6.B.2 OpenMP Use: Parallelization Scheme

• Split into p parallel tasks

  (p = number of processors)
  

Program Seismic

initialization

C$OMP PARALLEL

call main_subroutine()

C$OMP END PARALLEL

initialization done by master processor only

computation enclosed in a parallel region

® SPMD execution scheme

Previous slide

Next slide

Back to first slide

View graphic version