An ordered loop is not optimized well, which causes sometimes a timeout in test/worksharing/for/omp_for_collapse.c if execution streams are oversubscribed.
#pragma parallel omp for ordered
for (i = 1; i < 10000; i++) {
#pragma omp ordered
comp(i);
}Although order is not widely used, it should be optimized by using synchronization primitives.
An ordered loop is not optimized well, which causes sometimes a timeout in
test/worksharing/for/omp_for_collapse.cif execution streams are oversubscribed.Although
orderis not widely used, it should be optimized by using synchronization primitives.