LLVM OpenMP* Runtime Library
|
Classes | |
struct | bounds_infoXX_template< T > |
struct | bounds_info_t |
Enumerations | |
enum | sched_type : kmp_int32 { kmp_sch_lower = 32 , kmp_sch_static_chunked = 33 , kmp_sch_static = 34 , kmp_sch_dynamic_chunked = 35 , kmp_sch_guided_chunked = 36 , kmp_sch_runtime = 37 , kmp_sch_auto = 38 , kmp_sch_trapezoidal = 39 , kmp_sch_static_greedy = 40 , kmp_sch_static_balanced = 41 , kmp_sch_guided_iterative_chunked = 42 , kmp_sch_guided_analytical_chunked = 43 , kmp_sch_static_steal = 44 , kmp_sch_static_balanced_chunked = 45 , kmp_sch_guided_simd = 46 , kmp_sch_runtime_simd = 47 , kmp_sch_upper , kmp_ord_lower = 64 , kmp_ord_static_chunked = 65 , kmp_ord_static = 66 , kmp_ord_dynamic_chunked = 67 , kmp_ord_guided_chunked = 68 , kmp_ord_runtime = 69 , kmp_ord_auto = 70 , kmp_ord_trapezoidal = 71 , kmp_ord_upper , kmp_distribute_static_chunked = 91 , kmp_distribute_static = 92 , kmp_nm_lower = 160 , kmp_nm_static_chunked , kmp_nm_static = 162 , kmp_nm_dynamic_chunked = 163 , kmp_nm_guided_chunked = 164 , kmp_nm_runtime = 165 , kmp_nm_auto = 166 , kmp_nm_trapezoidal = 167 , kmp_nm_static_greedy = 168 , kmp_nm_static_balanced = 169 , kmp_nm_guided_iterative_chunked = 170 , kmp_nm_guided_analytical_chunked = 171 , kmp_nm_static_steal , kmp_nm_ord_static_chunked = 193 , kmp_nm_ord_static = 194 , kmp_nm_ord_dynamic_chunked = 195 , kmp_nm_ord_guided_chunked = 196 , kmp_nm_ord_runtime = 197 , kmp_nm_ord_auto = 198 , kmp_nm_ord_trapezoidal = 199 , kmp_nm_upper , kmp_sch_modifier_monotonic , kmp_sch_modifier_nonmonotonic , kmp_sch_default = kmp_sch_static } |
Functions | |
kmp_int32 | __kmpc_master (ident_t *loc, kmp_int32 global_tid) |
void | __kmpc_end_master (ident_t *loc, kmp_int32 global_tid) |
kmp_int32 | __kmpc_masked (ident_t *loc, kmp_int32 global_tid, kmp_int32 filter) |
void | __kmpc_end_masked (ident_t *loc, kmp_int32 global_tid) |
void | __kmpc_ordered (ident_t *loc, kmp_int32 gtid) |
void | __kmpc_end_ordered (ident_t *loc, kmp_int32 gtid) |
void | __kmpc_critical (ident_t *loc, kmp_int32 global_tid, kmp_critical_name *crit) |
void | __kmpc_end_critical (ident_t *loc, kmp_int32 global_tid, kmp_critical_name *crit) |
kmp_int32 | __kmpc_single (ident_t *loc, kmp_int32 global_tid) |
void | __kmpc_end_single (ident_t *loc, kmp_int32 global_tid) |
void | __kmpc_for_static_fini (ident_t *loc, kmp_int32 global_tid) |
void | __kmpc_doacross_init (ident_t *loc, int gtid, int num_dims, const struct kmp_dim *dims) |
kmp_int32 | __kmpc_sections_init (ident_t *loc, kmp_int32 gtid) |
kmp_int32 | __kmpc_next_section (ident_t *loc, kmp_int32 gtid, kmp_int32 numberOfSections) |
void | __kmpc_end_sections (ident_t *loc, kmp_int32 gtid) |
These functions are used for implementing #pragma omp for
, #pragma omp sections
, #pragma omp single
and #pragma omp master
constructs.
When handling loops, there are different functions for each of the signed and unsigned 32 and 64 bit integer types which have the name suffixes _4
, _4u
, _8
and _8u
. The semantics of each of the functions is the same, so they are only described once.
