diff --git a/openmp/runtime/src/kmp.h b/openmp/runtime/src/kmp.h
--- a/openmp/runtime/src/kmp.h
+++ b/openmp/runtime/src/kmp.h
@@ -2334,7 +2334,8 @@
   unsigned priority_specified : 1; /* set if the compiler provides priority
                                       setting for the task */
   unsigned detachable : 1; /* 1 == can detach */
-  unsigned reserved : 9; /* reserved for compiler use */
+  unsigned hidden_helper : 1; /* 1 == hidden helper task */
+  unsigned reserved : 8; /* reserved for compiler use */
 
   /* Library flags */ /* Total library flags must be 16 bits */
   unsigned tasktype : 1; /* task is either explicit(1) or implicit (0) */
@@ -2382,6 +2383,13 @@
   kmp_depnode_t
       *td_depnode; // Pointer to graph node if this task has dependencies
   kmp_task_team_t *td_task_team;
+  // The global thread id of the encountering thread. We need it because when a
+  // regular task depends on a hidden helper task, and the hidden helper task
+  // is finished on a hidden helper thread, it will call __kmp_release_deps to
+  // release all dependences. If now the task is a regular task, we need to pass
+  // the encountering gtid such that the task will be picked up and executed by
+  // its encountering team instead of hidden helper team.
+  kmp_int32 encountering_gtid;
   size_t td_size_alloc; // Size of task structure, including shareds etc.
 #if defined(KMP_GOMP_COMPAT)
   // 4 or 8 byte integers for the loop bounds in GOMP_taskloop
@@ -2449,6 +2457,9 @@
   kmp_int32 tt_max_threads; // # entries allocated for threads_data array
   kmp_int32 tt_found_proxy_tasks; // found proxy tasks since last barrier
   kmp_int32 tt_untied_task_encountered;
+  // There is hidden helper thread encountered in this task team so that we must
+  // wait when waiting on task team
+  kmp_int32 tt_hidden_helper_task_encountered;
 
   KMP_ALIGN_CACHE
   std::atomic<kmp_int32> tt_unfinished_threads; /* #threads still active */
@@ -2917,6 +2928,7 @@
 extern volatile int __kmp_init_monitor;
 #endif
 extern volatile int __kmp_init_user_locks;
+extern volatile int __kmp_init_hidden_helper_threads;
 extern int __kmp_init_counter;
 extern int __kmp_root_counter;
 extern int __kmp_version;
@@ -3985,6 +3997,45 @@
 
 extern void __kmp_omp_display_env(int verbose);
 
+// 1: it is initializing hidden helper team
+extern volatile int __kmp_init_hidden_helper;
+// 1: the hidden helper team is done
+extern volatile int __kmp_hidden_helper_team_done;
+// 1: enable hidden helper task
+extern kmp_int32 __kmp_enable_hidden_helper;
+// Main thread of hidden helper team
+extern kmp_info_t *__kmp_hidden_helper_main_thread;
+// Descriptors for the hidden helper threads
+extern kmp_info_t **__kmp_hidden_helper_threads;
+// Number of hidden helper threads
+extern kmp_int32 __kmp_hidden_helper_threads_num;
+// Number of hidden helper tasks that have not been executed yet
+extern std::atomic<kmp_int32> __kmp_unexecuted_hidden_helper_tasks;
+
+extern void __kmp_hidden_helper_initialize();
+extern void __kmp_hidden_helper_threads_initz_routine();
+extern void __kmp_do_initialize_hidden_helper_threads();
+extern void __kmp_hidden_helper_threads_initz_wait();
+extern void __kmp_hidden_helper_initz_release();
+extern void __kmp_hidden_helper_threads_deinitz_wait();
+extern void __kmp_hidden_helper_threads_deinitz_release();
+extern void __kmp_hidden_helper_main_thread_wait();
+extern void __kmp_hidden_helper_worker_thread_wait();
+extern void __kmp_hidden_helper_worker_thread_signal();
+extern void __kmp_hidden_helper_main_thread_release();
+
+// Check whether a given thread is a hidden helper thread
+#define KMP_HIDDEN_HELPER_THREAD(gtid)                                         \
+  ((gtid) >= 1 && (gtid) <= __kmp_hidden_helper_threads_num)
+
+#define KMP_HIDDEN_HELPER_WORKER_THREAD(gtid)                                  \
+  ((gtid) > 1 && (gtid) <= __kmp_hidden_helper_threads_num)
+
+// Map a gtid to a hidden helper thread. The first hidden helper thread, a.k.a
+// main thread, is skipped.
+#define KMP_GTID_TO_SHADOW_GTID(gtid)                                          \
+  ((gtid) % (__kmp_hidden_helper_threads_num - 1) + 2)
+
 #ifdef __cplusplus
 }
 #endif
diff --git a/openmp/runtime/src/kmp_global.cpp b/openmp/runtime/src/kmp_global.cpp
--- a/openmp/runtime/src/kmp_global.cpp
+++ b/openmp/runtime/src/kmp_global.cpp
@@ -46,6 +46,9 @@
 volatile int __kmp_init_common = FALSE;
 volatile int __kmp_init_middle = FALSE;
 volatile int __kmp_init_parallel = FALSE;
+volatile int __kmp_init_hidden_helper = FALSE;
+volatile int __kmp_init_hidden_helper_threads = FALSE;
+volatile int __kmp_hidden_helper_team_done = FALSE;
 #if KMP_USE_MONITOR
 volatile int __kmp_init_monitor =
     0; /* 1 - launched, 2 - actually started (Windows* OS only) */
diff --git a/openmp/runtime/src/kmp_runtime.cpp b/openmp/runtime/src/kmp_runtime.cpp
--- a/openmp/runtime/src/kmp_runtime.cpp
+++ b/openmp/runtime/src/kmp_runtime.cpp
@@ -3644,15 +3644,37 @@
     }
   }
 
-  /* find an available thread slot */
-  /* Don't reassign the zero slot since we need that to only be used by initial
-     thread */
-  for (gtid = (initial_thread ? 0 : 1); TCR_PTR(__kmp_threads[gtid]) != NULL;
-       gtid++)
-    ;
-  KA_TRACE(1,
-           ("__kmp_register_root: found slot in threads array: T#%d\n", gtid));
-  KMP_ASSERT(gtid < __kmp_threads_capacity);
+  // When hidden helper task is enabled, __kmp_threads is organized as follows:
+  // 0: initial thread, also a regular OpenMP thread.
+  // [1, __kmp_hidden_helper_threads_num]: slots for hidden helper threads.
+  // [__kmp_hidden_helper_threads_num + 1, __kmp_threads_capacity): slots for
+  // regular OpenMP threads.
+  if (TCR_4(__kmp_init_hidden_helper_threads)) {
+    // Find an available thread slot for hidden helper thread. Slots for hidden
+    // helper threads start from 1 to __kmp_hidden_helper_threads_num.
+    for (gtid = 1; TCR_PTR(__kmp_threads[gtid]) != NULL &&
+                   gtid <= __kmp_hidden_helper_threads_num;
+         gtid++)
+      ;
+    KMP_ASSERT(gtid <= __kmp_hidden_helper_threads_num);
+    KA_TRACE(1, ("__kmp_register_root: found slot in threads array for "
+                 "hidden helper thread: T#%d\n",
+                 gtid));
+  } else {
+    /* find an available thread slot */
+    // Don't reassign the zero slot since we need that to only be used by
+    // initial thread. Slots for hidden helper threads should also be skipped.
+    if (initial_thread && __kmp_threads[0] == NULL) {
+      gtid = 0;
+    } else {
+      for (gtid = __kmp_hidden_helper_threads_num + 1;
+           TCR_PTR(__kmp_threads[gtid]) != NULL; gtid++)
+        ;
+    }
+    KA_TRACE(
+        1, ("__kmp_register_root: found slot in threads array: T#%d\n", gtid));
+    KMP_ASSERT(gtid < __kmp_threads_capacity);
+  }
 
