/*
* This file is part of the StarPU Handbook.
* Copyright (C) 2014 INRIA
* See the file version.doxy for copying conditions.
*/
/*! \defgroup API_OpenMP_Runtime_Support OpenMP Runtime Support
\brief This section describes the interface provided for implementing OpenMP runtimes on top of StarPU.
\struct starpu_omp_lock_t
\ingroup API_OpenMP_Runtime_Support
Opaque Simple Lock object (\ref SimpleLock) for inter-task synchronization operations.
\sa starpu_omp_init_lock()
\sa starpu_omp_destroy_lock()
\sa starpu_omp_set_lock()
\sa starpu_omp_unset_lock()
\sa starpu_omp_test_lock()
\var starpu_omp_lock_t::internal
Is an opaque pointer for internal use.
\struct starpu_omp_nest_lock_t
\ingroup API_OpenMP_Runtime_Support
Opaque Nestable Lock object (\ref NestableLock) for inter-task synchronization operations.
\sa starpu_omp_init_nest_lock()
\sa starpu_omp_destroy_nest_lock()
\sa starpu_omp_set_nest_lock()
\sa starpu_omp_unset_nest_lock()
\sa starpu_omp_test_nest_lock()
\var starpu_omp_nest_lock_t::internal
Is an opaque pointer for internal use.
\enum starpu_omp_sched_value
\ingroup API_OpenMP_Runtime_Support
Set of constants for selecting the for loop iteration scheduling algorithm
(\ref OMPFor) as defined by the OpenMP specification.
\var starpu_omp_sched_value::starpu_omp_sched_undefined
\ingroup API_OpenMP_Runtime_Support
Undefined iteration scheduling algorithm.
\var starpu_omp_sched_value::starpu_omp_sched_static
\ingroup API_OpenMP_Runtime_Support
\b Static iteration scheduling algorithm.
\var starpu_omp_sched_value::starpu_omp_sched_dynamic
\ingroup API_OpenMP_Runtime_Support
\b Dynamic iteration scheduling algorithm.
\var starpu_omp_sched_value::starpu_omp_sched_guided
\ingroup API_OpenMP_Runtime_Support
\b Guided iteration scheduling algorithm.
\var starpu_omp_sched_value::starpu_omp_sched_auto
\ingroup API_OpenMP_Runtime_Support
\b Automatically choosen iteration scheduling algorithm.
\var starpu_omp_sched_value::starpu_omp_sched_runtime
\ingroup API_OpenMP_Runtime_Support
Choice of iteration scheduling algorithm deferred at \b runtime.
\sa starpu_omp_for()
\sa starpu_omp_for_inline_first()
\sa starpu_omp_for_inline_next()
\sa starpu_omp_for_alt()
\sa starpu_omp_for_inline_first_alt()
\sa starpu_omp_for_inline_next_alt()
\enum starpu_omp_proc_bind_value
\ingroup API_OpenMP_Runtime_Support
Set of constants for selecting the processor binding method, as defined in the
OpenMP specification.
\var starpu_omp_proc_bind_value::starpu_omp_proc_bind_undefined
\ingroup API_OpenMP_Runtime_Support
Undefined processor binding method.
\var starpu_omp_proc_bind_value::starpu_omp_proc_bind_false
\ingroup API_OpenMP_Runtime_Support
Team threads may be moved between places at any time.
\var starpu_omp_proc_bind_value::starpu_omp_proc_bind_true
\ingroup API_OpenMP_Runtime_Support
Team threads may not be moved between places.
\var starpu_omp_proc_bind_value::starpu_omp_proc_bind_master
\ingroup API_OpenMP_Runtime_Support
Assign every thread in the team to the same place as the \b master thread.
\var starpu_omp_proc_bind_value::starpu_omp_proc_bind_close
\ingroup API_OpenMP_Runtime_Support
Assign every thread in the team to a place \b close to the parent thread.
\var starpu_omp_proc_bind_value::starpu_omp_proc_bind_spread
\ingroup API_OpenMP_Runtime_Support
Assign team threads as a sparse distribution over the selected places.
\sa starpu_omp_get_proc_bind()
\struct starpu_omp_parallel_region_attr
\ingroup API_OpenMP_Runtime_Support
Set of attributes used for creating a new parallel region.
\sa starpu_omp_parallel_region()
\var struct starpu_codelet starpu_omp_parallel_region_attr::cl
Is a ::starpu_codelet (\ref API_Codelet_And_Tasks) to use for the parallel region
implicit tasks. The codelet must provide a CPU implementation function.
\var starpu_data_handle_t *starpu_omp_parallel_region_attr::handles
Is an array of zero or more ::starpu_data_handle_t data handle to be passed to
the parallel region implicit tasks.
