/* * kmp_taskdeps.cpp */ //===----------------------------------------------------------------------===// // // The LLVM Compiler Infrastructure // // This file is dual licensed under the MIT and the University of Illinois Open // Source Licenses. See LICENSE.txt for details. // //===----------------------------------------------------------------------===// //#define KMP_SUPPORT_GRAPH_OUTPUT 1 #include "kmp.h" #include "kmp_io.h" #include "kmp_wait_release.h" #if OMP_40_ENABLED // TODO: Improve memory allocation? keep a list of pre-allocated structures? // allocate in blocks? re-use list finished list entries? // TODO: don't use atomic ref counters for stack-allocated nodes. // TODO: find an alternate to atomic refs for heap-allocated nodes? // TODO: Finish graph output support // TODO: kmp_lock_t seems a tad to big (and heavy weight) for this. Check other // runtime locks // TODO: Any ITT support needed? #ifdef KMP_SUPPORT_GRAPH_OUTPUT static kmp_int32 kmp_node_id_seed = 0; #endif static void __kmp_init_node(kmp_depnode_t *node) { node->dn.task = NULL; // set to null initially, it will point to the right // task once dependences have been processed node->dn.successors = NULL; __kmp_init_lock(&node->dn.lock); node->dn.nrefs = 1; // init creates the first reference to the node #ifdef KMP_SUPPORT_GRAPH_OUTPUT node->dn.id = KMP_TEST_THEN_INC32(&kmp_node_id_seed); #endif } static inline kmp_depnode_t *__kmp_node_ref(kmp_depnode_t *node) { KMP_TEST_THEN_INC32(CCAST(kmp_int32 *, &node->dn.nrefs)); return node; } static inline void __kmp_node_deref(kmp_info_t *thread, kmp_depnode_t *node) { if (!node) return; kmp_int32 n = KMP_TEST_THEN_DEC32(CCAST(kmp_int32 *, &node->dn.nrefs)) - 1; if (n == 0) { KMP_ASSERT(node->dn.nrefs == 0); #if USE_FAST_MEMORY __kmp_fast_free(thread, node); #else __kmp_thread_free(thread, node); #endif } } #define KMP_ACQUIRE_DEPNODE(gtid, n) __kmp_acquire_lock(&(n)->dn.lock, (gtid)) #define KMP_RELEASE_DEPNODE(gtid, n) __kmp_release_lock(&(n)->dn.lock, (gtid)) static void __kmp_depnode_list_free(kmp_info_t *thread, kmp_depnode_list *list); enum { KMP_DEPHASH_OTHER_SIZE = 97, KMP_DEPHASH_MASTER_SIZE = 997 }; static inline kmp_int32 __kmp_dephash_hash(kmp_intptr_t addr, size_t hsize) { // TODO alternate to try: set = (((Addr64)(addrUsefulBits * 9.618)) % // m_num_sets ); return ((addr >> 6) ^ (addr >> 2)) % hsize; } static kmp_dephash_t *__kmp_dephash_create(kmp_info_t *thread, kmp_taskdata_t *current_task) { kmp_dephash_t *h; size_t h_size; if (current_task->td_flags.tasktype == TASK_IMPLICIT) h_size = KMP_DEPHASH_MASTER_SIZE; else h_size = KMP_DEPHASH_OTHER_SIZE; kmp_int32 size = h_size * sizeof(kmp_dephash_entry_t *) + sizeof(kmp_dephash_t); #if USE_FAST_MEMORY h = (kmp_dephash_t *)__kmp_fast_allocate(thread, size); #else h = (kmp_dephash_t *)__kmp_thread_malloc(thread, size); #endif h->size = h_size; #ifdef KMP_DEBUG h->nelements = 0; h->nconflicts = 0; #endif h->buckets = (kmp_dephash_entry **)(h + 1); for (size_t i = 0; i < h_size; i++) h->buckets[i] = 0; return h; } void __kmp_dephash_free_entries(kmp_info_t *thread, kmp_dephash_t *h) { for (size_t i = 0; i < h->size; i++) { if (h->buckets[i]) { kmp_dephash_entry_t *next; for (kmp_dephash_entry_t *entry = h->buckets[i]; entry; entry = next) { next = entry->next_in_bucket; __kmp_depnode_list_free(thread, entry->last_ins); __kmp_node_deref(thread, entry->last_out); #if USE_FAST_MEMORY __kmp_fast_free(thread, entry); #else __kmp_thread_free(thread, entry); #endif } h->buckets[i] = 0; } } } void __kmp_dephash_free(kmp_info_t *thread, kmp_dephash_t *h) { __kmp_dephash_free_entries(thread, h); #if USE_FAST_MEMORY __kmp_fast_free(thread, h); #else __kmp_thread_free(thread, h); #endif } static kmp_dephash_entry * __kmp_dephash_find(kmp_info_t *thread, kmp_dephash_t *h, kmp_intptr_t addr) { kmp_int32 bucket = __kmp_dephash_hash(addr, h->size); kmp_dephash_entry_t *entry; for (entry = h->buckets[bucket]; entry; entry = entry->next_in_bucket) if (entry->addr == addr) break; if (entry == NULL) { // create entry. This is only done by one thread so no locking required #if USE_FAST_MEMORY entry = (kmp_dephash_entry_t *)__kmp_fast_allocate( thread, sizeof(kmp_dephash_entry_t)); #else entry = (kmp_dephash_entry_t *)__kmp_thread_malloc( thread, sizeof(kmp_dephash_entry_t)); #endif entry->addr = addr; entry->last_out = NULL; entry->last_ins = NULL; entry->next_in_bucket = h->buckets[bucket]; h->buckets[bucket] = entry; #ifdef KMP_DEBUG h->nelements++; if (entry->next_in_bucket) h->nconflicts++; #endif } return entry; } static kmp_depnode_list_t *__kmp_add_node(kmp_info_t *thread, kmp_depnode_list_t *list, kmp_depnode_t *node) { kmp_depnode_list_t *new_head; #if USE_FAST_MEMORY new_head = (kmp_depnode_list_t *)__kmp_fast_allocate( thread, sizeof(kmp_depnode_list_t)); #else new_head = (kmp_depnode_list_t *)__kmp_thread_malloc( thread, sizeof(kmp_depnode_list_t)); #endif new_head->node = __kmp_node_ref(node); new_head->next = list; return new_head; } static void __kmp_depnode_list_free(kmp_info_t *thread, kmp_depnode_list *list) { kmp_depnode_list *next; for (; list; list = next) { next = list->next; __kmp_node_deref(thread, list->node); #if USE_FAST_MEMORY __kmp_fast_free(thread, list); #else __kmp_thread_free(thread, list); #endif } } static inline void __kmp_track_dependence(kmp_depnode_t *source, kmp_depnode_t *sink, kmp_task_t *sink_task) { #ifdef KMP_SUPPORT_GRAPH_OUTPUT kmp_taskdata_t *task_source = KMP_TASK_TO_TASKDATA(source->dn.task); // do not use sink->dn.task as that is only filled after the dependencies // are already processed! kmp_taskdata_t *task_sink = KMP_TASK_TO_TASKDATA(sink_task); __kmp_printf("%d(%s) -> %d(%s)\n", source->dn.id, task_source->td_ident->psource, sink->dn.id, task_sink->td_ident->psource); #endif #if OMPT_SUPPORT && OMPT_TRACE // OMPT tracks dependences between task (a=source, b=sink) in which // task a blocks the execution of b through the ompt_new_dependence_callback if (ompt_enabled && ompt_callbacks.ompt_callback(ompt_event_task_dependence_pair)) { kmp_taskdata_t *task_source = KMP_TASK_TO_TASKDATA(source->dn.task); kmp_taskdata_t *task_sink = KMP_TASK_TO_TASKDATA(sink_task); ompt_callbacks.ompt_callback(ompt_event_task_dependence_pair)( task_source->ompt_task_info.task_id, task_sink->ompt_task_info.task_id); } #endif /* OMPT_SUPPORT && OMPT_TRACE */ } template static inline kmp_int32 __kmp_process_deps(kmp_int32 gtid, kmp_depnode_t *node, kmp_dephash_t *hash, bool dep_barrier, kmp_int32 ndeps, kmp_depend_info_t *dep_list, kmp_task_t *task) { KA_TRACE(30, ("__kmp_process_deps<%d>: T#%d processing %d dependencies : " "dep_barrier = %d\n", filter, gtid, ndeps, dep_barrier)); kmp_info_t *thread = __kmp_threads[gtid]; kmp_int32 npredecessors = 0; for (kmp_int32 i = 0; i < ndeps; i++) { const kmp_depend_info_t *dep = &dep_list[i]; KMP_DEBUG_ASSERT(dep->flags.in); if (filter && dep->base_addr == 0) continue; // skip filtered entries kmp_dephash_entry_t *info = __kmp_dephash_find(thread, hash, dep->base_addr); kmp_depnode_t *last_out = info->last_out; if (dep->flags.out && info->last_ins) { for (kmp_depnode_list_t *p = info->last_ins; p; p = p->next) { kmp_depnode_t *indep = p->node; if (indep->dn.task) { KMP_ACQUIRE_DEPNODE(gtid, indep); if (indep->dn.task) { __kmp_track_dependence(indep, node, task); indep->dn.successors = __kmp_add_node(thread, indep->dn.successors, node); KA_TRACE(40, ("__kmp_process_deps<%d>: T#%d adding dependence from " "%p to %p\n", filter, gtid, KMP_TASK_TO_TASKDATA(indep->dn.task), KMP_TASK_TO_TASKDATA(task))); npredecessors++; } KMP_RELEASE_DEPNODE(gtid, indep); } } __kmp_depnode_list_free(thread, info->last_ins); info->last_ins = NULL; } else if (last_out && last_out->dn.task) { KMP_ACQUIRE_DEPNODE(gtid, last_out); if (last_out->dn.task) { __kmp_track_dependence(last_out, node, task); last_out->dn.successors = __kmp_add_node(thread, last_out->dn.successors, node); KA_TRACE( 40, ("__kmp_process_deps<%d>: T#%d adding dependence from %p to %p\n", filter, gtid, KMP_TASK_TO_TASKDATA(last_out->dn.task), KMP_TASK_TO_TASKDATA(task))); npredecessors++; } KMP_RELEASE_DEPNODE(gtid, last_out); } if (dep_barrier) { // if this is a sync point in the serial sequence, then the previous // outputs are guaranteed to be completed after // the execution of this task so the previous output nodes can be cleared. __kmp_node_deref(thread, last_out); info->last_out = NULL; } else { if (dep->flags.out) { __kmp_node_deref(thread, last_out); info->last_out = __kmp_node_ref(node); } else info->last_ins = __kmp_add_node(thread, info->last_ins, node); } } KA_TRACE(30, ("__kmp_process_deps<%d>: T#%d found %d predecessors\n", filter, gtid, npredecessors)); return npredecessors; } #define NO_DEP_BARRIER (false) #define DEP_BARRIER (true) // returns true if the task has any outstanding dependence static bool __kmp_check_deps(kmp_int32 gtid, kmp_depnode_t *node, kmp_task_t *task, kmp_dephash_t *hash, bool dep_barrier, kmp_int32 ndeps, kmp_depend_info_t *dep_list, kmp_int32 ndeps_noalias, kmp_depend_info_t *noalias_dep_list) { int i; #if KMP_DEBUG kmp_taskdata_t *taskdata = KMP_TASK_TO_TASKDATA(task); #endif KA_TRACE(20, ("__kmp_check_deps: T#%d checking dependencies for task %p : %d " "possibly aliased dependencies, %d non-aliased depedencies : " "dep_barrier=%d .\n", gtid, taskdata, ndeps, ndeps_noalias, dep_barrier)); // Filter deps in dep_list // TODO: Different algorithm for large dep_list ( > 10 ? ) for (i = 0; i < ndeps; i++) { if (dep_list[i].base_addr != 0) for (int j = i + 1; j < ndeps; j++) if (dep_list[i].base_addr == dep_list[j].base_addr) { dep_list[i].flags.in |= dep_list[j].flags.in; dep_list[i].flags.out |= dep_list[j].flags.out; dep_list[j].base_addr = 0; // Mark j element as void } } // doesn't need to be atomic as no other thread is going to be accessing this // node just yet. // npredecessors is set -1 to ensure that none of the releasing tasks queues // this task before we have finished processing all the dependencies node->dn.npredecessors = -1; // used to pack all npredecessors additions into a single atomic operation at // the end int npredecessors; npredecessors = __kmp_process_deps(gtid, node, hash, dep_barrier, ndeps, dep_list, task); npredecessors += __kmp_process_deps( gtid, node, hash, dep_barrier, ndeps_noalias, noalias_dep_list, task); node->dn.task = task; KMP_MB(); // Account for our initial fake value npredecessors++; // Update predecessors and obtain current value to check if there are still // any outstandig dependences (some tasks may have finished while we processed // the dependences) npredecessors = KMP_TEST_THEN_ADD32(CCAST(kmp_int32 *, &node->dn.npredecessors), npredecessors) + npredecessors; KA_TRACE(20, ("__kmp_check_deps: T#%d found %d predecessors for task %p \n", gtid, npredecessors, taskdata)); // beyond this point the task could be queued (and executed) by a releasing // task... return npredecessors > 0 ? true : false; } void __kmp_release_deps(kmp_int32 gtid, kmp_taskdata_t *task) { kmp_info_t *thread = __kmp_threads[gtid]; kmp_depnode_t *node = task->td_depnode; if (task->td_dephash) { KA_TRACE( 40, ("__kmp_release_deps: T#%d freeing dependencies hash of task %p.\n", gtid, task)); __kmp_dephash_free(thread, task->td_dephash); task->td_dephash = NULL; } if (!node) return; KA_TRACE(20, ("__kmp_release_deps: T#%d notifying successors of task %p.\n", gtid, task)); KMP_ACQUIRE_DEPNODE(gtid, node); node->dn.task = NULL; // mark this task as finished, so no new dependencies are generated KMP_RELEASE_DEPNODE(gtid, node); kmp_depnode_list_t *next; for (kmp_depnode_list_t *p = node->dn.successors; p; p = next) { kmp_depnode_t *successor = p->node; kmp_int32 npredecessors = KMP_TEST_THEN_DEC32(CCAST(kmp_int32 *, &successor->dn.npredecessors)) - 1; // successor task can be NULL for wait_depends or because deps are still // being processed if (npredecessors == 0) { KMP_MB(); if (successor->dn.task) { KA_TRACE(20, ("__kmp_release_deps: T#%d successor %p of %p scheduled " "for execution.\n", gtid, successor->dn.task, task)); __kmp_omp_task(gtid, successor->dn.task, false); } } next = p->next; __kmp_node_deref(thread, p->node); #if USE_FAST_MEMORY __kmp_fast_free(thread, p); #else __kmp_thread_free(thread, p); #endif } __kmp_node_deref(thread, node); KA_TRACE( 20, ("__kmp_release_deps: T#%d all successors of %p notified of completion\n", gtid, task)); } /*! @ingroup TASKING @param loc_ref location of the original task directive @param gtid Global Thread ID of encountering thread @param new_task task thunk allocated by __kmp_omp_task_alloc() for the ''new