   /* update global accounting */
   __kmp_all_nth++;
@@ -4303,8 +4325,20 @@
 #endif
 
   KMP_MB();
-  for (new_gtid = 1; TCR_PTR(__kmp_threads[new_gtid]) != NULL; ++new_gtid) {
-    KMP_DEBUG_ASSERT(new_gtid < __kmp_threads_capacity);
+
+  {
+    int new_start_gtid = TCR_4(__kmp_init_hidden_helper_threads)
+                             ? 1
+                             : __kmp_hidden_helper_threads_num + 1;
+
+    for (new_gtid = new_start_gtid; TCR_PTR(__kmp_threads[new_gtid]) != NULL;
+         ++new_gtid) {
+      KMP_DEBUG_ASSERT(new_gtid < __kmp_threads_capacity);
+    }
+
+    if (TCR_4(__kmp_init_hidden_helper_threads)) {
+      KMP_DEBUG_ASSERT(new_gtid <= __kmp_hidden_helper_threads_num);
+    }
   }
 
   /* allocate space for it. */
@@ -6249,6 +6283,15 @@
     return;
   }
 
+  // If hidden helper team has been initialized, we need to deinit it
+  if (TCR_4(__kmp_init_hidden_helper)) {
+    TCW_SYNC_4(__kmp_hidden_helper_team_done, TRUE);
+    // First release the main thread to let it continue its work
+    __kmp_hidden_helper_main_thread_release();
+    // Wait until the hidden helper team has been destroyed
+    __kmp_hidden_helper_threads_deinitz_wait();
+  }
+
   KMP_MB(); /* Flush all pending memory write invalidates.  */
 
   /* find out who we are and what we should do */
@@ -7125,6 +7168,41 @@
   __kmp_release_bootstrap_lock(&__kmp_initz_lock);
 }
 
+void __kmp_hidden_helper_initialize() {
+  if (TCR_4(__kmp_init_hidden_helper))
+    return;
+
+  // __kmp_parallel_initialize is required before we initialize hidden helper
+  if (!TCR_4(__kmp_init_parallel))
+    __kmp_parallel_initialize();
+
+  // Double check. Note that this double check should not be placed before
+  // __kmp_parallel_initialize as it will cause dead lock.
+  __kmp_acquire_bootstrap_lock(&__kmp_initz_lock);
+  if (TCR_4(__kmp_init_hidden_helper)) {
+    __kmp_release_bootstrap_lock(&__kmp_initz_lock);
+    return;
+  }
+
+  // Set the count of hidden helper tasks to be executed to zero
+  KMP_ATOMIC_ST_REL(&__kmp_unexecuted_hidden_helper_tasks, 0);
+
+  // Set the global variable indicating that we're initializing hidden helper
+  // team/threads
+  TCW_SYNC_4(__kmp_init_hidden_helper_threads, TRUE);
+
+  // Platform independent initialization
+  __kmp_do_initialize_hidden_helper_threads();
+
+  // Wait here for the finish of initialization of hidden helper teams
+  __kmp_hidden_helper_threads_initz_wait();
+
+  // We have finished hidden helper initialization
+  TCW_SYNC_4(__kmp_init_hidden_helper, TRUE);
+
+  __kmp_release_bootstrap_lock(&__kmp_initz_lock);
+}
+
 /* ------------------------------------------------------------------------ */
 
 void __kmp_run_before_invoked_task(int gtid, int tid, kmp_info_t *this_thr,
@@ -8470,7 +8548,6 @@
   }
 }
 
-
 void __kmp_omp_display_env(int verbose) {
   __kmp_acquire_bootstrap_lock(&__kmp_initz_lock);
   if (__kmp_init_serial == 0)
@@ -8478,3 +8555,59 @@
   __kmp_display_env_impl(!verbose, verbose);
   __kmp_release_bootstrap_lock(&__kmp_initz_lock);
 }
+
+// Globals and functions for hidden helper task
+kmp_info_t **__kmp_hidden_helper_threads;
+kmp_info_t *__kmp_hidden_helper_main_thread;
+kmp_int32 __kmp_hidden_helper_threads_num = 8;
+std::atomic<kmp_int32> __kmp_unexecuted_hidden_helper_tasks;
+#if KMP_OS_LINUX
+kmp_int32 __kmp_enable_hidden_helper = TRUE;
+#else
+kmp_int32 __kmp_enable_hidden_helper = FALSE;
+#endif
+
+namespace {
+std::atomic<kmp_int32> __kmp_hit_hidden_helper_threads_num;
+
+void __kmp_hidden_helper_wrapper_fn(int *gtid, int *, ...) {
+  // This is an explicit synchronization on all hidden helper threads in case
+  // that when a regular thread pushes a hidden helper task to one hidden
+  // helper thread, the thread has not been awaken once since they're released
+  // by the main thread after creating the team.
+  KMP_ATOMIC_INC(&__kmp_hit_hidden_helper_threads_num);
+  while (KMP_ATOMIC_LD_ACQ(&__kmp_hit_hidden_helper_threads_num) !=
+         __kmp_hidden_helper_threads_num)
+    ;
+
+  // If main thread, then wait for signal
+  if (__kmpc_master(nullptr, *gtid)) {
+    // First, unset the initial state and release the initial thread
+    TCW_4(__kmp_init_hidden_helper_threads, FALSE);
+    __kmp_hidden_helper_initz_release();
+    __kmp_hidden_helper_main_thread_wait();
+    // Now wake up all worker threads
+    for (int i = 1; i < __kmp_hit_hidden_helper_threads_num; ++i) {
+      __kmp_hidden_helper_worker_thread_signal();
+    }
+  }
+}
+} // namespace
+
+void __kmp_hidden_helper_threads_initz_routine() {
+  // Create a new root for hidden helper team/threads
+  const int gtid = __kmp_register_root(TRUE);
+  __kmp_hidden_helper_main_thread = __kmp_threads[gtid];
+  __kmp_hidden_helper_threads = &__kmp_threads[gtid];
+  __kmp_hidden_helper_main_thread->th.th_set_nproc =
+      __kmp_hidden_helper_threads_num;
+
+  KMP_ATOMIC_ST_REL(&__kmp_hit_hidden_helper_threads_num, 0);
+
+  __kmpc_fork_call(nullptr, 0, __kmp_hidden_helper_wrapper_fn);
+
+  // Set the initialization flag to FALSE
+  TCW_SYNC_4(__kmp_init_hidden_helper, FALSE);
+
+  __kmp_hidden_helper_threads_deinitz_release();
+}
diff --git a/openmp/runtime/src/kmp_settings.cpp b/openmp/runtime/src/kmp_settings.cpp
--- a/openmp/runtime/src/kmp_settings.cpp
+++ b/openmp/runtime/src/kmp_settings.cpp
@@ -503,6 +503,11 @@
   if (nth < (4 * __kmp_xproc))
     nth = (4 * __kmp_xproc);
 
