Search:

Type: Posts; User: bwatt

Page 1 of 3 1 2 3

Search: Search took 0.00 seconds.

  1. Replies
    1
    Views
    1,160

    Re: OpenCL wrapper Overhead

    What implementation are you talking about?
  2. Replies
    6
    Views
    2,342

    Re: opencl ! too much leak memory problem

    Thanks, I'm not seeing the leaks with my implementation. P.S. I had to make a number of changes, because you used a number of AMD specifics. Oh, and you also didn't provide the kernel source so I...
  3. Replies
    6
    Views
    2,342

    Re: opencl ! too much leak memory problem

    Please post your leaking program here. I'd like to run it on my implementation. Thanks.
  4. Replies
    2
    Views
    1,473

    Re: Callback When Event Already Set?

    The spec says "All callbacks registered for an event object must be called. All enqueued callbacks shall be called before the event object is destroyed." Bottom line, even though the event may have...
  5. Replies
    13
    Views
    2,814

    Re: problems with memories

    Don't have a GPU, using a CPU device. Sorry.
  6. Replies
    13
    Views
    2,814

    Re: problems with memories

    BTW, I can run your code (with minor mods to adapt it) on my system, that is, "this locks my computer and i need reboot system" doesn't happen on my implementation.
  7. Replies
    13
    Views
    2,814

    Re: problems with memories

    Also the for-loop rarely runs, do you really want "k=i*n; k<i+n; k++", for example, k=i*n where i*n=4300 and k<i+n where i+n=143 would never execute.
  8. Replies
    13
    Views
    2,814

    Re: problems with memories

    Dunno if this helps, but you need to change
    unsigned long int count = DATA_SIZE;
    to
    unsigned int count = DATA_SIZE;
    to match
    err |= clSetKernelArg(kernel, 2, sizeof(unsigned int), &count);...
  9. Replies
    3
    Views
    1,886

    Re: Getting Started with OpenCL

    One tutorial is using the C-language and the other is using the C++ language.
  10. Re: Measuring, and discrepancy of sequential and parallel ti

    Based upon what you have provided I am unable to answer your post. Please post a link to sample code that shows your application. If you application is private then include a simplified sample....
  11. Re: OpenCL man page sources available in public Subversion

    File a bugzilla please.
  12. Re: Problems in performing Saxpy with mapping/unmapping

    Yes, declare the buffers as local variables inside the loop so that upon leaving the scope of the loop those local variables, that is, the buffers, are released.
  13. Re: Problems in performing Saxpy with mapping/unmapping

    From the expert...

    "The way to achieve this behavior is to do the allocation of the buffer, at B, within a nested scope. So:


    A
    {
    B
    C
    D
  14. Re: Problems in performing Saxpy with mapping/unmapping

    Yup, what version are you running? Are you using the 'read' or 'map' version when you get bad results?
  15. Re: Problems in performing Saxpy with mapping/unmapping

    I'm using your code (that is, I'm using the C++ OpenCL wrapper too) and can do it (on Linux in my case), what error do you get?
  16. Re: Problems in performing Saxpy with mapping/unmapping

    Before calling your cleanupHost() function set the buffers to NULL which will release them.

    bufX = NULL;
    bufY = NULL;
    cleanupHost();
  17. Re: Problems in performing Saxpy with mapping/unmapping

    I ran you code (with some non-OpenCL adjustments for my implementation and OS :)) and got


    Testing GPU vs CPU with 128 elements
    Verification SUCCESS
    Testing GPU vs CPU with 256 elements...
  18. Re: Problems in performing Saxpy with mapping/unmapping

    Oh, one other minor point... you don't need to issue a finish if all of the previous enqueued command set a blocking flag


    // queue.finish();
  19. Re: Problems in performing Saxpy with mapping/unmapping

    Everything looks OK in general, and you should only unmap if you map:


    if(BUFFER_MODE == CL_MEM_ALLOC_HOST_PTR)
    {
    queue.enqueueUnmapMemObject(bufY, pY);
    }

    You...
  20. Re: Problems in performing Saxpy with mapping/unmapping

    Yes, for every clEnqueueMapBuffer you need to do a clEnqueueUnmapBuffer. It's just good 'balanced' programming to do this.
  21. Replies
    2
    Views
    1,143

    Re: How to read a clCreateImage3D

    Try using clEnqueueReadImage instead
  22. Replies
    7
    Views
    2,354

    Re: 2D FFT Routine Needed

    FYI, In IE8 the button doesn't work, use another browser such as FireFox.
  23. Replies
    7
    Views
    2,354

    Re: 2D FFT Routine Needed

    The button is there at the top of the page (somewhat hidden and in an awkward place to the right of the "Table of Contents" title) it will allow you to download all the files as a zip file named...
  24. Thread: vector datatypes

    by bwatt
    Replies
    5
    Views
    1,659

    Re: vector datatypes

    You also may need to have the buffer aligned so you could use posix_memalign in place of malloc in the above to insure that the float4 is on a 16-byte boundary.
  25. Replies
    7
    Views
    2,354

    Re: 2D FFT Routine Needed

    Or maybe the Ocean sample from the Apple developer site http://developer.apple.com/library/mac/#samplecode/OpenCL_OceanWave/Introduction/Intro.html#//apple_ref/doc/uid/DTS40009447 and click the...
Results 1 to 25 of 54
Page 1 of 3 1 2 3