Search:

Type: Posts; User: VincentH

Search: Search took 0.00 seconds.

  1. Replies
    2
    Views
    459

    manual caching

    The second option is what I call "manual caching". So when the accelerator has no cache, the second one runs faster. When the GPU is enough cache, then it won't really matter. Not tested in a while,...
  2. What can be implemented in what? OpenCL, OpenGL, HSA, SPIR, LLVM IR, HSA, Mantle, etc

    How we program parallel processors is not a final answer, but a global research project. CUDA/OpenCL was a good direction, but we're certainly not there! HSA, OpenCL SPIR and such are also...
  3. Replies
    32
    Views
    15,729

    Sticky: I'd like a name for the architecture and...

    I'd like a name for the architecture and sub-architecture via clGetDeviceInfo. Optionally also Series, Type and Version to split up the now useless CL_DEVICE_NAME. For CUDA you have compute...
  4. Replies
    0
    Views
    1,974

    Public extensions and add-ons on github

    Some goals are reinvented by each developer. For instance a header-file which makes interaction with OpenGL better and shortens initialisation (select best device), or a better version of cl.hpp. Or...
  5. Thread: Tutorial

    by VincentH
    Replies
    11
    Views
    4,953

    Re: Tutorial

    When doing a make of the CUDA/OpenCL SDK, you not only needs 'build-essential', but also:
    mesa-common-dev
    libx11-dev
    libgl1-mesa-dev
    libglu1-mesa-dev
    libglut3-dev
    libxmu-dev

    For shell-users:...
  6. Thread: Tutorial

    by VincentH
    Replies
    11
    Views
    4,953

    Re: Tutorial

    I have explained installing the correct driver on my blog: http://www.streamcomputing.eu/blog/2010 ... l-on-linux

    Installing the SDK is really straightforward. Just make sure you do the...
  7. Replies
    0
    Views
    1,684

    Q1 2011 Benchmarks - discussion

    Do you have problems with some benchmarks? Do you think some brand underperforms, but shouldn't? Unexpected crashes? Bug-fixes? Other news? Discuss it here.

    Known OpenCL benchmarks:
    - SiSoft
    -...
  8. Replies
    1
    Views
    1,686

    Re: Q1 2011 Benchmarks - results

    Since the blog will appear Thursday, here are the links to the benchmarks:

    http://www.sisoftware.net/?d=qa&f=g ... &a=oca
    http://www.phoronix-test-suite.com/
    http://ft.ornl.gov/doku/shoc/start...
  9. Replies
    1
    Views
    1,686

    Q1 2011 Benchmarks - results

    At http://www.streamcomputing.eu/blog/2011 ... benchmarks I asked people to drop their Benchmark-results here. This way we can find out, what is the fastest hardware on the moment. I invite the...
  10. Replies
    7
    Views
    4,432

    Re: Specification 1.0 to 1.1 diff

    I did a blog on it last week: http://www.streamcomputing.nl/blog/2010 ... -1-changes
    Hope that helps?
  11. Replies
    0
    Views
    2,624

    Automatic conversion on big-endian hardware

    While I love the low-level way of OpenCL, I don't like to be aware of endianess since I cannot easily test it. Can't OpenCL provide a little-endian interface and leave the actual endianess to the...
Results 1 to 11 of 11