Search:

Type: Posts; User: chai

Page 1 of 3 1 2 3

Search: Search took 0.00 seconds.

  1. Thread: for loop problem

    by chai
    Replies
    4
    Views
    1,085

    Re: for loop problem

    if you do



    #pragma unroll
    for(int l=0 ; l<L ; l++)
    {
    // INIT

    // fill local memory
  2. Replies
    1
    Views
    610

    Re: reg kernel function(linear search)

    Almost! You need to get the output through a buffer, however. Actually, I'm not sure if this is the only way, but definitely the standard:



    __kernel void lin_search (__global const float...
  3. Replies
    1
    Views
    1,180

    Re: simpe 2D triangle fill vs. built in 3D graphics

    From what I understand (and this should be verified), the graphics pipeline has fixed functions for rasterization, which includes triangle filling. So it might actually be faster to do simple 2d...
  4. Replies
    6
    Views
    2,504

    Re: When is a command queue actually executed?

    hah ok, so basically your program should rely on queues and blocking events, and the "under-the-hood" of OCL will take care of the rest in an optimal fashion...

    Thanks again!
  5. Re: How can i create or use a critical section in OpenCL?

    You can use the atomic functions (if your platform/device supports the extension) to implement a critical section (look at wikipedias critical section page, you can use atomic instructions to create...
  6. Replies
    6
    Views
    2,504

    Re: When is a command queue actually executed?

    Why is this not recommended? And what's the preferred alternative (clFinish, clWaitForEvents, ... ) ?
  7. Replies
    2
    Views
    1,369

    Re: Need to transfer a huge amount of data

    You want an OpenCL kernel to run on data that is continuously streaming into a smaller static buffer? A cool idea, but I don't think it's possible. You'd need the kernel to indicate what has been...
  8. Replies
    5
    Views
    2,356

    Re: Looking for multi-GPU applications

    Not a complete solution, but there is a multi-gpu tutorial for n-body simulation here:

    http://www.browndeertechnology.com/docs ... -rev3.html

    There are simpler ways to play with metrics for...
  9. Replies
    5
    Views
    2,155

    Re: Can i allocate memory inside kernel?

    That means you're using a host pointer, or CPU memory, for your buffer. But right now you can't allocate any new memory (or free it) from within a kernel, only through the application API calls....
  10. Replies
    1
    Views
    1,305

    Re: Reboot or killing jobs remotly after crash?

    ATI made a guide for using stream (and therefore, their OpenCL implementation) through ssh:
    http://developer.amd.com/gpu_assets/App ... motely.pdf



    Chances are, you won't be able to just kill...
  11. Re: What happens if one of kernel argument is const int ?

    Ehrm... okay. There's a bug in that compiler :)

    [/quote]

    Could you elaborate? I have plenty of "const unsigned int X" as parameters that work on several OCL compilers, do you mean that any...
  12. Replies
    15
    Views
    3,182

    Re: calculation of a float value

    Intel GPUs DO have vertex shaders. Regardless, shaders aren't in question unless you are using heavy OpenGL interoperation.

    However, the current Intel OpenCL implementation supports only Intel...
  13. Re: OpenCL Image Rotate/Scale/Translate, Affine Transform, .

    As far as resolution preservation goes, your biggest problem will be when you transform the pixel coordinates. Suppose you have a point (0,0) and after the transformation it becomes (0.23,5.6), then...
  14. Replies
    9
    Views
    4,519

    Re: Example for Random Number Generator?

    Look up mersenne twister algorithms for OpenCL (or you could port a CUDA or DirectX11 compute shader "kernel"). There are various implementations available.

    You could also pass in a buffer of...
  15. Replies
    1
    Views
    1,323

    Re: accessing elements of cl::Buffer?

    You'd have to read the buffer into program memory first--i.e. a cl_float array--using clEnqueueReadBuffer(...) . Look at any openCL sample that uses this function for an example, but direct buffer...
  16. Re: OpenCL Image Rotate/Scale/Translate, Affine Transform, .

    I can't give you a thorough or tested sample, but I can suggest the "proper" way to develop the algorithm using a transformation matrix.

    The parameters would be an input image and an output...
  17. Replies
    3
    Views
    2,307

    Re: Mipmaps and Image Pyramid

    OpenCL supports linear filtering (see the sampler_t documentation ), in 2 and 3 dimensional images, but not miplevels...

    Hmm, thinking about it now, i have a weird alternative--if you could...
  18. Replies
    2
    Views
    2,060

    Re: 3rd-party SDKs : a chaos

    I have a very similar framework to your example (dynamically loading the library and functions), and having used some of the available sdk's, agree entirely with your point. A lot of the...
  19. Replies
    5
    Views
    2,155

    Re: Can i allocate memory inside kernel?

    short answer: no

    long answer: not yet ;)
  20. Re: which format supported in clCreateFromD3D10Texture2DKHR?

    What is the error you are getting and what is the call which returns the error?

    I can't help you with the reference count, but if you have the DirectX debug layer it should provide output for GPU...
  21. Replies
    3
    Views
    2,307

    Re: Mipmaps and Image Pyramid

    You can use the gl sharing extension to load a specific mipmap level as described in the clCreateFromGLTexture2D Documentation, so you could iterate over all mip levels and create a cl_image for...
  22. Thread: OpenCl and Linux

    by chai
    Replies
    8
    Views
    3,928

    Re: OpenCl and Linux

    I've always glossed over the return value and set it to "void", but is there a way to check the return value of a kernel? Not sure how this would be very useful, curious nonetheless...
  23. Replies
    10
    Views
    3,554

    Re: Error using my .h files in Kernel OpenCL

    sounds like you included the .h file from several sources. try this:



    #ifndef INTERVALO
    #define INTERVALO
    struct _intervalo
    {
    // your struct here
    };
  24. Re: Multiple platform implementations in one application?

    :P looking into this more after reading responses, the dynamic functions were definitely unnecessary for multiple platforms. I'm keeping it for the uses mentioned above though.

    I'm still getting...
  25. Replies
    5
    Views
    3,173

    Re: OpenCL on MacBook Pro with NVidia 320m

    thanks for the very thorough follow-up! glad the suggestions helped.

    also, since you're on NVIDIA, you might want to try the loop unrolling extension. It would take way too long to fully unroll...
Results 1 to 25 of 60
Page 1 of 3 1 2 3