+  // If hidden helper task is enabled, we initialize the thread capacity with
+  // extra
+  // __kmp_hidden_helper_threads_num.
+  nth += __kmp_hidden_helper_threads_num;
+
   if (nth > __kmp_max_nth)
     nth = __kmp_max_nth;
 
@@ -1161,6 +1166,39 @@
   K_DIAG(1, ("__kmp_dflt_team_nth == %d\n", __kmp_dflt_team_nth));
 } // __kmp_stg_parse_num_threads
 
+static void __kmp_stg_parse_num_hidden_helper_threads(char const *name,
+                                                      char const *value,
+                                                      void *data) {
+  __kmp_stg_parse_int(name, value, 0, 16, &__kmp_hidden_helper_threads_num);
+  // If the number of hidden helper threads is zero, we disable hidden helper
+  // task
+  if (__kmp_hidden_helper_threads_num == 0) {
+    __kmp_enable_hidden_helper = FALSE;
+  }
+} // __kmp_stg_parse_num_hidden_helper_threads
+
+static void __kmp_stg_print_num_hidden_helper_threads(kmp_str_buf_t *buffer,
+                                                      char const *name,
+                                                      void *data) {
+  __kmp_stg_print_int(buffer, name, __kmp_hidden_helper_threads_num);
+} // __kmp_stg_print_num_hidden_helper_threads
+
+static void __kmp_stg_parse_use_hidden_helper(char const *name,
+                                              char const *value, void *data) {
+  __kmp_stg_parse_bool(name, value, &__kmp_enable_hidden_helper);
+#if !KMP_OS_LINUX
+  __kmp_enable_hidden_helper = FALSE;
+  K_DIAG(1,
+         ("__kmp_stg_parse_use_hidden_helper: Disable hidden helper task on "
+          "non-Linux platform although it is enabled by user explicitly.\n"));
+#endif
+} // __kmp_stg_parse_use_hidden_helper
+
+static void __kmp_stg_print_use_hidden_helper(kmp_str_buf_t *buffer,
+                                              char const *name, void *data) {
+  __kmp_stg_print_bool(buffer, name, __kmp_enable_hidden_helper);
+} // __kmp_stg_print_use_hidden_helper
+
 static void __kmp_stg_print_num_threads(kmp_str_buf_t *buffer, char const *name,
                                         void *data) {
   if (__kmp_env_format) {
@@ -4954,6 +4992,11 @@
      __kmp_stg_print_omp_cancellation, NULL, 0, 0},
     {"OMP_ALLOCATOR", __kmp_stg_parse_allocator, __kmp_stg_print_allocator,
      NULL, 0, 0},
+    {"LIBOMP_USE_HIDDEN_HELPER_TASK", __kmp_stg_parse_use_hidden_helper,
+     __kmp_stg_print_use_hidden_helper, NULL, 0, 0},
+    {"LIBOMP_NUM_HIDDEN_HELPER_THREADS",
+     __kmp_stg_parse_num_hidden_helper_threads,
+     __kmp_stg_print_num_hidden_helper_threads, NULL, 0, 0},
 
 #if OMPT_SUPPORT
     {"OMP_TOOL", __kmp_stg_parse_omp_tool, __kmp_stg_print_omp_tool, NULL, 0,
diff --git a/openmp/runtime/src/kmp_taskdeps.h b/openmp/runtime/src/kmp_taskdeps.h
--- a/openmp/runtime/src/kmp_taskdeps.h
+++ b/openmp/runtime/src/kmp_taskdeps.h
@@ -119,6 +119,7 @@
   KMP_RELEASE_DEPNODE(gtid, node);
 
   kmp_depnode_list_t *next;
+  kmp_taskdata_t *next_taskdata;
   for (kmp_depnode_list_t *p = node->dn.successors; p; p = next) {
     kmp_depnode_t *successor = p->node;
     kmp_int32 npredecessors = KMP_ATOMIC_DEC(&successor->dn.npredecessors) - 1;
@@ -131,7 +132,24 @@
         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);
+        // If a regular task depending on a hidden helper task, when the
+        // hidden helper task is done, the regular task should be executed by
+        // its encountering team.
+        if (KMP_HIDDEN_HELPER_THREAD(gtid)) {
+          // Hidden helper thread can only execute hidden helper tasks
+          KMP_ASSERT(task->td_flags.hidden_helper);
+          next_taskdata = KMP_TASK_TO_TASKDATA(successor->dn.task);
+          // If the dependent task is a regular task, we need to push to its
+          // encountering thread's queue; otherwise, it can be pushed to its own
+          // queue.
+          if (!next_taskdata->td_flags.hidden_helper) {
+            __kmp_omp_task(task->encountering_gtid, successor->dn.task, false);
+          } else {
+            __kmp_omp_task(gtid, successor->dn.task, false);
+          }
+        } else {
+          __kmp_omp_task(gtid, successor->dn.task, false);
+        }
       }
     }
 
diff --git a/openmp/runtime/src/kmp_taskdeps.cpp b/openmp/runtime/src/kmp_taskdeps.cpp
--- a/openmp/runtime/src/kmp_taskdeps.cpp
+++ b/openmp/runtime/src/kmp_taskdeps.cpp
@@ -588,7 +588,9 @@
                 current_task->td_flags.tasking_ser ||
                 current_task->td_flags.final;
   kmp_task_team_t *task_team = thread->th.th_task_team;
-  serial = serial && !(task_team && task_team->tt.tt_found_proxy_tasks);
+  serial = serial &&
+           !(task_team && (task_team->tt.tt_found_proxy_tasks ||
+                           task_team->tt.tt_hidden_helper_task_encountered));
 
   if (!serial && (ndeps > 0 || ndeps_noalias > 0)) {
     /* if no dependencies have been tracked yet, create the dependence hash */
diff --git a/openmp/runtime/src/kmp_tasking.cpp b/openmp/runtime/src/kmp_tasking.cpp
--- a/openmp/runtime/src/kmp_tasking.cpp
+++ b/openmp/runtime/src/kmp_tasking.cpp
@@ -325,6 +325,12 @@
 static kmp_int32 __kmp_push_task(kmp_int32 gtid, kmp_task_t *task) {
   kmp_info_t *thread = __kmp_threads[gtid];
   kmp_taskdata_t *taskdata = KMP_TASK_TO_TASKDATA(task);
+
+  if (taskdata->td_flags.hidden_helper) {
+    gtid = KMP_GTID_TO_SHADOW_GTID(gtid);
+    thread = __kmp_threads[gtid];
+  }
+
   kmp_task_team_t *task_team = thread->th.th_task_team;
   kmp_int32 tid = __kmp_tid_from_gtid(gtid);
   kmp_thread_data_t *thread_data;
@@ -363,7 +369,9 @@
   // Find tasking deque specific to encountering thread
   thread_data = &task_team->tt.tt_threads_data[tid];
 
