Implementing SPMD

program work

C$OMP PARALLEL DEFAULT(PRIVATE)

C$OMP& SHARED(N,global)

nthreads = omp_get_num_threads()

iam = omp_get_thread_num()

ichunk = N/nthreads

istart = iam*ichunk

iend = iam+1*ichunk -1

call my_sum(istart, iend, local)

C$OMP ATOMIC

global = global + local

C$OMP END PARALLEL

print*, global

end

Each thread works on its piece.

Previous slide

Next slide

Back to first slide

View graphic version