\var void *starpu_omp_parallel_region_attr::cl_arg
Is an optional pointer to an inline argument to be passed to the region implicit tasks.
\var size_t starpu_omp_parallel_region_attr::cl_arg_size
Is the size of the optional inline argument to be passed to the region implicit tasks, or 0 if unused.
\var unsigned starpu_omp_parallel_region_attr::cl_arg_free
Is a boolean indicating whether the optional inline argument should be automatically freed (true), or not (false).
\var int starpu_omp_parallel_region_attr::if_clause
Is a boolean indicating whether the \b if clause of the corresponding pragma
omp parallel is true or false.
\var int starpu_omp_parallel_region_attr::num_threads
Is an integer indicating the requested number of threads in the team of the
newly created parallel region, or 0 to let the runtime choose the number of
threads alone. This attribute may be ignored by the runtime system if the
requested number of threads is higher than the number of threads that the
runtime can create.
\struct starpu_omp_task_region_attr
\ingroup API_OpenMP_Runtime_Support
Set of attributes used for creating a new task region.
\sa starpu_omp_task_region()
\var struct starpu_codelet starpu_omp_task_region_attr::cl
Is a ::starpu_codelet (\ref API_Codelet_And_Tasks) to use for the task region
explicit task. The codelet must provide a CPU implementation function or an
accelerator implementation for offloaded target regions.
\var starpu_data_handle_t *starpu_omp_task_region_attr::handles
Is an array of zero or more ::starpu_data_handle_t data handle to be passed to
the task region explicit tasks.
\var void *starpu_omp_task_region_attr::cl_arg
Is an optional pointer to an inline argument to be passed to the region implicit tasks.
\var size_t starpu_omp_task_region_attr::cl_arg_size
Is the size of the optional inline argument to be passed to the region implicit
tasks, or 0 if unused.
\var unsigned starpu_omp_task_region_attr::cl_arg_free
Is a boolean indicating whether the optional inline argument should be
automatically freed (true), or not (false).
\var int starpu_omp_task_region_attr::if_clause
Is a boolean indicating whether the \b if clause of the corresponding pragma
omp task is true or false.
\var int starpu_omp_task_region_attr::final_clause
Is a boolean indicating whether the \b final clause of the corresponding pragma
omp task is true or false.
\var int starpu_omp_task_region_attr::untied_clause
Is a boolean indicating whether the \b untied clause of the corresponding pragma
omp task is true or false.
\var int starpu_omp_task_region_attr::mergeable_clause
Is a boolean indicating whether the \b mergeable clause of the corresponding pragma
omp task is true or false.
@name Initialisation
\ingroup API_OpenMP_Runtime_Support
\def STARPU_OPENMP
\ingroup API_OpenMP_Runtime_Support
This macro is defined when StarPU has been installed with OpenMP Runtime
support. It should be used in your code to detect the availability of
the runtime support for OpenMP.
\fn int starpu_omp_init(void)
\ingroup API_OpenMP_Runtime_Support
Initializes StarPU and its OpenMP Runtime support.
\fn void starpu_omp_shutdown(void)
\ingroup API_OpenMP_Runtime_Support
Shutdown StarPU and its OpenMP Runtime support.
@name Parallel
\anchor ORS_Parallel
\ingroup API_OpenMP_Runtime_Support
\fn void starpu_omp_parallel_region(const struct starpu_omp_parallel_region_attr *attr)
\ingroup API_OpenMP_Runtime_Support
Generates and launch an OpenMP parallel region and return after its
completion. \p attr specifies the attributes for the generated parallel region.
If this function is called from inside another, generating, parallel region, the
generated parallel region is nested within the generating parallel region.
This function can be used to implement \#pragma omp parallel.
\fn void starpu_omp_master(void (*f)(void *arg), void *arg)
\ingroup API_OpenMP_Runtime_Support
Executes a function only on the master thread of the OpenMP
parallel region it is called from. When called from a thread that is not the
master of the parallel region it is called from, this function does nothing. \p
f is the function to be called. \p arg is an argument passed to function \p f.
This function can be used to implement \#pragma omp master.
\fn int starpu_omp_master_inline(void)
\ingroup API_OpenMP_Runtime_Support
Determines whether the calling thread is the master of the OpenMP parallel region
it is called from or not.
This function can be used to implement \#pragma omp master without code
outlining.
\return !0 if called by the region's master thread.
\return 0 if not called by the region's master thread.
@name Synchronization
\anchor ORS_Synchronization
\ingroup API_OpenMP_Runtime_Support
\fn void starpu_omp_barrier(void)
\ingroup API_OpenMP_Runtime_Support
Waits until each participating thread of the innermost OpenMP parallel region
has reached the barrier and each explicit OpenMP task bound to this region has
completed its execution.