Static loop scheduling is handled by __kmpc_for_static_init_4 and friends. Only a single call is needed, since the iterations to be executed by any give thread can be determined as soon as the loop parameters are known.
Dynamic scheduling is handled by the __kmpc_dispatch_init_4 and __kmpc_dispatch_next_4 functions. The init function is called once in each thread outside the loop, while the next function is called each time that the previous chunk of work has been exhausted.
enum sched_type : kmp_int32 |
Describes the loop schedule to be used for a parallel for loop.
void __kmpc_critical | ( | ident_t * | loc, |
kmp_int32 | global_tid, | ||
kmp_critical_name * | crit | ||
) |
loc | source location information. |
global_tid | global thread number. |
crit | identity of the critical section. This could be a pointer to a lock associated with the critical section, or some other suitably unique value. |
Enter code protected by a critical
construct. This function blocks until the executing thread can enter the critical section.
Definition at line 1287 of file kmp_csupport.cpp.
void __kmpc_dispatch_fini_4 | ( | ident_t * | loc, |
kmp_int32 | gtid | ||
) |
loc | Source code location |
gtid | Global thread id |
Mark the end of a dynamic loop.
Definition at line 2986 of file kmp_dispatch.cpp.
void __kmpc_dispatch_fini_4u | ( | ident_t * | loc, |
kmp_int32 | gtid | ||
) |
See __kmpc_dispatch_fini_4
Definition at line 3000 of file kmp_dispatch.cpp.
void __kmpc_dispatch_fini_8 | ( | ident_t * | loc, |
kmp_int32 | gtid | ||
) |
See __kmpc_dispatch_fini_4
Definition at line 2993 of file kmp_dispatch.cpp.
void __kmpc_dispatch_fini_8u | ( | ident_t * | loc, |
kmp_int32 | gtid | ||
) |
See __kmpc_dispatch_fini_4
Definition at line 3007 of file kmp_dispatch.cpp.
void __kmpc_dispatch_init_4 | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
enum sched_type | schedule, | ||
kmp_int32 | lb, | ||
kmp_int32 | ub, | ||
kmp_int32 | st, | ||
kmp_int32 | chunk | ||
) |
loc | Source location |
gtid | Global thread id |
schedule | Schedule type |
lb | Lower bound |
ub | Upper bound |
st | Step (or increment if you prefer) |
chunk | The chunk size to block with |
This function prepares the runtime to start a dynamically scheduled for loop, saving the loop arguments. These functions are all identical apart from the types of the arguments.
Definition at line 2799 of file kmp_dispatch.cpp.
void __kmpc_dispatch_init_4u | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
enum sched_type | schedule, | ||
kmp_uint32 | lb, | ||
kmp_uint32 | ub, | ||
kmp_int32 | st, | ||
kmp_int32 | chunk | ||
) |
Definition at line 2811 of file kmp_dispatch.cpp.
void __kmpc_dispatch_init_8 | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
enum sched_type | schedule, | ||
kmp_int64 | lb, | ||
kmp_int64 | ub, | ||
kmp_int64 | st, | ||
kmp_int64 | chunk | ||
) |
Definition at line 2824 of file kmp_dispatch.cpp.
void __kmpc_dispatch_init_8u | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
enum sched_type | schedule, | ||
kmp_uint64 | lb, | ||
kmp_uint64 | ub, | ||
kmp_int64 | st, | ||
kmp_int64 | chunk | ||
) |
Definition at line 2837 of file kmp_dispatch.cpp.
int __kmpc_dispatch_next_4 | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
kmp_int32 * | p_last, | ||
kmp_int32 * | p_lb, | ||
kmp_int32 * | p_ub, | ||
kmp_int32 * | p_st | ||
) |
loc | Source code location |
gtid | Global thread id |
p_last | Pointer to a flag set to one if this is the last chunk or zero otherwise |
p_lb | Pointer to the lower bound for the next chunk of work |
p_ub | Pointer to the upper bound for the next chunk of work |
p_st | Pointer to the stride for the next chunk of work |
Get the next dynamically allocated chunk of work for this thread. If there is no more work, then the lb,ub and stride need not be modified.