-  // No lock needed since only owner can allocate
+  // No lock needed since only owner can allocate. If the task is hidden_helper,
+  // we don't need it either because we have initialized the dequeue for hidden
+  // helper thread data.
   if (UNLIKELY(thread_data->td.td_deque == NULL)) {
     __kmp_alloc_task_deque(thread, thread_data);
   }
@@ -429,6 +437,12 @@
 
   __kmp_release_bootstrap_lock(&thread_data->td.td_deque_lock);
 
+  // Signal one worker thread to execute the task
+  if (taskdata->td_flags.hidden_helper) {
+    // Wake hidden helper threads up if they're sleeping
+    __kmp_hidden_helper_worker_thread_signal();
+  }
+
   return TASK_SUCCESSFULLY_PUSHED;
 }
 
@@ -721,7 +735,6 @@
 #else /* ! USE_FAST_MEMORY */
   __kmp_thread_free(thread, taskdata);
 #endif
-
   KA_TRACE(20, ("__kmp_free_task: T#%d freed task %p\n", gtid, taskdata));
 }
 
@@ -922,7 +935,8 @@
     // Only need to keep track of count if team parallel and tasking not
     // serialized, or task is detachable and event has already been fulfilled
     if (!(taskdata->td_flags.team_serial || taskdata->td_flags.tasking_ser) ||
-        taskdata->td_flags.detachable == TASK_DETACHABLE) {
+        taskdata->td_flags.detachable == TASK_DETACHABLE ||
+        taskdata->td_flags.hidden_helper) {
       // Predecrement simulated by "- 1" calculation
       children =
           KMP_ATOMIC_DEC(&taskdata->td_parent->td_incomplete_child_tasks) - 1;
@@ -1171,6 +1185,7 @@
   kmp_task_t *task;
   kmp_taskdata_t *taskdata;
   kmp_info_t *thread = __kmp_threads[gtid];
+  kmp_info_t *encountering_thread = thread;
   kmp_team_t *team = thread->th.th_team;
   kmp_taskdata_t *parent_task = thread->th.th_current_task;
   size_t shareds_offset;
@@ -1178,6 +1193,25 @@
   if (UNLIKELY(!TCR_4(__kmp_init_middle)))
     __kmp_middle_initialize();
 
+  if (flags->hidden_helper) {
+    if (__kmp_enable_hidden_helper) {
+      if (!TCR_4(__kmp_init_hidden_helper))
+        __kmp_hidden_helper_initialize();
+
+      // For a hidden helper task encountered by a regular thread, we will push
+      // the task to the (gtid%__kmp_hidden_helper_threads_num)-th hidden helper
+      // thread.
+      if (!KMP_HIDDEN_HELPER_THREAD(gtid)) {
+        thread = __kmp_threads[KMP_GTID_TO_SHADOW_GTID(gtid)];
+        // We don't change the parent-child relation for hidden helper task as
+        // we need that to do per-task-region synchronization.
+      }
+    } else {
+      // If the hidden helper task is not enabled, reset the flag to FALSE.
+      flags->hidden_helper = FALSE;
+    }
+  }
+
   KA_TRACE(10, ("__kmp_task_alloc(enter): T#%d loc=%p, flags=(0x%x) "
                 "sizeof_task=%ld sizeof_shared=%ld entry=%p\n",
                 gtid, loc_ref, *((kmp_int32 *)flags), sizeof_kmp_task_t,
@@ -1188,24 +1222,27 @@
     }
     flags->final = 1;
   }
+
   if (flags->tiedness == TASK_UNTIED && !team->t.t_serialized) {
     // Untied task encountered causes the TSC algorithm to check entire deque of
     // the victim thread. If no untied task encountered, then checking the head
     // of the deque should be enough.
-    KMP_CHECK_UPDATE(thread->th.th_task_team->tt.tt_untied_task_encountered, 1);
+    KMP_CHECK_UPDATE(
+        encountering_thread->th.th_task_team->tt.tt_untied_task_encountered, 1);
   }
 
   // Detachable tasks are not proxy tasks yet but could be in the future. Doing
   // the tasking setup
   // when that happens is too late.
-  if (flags->proxy == TASK_PROXY || flags->detachable == TASK_DETACHABLE) {
+  if (flags->proxy == TASK_PROXY || flags->detachable == TASK_DETACHABLE ||
+      flags->hidden_helper) {
     if (flags->proxy == TASK_PROXY) {
       flags->tiedness = TASK_UNTIED;
       flags->merged_if0 = 1;
     }
     /* are we running in a sequential parallel or tskm_immediate_exec... we need
        tasking support enabled */
-    if ((thread->th.th_task_team) == NULL) {
+    if ((encountering_thread->th.th_task_team) == NULL) {
       /* This should only happen if the team is serialized
           setup a task team and propagate it to the thread */
       KMP_DEBUG_ASSERT(team->t.t_serialized);
@@ -1213,28 +1250,33 @@
                ("T#%d creating task team in __kmp_task_alloc for proxy task\n",
                 gtid));
       __kmp_task_team_setup(
-          thread, team,
+          encountering_thread, team,
           1); // 1 indicates setup the current team regardless of nthreads
-      thread->th.th_task_team = team->t.t_task_team[thread->th.th_task_state];
+      encountering_thread->th.th_task_team =
+          team->t.t_task_team[encountering_thread->th.th_task_state];
     }
-    kmp_task_team_t *task_team = thread->th.th_task_team;
+    kmp_task_team_t *task_team = encountering_thread->th.th_task_team;
 
     /* tasking must be enabled now as the task might not be pushed */
     if (!KMP_TASKING_ENABLED(task_team)) {
       KA_TRACE(
           30,
           ("T#%d enabling tasking in __kmp_task_alloc for proxy task\n", gtid));
-      __kmp_enable_tasking(task_team, thread);
-      kmp_int32 tid = thread->th.th_info.ds.ds_tid;
+      __kmp_enable_tasking(task_team, encountering_thread);
+      kmp_int32 tid = encountering_thread->th.th_info.ds.ds_tid;
       kmp_thread_data_t *thread_data = &task_team->tt.tt_threads_data[tid];
       // No lock needed since only owner can allocate
       if (thread_data->td.td_deque == NULL) {
-        __kmp_alloc_task_deque(thread, thread_data);
+        __kmp_alloc_task_deque(encountering_thread, thread_data);
       }
     }
 
-    if (task_team->tt.tt_found_proxy_tasks == FALSE)
+    if (flags->proxy == TASK_PROXY &&
+        task_team->tt.tt_found_proxy_tasks == FALSE)
       TCW_4(task_team->tt.tt_found_proxy_tasks, TRUE);
+    if (flags->hidden_helper &&
+        task_team->tt.tt_hidden_helper_task_encountered == FALSE)
+      TCW_4(task_team->tt.tt_hidden_helper_task_encountered, TRUE);
   }
 
   // Calculate shared structure offset including padding after kmp_task_t struct
@@ -1248,13 +1290,13 @@
   KA_TRACE(30, ("__kmp_task_alloc: T#%d Second malloc size: %ld\n", gtid,
                 sizeof_shareds));
 