This function can be used to implement \#pragma omp barrier.
\fn void starpu_omp_critical(void (*f)(void *arg), void *arg, const char *name)
\ingroup API_OpenMP_Runtime_Support
Waits until no other thread is executing within the context of the selected
critical section, then proceeds to the exclusive execution of a function within
the critical section. \p f is the function to be executed in the critical
section. \p arg is an argument passed to function \p f. \p name is the name of
the selected critical section. If name == NULL, the selected critical
section is the unique anonymous critical section.
This function can be used to implement \#pragma omp critical.
\fn void starpu_omp_critical_inline_begin(const char *name)
\ingroup API_OpenMP_Runtime_Support
Waits until execution can proceed exclusively within the context of the
selected critical section. \p name is the name of the selected critical
section. If name == NULL, the selected critical section is the unique
anonymous critical section.
This function together with #starpu_omp_critical_inline_end can be used to
implement \#pragma omp critical without code outlining.
\fn void starpu_omp_critical_inline_end(const char *name)
\ingroup API_OpenMP_Runtime_Support
Ends the exclusive execution within the context of the selected critical
section. \p name is the name of the selected critical section. If
name==NULL, the selected critical section is the unique anonymous
critical section.
This function together with #starpu_omp_critical_inline_begin can be used to
implement \#pragma omp critical without code outlining.
@name Worksharing
\anchor ORS_Worksharing
\ingroup API_OpenMP_Runtime_Support
\fn void starpu_omp_single(void (*f)(void *arg), void *arg, int nowait)
\ingroup API_OpenMP_Runtime_Support
Ensures that a single participating thread of the innermost OpenMP parallel
region executes a function. \p f is the function to be executed by a single
thread. \p arg is an argument passed to function \p f. \p nowait is a flag
indicating whether an implicit barrier is requested after the single section
(nowait==0) or not (nowait==!0).
This function can be used to implement \#pragma omp single.
\fn int starpu_omp_single_inline(void)
\ingroup API_OpenMP_Runtime_Support
Decides whether the current thread is elected to run the following single
section among the participating threads of the innermost OpenMP parallel
region.
This function can be used to implement \#pragma omp single without code
outlining.
\return !0 if the calling thread has won the election.
\return 0 if the calling thread has lost the election.
\fn void starpu_omp_single_copyprivate(void (*f)(void *arg, void *data, unsigned long long data_size), void *arg, void *data, unsigned long long data_size)
\ingroup API_OpenMP_Runtime_Support
This function executes \p f on a single task of the current parallel region
task, and then broadcast the contents of the memory block pointed by the
copyprivate pointer \p data and of size \p data_size to the corresponding \p
data pointed memory blocks of all the other participating region tasks. This
function can be used to implement \#pragma omp single with a copyprivate
clause.
\sa starpu_omp_single_copyprivate_inline
\sa starpu_omp_single_copyprivate_inline_begin
\sa starpu_omp_single_copyprivate_inline_end
\fn void *starpu_omp_single_copyprivate_inline_begin(void *data)
\ingroup API_OpenMP_Runtime_Support
This function elects one task among the tasks of the current parallel region
task to execute the following single section, and then broadcast the
copyprivate pointer \p data to all the other participating region tasks. This
function can be used to implement \#pragma omp single with a copyprivate
clause without code outlining.
\sa starpu_omp_single_copyprivate_inline
\sa starpu_omp_single_copyprivate_inline_end
\fn void starpu_omp_single_copyprivate_inline_end(void)
\ingroup API_OpenMP_Runtime_Support
This function completes the execution of a single section and returns the
broadcasted copyprivate pointer for tasks that lost the election and NULL for
the task that won the election. This function can be used to implement
\#pragma omp single with a copyprivate clause without code outlining.
\return the copyprivate pointer for tasks that lost the election and therefore did not execute the code of the single section.
\return NULL for the task that won the election and executed the code of the single section.
\sa starpu_omp_single_copyprivate_inline
\sa starpu_omp_single_copyprivate_inline_begin
\fn void starpu_omp_for(void (*f)(unsigned long long _first_i, unsigned long long _nb_i, void *arg), void *arg, unsigned long long nb_iterations, unsigned long long chunk, int schedule, int ordered, int nowait)
\ingroup API_OpenMP_Runtime_Support
Executes a parallel loop together with the other threads participating to the
innermost parallel region. \p f is the function to be executed iteratively. \p
arg is an argument passed to function \p f. \p nb_iterations is the number of
iterations to be performed by the parallel loop. \p chunk is the number of
consecutive iterations that should be affected to the same thread when
scheduling the loop workshares, it follows the semantics of the \c modifier
argument in OpenMP \#pragma omp for specification. \p schedule is the
scheduling mode according to the OpenMP specification. \p ordered is a flag
indicating whether the loop region may contain an ordered section
(ordered==!0) or not (ordered==0). \p nowait is a flag
indicating whether an implicit barrier is requested after the for section
(nowait==0) or not (nowait==!0).
