Search:

Type: Posts; User: PhilTillet

Search: Search took 0.00 seconds.

  1. Replies
    4
    Views
    1,827

    Re: Minimal kernel name length

    I can feel your sarcasm :D
    Indeed I made quite a mistake in my question, confusing minimal and maximal.

    It was more a matter of knowing until when we can have a unique representation in the...
  2. Replies
    4
    Views
    1,827

    Re: Minimal kernel name length

    Hey,

    this matters when dynamically generate code, and when the kernel function name also has to be generated. It can be useful to know if we generate a name too long for the standards for...
  3. Replies
    4
    Views
    1,827

    Minimal kernel name length

    Hello,

    I have searched in the standards and couldn't find the minimum kernel length as defined by the standards...
    Does anybody know what it is?

    Thanks !
  4. Re: Is it safe to cast a pointer-to-vector to pointer-to-sca

    I have done some further tests. In case someone wonders, on my NVidia OpenCL 1.1 SDK,
    *((float*)x) = 0
    Results in CL_OUT_OF_RESOURCES, so I guess such a cast has undefined behavior.

    In case some...
  5. Is it safe to cast a pointer-to-vector to pointer-to-scalar?

    Hello!

    Let's say x, y and z are three cl_mem(sizeof(float)*9) ... I was wondering if these kind of kernels were safe/portable. I couldn't find information on the standard whether this was legal or...
  6. Replies
    10
    Views
    4,978

    Re: can we use structure in opencl

    One important thing to keep in mind is also that size_t and ptrdiff_t are likely not to be the same size on the host and on the GPU, which can cause misalignment

    //Host code
    struct my_struct{
    ...
  7. Replies
    2
    Views
    1,129

    Using a buffer on multiple devices.

    Hello everybody,

    I have searched the web for similar questions but couldn't find the answer I need...

    I am making a dynamic scheduler (following AMD's OpenCL Guide) for handling multiple GPUs....
  8. Replies
    1
    Views
    1,034

    Re: Buffer sharing on multiple queues

    Hello,

    As far as I know, the behavior of using one handle on multiple devices is vendor-specific... nothing is said about "broadcasting" a cl_mem object from the host to all the devices...
    I...
Results 1 to 8 of 9