hpx::execution::experimental::get_parallel_scheduler

hpx::execution::experimental::get_parallel_scheduler#

Defined in header hpx/execution.hpp.

See Public API for a list of names and headers that are part of the public HPX API.

See also:

parallel_scheduler hpx::execution::experimental::get_parallel_scheduler()#

Return a parallel_scheduler that schedules on the default HPX thread pool.

This is HPX’s implementation of the P2079 std::execution::get_parallel_scheduler() entry point. Work started on the returned scheduler uses the same worker threads as the default pool rather than a separate system thread pool.

Note

For a named resource-partitioner pool, use the overload that takes hpx::threads::thread_pool_base&.

Returns:

A scheduler bound to HPX’s default thread pool.

parallel_scheduler hpx::execution::experimental::get_parallel_scheduler(hpx::threads::thread_pool_base &pool)#

Return a parallel_scheduler that runs work on pool.

Look named pools up with hpx::resource::get_thread_pool. A misspelled or unknown name throws hpx::exception; it does not produce undefined behavior.

Parameters:

pool – HPX thread pool that should execute scheduled work.

Pre:

pool must outlive the returned scheduler and every operation started on it. The scheduler stores a non-owning pointer; it does not extend the pool’s lifetime. This precondition is not enforced at runtime (caller responsibility). Debug builds assert that pool currently reports at least one OS thread.

Returns:

A scheduler bound to pool.