Announcement

Collapse
No announcement yet.

Updated and Optimized Ubuntu Free Graphics Drivers

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

  • Updates

    * added ubuntu 15.04/vivid support
    * updated llvm 3.6 to rc2, including radeon fixes (building now)
    * added gallium-nine instructions here: https://launchpad.net/~oibaf/+archiv...aphics-drivers

    Comment


    • Did somebody try latest libvdpau/vdpauinfo with gtx 960+ or latest vainfo with broadwell?

      Comment


      • Originally posted by oibaf View Post
        * added ubuntu 15.04/vivid support
        * updated llvm 3.6 to rc2, including radeon fixes (building now)
        * added gallium-nine instructions here: https://launchpad.net/~oibaf/+archiv...aphics-drivers

        Can you add that lower kernel is supported (above 3.13 or 3.14, I'm not sure) if you add to the boot option drm.rnodes=1 That activates render-nodes.

        Also perhaps you can add the wine version can be installed via PlayOnLinux.

        Comment


        • Originally posted by oibaf View Post
          * updated llvm 3.6 to rc2, including radeon fixes (building now)
          Sounds great! IMO, outdated LLVM has been most pressing issue recently. Because in RadeonSI absolutely everything relies on shaders, even 2D, and older LLVMs were really crapy in this regard.

          Comment


          • Strange, lib has really updated. However, compiled binaries (e.g. OpenCL kernels) contain mention of -rc1 inside of them once they're compiled. Is it just as planned? (e.g. runtime from rc1?)

            Comment


            • Added info about older kernel. About llvm if there are still issues with 3.6-rc2 make sure to open a bug on freedesktop or it will never be fixed.

              Comment


              • Originally posted by SystemCrasher View Post
                Strange, lib has really updated. However, compiled binaries (e.g. OpenCL kernels) contain mention of -rc1 inside of them once they're compiled. Is it just as planned? (e.g. runtime from rc1?)
                I think that's just the string of version which was compiled and it's not meaningful, I just forced a rebuild with newer llvm of mesa and libclc anyway.

                Comment


                • Originally posted by oibaf View Post
                  updated llvm 3.6 to rc2, including radeon fixes (building now)
                  Yestardays, the comand output was:

                  Code:
                  $ glxinfo | grep OpenGL

                  '+vgpr-spilling' is not a recognized feature for this target (ignoring feature)
                  '+vgpr-spilling' is not a recognized feature for this target (ignoring feature)

                  OpenGL vendor string: X.Org
                  OpenGL renderer string: Gallium 0.4 on AMD BONAIRE
                  OpenGL core profile version string: 3.3 (Core Profile) Mesa 10.5.0-devel (git-6094619 2015-01-31 trusty-oibaf-ppa)
                  OpenGL core profile shading language version string: 3.30
                  OpenGL core profile context flags: (none)
                  OpenGL core profile profile mask: core profile
                  OpenGL core profile extensions:
                  OpenGL version string: 3.0 Mesa 10.5.0-devel (git-6094619 2015-01-31 trusty-oibaf-ppa)
                  OpenGL shading language version string: 1.30
                  OpenGL context flags: (none)
                  OpenGL extensions:
                  OpenGL ES profile version string: OpenGL ES 3.0 Mesa 10.5.0-devel (git-6094619 2015-01-31 trusty-oibaf-ppa)
                  OpenGL ES profile shading language version string: OpenGL ES GLSL ES 3.00
                  OpenGL ES profile extensions:
                  Today, the comand output is:

                  Code:
                  $ glxinfo | grep OpenGL
                  OpenGL vendor string: X.Org
                  OpenGL renderer string: Gallium 0.4 on AMD BONAIRE
                  OpenGL core profile version string: 3.3 (Core Profile) Mesa 10.5.0-devel (git-753c327 2015-02-02 trusty-oibaf-ppa)
                  OpenGL core profile shading language version string: 3.30
                  OpenGL core profile context flags: (none)
                  OpenGL core profile profile mask: core profile
                  OpenGL core profile extensions:
                  OpenGL version string: 3.0 Mesa 10.5.0-devel (git-753c327 2015-02-02 trusty-oibaf-ppa)
                  OpenGL shading language version string: 1.30
                  OpenGL context flags: (none)
                  OpenGL extensions:
                  OpenGL ES profile version string: OpenGL ES 3.0 Mesa 10.5.0-devel (git-753c327 2015-02-02 trusty-oibaf-ppa)
                  OpenGL ES profile shading language version string: OpenGL ES GLSL ES 3.00
                  OpenGL ES profile extensions:
                  The output "+vgpr-spilling' is not a recognized feature for this target (ignoring feature) ", is missing ..

