Announcement

Collapse
No announcement yet.

AMDGPU-PRO 16.60 Released

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

  • #91
    Originally posted by bridgman View Post
    Freesync instructions have been updated. The freesync_prop utility was an internal one...
    OK, cool

    Comment


    • #92
      Someone got it working with Radeon R7 M265?
      I've installed but can't access my user session as lightdm is always going back to the login screen.

      In Xorg.log i've found:
      [ 46.106] (EE) AIGLX error: dlopen of /usr/lib/x86_64-linux-gnu/dri/i965_dri.so failed (/usr/lib/x86_64-linux-gnu/dri/i965_dri.so: undefined symbol: is64bitelf) [ 46.106] (EE) AIGLX: reverting to software rendering [ 46.118] (EE) AIGLX error: dlopen of /usr/lib/x86_64-linux-gnu/dri/swrast_dri.so failed (/usr/lib/x86_64-linux-gnu/dri/swrast_dri.so: undefined symbol: is64bitelf) [ 46.118] (EE) GLX: could not load software renderer Not sure about how to fix it.

      Comment


      • #93
        AMDGPU does not install is kernel 4.4 supported?

        Loading new amdgpu-pro-16.60-379184 DKMS files...
        First Installation: checking all kernels...
        Building only for 4.4.46-040446-generic
        Building for architecture x86_64
        Building initial module for 4.4.46-040446-generic
        ERROR (dkms apport): kernel package linux-headers-4.4.46-040446-generic is not supported
        Error! Bad return status for module build on kernel: 4.4.46-040446-generic (x86_64)
        Consult /var/lib/dkms/amdgpu-pro/16.60-379184/build/make.log for more information.

        $ cat /var/lib/dkms/amdgpu-pro/16.60-379184/build/make.log
        DKMS make.log for amdgpu-pro-16.60-379184 for kernel 4.4.46-040446-generic (x86_64)
        Wed Feb 1 14:26:30 EST 2017
        make: Entering directory '/usr/src/linux-headers-4.4.46-040446-generic'
        LD /var/lib/dkms/amdgpu-pro/16.60-379184/build/built-in.o
        LD /var/lib/dkms/amdgpu-pro/16.60-379184/build/amd/amdgpu/built-in.o
        CC [M] /var/lib/dkms/amdgpu-pro/16.60-379184/build/amd/amdgpu/amdgpu_drv.o
        CC [M] /var/lib/dkms/amdgpu-pro/16.60-379184/build/amd/amdgpu/amdgpu_device.o
        /var/lib/dkms/amdgpu-pro/16.60-379184/build/amd/amdgpu/amdgpu_device.c: In function ‘amdgpu_gpu_reset’:
        /var/lib/dkms/amdgpu-pro/16.60-379184/build/amd/amdgpu/amdgpu_device.c:2327:11: error: implicit declaration of function ‘drm_atomic_helper_suspend’ [-Werror=implicit-function-declaration]
        state = drm_atomic_helper_suspend(adev->ddev);
        ^~~~~~~~~~~~~~~~~~~~~~~~~
        /var/lib/dkms/amdgpu-pro/16.60-379184/build/amd/amdgpu/amdgpu_device.c:2327:9: warning: assignment makes pointer from integer without a cast [-Wint-conversion]
        state = drm_atomic_helper_suspend(adev->ddev);
        ^
        /var/lib/dkms/amdgpu-pro/16.60-379184/build/amd/amdgpu/amdgpu_device.c:2438:7: error: implicit declaration of function ‘drm_atomic_helper_resume’ [-Werror=implicit-function-declaration]
        r = drm_atomic_helper_resume(adev->ddev, state);
        ^~~~~~~~~~~~~~~~~~~~~~~~
        /var/lib/dkms/amdgpu-pro/16.60-379184/build/amd/amdgpu/amdgpu_device.c: In function ‘amdgpu_get_pcie_info’:
        /var/lib/dkms/amdgpu-pro/16.60-379184/build/amd/amdgpu/amdgpu_device.c:2490:9: error: implicit declaration of function ‘drm_pcie_get_max_link_width’ [-Werror=implicit-function-declaration]
        ret = drm_pcie_get_max_link_width(adev->ddev, &mask);
        ^~~~~~~~~~~~~~~~~~~~~~~~~~~
        cc1: some warnings being treated as errors
        scripts/Makefile.build:258: recipe for target '/var/lib/dkms/amdgpu-pro/16.60-379184/build/amd/amdgpu/amdgpu_device.o' failed
        make[2]: *** [/var/lib/dkms/amdgpu-pro/16.60-379184/build/amd/amdgpu/amdgpu_device.o] Error 1
        scripts/Makefile.build:403: recipe for target '/var/lib/dkms/amdgpu-pro/16.60-379184/build/amd/amdgpu' failed
        make[1]: *** [/var/lib/dkms/amdgpu-pro/16.60-379184/build/amd/amdgpu] Error 2
        Makefile:1408: recipe for target '_module_/var/lib/dkms/amdgpu-pro/16.60-379184/build' failed
        make: *** [_module_/var/lib/dkms/amdgpu-pro/16.60-379184/build] Error 2
        make: Leaving directory '/usr/src/linux-headers-4.4.46-040446-generic'