The function \p f will be called with arguments \p _first_i, the first iteration
to perform, \p _nb_i, the number of consecutive iterations to perform before
returning, \p arg, the free \p arg argument.
This function can be used to implement \#pragma omp for.
\fn int starpu_omp_for_inline_first(unsigned long long nb_iterations, unsigned long long chunk, int schedule, int ordered, unsigned long long *_first_i, unsigned long long *_nb_i)
\ingroup API_OpenMP_Runtime_Support
Decides whether the current thread should start to execute a parallel loop
section. See #starpu_omp_for for the argument description.
This function together with #starpu_omp_for_inline_next can be used to
implement \#pragma omp for without code outlining.
\return !0 if the calling thread participates to the loop region and
should execute a first chunk of iterations. In that case, \p *_first_i will be
set to the first iteration of the chunk to perform and \p *_nb_i will be set to
the number of iterations of the chunk to perform.
\return 0 if the calling thread does not participate to the loop region
because all the available iterations have been affected to the other threads of
the parallel region.
\sa starpu_omp_for
\fn int starpu_omp_for_inline_next(unsigned long long nb_iterations, unsigned long long chunk, int schedule, int ordered, unsigned long long *_first_i, unsigned long long *_nb_i)
\ingroup API_OpenMP_Runtime_Support
Decides whether the current thread should continue to execute a parallel loop
section. See #starpu_omp_for for the argument description.
This function together with #starpu_omp_for_inline_first can be used to
implement \#pragma omp for without code outlining.
\return !0 if the calling thread should execute a next chunk of
iterations. In that case, \p *_first_i will be set to the first iteration of the
chunk to perform and \p *_nb_i will be set to the number of iterations of the
chunk to perform.
\return 0 if the calling thread does not participate anymore to the loop
region because all the available iterations have been affected to the other
threads of the parallel region.
\sa starpu_omp_for
\fn void starpu_omp_for_alt(void (*f)(unsigned long long _begin_i, unsigned long long _end_i, void *arg), void *arg, unsigned long long nb_iterations, unsigned long long chunk, int schedule, int ordered, int nowait)
\ingroup API_OpenMP_Runtime_Support
Alternative implementation of a parallel loop. This function differs from
#starpu_omp_for in the expected arguments of the loop function \c f.
The function \p f will be called with arguments \p _begin_i, the first iteration
to perform, \p _end_i, the first iteration not to perform before
returning, \p arg, the free \p arg argument.
This function can be used to implement \#pragma omp for.
\sa starpu_omp_for
\fn int starpu_omp_for_inline_first_alt(unsigned long long nb_iterations, unsigned long long chunk, int schedule, int ordered, unsigned long long *_begin_i, unsigned long long *_end_i)
\ingroup API_OpenMP_Runtime_Support
Inline version of the alternative implementation of a parallel loop.
This function together with #starpu_omp_for_inline_next_alt can be used to
implement \#pragma omp for without code outlining.
\sa starpu_omp_for
\sa starpu_omp_for_alt
\sa starpu_omp_for_inline_first
\fn int starpu_omp_for_inline_next_alt(unsigned long long nb_iterations, unsigned long long chunk, int schedule, int ordered, unsigned long long *_begin_i, unsigned long long *_end_i)
\ingroup API_OpenMP_Runtime_Support
Inline version of the alternative implementation of a parallel loop.
This function together with #starpu_omp_for_inline_first_alt can be used to
implement \#pragma omp for without code outlining.
\sa starpu_omp_for
\sa starpu_omp_for_alt
\sa starpu_omp_for_inline_next
\fn void starpu_omp_ordered(void (*f)(void *arg), void *arg)
\ingroup API_OpenMP_Runtime_Support
Ensures that a function is sequentially executed once for each iteration in
order within a parallel loop, by the thread that own the iteration. \p f is the
function to be executed by the thread that own the current iteration. \p arg is
an argument passed to function \p f.
This function can be used to implement \#pragma omp ordered.
\fn void starpu_omp_ordered_inline_begin(void)
\ingroup API_OpenMP_Runtime_Support
Waits until all the iterations of a parallel loop below the iteration owned by
the current thread have been executed.
