How to use Lock Functions (Lock Routines) from Openmp?

Question

I need to create 5 easy-to-understand algorithms for each of the blocking functions below in C or C++ to be able to exemplify how each one works and present to my office staff.

• omp_init_lock_with_hint
• omp_init_nest_lock_with_hint

The problem is that I barely understood the functioning of the blocking functions. I’m using as a reference the most updated Openmp documentation available in the link: http://www.openmp.org/wp-content/uploads/openmp-4.5.pdf (pages 270 to 275).

I wonder if you could explain to me how the blocking functions work and help create an example using one of the functions so that I understand the functioning and can create other examples to present to my class?

Grateful from the start.

Answer 1

The difference between the two specified functions, omp_init_lock_with_hint() and omp_init_nest_lock_with_hint() is that it initializes a lock simple and this initializes a lock nestled.

The difference between the two (apart from the obvious that one receives a omp_lock_t * and the other one omp_nest_lock_t *) is that a task that gets a simple lock via omp_set_lock() can’t call omp_set_lock() in the same lock without first releasing it using omp_unset_lock(), under penalty of generating a deadlock.

Other Openmp tasks can call omp_set_lock() in a latch already belonging to a different task; they will latch until the task that currently holds the latch releases it using omp_unset_lock() (and that the other tasks in front of it in the queue waiting for the latch also get and release the latch).

Already to a omp_nest_lock_t, if the task owner of the lock calls omp_set_nest_lock() in it, it will merely increase an internal counter, and by calling omp_unset_nest_lock() this counter will be decreased. If the value of the counter reaches zero, the lock is released. In other words, although the call to omp_set_nest_lock() in a lock already belonging to the task does not cause deadlock, the task has to issue so many omp_unset_nest_lock() how much issued omp_set_nest_lock() before the lock is released.

As for the usage examples, I will need some time to restore my knowledge in Openmp, but the general idea is that with the simple lock, each task gets the lock with omp_set_lock(), perform the synchronous and infallible operations (no exceptions) in the critical region, and then release the lock with omp_unset_lock(). In the case of nested_lock, the task can do more complex things, such as getting the lock with omp_set_nest_lock(), wait for an event and call omp_unset_nest_lock() in a callback.