Definition at line 2917 of file kmp_dispatch.cpp.
int __kmpc_dispatch_next_4u | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
kmp_int32 * | p_last, | ||
kmp_uint32 * | p_lb, | ||
kmp_uint32 * | p_ub, | ||
kmp_int32 * | p_st | ||
) |
See __kmpc_dispatch_next_4
Definition at line 2933 of file kmp_dispatch.cpp.
int __kmpc_dispatch_next_8 | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
kmp_int32 * | p_last, | ||
kmp_int64 * | p_lb, | ||
kmp_int64 * | p_ub, | ||
kmp_int64 * | p_st | ||
) |
See __kmpc_dispatch_next_4
Definition at line 2950 of file kmp_dispatch.cpp.
int __kmpc_dispatch_next_8u | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
kmp_int32 * | p_last, | ||
kmp_uint64 * | p_lb, | ||
kmp_uint64 * | p_ub, | ||
kmp_int64 * | p_st | ||
) |
See __kmpc_dispatch_next_4
Definition at line 2966 of file kmp_dispatch.cpp.
void __kmpc_dist_dispatch_init_4 | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
enum sched_type | schedule, | ||
kmp_int32 * | p_last, | ||
kmp_int32 | lb, | ||
kmp_int32 | ub, | ||
kmp_int32 | st, | ||
kmp_int32 | chunk | ||
) |
Difference from __kmpc_dispatch_init set of functions is these functions are called for composite distribute parallel for construct. Thus before regular iterations dispatching we need to calc per-team iteration space.
These functions are all identical apart from the types of the arguments.
Definition at line 2856 of file kmp_dispatch.cpp.
void __kmpc_dist_for_static_init_4 | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
kmp_int32 | schedule, | ||
kmp_int32 * | plastiter, | ||
kmp_int32 * | plower, | ||
kmp_int32 * | pupper, | ||
kmp_int32 * | pupperD, | ||
kmp_int32 * | pstride, | ||
kmp_int32 | incr, | ||
kmp_int32 | chunk | ||
) |
loc | Source code location |
gtid | Global thread id of this thread |
schedule | Scheduling type for the parallel loop |
plastiter | Pointer to the "last iteration" flag |
plower | Pointer to the lower bound |
pupper | Pointer to the upper bound of loop chunk |
pupperD | Pointer to the upper bound of dist_chunk |
pstride | Pointer to the stride for parallel loop |
incr | Loop increment |
chunk | The chunk size for the parallel loop |
Each of the four functions here are identical apart from the argument types.
The functions compute the upper and lower bounds and strides to be used for the set of iterations to be executed by the current thread from the statically scheduled loop that is described by the initial values of the bounds, strides, increment and chunks for parallel loop and distribute constructs.
Definition at line 966 of file kmp_sched.cpp.
void __kmpc_dist_for_static_init_4u | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
kmp_int32 | schedule, | ||
kmp_int32 * | plastiter, | ||
kmp_uint32 * | plower, | ||
kmp_uint32 * | pupper, | ||
kmp_uint32 * | pupperD, | ||
kmp_int32 * | pstride, | ||
kmp_int32 | incr, | ||
kmp_int32 | chunk | ||
) |
See __kmpc_dist_for_static_init_4
Definition at line 979 of file kmp_sched.cpp.
void __kmpc_dist_for_static_init_8 | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
kmp_int32 | schedule, | ||
kmp_int32 * | plastiter, | ||
kmp_int64 * | plower, | ||
kmp_int64 * | pupper, | ||
kmp_int64 * | pupperD, | ||
kmp_int64 * | pstride, | ||
kmp_int64 | incr, | ||
kmp_int64 | chunk | ||
) |
See __kmpc_dist_for_static_init_4
Definition at line 992 of file kmp_sched.cpp.
void __kmpc_dist_for_static_init_8u | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
kmp_int32 | schedule, | ||
kmp_int32 * | plastiter, | ||
kmp_uint64 * | plower, | ||
kmp_uint64 * | pupper, | ||
kmp_uint64 * | pupperD, | ||
kmp_int64 * | pstride, | ||
kmp_int64 | incr, | ||
kmp_int64 | chunk | ||
) |
See __kmpc_dist_for_static_init_4
Definition at line 1005 of file kmp_sched.cpp.
void __kmpc_doacross_init | ( | ident_t * | loc, |
int | gtid, | ||
int | num_dims, | ||
const struct kmp_dim * | dims | ||
) |
loc | source location information. |
gtid | global thread number. |
num_dims | number of associated doacross loops. |
dims | info on loops bounds. |
Initialize doacross loop information. Expect compiler send us inclusive bounds, e.g. for(i=2;i<9;i+=2) lo=2, up=8, st=2.