-// Avoid double allocation here by combining shareds with taskdata
+  // Avoid double allocation here by combining shareds with taskdata
 #if USE_FAST_MEMORY
-  taskdata = (kmp_taskdata_t *)__kmp_fast_allocate(thread, shareds_offset +
-                                                               sizeof_shareds);
+  taskdata = (kmp_taskdata_t *)__kmp_fast_allocate(
+      encountering_thread, shareds_offset + sizeof_shareds);
 #else /* ! USE_FAST_MEMORY */
-  taskdata = (kmp_taskdata_t *)__kmp_thread_malloc(thread, shareds_offset +
-                                                               sizeof_shareds);
+  taskdata = (kmp_taskdata_t *)__kmp_thread_malloc(
+      encountering_thread, shareds_offset + sizeof_shareds);
 #endif /* USE_FAST_MEMORY */
   ANNOTATE_HAPPENS_AFTER(taskdata);
 
@@ -1281,8 +1323,8 @@
   task->part_id = 0; // AC: Always start with 0 part id
 
   taskdata->td_task_id = KMP_GEN_TASK_ID();
-  taskdata->td_team = team;
-  taskdata->td_alloc_thread = thread;
+  taskdata->td_team = thread->th.th_team;
+  taskdata->td_alloc_thread = encountering_thread;
   taskdata->td_parent = parent_task;
   taskdata->td_level = parent_task->td_level + 1; // increment nesting level
   KMP_ATOMIC_ST_RLX(&taskdata->td_untied_count, 0);
@@ -1301,6 +1343,8 @@
   taskdata->td_flags.destructors_thunk = flags->destructors_thunk;
   taskdata->td_flags.proxy = flags->proxy;
   taskdata->td_flags.detachable = flags->detachable;
+  taskdata->td_flags.hidden_helper = flags->hidden_helper;
+  taskdata->encountering_gtid = gtid;
   taskdata->td_task_team = thread->th.th_task_team;
   taskdata->td_size_alloc = shareds_offset + sizeof_shareds;
   taskdata->td_flags.tasktype = TASK_EXPLICIT;
@@ -1342,12 +1386,11 @@
   if (UNLIKELY(ompt_enabled.enabled))
     __ompt_task_init(taskdata, gtid);
 #endif
-// Only need to keep track of child task counts if team parallel and tasking not
-// serialized or if it is a proxy or detachable task
-  if (flags->proxy == TASK_PROXY ||
-      flags->detachable == TASK_DETACHABLE ||
-      !(taskdata->td_flags.team_serial || taskdata->td_flags.tasking_ser))
-  {
+  // Only need to keep track of child task counts if team parallel and tasking
+  // not serialized or if it is a proxy or detachable or hidden helper task
+  if (flags->proxy == TASK_PROXY || flags->detachable == TASK_DETACHABLE ||
+      flags->hidden_helper ||
+      !(taskdata->td_flags.team_serial || taskdata->td_flags.tasking_ser)) {
     KMP_ATOMIC_INC(&parent_task->td_incomplete_child_tasks);
     if (parent_task->td_taskgroup)
       KMP_ATOMIC_INC(&parent_task->td_taskgroup->count);
@@ -1358,6 +1401,12 @@
     }
   }
 
+  if (flags->hidden_helper) {
+    taskdata->td_flags.task_serial = FALSE;
+    // Increment the number of hidden helper tasks to be executed
+    KMP_ATOMIC_INC(&__kmp_unexecuted_hidden_helper_tasks);
+  }
+
   KA_TRACE(20, ("__kmp_task_alloc(exit): T#%d created task %p parent=%p\n",
                 gtid, taskdata, taskdata->td_parent));
   ANNOTATE_HAPPENS_BEFORE(task);
@@ -1395,6 +1444,11 @@
                                          size_t sizeof_shareds,
                                          kmp_routine_entry_t task_entry,
                                          kmp_int64 device_id) {
+  if (__kmp_enable_hidden_helper) {
+    auto &input_flags = reinterpret_cast<kmp_tasking_flags_t &>(flags);
+    input_flags.hidden_helper = TRUE;
+  }
+
   return __kmpc_omp_task_alloc(loc_ref, gtid, flags, sizeof_kmp_task_t,
                                sizeof_shareds, task_entry);
 }
@@ -1467,6 +1521,13 @@
   }
 #endif
 
+  // Decreament the counter of hidden helper tasks to be executed
+  if (taskdata->td_flags.hidden_helper) {
+    // Hidden helper tasks can only be executed by hidden helper threads
+    KMP_ASSERT(KMP_HIDDEN_HELPER_THREAD(gtid));
+    KMP_ATOMIC_DEC(&__kmp_unexecuted_hidden_helper_tasks);
+  }
+
   // Proxy tasks are not handled by the runtime
   if (taskdata->td_flags.proxy != TASK_PROXY) {
     ANNOTATE_HAPPENS_AFTER(task);
@@ -1864,6 +1925,12 @@
 
     must_wait = must_wait || (thread->th.th_task_team != NULL &&
                               thread->th.th_task_team->tt.tt_found_proxy_tasks);
+    // If hidden helper thread is encountered, we must enable wait here.
+    must_wait =
+        must_wait ||
+        (__kmp_enable_hidden_helper && thread->th.th_task_team != NULL &&
+         thread->th.th_task_team->tt.tt_hidden_helper_task_encountered);
+
     if (must_wait) {
       kmp_flag_32<false, false> flag(
           RCAST(std::atomic<kmp_uint32> *,
@@ -2830,11 +2897,13 @@
 
   thread->th.th_reap_state = KMP_NOT_SAFE_TO_REAP;
   threads_data = (kmp_thread_data_t *)TCR_PTR(task_team->tt.tt_threads_data);
+
   KMP_DEBUG_ASSERT(threads_data != NULL);
 
   nthreads = task_team->tt.tt_nproc;
   unfinished_threads = &(task_team->tt.tt_unfinished_threads);
-  KMP_DEBUG_ASSERT(nthreads > 1 || task_team->tt.tt_found_proxy_tasks);
+  KMP_DEBUG_ASSERT(nthreads > 1 || task_team->tt.tt_found_proxy_tasks ||
+                   task_team->tt.tt_hidden_helper_task_encountered);
   KMP_DEBUG_ASSERT(*unfinished_threads >= 0);
 
   while (1) { // Outer loop keeps trying to find tasks in case of single thread
@@ -2914,8 +2983,8 @@
         }
       }
 
-      if (task == NULL) // break out of tasking loop
-        break;
+      if (task == NULL)
+        break; // break out of tasking loop
 
 // Found a task; execute it
 #if USE_ITT_BUILD && USE_ITT_NOTIFY
@@ -3001,7 +3070,8 @@
 
     // We could be getting tasks from target constructs; if this is the only
     // thread, keep trying to execute tasks from own queue
-    if (nthreads == 1)
+    if (nthreads == 1 &&
+        KMP_ATOMIC_LD_ACQ(&current_task->td_incomplete_child_tasks))
       use_own_tasks = 1;
     else {
       KA_TRACE(15,
@@ -3389,6 +3459,7 @@
   task_team->tt.tt_nproc = nthreads = team->t.t_nproc;
 
