI2 using monitor section aticonfig-Monitor[0]-0
            [  2496.356] (II) fglrx(0): Output HDMI3 using monitor section aticonfig-Monitor[0]-0
            [  2496.396] (II) fglrx(0): Output DP2 using monitor section aticonfig-Monitor[0]-0
            [  2496.436] (II) fglrx(0): Output DP3 using monitor section aticonfig-Monitor[0]-0
            [  2496.436] (WW) fglrx(0): No outputs definitely connected, trying again...
            [  2496.436] (WW) fglrx(0): Unable to find connected outputs - setting 1024x768 initial framebuffer
            [  2496.436] (II) fglrx(0): Using default gamma of (1.0, 1.0, 1.0) unless otherwise stated.
            [  2496.436] (II) fglrx(0): Kernel page flipping support detected, enabling
            [  2496.436] (EE) fglrx(0): No modes.
            [  2496.436] (WW) fglrx(0): PreInit PowerXpress failed!(EE) 
            [  2496.436] (EE) Backtrace:
            [  2496.436] (EE) 0: /usr/bin/X (xorg_backtrace+0x36) [0x58a426]
            [  2496.436] (EE) 1: /usr/bin/X (0x400000+0x18e279) [0x58e279]
            [  2496.437] (EE) 2: /usr/lib/libpthread.so.0 (0x7fabdbc24000+0xf1e0) [0x7fabdbc331e0]
            [  2496.437] (EE) 3: /usr/lib/xorg/modules/drivers/fglrx_drv.so (xdl_xs113_atiddxFreeScreen+0xd4) [0x7fabd8e32114]
            [  2496.437] (EE) 4: /usr/bin/X (xf86DeleteScreen+0x84) [0x480c74]
            [  2496.437] (EE) 5: /usr/bin/X (InitOutput+0xc4e) [0x47a0fe]
            [  2496.437] (EE) 6: /usr/bin/X (0x400000+0x26786) [0x426786]
            [  2496.437] (EE) 7: /usr/lib/libc.so.6 (__libc_start_main+0xf5) [0x7fabda8a8a15]
            [  2496.437] (EE) 8: /usr/bin/X (0x400000+0x26cad) [0x426cad]
            [  2496.437] (EE) 
            [  2496.437] (EE) Segmentation fault at address 0x38
            [  2496.437] 
            Fatal server error:
            [  2496.437] Caught signal 11 (Segmentation fault). Server aborting
            Is there any documentation on what configuration is supposed to work?
            Last edited by ChrisXY; 01-18-2013, 07:02 AM.

            Comment


            • #7
              I've just updated ebuild for it on my Gentoo ( 64-bit on Phenom 955BE/HD6850 - 3 monitor setup) and replaced last 12.11-beta11

              A couple annoying bugs seem to have been fixed. Otherwise, no big differences. Not that I would even notice 3D performance changes as I'm not big gamer.

              Comment


              • #8
                No news about HD Audio bitstream?

                Comment


                • #9
                  One positive thing I noticed is that ground in triggerrally.com is no longer transparent, like it was with 12.10 drivers. Graphical glitches in Cinnamon are still there.

                  Another major pleasant change is that card now automatically clocks down, to 300 and 150 Mhz for Core and Memory, respectively, when there's no heavy load. With 12.10 drivers it was 500, 1250 minimum, and I used to get annoying fan noise.

                  Ubuntu 12.04, Radeon 7950.

                  Comment


                  • #10
                    Thank you, AMD, for making an effort to publish a Linux release log. I recently upgraded my AMD Radeon HD 4870 to a HD 7850 due to the legacy graphics drivers (for both Linux & Window). I wasn't please that my old card became "unsupported" after only 3 years (and I still am not), but I am happy that my new 7850 runs Arch + GnomeShell very smoothly. There's no difference between the framerate of the Unigine Heaven Tech Demo running on Windows 8 + DirectX 11 and Linux 3.6 + OpenGL, even at very high settings with Tesselation. Blender + Dyntopo is running very smooth as well. Awesome

                    Comment


                    • #11
                      Also, luxmark score is still at around 1200 compared to 1700 for W8 with 12.10 drivers, and 1800 for W8 beta drivers.

                      Comment


                      • #12
                        From the changelog it seems that they have fixed every last bug I encountered with the last beta version.

                        Installing when an ebuild is in tree/overlay.

                        Comment


                        • #13
                          Linux kernel 2.6 or above (up to 3.5)
                          So the latest kernel they support that is not End Of Life is 3.4?
                          3.5 (released July 21, 2012) and 3.6 (released September 30, 2012) are already EOL, version 3.7 was released on December 11, 2012, more than a month ago, and has already the third point release.
                          At this rate, can we expect to have support for 3.7 before summer (when we most likely are at 3.9 or 3.10 for the latest stable kernel) ?

                          Comment


                          • #14
                            Did no one notice:
                            [366820] Performance of Valve Linux games

                            Although vaque can we get some gamers or Michael to test? I have an unsupported HD 3870 or I would...

                            Comment


                            • #15
                              Originally posted by ChrisXY View Post
                              I tried it with an AMD Enduro notebook. "Downgraded" to all the stable versions, X.Org X Server 1.13.1, mesa 9.0.1 and xf86-video-intel 2.20.18.

                              Didn't work. X start looks ok, until…
                              Code:
                              [   749.149] (EE) fglrx(0): Map intel primary surface to user space failed!
                              [   749.149] 
                              Fatal server error:
                              [   749.149] AddScreen/ScreenInit failed for driver 0
                              Sorry, I was too quick. It was still using sna. I switched to uxa and it works now, even with xf86-video-intel from git.

                              Kind of.

                              Multiple screens suck. X often crashes when enabling or disabling screens. With three screens, sometimes screens that are supposed to be side-by-side get weirdly cloned, they display the same thing but the cursor can move between them as if they were side by side.
                              When moving windows they are weirdly "wobbly" and often pixel rows seem to seem to be missing or have some other weird effect. It's not visible on video captured with ffcast.

                              With one screen it is kind of ok. Rendering seems to work as good as catalyst's rendering works. Serious Sam 3 BFE on Ultra settings is very nice. Example of one problem: The steam overlay is unclickable, vanishes after about 5 seconds and then the game hangs.

                              Originally posted by krasnoglaz View Post
                              Also, luxmark score is still at around 1200 compared to 1700 for W8 with 12.10 drivers, and 1800 for W8 beta drivers.
                              1127 on HD 7970M.
                              Last edited by ChrisXY; 01-18-2013, 10:14 AM.

                              Comment

                              Working...
                              X