Search:

Type: Posts; User: SabinManiac

Search: Search took 0.00 seconds.

  1. Re: Difference between several runs of opencl code

    It seems ok now. Curious thing :D .Thanks
  2. Difference between several runs of opencl code

    Hello,

    I have this opencl code:




    unsigned long permute_pc1(unsigned long src,int tid){
    unsigned long dst = 0;
    int srcPos = 0;
  3. Replies
    3
    Views
    1,236

    Re: Problem with bitwise or

    I have enumerated my device extension and int64_base_atomics aren't among them:



    CL_DEVICE_EXTENSIONS: cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing cl_nv_d3d9_sharing...
  4. Replies
    3
    Views
    1,236

    Re: Problem with bitwise or

    Thank you for the reply. The sugested solution works and I understand why, now. But I tried to improve things with atomic_or for 64 bits long operands. I added the pragma specifying the extension to...
  5. Replies
    3
    Views
    1,236

    Problem with bitwise or

    Hello again,
    I have the following code:

    #include <oclUtils.h>
    #include <stdio.h>
    #include "Timer.h"


    int main()
    {
  6. Replies
    5
    Views
    2,850

    Re: CPU faster in vector addition than GPU

    Thank you for your replies.
  7. Replies
    5
    Views
    2,850

    Re: CPU faster in vector addition than GPU

    Ok I understand what you are saying, so I fed more data for the GPU processing. The result is the same. CPU outruns the GPU by an order of size.



    #include <oclUtils.h>
    #include "Timer.h"...
  8. Replies
    5
    Views
    2,850

    CPU faster in vector addition than GPU

    Hello,

    I have recently started working with OpenCL, but at my first program I ran into some problems.

    This is the code:




    #include <oclUtils.h>
Results 1 to 8 of 11