   KMP_ATOMIC_ST_REL(&task_team->tt.tt_unfinished_threads, nthreads);
+  TCW_4(task_team->tt.tt_hidden_helper_task_encountered, FALSE);
   TCW_4(task_team->tt.tt_active, TRUE);
 
   KA_TRACE(20, ("__kmp_allocate_task_team: T#%d exiting; task_team = %p "
@@ -3561,6 +3632,26 @@
                     ((team != NULL) ? team->t.t_id : -1), other_team));
     }
   }
+
+  // For regular thread, task enabling should be called when the task is going
+  // to be pushed to a dequeue. However, for the hidden helper thread, we need
+  // it ahead of time so that some operations can be performed without race
+  // condition.
+  if (this_thr == __kmp_hidden_helper_main_thread) {
+    for (int i = 0; i < 2; ++i) {
+      kmp_task_team_t *task_team = team->t.t_task_team[i];
+      if (KMP_TASKING_ENABLED(task_team)) {
+        continue;
+      }
+      __kmp_enable_tasking(task_team, this_thr);
+      for (int j = 0; j < task_team->tt.tt_nproc; ++j) {
+        kmp_thread_data_t *thread_data = &task_team->tt.tt_threads_data[j];
+        if (thread_data->td.td_deque == NULL) {
+          __kmp_alloc_task_deque(__kmp_hidden_helper_threads[j], thread_data);
+        }
+      }
+    }
+  }
 }
 
 // __kmp_task_team_sync: Propagation of task team data from team to threads
diff --git a/openmp/runtime/src/kmp_wait_release.h b/openmp/runtime/src/kmp_wait_release.h
--- a/openmp/runtime/src/kmp_wait_release.h
+++ b/openmp/runtime/src/kmp_wait_release.h
@@ -389,6 +389,26 @@
         break;
     }
 
+    // For hidden helper thread, if task_team is nullptr, it means the main
+    // thread has not released the barrier. We cannot wait here because once the
+    // main thread releases all children barriers, all hidden helper threads are
+    // still sleeping. This leads to a problem that following configuration,
+    // such as task team sync, will not be performed such that this thread does
+    // not have task team. Usually it is not bad. However, a corner case is,
+    // when the first task encountered is an untied task, the check in
+    // __kmp_task_alloc will crash because it uses the task team pointer without
+    // checking whether it is nullptr. It is probably under some kind of
+    // assumption.
+    if (task_team && KMP_HIDDEN_HELPER_WORKER_THREAD(th_gtid) &&
+        !TCR_4(__kmp_hidden_helper_team_done)) {
+      // If there is still hidden helper tasks to be executed, the hidden helper
+      // thread will not enter a waiting status.
+      if (KMP_ATOMIC_LD_ACQ(&__kmp_unexecuted_hidden_helper_tasks) == 0) {
+        __kmp_hidden_helper_worker_thread_wait();
+      }
+      continue;
+    }
+
     // Don't suspend if KMP_BLOCKTIME is set to "infinite"
     if (__kmp_dflt_blocktime == KMP_MAX_BLOCKTIME &&
         __kmp_pause_status != kmp_soft_paused)
diff --git a/openmp/runtime/src/z_Linux_util.cpp b/openmp/runtime/src/z_Linux_util.cpp
--- a/openmp/runtime/src/z_Linux_util.cpp
+++ b/openmp/runtime/src/z_Linux_util.cpp
@@ -25,6 +25,7 @@
 #include <alloca.h>
 #endif
 #include <math.h> // HUGE_VAL.
+#include <semaphore.h>
 #include <sys/resource.h>
 #include <sys/syscall.h>
 #include <sys/time.h>
@@ -2447,7 +2448,7 @@
                            ,
                            void **exit_frame_ptr
 #endif
-                           ) {
+) {
 #if OMPT_SUPPORT
   *exit_frame_ptr = OMPT_GET_FRAME_ADDRESS(0);
 #endif
@@ -2526,4 +2527,165 @@
 
