Search:

Type: Posts; User: mustang

Page 1 of 2 1 2

Search: Search took 0.00 seconds.

  1. Replies
    2
    Views
    1,136

    Re: questions about gpu computing

    Thank you for answering!!
  2. Replies
    2
    Views
    1,136

    questions about gpu computing

    Hi,
    I would like to ask you the following questions:
    1- do you know why the gpu is good for data parallel but not for task parallel? the arquitecture? why?
    2- I heard the GPU is much slower...
  3. Thread: dot product

    by mustang
    Replies
    6
    Views
    1,819

    Re: dot product

    thanks, anyway the idea up to the moment is trying to make the method to run faster in the GPU and maybe then do the dynamic termination!!

    Pablo
  4. Thread: dot product

    by mustang
    Replies
    6
    Views
    1,819

    Re: dot product

    I changed the program and I do all the operations in kernels now even the escalar operations. I guess that something that currently Im not doing but in a better version should be necessary to...
  5. Thread: dot product

    by mustang
    Replies
    6
    Views
    1,819

    Re: dot product

    Thanks!! I will search for it, the idea would be a kernel to multiply position per position and sum all, am I right? the problem is that a time ago I tried just doing the first part and only that...
  6. Thread: dot product

    by mustang
    Replies
    6
    Views
    1,819

    dot product

    Hi,
    I would like to know if someone knows how to implement the dot product in a way that is efficient, at least not slower or not much slower than doing it in a CPU. My idea is to implement the...
  7. Replies
    9
    Views
    2,697

    Re: diferent time in openCL programs execution

    Thanks, I guess it is not that I forgot to allocate a buffer because I have used the same program with smaller matrix and it seemed to work fine!! and I have centralized the points where I change the...
  8. Replies
    2
    Views
    1,017

    Re: inconsistent data

    thanks again!! yes I know but I saw it used sometimes where I thought it would not be necessary so I started to have doubts!!

    Pablo
  9. Replies
    2
    Views
    1,017

    inconsistent data

    Hi,
    I have a doubt: if Im working with data parallel (using ClEnqueueNDRangeKernel..) and Im launching many kernels and in some cases the data that is the input of a kernel is the output of a...
  10. Replies
    9
    Views
    2,697

    Re: diferent time in openCL programs execution

    Hi,
    I know this is an off topic but I prefer not to open another topic, but my question is if it is posible that I cant handle a 19713X19713 matriz (it is not sparse). it throws segment fault (I...
  11. Replies
    9
    Views
    2,697

    Re: diferent time in openCL programs execution

    thanks for the reply!!!
    Im worried about setting arguments and creating the buffers inside because I made the conjugate gradient method with space matrix in both gpu and cpu and the cpu is much...
  12. Replies
    9
    Views
    2,697

    Re: diferent time in openCL programs execution

    thanks for answering!!!
    I guess Im forced to pass read to the cpu some results cause I have to do for example a few scalar divisions (for example a float divided by another float and the result is...
  13. Replies
    9
    Views
    2,697

    Re: diferent time in openCL programs execution

    Thanks!! So I guess the real time is not first but the following!! thats good! but what if I had a program that takes a lot of time? should I execute a smaller program before executing the main...
  14. Replies
    9
    Views
    2,697

    diferent time in openCL programs execution

    Hi again,
    I have done a simple matrix vector product with sparse matrix, and it works but the problem is that the first time I executed it takes much more time than the following times, for...
  15. Replies
    9
    Views
    2,564

    Re: doubts with work items and groups

    Thanks again!! I was using gettimeofday() but it didnt seem very precise, probably Im wrong. What is bad for me is that if been trying to implement a version of conjugate gradient method, but if...
  16. Replies
    9
    Views
    2,564

    Re: doubts with work items and groups

    Ive been reading and as usually some questions are answered and others rise!! For example, it is not clear, at least for me, what executes the work-items, as it seems to be the CUDA cores, but there...
  17. Replies
    9
    Views
    2,564

    Re: doubts with work items and groups

    Thanks once again!! Ill read those sections!!

    Thanks!!
    Pablo
  18. Replies
    9
    Views
    2,564

    Re: doubts with work items and groups

    Thank you again!!! And regarding execution, I have an nvidia card (it is a 9400/ION), is it correct that each work group is executed by only one CUDA core (in test deviceQuery it says that it has...
  19. Replies
    9
    Views
    2,564

    Re: doubts with work items and groups

    thanks for answering!! but then I have a doubt, in clEnqueueNDRangeKernel() there are two parameters which are const size_t global_work_size and const size_t local_work_size, for example if I...
  20. Replies
    9
    Views
    2,564

    doubts with work items and groups

    Hello, Im new in the forum and I would like to ask something: I would like to know if the max amount of work items is the one that says max number of the dimension, in my case 512x512x64 or if in...
Results 1 to 20 of 30
Page 1 of 2 1 2