1 / 13

CS470/570 Lecture 5

CS470/570 Lecture 5. Introduction to OpenMP Compute Pi example OpenMP directives and options. Introduction to OpenMP. A set of compiler directives and library routines to support parallel programming in a share memory architecture Can be used with C, C++ and ForTran

bernie
Télécharger la présentation

CS470/570 Lecture 5

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. CS470/570 Lecture 5 • Introduction to OpenMP • Compute Pi example • OpenMP directives and options

  2. Introduction to OpenMP • A set of compiler directives and library routines to support parallel programming in a share memory architecture • Can be used with C, C++ and ForTran • A program created with OpenMP starts with a master thread • Various directives cause the program to created and execute multiple threads

  3. Introduction to OpenMP Additional Threads Execute Master Thread New Threads created with #pragma omp parallel Threads joined Only master thread remains

  4. Introduction to OpenMp • Directives have the following form • #pragma omp construct [clauses] • Directives are applied to a structured block of code • A structured block has one entry point (at the top) and one exit point (at the bottom)

  5. Introduction to OpenMp • Supports explicit parallelism • Programmer needs to specify where parallel threads should be used • Programmer needs to think about synchronization issues

  6. Simple Example • Sequential algorithm to calculate pi • Find the area under the curve 4/(1+x2) between 0 and 1. • Example of numerical integration • Parallelization of the algorithm using OpenMP

  7. Compute Pi

  8. Sequential Algorithm to Computer Pi int main(int argc, char *argv[]) { double midpoint, width, pi, sum; int numIntervals, I; numIntervals = atoi(argv[1]); width = 1.0 / numIntervals; pi = 0.0; for(i = 0; i < numIntervals; i++) { midpoint = (i + 0.5) * width; sum = sum +4.0 / (1.0 + midpoint * midpoint); } pi = width*sum; printf("%1.20f \n", pi); }

  9. Parallelized Algorithm to Computer Pi int main(int argc, char *argv[]) { double midpoint, width, pi, sum; int numIntervals, I; numIntervals = atoi(argv[1]); width = 1.0 / numIntervals; pi = 0.0; #pragma omp parallel for default(shared) private(i, midpoint)reduction(+:sum) for(i = 0; i < numIntervals; i++) { midpoint = (i + 0.5) * width; sum = sum +4.0 / (1.0 + midpoint * midpoint); } pi = width*sum; printf("%1.20f \n", pi); }

  10. OpenMP Directives • #pragma omp parallel • #pragma omp for • #pragma omp sections

  11. #pragma omp parallelExample void printId(int myId) { printf("%d\n", myId); } int main(int argc, char *argv[]) { #pragma omp parallel { printId(omp_get_thread_num()); } }

  12. #pragma omp for int main(int argc, char *argv[]) { int i; #pragma omp parallel for num_threads(4) for (i = 0; i < 10; i++) { printf("%d %d\n",omp_get_thread_num(),i); } }

  13. #pragma omp sections int main(int argc, char *argv[]) { #pragma omp parallel num_threads(4) { #pragma omp sections { #pragma omp section {printf("%d\n",omp_get_thread_num());} #pragma omp section {printf("%d\n",omp_get_thread_num());} #pragma omp section {printf("%d\n",omp_get_thread_num());} #pragma omp section {printf("%d\n",omp_get_thread_num());} } } }

More Related