This function together with #starpu_omp_ordered_inline_end can be used to
implement \#pragma omp ordered without code code outlining.
\fn void starpu_omp_ordered_inline_end(void)
\ingroup API_OpenMP_Runtime_Support
Notifies that the ordered section for the current iteration has been completed.
This function together with #starpu_omp_ordered_inline_begin can be used to
implement \#pragma omp ordered without code code outlining.
\fn void starpu_omp_sections(unsigned long long nb_sections, void (**section_f)(void *arg), void **section_arg, int nowait)
\ingroup API_OpenMP_Runtime_Support
Ensures that each function of a given array of functions is executed by one and
only one thread. \p nb_sections is the number of functions in the array \p
section_f. \p section_f is the array of functions to be executed as sections. \p
section_arg is an array of arguments to be passed to the corresponding function.
\p nowait is a flag indicating whether an implicit barrier is requested after
the execution of all the sections (nowait==0) or not (nowait==!0).
This function can be used to implement \#pragma omp sections and \#pragma omp section.
\fn void starpu_omp_sections_combined(unsigned long long nb_sections, void (*section_f)(unsigned long long section_num, void *arg), void *section_arg, int nowait)
\ingroup API_OpenMP_Runtime_Support
Alternative implementation of sections. This function differs from
#starpu_omp_sections in that all the sections are combined within a single
function in this version. \p section_f is the function implementing the combined
sections.
The function \p section_f will be called with arguments \p section_num, the
section number to be executed, \p arg, the entry of \p section_arg corresponding
to this section.
This function can be used to implement \#pragma omp sections and \#pragma omp section.
\sa starpu_omp_sections
@name Task
\anchor ORS_Task
\ingroup API_OpenMP_Runtime_Support
\fn void starpu_omp_task_region(const struct starpu_omp_task_region_attr *attr)
\ingroup API_OpenMP_Runtime_Support
Generates an explicit child task. The execution of the generated task is
asynchronous with respect to the calling code unless specified otherwise.
\p attr specifies the attributes for the generated task region.
This function can be used to implement \#pragma omp task.
\fn void starpu_omp_taskwait(void)
\ingroup API_OpenMP_Runtime_Support
Waits for the completion of the tasks generated by the current task. This
function does not wait for the descendants of the tasks generated by the current
task.
This function can be used to implement \#pragma omp taskwait.
\fn void starpu_omp_taskgroup(void (*f)(void *arg), void *arg)
\ingroup API_OpenMP_Runtime_Support
Launches a function and wait for the completion of every descendant task
generated during the execution of the function.
This function can be used to implement \#pragma omp taskgroup.
\sa starpu_omp_taskgroup_inline_begin
\sa starpu_omp_taskgroup_inline_end
\fn void starpu_omp_taskgroup_inline_begin(void)
\ingroup API_OpenMP_Runtime_Support
Launches a function and gets ready to wait for the completion of every descendant task
generated during the dynamic scope of the taskgroup.
This function can be used to implement \#pragma omp taskgroup without code outlining.
\sa starpu_omp_taskgroup
\sa starpu_omp_taskgroup_inline_end
\fn void starpu_omp_taskgroup_inline_end(void)
\ingroup API_OpenMP_Runtime_Support
Waits for the completion of every descendant task
generated during the dynamic scope of the taskgroup.
This function can be used to implement \#pragma omp taskgroup without code outlining.
\sa starpu_omp_taskgroup
\sa starpu_omp_taskgroup_inline_begin
@name API
\anchor ORS_API
\ingroup API_OpenMP_Runtime_Support
\fn void starpu_omp_set_num_threads(int threads)
\ingroup API_OpenMP_Runtime_Support
This function sets ICVS nthreads_var for the parallel regions to be created
with the current region.
Note: The StarPU OpenMP runtime support currently ignores
this setting for nested parallel regions.
\sa starpu_omp_get_num_threads
\sa starpu_omp_get_thread_num
\sa starpu_omp_get_max_threads
\sa starpu_omp_get_num_procs
\fn int starpu_omp_get_num_threads()
\ingroup API_OpenMP_Runtime_Support
This function returns the number of threads of the current region.
\return the number of threads of the current region.
\sa starpu_omp_set_num_threads
\sa starpu_omp_get_thread_num
\sa starpu_omp_get_max_threads
\sa starpu_omp_get_num_procs
\fn int starpu_omp_get_thread_num()
\ingroup API_OpenMP_Runtime_Support
This function returns the rank of the current thread among the threads
of the current region.
\return the rank of the current thread in the current region.
\sa starpu_omp_set_num_threads
\sa starpu_omp_get_num_threads
\sa starpu_omp_get_max_threads
\sa starpu_omp_get_num_procs
\fn int starpu_omp_get_max_threads()
\ingroup API_OpenMP_Runtime_Support
This function returns the maximum number of threads that can be used to
create a region from the current region.
\return the maximum number of threads that can be used to create a region from the current region.
\sa starpu_omp_set_num_threads
\sa starpu_omp_get_num_threads
\sa starpu_omp_get_thread_num
\sa starpu_omp_get_num_procs
\fn int starpu_omp_get_num_procs(void)
\ingroup API_OpenMP_Runtime_Support
This function returns the number of StarPU CPU workers.
\return the number of StarPU CPU workers.
\sa starpu_omp_set_num_threads
\sa starpu_omp_get_num_threads
\sa starpu_omp_get_thread_num
\sa starpu_omp_get_max_threads
\fn int starpu_omp_in_parallel(void)
\ingroup API_OpenMP_Runtime_Support
This function returns whether it is called from the scope of a parallel region or not.
\return !0 if called from a parallel region scope.
\return 0 otherwise.
\fn void starpu_omp_set_dynamic(int dynamic_threads)
\ingroup API_OpenMP_Runtime_Support
This function enables (1) or disables (0) dynamically adjusting the number of parallel threads.
Note: The StarPU OpenMP runtime support currently ignores the argument of this function.
\sa starpu_omp_get_dynamic
\fn int starpu_omp_get_dynamic(void)
\ingroup API_OpenMP_Runtime_Support
This function returns the state of dynamic thread number adjustment.
\return !0 if dynamic thread number adjustment is enabled.
\return 0 otherwise.
\sa starpu_omp_set_dynamic
\fn void starpu_omp_set_nested(int nested)
\ingroup API_OpenMP_Runtime_Support
This function enables (1) or disables (0) nested parallel regions.
Note: The StarPU OpenMP runtime support currently ignores the argument of this function.
\sa starpu_omp_get_nested
\sa starpu_omp_get_max_active_levels
\sa starpu_omp_set_max_active_levels
\sa starpu_omp_get_level
\sa starpu_omp_get_active_level
\fn int starpu_omp_get_nested(void)
\ingroup API_OpenMP_Runtime_Support
This function returns whether nested parallel sections are enabled or not.
\return !0 if nested parallel sections are enabled.
\return 0 otherwise.
\sa starpu_omp_set_nested
\sa starpu_omp_get_max_active_levels
\sa starpu_omp_set_max_active_levels
\sa starpu_omp_get_level
\sa starpu_omp_get_active_level
\fn int starpu_omp_get_cancellation(void)
\ingroup API_OpenMP_Runtime_Support
This function returns the state of the cancel ICVS var.
\fn void starpu_omp_set_schedule(enum starpu_omp_sched_value kind, int modifier)
\ingroup API_OpenMP_Runtime_Support
This function sets the default scheduling kind for upcoming loops within the
current parallel section. \p kind is the scheduler kind, \p modifier
complements the scheduler kind with informations such as the chunk size,
in accordance with the OpenMP specification.
\sa starpu_omp_get_schedule
\fn void starpu_omp_get_schedule(enum starpu_omp_sched_value *kind, int *modifier)
\ingroup API_OpenMP_Runtime_Support
This function returns the current selected default loop scheduler.
\return the kind and the modifier of the current default loop scheduler.
\sa starpu_omp_set_schedule
\fn int starpu_omp_get_thread_limit(void)
\ingroup API_OpenMP_Runtime_Support
This function returns the number of StarPU CPU workers.
\return the number of StarPU CPU workers.
\fn void starpu_omp_set_max_active_levels(int max_levels)
\ingroup API_OpenMP_Runtime_Support
This function sets the maximum number of allowed active parallel section levels.
Note: The StarPU OpenMP runtime support currently ignores the argument of this function and assume \p max_levels equals 1 instead.
\sa starpu_omp_set_nested
\sa starpu_omp_get_nested
\sa starpu_omp_get_max_active_levels
\sa starpu_omp_get_level
\sa starpu_omp_get_active_level
\fn int starpu_omp_get_max_active_levels(void)
\ingroup API_OpenMP_Runtime_Support
This function returns the current maximum number of allowed active parallel section levels
\return the current maximum number of allowed active parallel section levels.
\sa starpu_omp_set_nested
\sa starpu_omp_get_nested
\sa starpu_omp_set_max_active_levels
\sa starpu_omp_get_level
\sa starpu_omp_get_active_level
\fn int starpu_omp_get_level(void)
\ingroup API_OpenMP_Runtime_Support
This function returns the nesting level of the current parallel section.
\return the nesting level of the current parallel section.