Definition at line 4088 of file kmp_csupport.cpp.
void __kmpc_end_critical | ( | ident_t * | loc, |
kmp_int32 | global_tid, | ||
kmp_critical_name * | crit | ||
) |
loc | source location information. |
global_tid | global thread number . |
crit | identity of the critical section. This could be a pointer to a lock associated with the critical section, or some other suitably unique value. |
Leave a critical section, releasing any lock that was held during its execution.
Definition at line 1646 of file kmp_csupport.cpp.
void __kmpc_end_masked | ( | ident_t * | loc, |
kmp_int32 | global_tid | ||
) |
loc | source location information. |
global_tid | global thread number . |
Mark the end of a masked
region. This should only be called by the thread that executes the masked
region.
Definition at line 950 of file kmp_csupport.cpp.
void __kmpc_end_master | ( | ident_t * | loc, |
kmp_int32 | global_tid | ||
) |
loc | source location information. |
global_tid | global thread number . |
Mark the end of a master
region. This should only be called by the thread that executes the master
region.
Definition at line 862 of file kmp_csupport.cpp.
void __kmpc_end_ordered | ( | ident_t * | loc, |
kmp_int32 | gtid | ||
) |
loc | source location information. |
gtid | global thread number. |
End execution of an ordered
construct.
Definition at line 1052 of file kmp_csupport.cpp.
void __kmpc_end_sections | ( | ident_t * | loc, |
kmp_int32 | gtid | ||
) |
loc | source location information |
global_tid | global thread number |
End of "sections" construct. Don't need to wait here: barrier is added separately when needed.
Definition at line 2637 of file kmp_dispatch.cpp.
void __kmpc_end_single | ( | ident_t * | loc, |
kmp_int32 | global_tid | ||
) |
loc | source location information |
global_tid | global thread number |
Mark the end of a single
construct. This function should only be called by the thread that executed the block of code protected by the single
construct.
Definition at line 1919 of file kmp_csupport.cpp.
void __kmpc_for_static_fini | ( | ident_t * | loc, |
kmp_int32 | global_tid | ||
) |
loc | Source location |
global_tid | Global thread id |
Mark the end of a statically scheduled loop.
Definition at line 1946 of file kmp_csupport.cpp.
void __kmpc_for_static_init_4 | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
kmp_int32 | schedtype, | ||
kmp_int32 * | plastiter, | ||
kmp_int32 * | plower, | ||
kmp_int32 * | pupper, | ||
kmp_int32 * | pstride, | ||
kmp_int32 | incr, | ||
kmp_int32 | chunk | ||
) |
loc | Source code location |
gtid | Global thread id of this thread |
schedtype | Scheduling type |
plastiter | Pointer to the "last iteration" flag |
plower | Pointer to the lower bound |
pupper | Pointer to the upper bound |
pstride | Pointer to the stride |
incr | Loop increment |
chunk | The chunk size |
Each of the four functions here are identical apart from the argument types.
The functions compute the upper and lower bounds and stride to be used for the set of iterations to be executed by the current thread from the statically scheduled loop that is described by the initial values of the bounds, stride, increment and chunk size.
Definition at line 872 of file kmp_sched.cpp.
void __kmpc_for_static_init_4u | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
kmp_int32 | schedtype, | ||
kmp_int32 * | plastiter, | ||
kmp_uint32 * | plower, | ||
kmp_uint32 * | pupper, | ||
kmp_int32 * | pstride, | ||
kmp_int32 | incr, | ||
kmp_int32 | chunk | ||
) |
Definition at line 888 of file kmp_sched.cpp.
void __kmpc_for_static_init_8 | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
kmp_int32 | schedtype, | ||
kmp_int32 * | plastiter, | ||
kmp_int64 * | plower, | ||
kmp_int64 * | pupper, | ||
kmp_int64 * | pstride, | ||
kmp_int64 | incr, | ||
kmp_int64 | chunk | ||
) |
Definition at line 905 of file kmp_sched.cpp.
void __kmpc_for_static_init_8u | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
kmp_int32 | schedtype, | ||
kmp_int32 * | plastiter, | ||
kmp_uint64 * | plower, | ||
kmp_uint64 * | pupper, | ||
kmp_int64 * | pstride, | ||
kmp_int64 | incr, | ||
kmp_int64 | chunk | ||
) |
Definition at line 921 of file kmp_sched.cpp.