                  Thanks.

                  Comment


                  • Originally posted by oibaf View Post
                    I think that's just the string of version which was compiled and it's not meaningful, I just forced a rebuild with newer llvm of mesa and libclc anyway.
                    Well, looking on behavior I got impression it is really updated lib. Yet it presents itself as -rc1 in generated code. That's what puzzled me a bit.

                    As for bugs... as you can see, I've been running some internal checks. And I can admit LLVM 3.6 is first version which can cope with some OpenCL loads without crashing whole GPU (often means loss of graphics in whole system) or at least calling program. While there're some odds, most of them are already known to AMD devs as far as I know.

                    Comment


                    • Originally posted by SystemCrasher View Post
                      Well, looking on behavior I got impression it is really updated lib. Yet it presents itself as -rc1 in generated code. That's what puzzled me a bit.

                      As for bugs... as you can see, I've been running some internal checks. And I can admit LLVM 3.6 is first version which can cope with some OpenCL loads without crashing whole GPU (often means loss of graphics in whole system) or at least calling program. While there're some odds, most of them are already known to AMD devs as far as I know.
                      If you want you can test my PPA, which has latest llvm 3.7git with latest mesa, intel, etc AND opencl (i386/amd64) with amdgcn enabled for radeonsi.

                      https://launchpad.net/~paulo-miguel-...e/ubuntu/mesa/

                      opencl looks to be working too:

                      16:21 Seg Fev 02$ ./cl-demo 1000000 10
                      Choose platform:
                      [0] Mesa
                      Enter choice: 0
                      Choose device:
                      [0] AMD CAPE VERDE
                      Enter choice: 0
                      ---------------------------------------------------------------------
                      NAME: AMD CAPE VERDE
                      VENDOR: X.Org
                      PROFILE: FULL_PROFILE
                      VERSION: OpenCL 1.1 MESA 10.5.0-devel (padoka PPA)
                      EXTENSIONS:
                      DRIVER_VERSION: 10.5.0-devel (padoka PPA)

                      Type: GPU
                      EXECUTION_CAPABILITIES: Kernel
                      GLOBAL_MEM_CACHE_TYPE: None (0)
                      CL_DEVICE_LOCAL_MEM_TYPE: Local (1)
                      SINGLE_FP_CONFIG: 0x7
                      QUEUE_PROPERTIES: 0x2

                      VENDOR_ID: 4098
                      MAX_COMPUTE_UNITS: 10
                      MAX_WORK_ITEM_DIMENSIONS: 3
                      MAX_WORK_GROUP_SIZE: 256
                      PREFERRED_VECTOR_WIDTH_CHAR: 16
                      PREFERRED_VECTOR_WIDTH_SHORT: 8
                      PREFERRED_VECTOR_WIDTH_INT: 4
                      PREFERRED_VECTOR_WIDTH_LONG: 2
                      PREFERRED_VECTOR_WIDTH_FLOAT: 4
                      PREFERRED_VECTOR_WIDTH_DOUBLE: 2
                      MAX_CLOCK_FREQUENCY: 625
                      ADDRESS_BITS: 32
                      MAX_MEM_ALLOC_SIZE: 268435456
                      IMAGE_SUPPORT: 0
                      MAX_READ_IMAGE_ARGS: 32
                      MAX_WRITE_IMAGE_ARGS: 32
                      IMAGE2D_MAX_WIDTH: 32768
                      IMAGE2D_MAX_HEIGHT: 32768
                      IMAGE3D_MAX_WIDTH: 4096
                      IMAGE3D_MAX_HEIGHT: 4096
                      IMAGE3D_MAX_DEPTH: 4096
                      MAX_SAMPLERS: 0
                      MAX_PARAMETER_SIZE: 1024
                      MEM_BASE_ADDR_ALIGN: 128
                      MIN_DATA_TYPE_ALIGN_SIZE: 128
                      GLOBAL_MEM_CACHELINE_SIZE: 0
                      GLOBAL_MEM_CACHE_SIZE: 0
                      GLOBAL_MEM_SIZE: 1073741824
                      MAX_CONSTANT_BUFFER_SIZE: 268435456
                      MAX_CONSTANT_ARGS: 0
                      LOCAL_MEM_SIZE: 32768
                      ERROR_CORRECTION_SUPPORT: 0
                      PROFILING_TIMER_RESOLUTION: 0
                      ENDIAN_LITTLE: 1
                      AVAILABLE: 1
                      COMPILER_AVAILABLE: 1
                      MAX_WORK_GROUP_SIZES: 256 256 256
                      ---------------------------------------------------------------------
                      *** Kernel compilation resulted in non-empty log message.
                      *** Set environment variable CL_HELPER_PRINT_COMPILER_OUTPUT=1 to see more.
                      *** NOTE: this may include compiler warnings and other important messages
                      *** about your code.
                      *** Set CL_HELPER_NO_COMPILER_OUTPUT_NAG=1 to disable this message.
                      0.000870 s
                      27.597392 GB/s
                      GOOD

