Databases
StackOverflow2013
Postsdbo
POUsing openmp better
 

Note that there are some explanatory texts on larger screens.

plurals
  1. POUsing openmp better
    primarykey
    data
    text
    <p>I am trying to implement openMP, but like so many other posters before me, the result has simply been to slow the code down. Inspired by previous answers, I went from using <code>#pragma omp parallel for</code> to <code>#pragma omp task</code>, in the hope that I could avoid some overhead. Unfortunately, the parallelized code is still twice as slow as the serial. From other answers, it seems that the proper procedure depends on the specific demands of the code, which is why I thought I would have to ask a question myself.</p> <p>First the pseudo-code:</p> <pre><code>#pragma omp parallel { #pragma omp master while (will be run some hundreds of millions of times) { for (between 5 and 20 iterations) { #pragma omp task (something) } #pragma omp taskwait &lt;- it is important that all the above tasks are completed before going on (something) if (something) { (something) for (between 50 and 200 iterations) { #pragma omp task (something) } #pragma omp taskwait (something) } } } </code></pre> <p>Only the two for-loops can be parallelized, the rest must be done in the right order. I came up with putting the parallel and master directives outside the while-loop in an attempt at reducing the overhead of creating the team. </p> <p>I am also a bit curious whether I am using taskwait properly - the specification states that the "parent task" is put on hold until all child tasks have been executed, but it is not quite clear whether that terminology also applies here, where the task regions are not nested.</p> <p>Can anyone come up with a better way of using openMP, such that I may actually get a speed-up?</p> <p>EDIT: each step in the while-loop depends on all previous steps, and thus they have to be done serially, with an update at the end. It is an implementation of an "event-driven-algorithm" for simulating neural networks, if anyone was wondering.</p>
    singulars
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. PTQuestion
    1. USGrizzly
    1. USKaare
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. PO
      singulars
      1. PTAnswer
    2. PO
      singulars
      1. PTAnswer
    3. PO
      singulars
      1. PTAnswer
    1. VO
      singulars
      1. POUsing openmp better
      1. This table or related slice is empty.
      1. VTUpMod
    2. VO
      singulars
      1. POUsing openmp better
      1. This table or related slice is empty.
      1. VTUpMod
    1. COhow long does each iteration of the for loop take? if the tasks are to small it is quite possible that it is simply not possible to get a speedup here. Besides why would `#pragma omp task` be faster then `#pragma omp for`? Afterall the later should be able to get away with much less managing overhead. To me it seems that if it is faster you probably used the wrong scheduling mode for your situation. About taskwait: As I understand it the `master` section should be your parent task (or maybe the `parallel` section but that seems unlikely)
      singulars
      1. POUsing openmp better
      1. USGrizzly
    2. COI got the idea that tasks would be faster because an answer to an old question said something along the lines of "if there are too few iterations in the for-loop, you are better off using tasks instead". In the serial case, it is possible to go through 10000 iterations of the while loop in 1.7 seconds. Considering the other settings, a ball-park estimate would be 1.0-0.5 microseconds for each iteration of the second for-loop. I know it's short, but was told that I was underestimating the power of parallelization, and decided to give it a shot :)
      singulars
      1. POUsing openmp better
      1. USKaare
    3. COIt really sounds like you need to think about either a new algorithm, or a new parallel processing paradigm, or maybe even both.
      singulars
      1. POUsing openmp better
      1. UStalonmies
 

Querying!

 
Guidance
A row detail

Detail views are divided into sections. All the information in the data section comes from columns in the selected row. The other sections display data from other, related rows.

Related data can be related in a to-one or a to-many fashion. Captions of data related in a to-many fashion link to a list view showing a filtered view of the table.

Try moving around until you find a non-empty to-many entry and click on the label to get to one. You can move back to the root by clicking on the database name in the header.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload