xpra icon
Bug tracker and wiki

Changes between Version 41 and Version 42 of CSC


Ignore:
Timestamp:
01/10/14 11:52:28 (6 years ago)
Author:
Antoine Martin
Comment:

moved opencl and cuda to sub pages

Legend:

Unmodified
Added
Removed
Modified
  • CSC

    v41 v42  
    1717On the server side we currently (as of v0.11) support 4 modules for doing this step:
    1818* [/browser/xpra/trunk/src/xpra/codecs/csc_swscale csc_swscale] which uses [http://www.ffmpeg.org/ FFmpeg]'s swscale (CPU based)
    19 * [/wiki/CSC#csc_opencl csc_opencl] (GPU or CPU based via [http://www.khronos.org/opencl/ OpenCL])
    20 * [/wiki/CSC#csc_nvcuda csc_nvcuda] (GPU based via [http://www.nvidia.com/object/cuda_home_new.html CUDA])
     19* [/wiki/CSC/OpenCL csc_opencl] (GPU or CPU based via [http://www.khronos.org/opencl/ OpenCL])
     20* [/wiki/CSC/NVCUDA csc_nvcuda] (GPU based via [http://www.nvidia.com/object/cuda_home_new.html CUDA])
    2121* [/browser/xpra/trunk/src/xpra/codecs/csc_cython csc_cython] (a slower fallback module in Cython)
    2222
     
    2424[[BR]]
    2525
    26 On the client side, if a CSC step is needed, it will use {{{swscale}}}. The {{{OpenGL}}} mode does not need a CSC step as it can display YUV pixels directly on screen.
     26On the client side, if a CSC step is needed, it will use {{{swscale}}} only (for now). The {{{OpenGL}}} mode does not need a CSC step as it can display YUV pixels directly on screen, and do the scaling itself too.
    2727}}}
    2828
     
    5252
    5353{{{#!div class="box"
    54 == {{{csc_opencl}}} ==
    55 Links:
    56 * this module relies on [http://documen.tician.de/pyopencl/ PyOpenCL]
    57 * original feature request ticket with more information: #422
    58 * [/browser/xpra/trunk/src/xpra/codecs/csc_opencl csc_opencl module source]
    59 
    60 ----
    61 
    62 Options:
    63 * select which device type is used as backend using:
    64 {{{
    65 XPRA_OPENCL_DEVICE_TYPE=devicetype xpra start ...
    66 }}}
    67 Where `devicetype` is either {{{GPU}}} or {{{CPU}}}. [[BR]]
    68 * choose the {{{OpenCL}}} implementation to use when more than one are installed:
    69 {{{
    70 XPRA_OPENCL_PLATFORM="Intel" xpra ...
    71 }}}
    72 * select a specific {{{OpenCL}}} device when more than one is available:
    73 {{{
    74 XPRA_OPENCL_DEVICE_NAME="GeForce GTX 760" xpra ...
    75 }}}
    76 
    77 To get more information use:
    78 {{{
    79 XPRA_OPENCL_DEBUG=1 xpra ...
    80 }}}
    81 }}}
    82 
    83 {{{#!div class="box"
    84 == {{{csc_nvcuda}}} ==
    85 You may need to build xpra with the flag "{{{--with-csc_nvcuda}}}".
    86 You must use a sufficiently recent versionof the {{{nvidia}}} video driver, preferably obtained directly from nvidia, and not the {{{nouveau}}} drivers (which you should probably remove).
    87 [[BR]]
    88 
    89 Links:
    90 * this module relies on [http://documen.tician.de/pycuda/ PyCUDA]
    91 * original feature request ticket with more information: #384
    92 * [/browser/xpra/trunk/src/xpra/codecs/csc_nvcuda csc_nvcuda module source]
    93 
    94 ----
    95 
    96 Options:
    97 * choose the device used with:
    98 {{{
    99 XPRA_CUDA_DEVICE=N xpra ...
    100 }}}
    101 Where `N` is the device ID. Use:
    102 {{{
    103 XPRA_CUDA_DEBUG=1 xpra ...
    104 }}}
    105  to get more information.
    106 }}}
    107 
    108 
    109 {{{#!div class="box"
    11054== Scaling ==
    11155