        Last edited by joker5bb; 01 February 2017, 03:28 PM.

        Comment


        • #94
          The 4.4 kernel is definitely supported, not sure what is going on here.

          The drm_atomic_helper_suspend & _resume functions were not added until the 4.5 kernel and don't seem to be used by upstream amdgpu at all. I'll see if I can get someone familiar with the KCL code to look at this and understand why it is looking for those functions on your system.
          Test signature

          Comment


          • #95
            Originally posted by matszpk View Post
            Still, AMDGPU-PRO 16.60 doesn't support Pitcairn devices correctly. Just, I found message: "(EE) AMDGPU(G0): amdgpu_setup_kernel_mem failed" while running Xserver and any trial to run OpenCL application on this devices give OpenCL error -6 (OUT_OF_HOST_MEMORY) while trying creating command queue. I am testing drivers having two devices (Pitcairn and Bonaire, where Pitcairn is primary graphics card). Can anybody from AMD team fix support for Radeon HD 7x00?
            Quick question - does the same program work OK on the Bonaire ? I think that is what you are saying, just wanted to make sure.

            Thanks...
            Test signature

            Comment


            • #96
              Originally posted by bridgman View Post
              The 4.4 kernel is definitely supported, not sure what is going on here.

              The drm_atomic_helper_suspend & _resume functions were not added until the 4.5 kernel and don't seem to be used by upstream amdgpu at all. I'll see if I can get someone familiar with the KCL code to look at this and understand why it is looking for those functions on your system.
              People use random PPAs on Ubuntu, that is what happens
              That 4.46 kernel seems to be from PPA mainline, maybe ask him does it happen with regular one 4.4 kernel from main repo Further, if not chocolate then vanilla
              Last edited by dungeon; 01 February 2017, 09:48 PM.

              Comment


              • #97
                Originally posted by dungeon View Post
                maybe ask him does it happen with regular one 4.4 kernel from main repo
                That's a good question - joker5bb ?
                Test signature

                Comment


                • #98
                  Originally posted by bridgman View Post
                  Freesync instructions have been updated. The freesync_prop utility was an internal one - all the info is now available through randr:

                  http://support.amd.com/en-us/kb-arti...-IN-LINUX.aspx
                  Very interesting, thanks! This article mentions V-Sync, is it possible to enable V-Sync through xrandr?

                  Comment


                  • #99
                    I have an HP Z420 workstation with Firepro W9100 16GB and 32" Sharp pn-k321 4k monitor. The Sharp monitor supports multi-stream-transport (MST) allowing 4k resolution at 60 hz via display-1.2 port. I have several disks with one OS per disk: redhat 7.3, debian 8.7, ubuntu 16.04, windows 10 pro, suse 12 sp2, kaOS and qubes-os 3.2. I am running a not-yet-released math software that depends on openGL for 2d/3d graphics and I do testing using above setup.

                    Currently, I run fglrx on debian 8.7 because that works with MST at 60 hz. My windows 10 pro works with amdgpu-pro with MST at 60 hz.

                    A few days ago, I finally successfully installed the 1-25-2017 amdgpu-pro driver on redhat 7.3 and ubuntu 16.04. It works so far on both OS, except that it only gives me 1024x768 resolution if I enable MST on my Sharp monitor. If my monitor is configured for single-stream-transport (SST), I do get 4k resolution but at 30 hz instead of 60 hz.

                    Can I use xrandr to set the correct resolution? If so, what would be the parameters? If not, what would be a solution?

                    Thanks in advance.

                    Comment


                    • thanks for the great tips and insults (i dont care). people on this forum rock anyway!!!

                      sadly the tests will have to wait here as im on some works at home and at the company place too. agenda is crazy here... later maybe..

                      Comment

                      Working...
                      X