 #endif
 
+// Functions for hidden helper task
+namespace {
+// Condition variable for initializing hidden helper team
+pthread_cond_t hidden_helper_threads_initz_cond_var;
+pthread_mutex_t hidden_helper_threads_initz_lock;
+volatile int hidden_helper_initz_signaled = FALSE;
+
+// Condition variable for deinitializing hidden helper team
+pthread_cond_t hidden_helper_threads_deinitz_cond_var;
+pthread_mutex_t hidden_helper_threads_deinitz_lock;
+volatile int hidden_helper_deinitz_signaled = FALSE;
+
+// Condition variable for the wrapper function of main thread
+pthread_cond_t hidden_helper_main_thread_cond_var;
+pthread_mutex_t hidden_helper_main_thread_lock;
+volatile int hidden_helper_main_thread_signaled = FALSE;
+
+// Semaphore for worker threads. We don't use condition variable here in case
+// that when multiple signals are sent at the same time, only one thread might
+// be waken.
+sem_t hidden_helper_task_sem;
+} // namespace
+
+void __kmp_hidden_helper_worker_thread_wait() {
+  int status = sem_wait(&hidden_helper_task_sem);
+  KMP_CHECK_SYSFAIL("sem_wait", status);
+}
+
+void __kmp_do_initialize_hidden_helper_threads() {
+  // Initialize condition variable
+  int status =
+      pthread_cond_init(&hidden_helper_threads_initz_cond_var, nullptr);
+  KMP_CHECK_SYSFAIL("pthread_cond_init", status);
+
+  status = pthread_cond_init(&hidden_helper_threads_deinitz_cond_var, nullptr);
+  KMP_CHECK_SYSFAIL("pthread_cond_init", status);
+
+  status = pthread_cond_init(&hidden_helper_main_thread_cond_var, nullptr);
+  KMP_CHECK_SYSFAIL("pthread_cond_init", status);
+
+  status = pthread_mutex_init(&hidden_helper_threads_initz_lock, nullptr);
+  KMP_CHECK_SYSFAIL("pthread_mutex_init", status);
+
+  status = pthread_mutex_init(&hidden_helper_threads_deinitz_lock, nullptr);
+  KMP_CHECK_SYSFAIL("pthread_mutex_init", status);
+
+  status = pthread_mutex_init(&hidden_helper_main_thread_lock, nullptr);
+  KMP_CHECK_SYSFAIL("pthread_mutex_init", status);
+
+  // Initialize the semaphore
+  status = sem_init(&hidden_helper_task_sem, 0, 0);
+  KMP_CHECK_SYSFAIL("sem_init", status);
+
+  // Create a new thread to finish initialization
+  pthread_t handle;
+  status = pthread_create(
+      &handle, nullptr,
+      [](void *) -> void * {
+        __kmp_hidden_helper_threads_initz_routine();
+        return nullptr;
+      },
+      nullptr);
+  KMP_CHECK_SYSFAIL("pthread_create", status);
+}
+
+void __kmp_hidden_helper_threads_initz_wait() {
+  // Initial thread waits here for the completion of the initialization. The
+  // condition variable will be notified by main thread of hidden helper teams.
+  int status = pthread_mutex_lock(&hidden_helper_threads_initz_lock);
+  KMP_CHECK_SYSFAIL("pthread_mutex_lock", status);
+
+  if (!TCR_4(hidden_helper_initz_signaled)) {
+    status = pthread_cond_wait(&hidden_helper_threads_initz_cond_var,
+                               &hidden_helper_threads_initz_lock);
+    KMP_CHECK_SYSFAIL("pthread_cond_wait", status);
+  }
+
+  status = pthread_mutex_unlock(&hidden_helper_threads_initz_lock);
+  KMP_CHECK_SYSFAIL("pthread_mutex_unlock", status);
+}
+
+void __kmp_hidden_helper_initz_release() {
+  // After all initialization, reset __kmp_init_hidden_helper_threads to false.
+  int status = pthread_mutex_lock(&hidden_helper_threads_initz_lock);
+  KMP_CHECK_SYSFAIL("pthread_mutex_lock", status);
+
+  status = pthread_cond_signal(&hidden_helper_threads_initz_cond_var);
+  KMP_CHECK_SYSFAIL("pthread_cond_wait", status);
+
+  TCW_SYNC_4(hidden_helper_initz_signaled, TRUE);
+
+  status = pthread_mutex_unlock(&hidden_helper_threads_initz_lock);
+  KMP_CHECK_SYSFAIL("pthread_mutex_unlock", status);
+}
+
+void __kmp_hidden_helper_main_thread_wait() {
+  // The main thread of hidden helper team will be blocked here. The
+  // condition variable can only be signal in the destructor of RTL.
+  int status = pthread_mutex_lock(&hidden_helper_main_thread_lock);
+  KMP_CHECK_SYSFAIL("pthread_mutex_lock", status);
+
+  if (!TCR_4(hidden_helper_main_thread_signaled)) {
+    status = pthread_cond_wait(&hidden_helper_main_thread_cond_var,
+                               &hidden_helper_main_thread_lock);
+    KMP_CHECK_SYSFAIL("pthread_cond_wait", status);
+  }
+
+  status = pthread_mutex_unlock(&hidden_helper_main_thread_lock);
+  KMP_CHECK_SYSFAIL("pthread_mutex_unlock", status);
+}
+
+void __kmp_hidden_helper_main_thread_release() {
+  // The initial thread of OpenMP RTL should call this function to wake up the
+  // main thread of hidden helper team.
+  int status = pthread_mutex_lock(&hidden_helper_main_thread_lock);
+  KMP_CHECK_SYSFAIL("pthread_mutex_lock", status);
+
+  status = pthread_cond_signal(&hidden_helper_main_thread_cond_var);
+  KMP_CHECK_SYSFAIL("pthread_cond_signal", status);
+
+  // The hidden helper team is done here
+  TCW_SYNC_4(hidden_helper_main_thread_signaled, TRUE);
+
+  status = pthread_mutex_unlock(&hidden_helper_main_thread_lock);
+  KMP_CHECK_SYSFAIL("pthread_mutex_unlock", status);
+}
+
+void __kmp_hidden_helper_worker_thread_signal() {
+  int status = sem_post(&hidden_helper_task_sem);
+  KMP_CHECK_SYSFAIL("sem_post", status);
+}
+
+void __kmp_hidden_helper_threads_deinitz_wait() {
+  // Initial thread waits here for the completion of the deinitialization. The
+  // condition variable will be notified by main thread of hidden helper teams.
+  int status = pthread_mutex_lock(&hidden_helper_threads_deinitz_lock);
+  KMP_CHECK_SYSFAIL("pthread_mutex_lock", status);
+
+  if (!TCR_4(hidden_helper_deinitz_signaled)) {
+    status = pthread_cond_wait(&hidden_helper_threads_deinitz_cond_var,
+                               &hidden_helper_threads_deinitz_lock);
+    KMP_CHECK_SYSFAIL("pthread_cond_wait", status);
+  }
+
+  status = pthread_mutex_unlock(&hidden_helper_threads_deinitz_lock);
+  KMP_CHECK_SYSFAIL("pthread_mutex_unlock", status);
+}
+
+void __kmp_hidden_helper_threads_deinitz_release() {
+  int status = pthread_mutex_lock(&hidden_helper_threads_deinitz_lock);
+  KMP_CHECK_SYSFAIL("pthread_mutex_lock", status);
+
+  status = pthread_cond_signal(&hidden_helper_threads_deinitz_cond_var);
+  KMP_CHECK_SYSFAIL("pthread_cond_wait", status);
+
+  TCW_SYNC_4(hidden_helper_deinitz_signaled, TRUE);
+
+  status = pthread_mutex_unlock(&hidden_helper_threads_deinitz_lock);
+  KMP_CHECK_SYSFAIL("pthread_mutex_unlock", status);
+}
+
 // end of file //
diff --git a/openmp/runtime/src/z_Windows_NT_util.cpp b/openmp/runtime/src/z_Windows_NT_util.cpp
--- a/openmp/runtime/src/z_Windows_NT_util.cpp
+++ b/openmp/runtime/src/z_Windows_NT_util.cpp
@@ -1629,3 +1629,40 @@
 
