Parallel Computing on Heterogeneous Networks (Wiley Series on Parallel and Distributed Computing)

O

OpenMP, 80–94

construct, 81

directive, 80–81

example

multithreaded dot product, 94–94

parallel regions, 81–82

private variables, 88–90

reduction operations, 91

runtime library functions, 91–93

omp_get_max_threads, 92

omp_get_num_procs, 92

omp_get_num_threads, 92

omp_get_thread_num, 82, 92

omp_set_num_threads, 81–82, 92

shared variables, 88, 90

synchronization constructs

atomic, 87

critical, 86–87

ordered, 85

synchronization directives

barrier, 85

flush, 88

work-sharing constructs, 82–86

for, 82–84

master, 85

sections, 84–85

single, 85

Optimizing compilers

for shared memory multiprocessors, 67

for vector and superscalar processors, 22–33

Orbix, 332, 336, 338