Search:

Type: Posts; User: ljbade

Search: Search took 0.00 seconds.

  1. Thread: opencl debugger

    by ljbade
    Replies
    1
    Views
    1,443

    Re: opencl debugger

    It depends on what OpenCL platform you are using.

    For AMD/ATI:
    http://developer.amd.com/tools/gdebugge ... fault.aspx

    For NVIDIA:
    http://developer.nvidia.com/nvidia-parallel-nsight

    Both of...
  2. Replies
    5
    Views
    3,050

    Re: 16 bit integer Atomics

    You can always use masking combined with 32bit atomic ops to work with a 16bit buffer.
  3. Replies
    6
    Views
    1,891

    Re: driver falls while kernel is being executed

    What video cards are you using? NVIDIA or ATI?

    Are you correctly allocating the right size of device memory buffers?
    Is your code somehow attempting to read memory past the end of the memory...
  4. Re: getting number of CL-GL interop devices creates sigsegv

    Well the only other input to that function is the_context_properties.

    Have you used the exact same context properties as the sample? Perhaps you have given the function an incorrect property that...
  5. Re: getting number of CL-GL interop devices creates sigsegv

    Does the SDK sample query the return size of clGetGLContextInfoKHR? If it just assumes the result is always sizeof(cl_device_id) then it may be avoiding a bug in the driver.
  6. Replies
    2
    Views
    1,501

    Re: Mistakes in CL man pages

    Thanks, I have created bug no. 422.
  7. Replies
    2
    Views
    1,501

    Mistakes in CL man pages

    I have found some mistakes in the OpenCL 1.1 (and 1.0) man pages for the clEnqueueRead/Write* functions.

    I have edited the SVN version of the man pages and uploaded them here.

    The 1.0 pages...
  8. Replies
    5
    Views
    2,234

    Re: Can i allocate memory inside kernel?

    One trick I have seen is to allocate a large block of __global memory using the host API, then use their favorite heap allocation algorithm inside the kernel to carve this memory up into smaller...
  9. Replies
    3
    Views
    3,015

    Re: Kernel's source code protection

    It would be nice if Khronos created some sort of 'generic' portable asm like Microsoft has in DirectCompute to avoid these problems. With OpenGL 2 a lot of game companies use the ARB shader asm...
  10. Replies
    5
    Views
    2,672

    Re: Communication between OpenCL and CUDA

    The compiler is comes with NVIDIA's drivers so you wouldn't need to distribute the compiler. I doubt NVIDIA would let you distribute the compiler as it would create conflicts etc. and new CUDA cards...
  11. Replies
    13
    Views
    4,880

    Re: How to reduce the read buffer time

    You can disable the timeout:
    http://www.microsoft.com/whdc/device/display/wddm_timeout.mspx
  12. Replies
    1
    Views
    2,574

    Re: Bring this DirectCompute features to OpenCL!

    Other useful stuff with the texture objects, 1d textures, and 1d array textures
  13. Ability to change border color with image objects

    Currently on pg 238 of the 1.1 spec it says this:


    I would like to see the ability to change the border color added. All current GPUs support this in their hardware (since OpenGL has it).
    ...
  14. Replies
    8
    Views
    4,241

    Re: Correct way to declare a global sampler_t

    Interesting... bug 366.
    I noticed that too when I was reading about __constant.

    Have you read the other discussion? In post #20 I have a suggeston for how this could be resolved: by introducing a...
  15. Replies
    8
    Views
    4,241

    Re: Correct way to declare a global sampler_t

    So which was is correct?

    I found this in the AMD Stream 2.3 Release Notes:


    So AMD might have already discovered this bug?
  16. Replies
    8
    Views
    4,241

    Correct way to declare a global sampler_t

    I have been having a discussion with eyebex on the NVIDIA OpenCL forum about the image support.

    One interesting thing came up:
    What is the correct way to declare a global sampler_t object?

    In...
  17. Replies
    5
    Views
    3,190

    Re: Make it easier to get the CL version number

    Thanks for that tip. I will put that into the code tomorrow.
    I still reckon an integer version number is easier than strings.
  18. Replies
    5
    Views
    3,190

    Make it easier to get the CL version number

    I currently have this to check for OpenCL 1.1:

    // get the version
    cl.platform.version_major = atoi(strchr(cl.platform.version, ' ') + 1);
    cl.platform.version_minor =...
Results 1 to 18 of 19