kmp_int32 __kmpc_masked | ( | ident_t * | loc, |
kmp_int32 | global_tid, | ||
kmp_int32 | filter | ||
) |
loc | source location information. |
global_tid | global thread number. |
filter | result of evaluating filter clause on thread global_tid, or zero if no filter clause present |
masked
block, 0 otherwise. Definition at line 894 of file kmp_csupport.cpp.
kmp_int32 __kmpc_master | ( | ident_t * | loc, |
kmp_int32 | global_tid | ||
) |
loc | source location information. |
global_tid | global thread number . |
master
block, 0 otherwise. Definition at line 805 of file kmp_csupport.cpp.
Referenced by __kmpc_barrier_master_nowait().
kmp_int32 __kmpc_next_section | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
kmp_int32 | numberOfSections | ||
) |
loc | source location information |
global_tid | global thread number |
numberOfSections | number of sections in the 'sections' construct |
Definition at line 2548 of file kmp_dispatch.cpp.
void __kmpc_ordered | ( | ident_t * | loc, |
kmp_int32 | gtid | ||
) |
loc | source location information. |
gtid | global thread number. |
Start execution of an ordered
construct.
Definition at line 979 of file kmp_csupport.cpp.
kmp_int32 __kmpc_sections_init | ( | ident_t * | loc, |
kmp_int32 | gtid | ||
) |
loc | source location information |
global_tid | global thread number |
Beginning of sections construct. There are no implicit barriers in the "sections" calls, rather the compiler should introduce an explicit barrier if it is required.
This implementation is based on __kmp_dispatch_init, using same constructs for shared data (we can't have sections nested directly in omp for loop, there should be a parallel region in between)
Definition at line 2464 of file kmp_dispatch.cpp.
kmp_int32 __kmpc_single | ( | ident_t * | loc, |
kmp_int32 | global_tid | ||
) |
loc | source location information |
global_tid | global thread number |
Test whether to execute a single
construct. There are no implicit barriers in the two "single" calls, rather the compiler should introduce an explicit barrier if it is required.
Definition at line 1866 of file kmp_csupport.cpp.
void __kmpc_team_static_init_4 | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
kmp_int32 * | p_last, | ||
kmp_int32 * | p_lb, | ||
kmp_int32 * | p_ub, | ||
kmp_int32 * | p_st, | ||
kmp_int32 | incr, | ||
kmp_int32 | chunk | ||
) |
loc | Source location |
gtid | Global thread id |
p_last | pointer to last iteration flag |
p_lb | pointer to Lower bound |
p_ub | pointer to Upper bound |
p_st | Step (or increment if you prefer) |
incr | Loop increment |
chunk | The chunk size to block with |
The functions compute the upper and lower bounds and stride to be used for the set of iterations to be executed by the current team from the statically scheduled loop that is described by the initial values of the bounds, stride, increment and chunk for the distribute construct as part of composite distribute parallel loop construct. These functions are all identical apart from the types of the arguments.
Definition at line 1044 of file kmp_sched.cpp.
void __kmpc_team_static_init_4u | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
kmp_int32 * | p_last, | ||
kmp_uint32 * | p_lb, | ||
kmp_uint32 * | p_ub, | ||
kmp_int32 * | p_st, | ||
kmp_int32 | incr, | ||
kmp_int32 | chunk | ||
) |
Definition at line 1056 of file kmp_sched.cpp.
void __kmpc_team_static_init_8 | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
kmp_int32 * | p_last, | ||
kmp_int64 * | p_lb, | ||
kmp_int64 * | p_ub, | ||
kmp_int64 * | p_st, | ||
kmp_int64 | incr, | ||
kmp_int64 | chunk | ||
) |
Definition at line 1068 of file kmp_sched.cpp.
void __kmpc_team_static_init_8u | ( | ident_t * | loc, |
kmp_int32 | gtid, | ||
kmp_int32 * | p_last, | ||
kmp_uint64 * | p_lb, | ||
kmp_uint64 * | p_ub, | ||
kmp_int64 * | p_st, | ||
kmp_int64 | incr, | ||
kmp_int64 | chunk | ||
) |
Definition at line 1080 of file kmp_sched.cpp.