\sa starpu_omp_set_nested
\sa starpu_omp_get_nested
\sa starpu_omp_get_max_active_levels
\sa starpu_omp_set_max_active_levels
\sa starpu_omp_get_active_level
\fn int starpu_omp_get_ancestor_thread_num(int level)
\ingroup API_OpenMP_Runtime_Support
This function returns the number of the ancestor of the current parallel section.
\return the number of the ancestor of the current parallel section.
\fn int starpu_omp_get_team_size(int level)
\ingroup API_OpenMP_Runtime_Support
This function returns the size of the team of the current parallel section.
\return the size of the team of the current parallel section.
\fn int starpu_omp_get_active_level(void)
\ingroup API_OpenMP_Runtime_Support
This function returns the nestinglevel of the current innermost active parallel section.
\return the nestinglevel of the current innermost active parallel section.
\sa starpu_omp_set_nested
\sa starpu_omp_get_nested
\sa starpu_omp_get_max_active_levels
\sa starpu_omp_set_max_active_levels
\sa starpu_omp_get_level
\fn int starpu_omp_in_final(void)
\ingroup API_OpenMP_Runtime_Support
This function checks whether the current task is final or not.
\return !0 if called from a final task.
\return 0 otherwise.
\fn enum starpu_omp_proc_bind_value starpu_omp_get_proc_bind(void)
\ingroup API_OpenMP_Runtime_Support
This function returns the proc_bind setting of the current parallel region.
\return the proc_bind setting of the current parallel region.
\fn void starpu_omp_set_default_device(int device_num)
\ingroup API_OpenMP_Runtime_Support
This function sets the number of the device to use as default.
Note: The StarPU OpenMP runtime support currently ignores the argument of this function.
\sa starpu_omp_get_default_device
\sa starpu_omp_is_initial_device
\fn int starpu_omp_get_default_device(void)
\ingroup API_OpenMP_Runtime_Support
This function returns the number of the device used as default.
\return the number of the device used as default.
\sa starpu_omp_set_default_device
\sa starpu_omp_is_initial_device
\fn int starpu_omp_get_num_devices(void)
\ingroup API_OpenMP_Runtime_Support
This function returns the number of the devices.
\return the number of the devices.
\fn int starpu_omp_get_num_teams(void)
\ingroup API_OpenMP_Runtime_Support
This function returns the number of teams in the current teams region.
\return the number of teams in the current teams region.
\sa starpu_omp_get_num_teams
\fn int starpu_omp_get_team_num(void)
\ingroup API_OpenMP_Runtime_Support
This function returns the team number of the calling thread.
\return the team number of the calling thread.
\sa starpu_omp_get_num_teams
\fn int starpu_omp_is_initial_device(void)
\ingroup API_OpenMP_Runtime_Support
This function checks whether the current device is the initial device or not.
\fn int starpu_omp_get_max_task_priority
\ingroup API_OpenMP_Runtime_Support
The omp_get_max_task_priority routine returns the maximum value that can be
specified in the priority clause.
\return !0 if called from the host device.
\return 0 otherwise.
\sa starpu_omp_set_default_device
\sa starpu_omp_get_default_device
\fn void starpu_omp_init_lock(starpu_omp_lock_t *lock)
\ingroup API_OpenMP_Runtime_Support
This function initializes an opaque lock object.
\sa starpu_omp_destroy_lock
\sa starpu_omp_set_lock
\sa starpu_omp_unset_lock
\sa starpu_omp_test_lock
\fn void starpu_omp_destroy_lock(starpu_omp_lock_t *lock)
\ingroup API_OpenMP_Runtime_Support
This function destroys an opaque lock object.
\sa starpu_omp_init_lock
\sa starpu_omp_set_lock
\sa starpu_omp_unset_lock
\sa starpu_omp_test_lock
\fn void starpu_omp_set_lock(starpu_omp_lock_t *lock)
\ingroup API_OpenMP_Runtime_Support
This function locks an opaque lock object. If the lock is already locked, the
function will block until it succeeds in exclusively acquiring the lock.
\sa starpu_omp_init_lock
\sa starpu_omp_destroy_lock
\sa starpu_omp_unset_lock
\sa starpu_omp_test_lock
\fn void starpu_omp_unset_lock(starpu_omp_lock_t *lock)
\ingroup API_OpenMP_Runtime_Support
This function unlocks a previously locked lock object. The behaviour of this
function is unspecified if it is called on an unlocked lock object.
\sa starpu_omp_init_lock
\sa starpu_omp_destroy_lock
\sa starpu_omp_set_lock
\sa starpu_omp_test_lock
\fn int starpu_omp_test_lock(starpu_omp_lock_t *lock)
\ingroup API_OpenMP_Runtime_Support
This function unblockingly attempts to lock a lock object and returns whether
it succeeded or not.
