Search:

Type: Posts; User: pplaszew

Search: Search took 0.00 seconds.

  1. Replies
    1
    Views
    2,227

    ATI and NVIDIA from OpenCL perspective

    Hi,
    I have some experience in NV CUDA and recently switched to OpenCL. So far I always targeted Nvidia architecture and optimized my kernels accordingly: coalesced memory access, no divergent...
  2. Replies
    1
    Views
    1,507

    Re: help with work-group

    Data parallelism, SPMD - you definitelly should google for those.

    Just quick hint:

    1.Sequential code:


    float * in = new float(100);
    float * out = new float(100);
  3. Thread: work_dim > 3?

    by pplaszew
    Replies
    1
    Views
    1,504

    Re: work_dim > 3?

    "4D code" - what a times... 8)

    I doubt dim > 3 will be there soon. Look at CUDA - 3 years on the market and still dim <= 3.
  4. Replies
    2
    Views
    1,550

    Re: Constraints of limited RAM on GPGPUs

    Before executing the kernel you need to copy input data to global memory and allocate global memory for output data. If total of both allocations is bigger than available gpu global memory you'll...
  5. Replies
    1
    Views
    1,386

    Re: functions called in a kernel

    hi,
    You need to change the signature of your functions to pass pointers to local mem:


    void Fun(__local float2* t1, __local float* t2, float eta, unsigned int v)


    If I understand you...
  6. Re: Building programs with pre-compiled binaries

    [/quote]

    I don't have customers :wink:
    But yeah, generally you're right. So actually if you want to have reliable app there's no way yet to hide your kernels source.
  7. Replies
    4
    Views
    4,357

    Re: Local memory allocation

    You mean when allocating statically? In kernel. Like this:


    __kernel void K(
    //.. kernel args
    ){
    //definition of s_el
    __local float s_el[32];
    //.. download data from global...
  8. Re: Higher register usage after migration from cuda

    EDIT: Forgot to mention: Cuda compiles my kernels for arch 11, opencl only to arch 10. Since I dont know how to force opencl to compile for particular architecture (is it possible?) I compiled with...
  9. Higher register usage after migration from cuda

    Hi,

    I've just migrated my program from cuda to opencl. It involved a bit work to change all the host code, like device initialization, memory allocation, kernel execution etc.

    For the device...
  10. Replies
    4
    Views
    4,357

    Local memory allocation

    Hi everyone,
    I've read somewhere (some forum I cannot recall right now) that allocating local ("shared" in nvidia cuda nomenclature) memory statically like below should be avoided since it's...
  11. Re: Building programs with pre-compiled binaries

    Hi,
    I think trying to build and recompiling on error quite ugly.
    I myself developed primitive configuration file solution. File is not meant to be edited manually - it stores info about builds. For...
Results 1 to 11 of 13