task'' @param ndeps Number of depend items with possible aliasing @param dep_list List of depend items with possible aliasing @param ndeps_noalias Number of depend items with no aliasing @param noalias_dep_list List of depend items with no aliasing @return Returns either TASK_CURRENT_NOT_QUEUED if the current task was not suspendend and queued, or TASK_CURRENT_QUEUED if it was suspended and queued Schedule a non-thread-switchable task with dependences for execution */ kmp_int32 __kmpc_omp_task_with_deps(ident_t *loc_ref, kmp_int32 gtid, kmp_task_t *new_task, kmp_int32 ndeps, kmp_depend_info_t *dep_list, kmp_int32 ndeps_noalias, kmp_depend_info_t *noalias_dep_list) { kmp_taskdata_t *new_taskdata = KMP_TASK_TO_TASKDATA(new_task); KA_TRACE(10, ("__kmpc_omp_task_with_deps(enter): T#%d loc=%p task=%p\n", gtid, loc_ref, new_taskdata)); kmp_info_t *thread = __kmp_threads[gtid]; kmp_taskdata_t *current_task = thread->th.th_current_task; #if OMPT_SUPPORT && OMPT_TRACE /* OMPT grab all dependences if requested by the tool */ if (ompt_enabled && ndeps + ndeps_noalias > 0 && ompt_callbacks.ompt_callback(ompt_event_task_dependences)) { kmp_int32 i; new_taskdata->ompt_task_info.ndeps = ndeps + ndeps_noalias; new_taskdata->ompt_task_info.deps = (ompt_task_dependence_t *)KMP_OMPT_DEPS_ALLOC( thread, (ndeps + ndeps_noalias) * sizeof(ompt_task_dependence_t)); KMP_ASSERT(new_taskdata->ompt_task_info.deps != NULL); for (i = 0; i < ndeps; i++) { new_taskdata->ompt_task_info.deps[i].variable_addr = (void *)dep_list[i].base_addr; if (dep_list[i].flags.in && dep_list[i].flags.out) new_taskdata->ompt_task_info.deps[i].dependence_flags = ompt_task_dependence_type_inout; else if (dep_list[i].flags.out) new_taskdata->ompt_task_info.deps[i].dependence_flags = ompt_task_dependence_type_out; else if (dep_list[i].flags.in) new_taskdata->ompt_task_info.deps[i].dependence_flags = ompt_task_dependence_type_in; } for (i = 0; i < ndeps_noalias; i++) { new_taskdata->ompt_task_info.deps[ndeps + i].variable_addr = (void *)noalias_dep_list[i].base_addr; if (noalias_dep_list[i].flags.in && noalias_dep_list[i].flags.out) new_taskdata->ompt_task_info.deps[ndeps + i].dependence_flags = ompt_task_dependence_type_inout; else if (noalias_dep_list[i].flags.out) new_taskdata->ompt_task_info.deps[ndeps + i].dependence_flags = ompt_task_dependence_type_out; else if (noalias_dep_list[i].flags.in) new_taskdata->ompt_task_info.deps[ndeps + i].dependence_flags = ompt_task_dependence_type_in; } } #endif /* OMPT_SUPPORT && OMPT_TRACE */ bool serial = current_task->td_flags.team_serial || current_task->td_flags.tasking_ser || current_task->td_flags.final; #if OMP_45_ENABLED kmp_task_team_t *task_team = thread->th.th_task_team; serial = serial && !