\return !0 if the function succeeded in acquiring the lock.
\return 0 if the lock was already locked.
\sa starpu_omp_init_lock
\sa starpu_omp_destroy_lock
\sa starpu_omp_set_lock
\sa starpu_omp_unset_lock
\fn void starpu_omp_init_nest_lock(starpu_omp_nest_lock_t *lock)
\ingroup API_OpenMP_Runtime_Support
This function initializes an opaque lock object supporting nested locking operations.
\sa starpu_omp_destroy_nest_lock
\sa starpu_omp_set_nest_lock
\sa starpu_omp_unset_nest_lock
\sa starpu_omp_test_nest_lock
\fn void starpu_omp_destroy_nest_lock(starpu_omp_nest_lock_t *lock)
\ingroup API_OpenMP_Runtime_Support
This function destroys an opaque lock object supporting nested locking operations.
\sa starpu_omp_init_nest_lock
\sa starpu_omp_set_nest_lock
\sa starpu_omp_unset_nest_lock
\sa starpu_omp_test_nest_lock
\fn void starpu_omp_set_nest_lock(starpu_omp_nest_lock_t *lock)
\ingroup API_OpenMP_Runtime_Support
This function locks an opaque lock object supporting nested locking operations.
If the lock is already locked by another task, the function will block until
it succeeds in exclusively acquiring the lock. If the lock is already taken by
the current task, the function will increase the nested locking level of the
lock object.
\sa starpu_omp_init_nest_lock
\sa starpu_omp_destroy_nest_lock
\sa starpu_omp_unset_nest_lock
\sa starpu_omp_test_nest_lock
\fn void starpu_omp_unset_nest_lock(starpu_omp_nest_lock_t *lock)
\ingroup API_OpenMP_Runtime_Support
This function unlocks a previously locked lock object supporting nested locking
operations. If the lock has been locked multiple times in nested fashion, the
nested locking level is decreased and the lock remains locked. Otherwise, if
the lock has only been locked once, it becomes unlocked. The behaviour of this
function is unspecified if it is called on an unlocked lock object. The
behaviour of this function is unspecified if it is called from a different task
than the one that locked the lock object.
\sa starpu_omp_init_nest_lock
\sa starpu_omp_destroy_nest_lock
\sa starpu_omp_set_nest_lock
\sa starpu_omp_test_nest_lock
\fn int starpu_omp_test_nest_lock(starpu_omp_nest_lock_t *lock)
\ingroup API_OpenMP_Runtime_Support
This function unblocking attempts to lock an opaque lock object supporting
nested locking operations and returns whether it succeeded or not. If the lock
is already locked by another task, the function will return without having
acquired the lock. If the lock is already taken by the current task, the
function will increase the nested locking level of the lock object.
\return !0 if the function succeeded in acquiring the lock.
\return 0 if the lock was already locked.
\sa starpu_omp_init_nest_lock
\sa starpu_omp_destroy_nest_lock
\sa starpu_omp_set_nest_lock
\sa starpu_omp_unset_nest_lock
\fn void starpu_omp_atomic_fallback_inline_begin(void)
\ingroup API_OpenMP_Runtime_Support
This function implements the entry point of a fallback global atomic region. It
blocks until it succeeds in acquiring exclusive access to the global atomic
region.
\sa starpu_omp_atomic_fallback_inline_end
\fn void starpu_omp_atomic_fallback_inline_end(void)
\ingroup API_OpenMP_Runtime_Support
This function implements the exit point of a fallback global atomic region. It
release the exclusive access to the global atomic region.
\sa starpu_omp_atomic_fallback_inline_begin
\fn double starpu_omp_get_wtime(void)
\ingroup API_OpenMP_Runtime_Support
This function returns the elapsed wallclock time in seconds.
\return the elapsed wallclock time in seconds.
\sa starpu_omp_get_wtick
\fn double starpu_omp_get_wtick(void)
\ingroup API_OpenMP_Runtime_Support
This function returns the precision of the time used by \p starpu_omp_get_wtime.
\return the precision of the time used by \p starpu_omp_get_wtime.
\sa starpu_omp_get_wtime
\fn void starpu_omp_vector_annotate(starpu_data_handle_t handle, uint32_t slice_base)
\ingroup API_OpenMP_Runtime_Support
This function enables setting additional vector metadata needed by the OpenMP Runtime Support.
\p handle is vector data handle.
\p slice_base is the base of an array slice, expressed in number of vector elements from the array base.
\sa STARPU_VECTOR_GET_SLICE_BASE
*/