Search:

Type: Posts; User: ibbles

Page 1 of 4 1 2 3 4

Search: Search took 0.02 seconds.

  1. Replies
    3
    Views
    971

    Your second suggestion will work; pass pointers...

    Your second suggestion will work; pass pointers to the function. Just remember that you may need to add address qualifiers to the function declaration.
  2. Replies
    3
    Views
    971

    Solution 1 will not work. The parameters a and b...

    Solution 1 will not work. The parameters a and b are local to the function and only contains copies of the variables passed when f is called.
  3. Replies
    6
    Views
    2,269

    Re: Do I need atomic_add in that case?

    This sounds very much like a common histogram calculation. In OpenCL this is a non-trivial operation if it's to be done efficiently, but I think that a few of the SDKs include examples of this in...
  4. Replies
    2
    Views
    1,427

    Re: Garbage output by array of struct in OpenCL

    When declaring the structure on the host side you should use the cl_* types to ensure that they are the same on the host and the device. cl_int in your case.

    If you suspect alignment problems,...
  5. Replies
    2
    Views
    995

    Re: Error in pointer arithmetic

    It is because


    currentPosition += 4;

    changes the value of the pointer currentPosition and not what the pointer points to. The operation becomes pointless since currentPosition is never used...
  6. Replies
    1
    Views
    1,292

    Re: OpenCL, NVIDIA on Windows 7 - Program freezes

    That is caused by a feature called the watchdog timer. Try searching for that and you will probably find a few threads about it, this one, for example.
  7. Thread: Runtime Error

    by ibbles
    Replies
    2
    Views
    881

    Re: Runtime Error

    I think we need to see the source code for the kernel in order to help.
  8. Replies
    2
    Views
    930

    Re: Distributed compilation

    Not as part of the OpenCL standard, but efforts are being made to support this with higher level APIs. See for example https://www.alpha-tierchen.de/~bkoenig/clara/ and ...
  9. Replies
    6
    Views
    1,892

    Re: driver falls while kernel is being executed

    Yeah, it can be true. But if it's so, how long kernel can stay in queue (executing)?
    Have we possibilty to turn off watchdog timer?[/quote]

    What OS are you using? If Windows, then this thread...
  10. Replies
    6
    Views
    1,892

    Re: driver falls while kernel is being executed

    What do you mean by driver fail? Do you get some error code back? Does the kernel take very long time to execute, which may trigger a watchdog timer?
  11. Replies
    3
    Views
    1,353

    Re: Memory test: cpu timers vs. gpu timer sample

    I would like to know the cause of this as well. Performance measurements are always important and this makes me question some of my own measurements.

    I noticed that you copy from a memory mapped...
  12. Replies
    3
    Views
    1,634

    Re: clCreateContext() question.

    We need to keep apart the physical device, i.e., the collections of transistors, from the device IDs, which is what is passed to clCreateContext(). While the same lump of transistors can be utilized...
  13. Replies
    2
    Views
    1,637

    Re: Practice exercises for learning OpenCL

    Perhaps a Mandelbrot fractal. I like that example because it gives visual results which are nice to look at, is very simple to implement naively but has loads of potential for all kinds of...
  14. Replies
    5
    Views
    1,843

    Re: two level of parallel

    Creating the command queue with the out-of-order flag will let the hardware run the three kernel invocations in parallel, but the actual parallelism achieved by this is up to the implementation. Not...
  15. Replies
    5
    Views
    1,843

    Re: two level of parallel

    I would say that running the kernel three times is conceptually better since it is easier to understand what is happening. It will also make it easier should you ever require to run other than three...
  16. Replies
    4
    Views
    1,980

    Re: How to draw an image from openCL?

    I began trying to make a really small c++ example that renders a OpenCL generated image using OpenGL, but it turned out to require quite a bit of code. More than I can write on an evening. A decent...
  17. Replies
    4
    Views
    1,980

    Re: How to draw an image from openCL?

    So this is more of a project for learning vector algebra, geometry, trigonometry, and compute science rather than with the purpose of producing something that will be useful^* later? I don't know...
  18. Replies
    4
    Views
    1,980

    Re: How to draw an image from openCL?

    I think that sharing a texture with OpenGL is the simple solution to rendering images generated in OpenCL.

    I'm wondering a bit about what you are really trying to do. You say you want to make a 3D...
  19. Re: Setup OpenCL for completely heterogeneous implementation

    In addition to the headers, you must also link to the OpenCL library supplied with the NVidia GpuComputing SDK.
  20. Replies
    1
    Views
    1,069

    Re: Returning a single value from kernel!

    That would be the most obvious way. You could write to an image instead, but that would hardly be any simpler.
  21. Re: How to read elements from a float16 (or other) using ind

    When writing ordinary host code, at least with gcc, then accessing data of some type through a pointer of another type gives warnings about type-punned pointers and strict aliasing rules. Is the same...
  22. Replies
    2
    Views
    1,100

    Re: Computing Mean values of a set of vectors ?

    This sounds like a straight-forward reduction. Many OpenCL sample sets includes an example of this, and you can also have a look at...
  23. Replies
    2
    Views
    1,338

    Re: opencl sample codes

    Also, AMD's OpenCL implementation works just fine on Intel CPUs as well.

    http://developer.amd.com/zones/OpenCLZone/pages/GettingStarted.aspx
  24. Re: Multiple platform implementations in one application?

    Is all this really necessary? Can't we just call clGetPlatformIDs and then clCreateContext once for each platform found?
  25. Replies
    10
    Views
    3,733

    Re: Error using my .h files in Kernel OpenCL

    Note that the compilation of the kernels is completely separate from the compilation of the host application. Kernel compilation normally happens during runtime and has nothing to do with the...
Results 1 to 25 of 80
Page 1 of 4 1 2 3 4