                      and

                      16:21 Seg Fev 02$ clinfo
                      Number of platforms: 1
                      Platform Profile: FULL_PROFILE
                      Platform Version: OpenCL 1.1 MESA 10.5.0-devel (padoka PPA)
                      Platform Name: Clover
                      Platform Vendor: Mesa
                      Platform Extensions: cl_khr_icd


                      Platform Name: Clover
                      Number of devices: 1
                      Device Type: CL_DEVICE_TYPE_GPU
                      Device ID: 4098
                      Max compute units: 10
                      Max work items dimensions: 3
                      Max work items[0]: 256
                      Max work items[1]: 256
                      Max work items[2]: 256
                      Max work group size: 256
                      Preferred vector width char: 16
                      Preferred vector width short: 8
                      Preferred vector width int: 4
                      Preferred vector width long: 2
                      Preferred vector width float: 4
                      Preferred vector width double: 2
                      Native vector width char: 16
                      Native vector width short: 8
                      Native vector width int: 4
                      Native vector width long: 2
                      Native vector width float: 4
                      Native vector width double: 2
                      Max clock frequency: 625Mhz
                      Address bits: 32
                      Max memory allocation: 268435456
                      Image support: No
                      Max number of images read arguments: 32
                      Max number of images write arguments: 32
                      Max image 2D width: 5
                      Max image 2D height: 0
                      Max image 3D width: 16884632
                      Max image 3D height: 139724036292136
                      Max image 3D depth: 140737199208032
                      Max samplers within kernel: 0
                      Max size of kernel argument: 1024
                      Alignment (bits) of base address: 128
                      Minimum alignment (bytes) for any datatype: 128
                      Single precision floating point capability
                      Denorms: Yes
                      Quiet NaNs: Yes
                      Round to nearest even: Yes
                      Round to zero: No
                      Round to +ve and infinity: No
                      IEEE754-2008 fused multiply-add: No
                      Cache type: None
                      Cache line size: 0
                      Cache size: 0
                      Global memory size: 1073741824
                      Constant buffer size: 268435456
                      Max number of constant args: 0
                      Local memory type: Local
                      Local memory size: 32768
                      Error correction support: 0
                      Unified memory for Host and Device: 1
                      Profiling timer resolution: 0
                      Device endianess: Little
                      Available: Yes
                      Compiler available: Yes
                      Execution capabilities:
                      Execute OpenCL kernels: Yes
                      Execute native function: No
                      Queue properties:
                      Out-of-Order: No
                      Profiling : Yes
                      Platform ID: 0x7f1408dd4ae0
                      Name: AMD CAPE VERDE
                      Vendor: X.Org
                      Device OpenCL C version: OpenCL C 1.1
                      Driver version: 10.5.0-devel (padoka PPA)
                      Profile: FULL_PROFILE
                      Version: OpenCL 1.1 MESA 10.5.0-devel (padoka PPA)

                      enjoy!

                      Comment

                      Working...
                      X