Search:

Type: Posts; User: notzed

Page 1 of 11 1 2 3 4

Search: Search took 0.01 seconds.

  1. Replies
    5
    Views
    2,384

    Re: Query Memory used on GPU

    The precise details would depend on the driver and implementation - and such details are proprietary.

    But both of those scenarios seem likely possible causes to me (of several).

    Note that...
  2. Replies
    5
    Views
    2,384

    Re: Query Memory used on GPU

    Err, virtualises that was supposed to be.
  3. Replies
    5
    Views
    2,384

    Re: Query Memory used on GPU

    I suspect there isn't such a query as it isn't something that can really be calculated.

    buffers and images only need to be 'on card' when a kernel using them is executing, they could otherwise be...
  4. Re: Is float point operation in OpenCL stochastic?

    No of course they aren't stochastic - floating point operations obviously work in a deterministic manner otherwise they wouldn't be much use would they? If they weren't deterministic in a GPU,...
  5. Re: When I want to use atomic_add, what should I do?

    IIRC opencl 1.2 includes atomics automatically.

    If invalid pragmas still work, well maybe that's a bug in the compiler (although i don't know if the spec requires them to fail or not).

    [/quote]
  6. Thread: Buffer Vs Image

    by notzed
    Replies
    7
    Views
    3,583

    Re: Buffer Vs Image

    Note I specifically said: "If you have regular grid access that can be fully coalesced at every access"

    At worst case, a single multiply is cheap compared to a non-fully-coalesced memory read, so...
  7. Re: When I want to use atomic_add, what should I do?

    I think 'yes' to all three.

    But really - these questions are easily tested by trying it. It wont compile or wont work if you get it wrong.
  8. Replies
    1
    Views
    1,585

    Re: Research topic for parallel programming

    I think if this is part of a course, then selecting the algorithm to demonstrate is part of the graded material too - because the answer to the proposition of "proving opencl is better than mpi" is...
  9. Replies
    3
    Views
    1,694

    Re: Array in global memory - strange results

    Just do an internets search for parallel reduction, parallel sum, and see where that leads you. It is a widely researched and documented topic.
  10. Thread: Buffer Vs Image

    by notzed
    Replies
    7
    Views
    3,583

    Re: Buffer Vs Image

    If you have regular grid access that can be fully coalesced at every access, and the data is stored as floats (requiring no conversion), and you don't need the image features (interpolation, etc),...
  11. Replies
    2
    Views
    1,422

    Re: 2 times sin() vs one local memory access

    The AMD APP programming guide has some fairly detailed numbers on memory/l1/lds/constants throughout chapter 4 - although it's not the same hardware it should be roughly comparable.

    You might...
  12. Replies
    6
    Views
    3,280

    Re: OpenCL 1.2 clCreateImage(...)

    You didn't set cl_image_desc.image_depth to anything.
  13. Replies
    5
    Views
    2,194

    Re: How Can I reduce the GPU Kernel Performance ?

    OpenCL 1.2 has some api's to partition the device. See section 4.3 of the opencl specification.
  14. Replies
    5
    Views
    2,194

    Re: How Can I reduce the GPU Kernel Performance ?

    You just have to break your code up so it runs in more but shorter kernels, at this time even when the hardware supports concurrent scheduling the load balancing isn't very good.

    Bugs can also...
  15. Replies
    3
    Views
    1,724

    Re: Best Practice for multiple dll

    I'm sure this is in the manual somewhere ...

    The opencl library IS exactly the same between all vendors - i believe khronos supplies it to them. But that is only a shell which discovers the...
  16. Replies
    2
    Views
    1,557

    Re: concurrent access to global memory

    Use an algorithm that doesn't require global sync?

    Unfortunately you don't give enough information to suggest more than that ...
  17. Re: Kernel produces vertical bars instead of solid color. Wh

    OpenCL implementations probably use GCD calculators to choose good values so it's only prime numbers that are problematic - but even if they chose 1x1 that kernel would still function fine. It's...
  18. Replies
    3
    Views
    2,904

    Re: OpenCL Synchronization between workgroups.

    FYI he's also asked it over here a few days earlier, and has more source + a ton of help.

    http://devgurus.amd.com/message/1284667

    One hopes he might summarise his findings should he get it to...
  19. Re: I beg u plzzz ,I really need help with this open cl code

    Is this homework or for marked study?
  20. Replies
    2
    Views
    1,647

    Re: global & local size in 2D problem

    I know you mentioned numbers in your question, but what are all these values actually set to here?

    Hint: global_size[x] needs to be an integer multiple of local_size[x]

    The error code you're...
  21. Replies
    6
    Views
    3,787

    Re: problem with openCL/multiple kernels

    You should probably just start a new topic.

    Since what you describe should work, you should include code, which will prevent misunderstandings or forgetting to mention important details.
  22. Re: CL_OUT_OF_RESOURCES when changing to OpenCL 1.1

    Look harder for bugs. Every time i move card vendors or technologies or operating systems i swear the vendor messed it up. Pretty much every time it was me that did.
  23. Re: CL/GL Interop, OSX -- ever shared a Renderbuffer or Text

    Ok, but, did you create gl buffer objects before creating the cl context, or after? I do the gl context, then all gl buffer objects, then the cl context (due to original compartmentalization; all...
  24. Replies
    4
    Views
    2,073

    Re: Kernel undeterministic behavior

    If you know "exactly what's wrong", then why did you ask why it isn't working?

    That's no way to get help.

    Anyway - masters or homework, any school work is yours and yours alone.
  25. Re: CL/GL Interop, OSX -- ever shared a Renderbuffer or Text

    Hmm, I created the gl context first then the cl context.

    I had to do some weird stuff to get the gl context out of the glut-like stuff provided by JOCL though, and/or run the cl init in the...
Results 1 to 25 of 271
Page 1 of 11 1 2 3 4