(task_team && task_team->tt.tt_found_proxy_tasks); #endif if (!serial && (ndeps > 0 || ndeps_noalias > 0)) { /* if no dependencies have been tracked yet, create the dependence hash */ if (current_task->td_dephash == NULL) current_task->td_dephash = __kmp_dephash_create(thread, current_task); #if USE_FAST_MEMORY kmp_depnode_t *node = (kmp_depnode_t *)__kmp_fast_allocate(thread, sizeof(kmp_depnode_t)); #else kmp_depnode_t *node = (kmp_depnode_t *)__kmp_thread_malloc(thread, sizeof(kmp_depnode_t)); #endif __kmp_init_node(node); new_taskdata->td_depnode = node; if (__kmp_check_deps(gtid, node, new_task, current_task->td_dephash, NO_DEP_BARRIER, ndeps, dep_list, ndeps_noalias, noalias_dep_list)) { KA_TRACE(10, ("__kmpc_omp_task_with_deps(exit): T#%d task had blocking " "dependencies: " "loc=%p task=%p, return: TASK_CURRENT_NOT_QUEUED\n", gtid, loc_ref, new_taskdata)); return TASK_CURRENT_NOT_QUEUED; } } else { KA_TRACE(10, ("__kmpc_omp_task_with_deps(exit): T#%d ignored dependencies " "for task (serialized)" "loc=%p task=%p\n", gtid, loc_ref, new_taskdata)); } KA_TRACE(10, ("__kmpc_omp_task_with_deps(exit): T#%d task had no blocking " "dependencies : " "loc=%p task=%p, transferring to __kmpc_omp_task\n", gtid, loc_ref, new_taskdata)); return __kmpc_omp_task(loc_ref, gtid, new_task); } /*! @ingroup TASKING @param loc_ref location of the original task directive @param gtid Global Thread ID of encountering thread @param ndeps Number of depend items with possible aliasing @param dep_list List of depend items with possible aliasing @param ndeps_noalias Number of depend items with no aliasing @param noalias_dep_list List of depend items with no aliasing Blocks the current task until all specifies dependencies have been fulfilled. */ void __kmpc_omp_wait_deps(ident_t *loc_ref, kmp_int32 gtid, kmp_int32 ndeps, kmp_depend_info_t *dep_list, kmp_int32 ndeps_noalias, kmp_depend_info_t *noalias_dep_list) { KA_TRACE(10, ("__kmpc_omp_wait_deps(enter): T#%d loc=%p\n", gtid, loc_ref)); if (ndeps == 0 && ndeps_noalias == 0) { KA_TRACE(10, ("__kmpc_omp_wait_deps(exit): T#%d has no dependencies to " "wait upon : loc=%p\n", gtid, loc_ref)); return; } kmp_info_t *thread = __kmp_threads[gtid]; kmp_taskdata_t *current_task = thread->th.th_current_task; // We can return immediately as: // - dependences are not computed in serial teams (except with proxy tasks) // - if the dephash is not yet created it means we have nothing to wait for bool ignore = current_task->td_flags.team_serial || current_task->td_flags.tasking_ser || current_task->td_flags.final; #if OMP_45_ENABLED ignore = ignore && thread->th.th_task_team != NULL && thread->th.th_task_team->tt.tt_found_proxy_tasks == FALSE; #endif ignore = ignore || current_task->td_dephash == NULL; if (ignore) { KA_TRACE(10, ("__kmpc_omp_wait_deps(exit): T#%d has no blocking " "dependencies : loc=%p\n", gtid, loc_ref)); return; } kmp_depnode_t node; __kmp_init_node(&node); if (!__kmp_check_deps(gtid, &node, NULL, current_task->td_dephash, DEP_BARRIER, ndeps, dep_list, ndeps_noalias, noalias_dep_list)) { KA_TRACE(10, ("__kmpc_omp_wait_deps(exit): T#%d has no blocking " "dependencies : loc=%p\n", gtid, loc_ref)); return; } int thread_finished = FALSE; kmp_flag_32 flag((volatile kmp_uint32 *)&(node.dn.npredecessors), 0U); while (node.dn.npredecessors > 0) { flag.execute_tasks(thread, gtid, FALSE, &thread_finished, #if USE_ITT_BUILD NULL, #endif __kmp_task_stealing_constraint); } KA_TRACE(10, ("__kmpc_omp_wait_deps(exit): T#%d finished waiting : loc=%p\n", gtid, loc_ref)); } #endif /* OMP_40_ENABLED */