   return running_threads;
 } //__kmp_get_load_balance()
+
+// Functions for hidden helper task
+void __kmp_hidden_helper_worker_thread_wait() {
+  KMP_ASSERT(0 && "Hidden helper task is not supported on Windows");
+}
+
+void __kmp_do_initialize_hidden_helper_threads() {
+  KMP_ASSERT(0 && "Hidden helper task is not supported on Windows");
+}
+
+void __kmp_hidden_helper_threads_initz_wait() {
+  KMP_ASSERT(0 && "Hidden helper task is not supported on Windows");
+}
+
+void __kmp_hidden_helper_initz_release() {
+  KMP_ASSERT(0 && "Hidden helper task is not supported on Windows");
+}
+
+void __kmp_hidden_helper_main_thread_wait() {
+  KMP_ASSERT(0 && "Hidden helper task is not supported on Windows");
+}
+
+void __kmp_hidden_helper_main_thread_release() {
+  KMP_ASSERT(0 && "Hidden helper task is not supported on Windows");
+}
+
+void __kmp_hidden_helper_worker_thread_signal() {
+  KMP_ASSERT(0 && "Hidden helper task is not supported on Windows");
+}
+
+void __kmp_hidden_helper_threads_deinitz_wait() {
+  KMP_ASSERT(0 && "Hidden helper task is not supported on Windows");
+}
+
+void __kmp_hidden_helper_threads_deinitz_release() {
+  KMP_ASSERT(0 && "Hidden helper task is not supported on Windows");
+}
diff --git a/openmp/runtime/test/tasking/hidden_helper_task/common.h b/openmp/runtime/test/tasking/hidden_helper_task/common.h
new file mode 100644
--- /dev/null
+++ b/openmp/runtime/test/tasking/hidden_helper_task/common.h
@@ -0,0 +1,59 @@
+#include <cassert>
+#include <iostream>
+#include <string>
+
+extern "C" {
+struct ident_t;
+
+using kmp_int32 = int32_t;
+using kmp_int64 = int64_t;
+using kmp_routine_entry_t = kmp_int32 (*)(kmp_int32, void *);
+using kmp_intptr_t = intptr_t;
+
+typedef struct kmp_depend_info {
+  kmp_intptr_t base_addr;
+  size_t len;
+  struct {
+    bool in : 1;
+    bool out : 1;
+    bool mtx : 1;
+  } flags;
+} kmp_depend_info_t;
+
+typedef union kmp_cmplrdata {
+  kmp_int32 priority;
+  kmp_routine_entry_t destructors;
+} kmp_cmplrdata_t;
+
+typedef struct kmp_task {
+  void *shareds;
+  kmp_routine_entry_t routine;
+  kmp_int32 part_id;
+  kmp_cmplrdata_t data1;
+  kmp_cmplrdata_t data2;
+} kmp_task_t;
+
+int32_t __kmpc_global_thread_num(void *);
+kmp_task_t *__kmpc_omp_task_alloc(ident_t *, kmp_int32, kmp_int32, size_t,
+                                  size_t, kmp_routine_entry_t);
+kmp_task_t *__kmpc_omp_target_task_alloc(ident_t *, kmp_int32, kmp_int32,
+                                         size_t, size_t, kmp_routine_entry_t,
+                                         kmp_int64);
+kmp_int32 __kmpc_omp_taskwait(ident_t *, kmp_int32);
+kmp_int32 __kmpc_omp_task(ident_t *, kmp_int32, kmp_task_t *);
+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);
+void __kmpc_taskgroup(ident_t *, kmp_int32);
+void __kmpc_end_taskgroup(ident_t *, kmp_int32);
+}
+
+static kmp_int32 get_num_hidden_helper_threads() {
+  static kmp_int32 __kmp_hidden_helper_threads_num = 8;
+  if (const char *env = std::getenv("LIBOMP_NUM_HIDDEN_HELPER_THREADS")) {
+    return std::stoi(env);
+  }
+  return __kmp_hidden_helper_threads_num;
+}
diff --git a/openmp/runtime/test/tasking/hidden_helper_task/depend.cpp b/openmp/runtime/test/tasking/hidden_helper_task/depend.cpp
new file mode 100644
--- /dev/null
+++ b/openmp/runtime/test/tasking/hidden_helper_task/depend.cpp
@@ -0,0 +1,131 @@
+// RUN: %libomp-cxx-compile-and-run
+
+/*
+ * This test aims to check whether hidden helper task can work with regular task
+ * in terms of dependences. It is equivalent to the following code:
+ *
+ * #pragma omp parallel
+ * for (int i = 0; i < N; ++i) {
+ *   int data = -1;
+ * #pragma omp task shared(data) depend(out: data)
+ *   {
+ *     data = 1;
+ *   }
+ * #pragma omp hidden helper task shared(data) depend(inout: data)
+ *   {
+ *     data += 2;
+ *   }
+ * #pragma omp hidden helper task shared(data) depend(inout: data)
+ *   {
+ *     data += 4;
+ *   }
+ * #pragma omp task shared(data) depend(inout: data)
+ *   {
+ *     data += 8;
+ *   }
+ * #pragma omp taskwait
+ *   assert(data == 15);
+ * }
+ */
+
+#include "common.h"
+
+extern "C" {
+struct kmp_task_t_with_privates {
+  kmp_task_t task;
+};
+
+struct anon {
+  int32_t *data;
+};
+}
+
+template <int I>
+kmp_int32 omp_task_entry(kmp_int32 gtid, kmp_task_t_with_privates *task) {
+  auto shareds = reinterpret_cast<anon *>(task->task.shareds);
+  auto p = shareds->data;
+  *p += I;
+  return 0;
+}
+
+int main(int argc, char *argv[]) {
+  constexpr const int N = 1024;
+#pragma omp parallel for
+  for (int i = 0; i < N; ++i) {
+    int32_t gtid = __kmpc_global_thread_num(nullptr);
+    int32_t data = 0;
+
+    // Task 1
+    auto task1 = __kmpc_omp_task_alloc(
+        nullptr, gtid, 1, sizeof(kmp_task_t_with_privates), sizeof(anon),
+        reinterpret_cast<kmp_routine_entry_t>(omp_task_entry<1>));
+
+    auto shareds = reinterpret_cast<anon *>(task1->shareds);
+    shareds->data = &data;
+
+    kmp_depend_info_t depinfo1;
+    depinfo1.base_addr = reinterpret_cast<intptr_t>(&data);
+    depinfo1.flags.out = 1;
+    depinfo1.len = 4;
+
+    __kmpc_omp_task_with_deps(nullptr, gtid, task1, 1, &depinfo1, 0, nullptr);
+
+    // Task 2
+    auto task2 = __kmpc_omp_target_task_alloc(
+        nullptr, gtid, 1, sizeof(kmp_task_t_with_privates), sizeof(anon),
+        reinterpret_cast<kmp_routine_entry_t>(omp_task_entry<2>), -1);
+
+    shareds = reinterpret_cast<anon *>(task2->shareds);
+    shareds->data = &data;
+
+    kmp_depend_info_t depinfo2;
+    depinfo2.base_addr = reinterpret_cast<intptr_t>(&data);
+    depinfo2.flags.in = 1;
+    depinfo2.flags.out = 1;
+    depinfo2.len = 4;
+
+    __kmpc_omp_task_with_deps(nullptr, gtid, task2, 1, &depinfo2, 0, nullptr);
+
+    // Task 3
+    auto task3 = __kmpc_omp_target_task_alloc(
+        nullptr, gtid, 1, sizeof(kmp_task_t_with_privates), sizeof(anon),
+        reinterpret_cast<kmp_routine_entry_t>(omp_task_entry<4>), -1);
+
+    shareds = reinterpret_cast<anon *>(task3->shareds);
+    shareds->data = &data;
+
+    kmp_depend_info_t depinfo3;
+    depinfo3.base_addr = reinterpret_cast<intptr_t>(&data);
+    depinfo3.flags.in = 1;
+    depinfo3.flags.out = 1;
+    depinfo3.len = 4;
+
+    __kmpc_omp_task_with_deps(nullptr, gtid, task3, 1, &depinfo3, 0, nullptr);
+
+    // Task 4
+    auto task4 = __kmpc_omp_task_alloc(
+        nullptr, gtid, 1, sizeof(kmp_task_t_with_privates), sizeof(anon),
+        reinterpret_cast<kmp_routine_entry_t>(omp_task_entry<8>));
+
+    shareds = reinterpret_cast<anon *>(task4->shareds);
+    shareds->data = &data;
+
+    kmp_depend_info_t depinfo4;
+    depinfo4.base_addr = reinterpret_cast<intptr_t>(&data);
+    depinfo4.flags.in = 1;
+    depinfo4.flags.out = 1;
+    depinfo4.len = 4;
+
+    __kmpc_omp_task_with_deps(nullptr, gtid, task4, 1, &depinfo4, 0, nullptr);
+
+    // Wait for all tasks
+    __kmpc_omp_taskwait(nullptr, gtid);
+
+    assert(data == 15);
+  }
+
+  std::cout << "PASS\n";
+  return 0;
+}
+
+// CHECK: PASS
diff --git a/openmp/runtime/test/tasking/hidden_helper_task/gtid.cpp b/openmp/runtime/test/tasking/hidden_helper_task/gtid.cpp
new file mode 100644
--- /dev/null
+++ b/openmp/runtime/test/tasking/hidden_helper_task/gtid.cpp
@@ -0,0 +1,132 @@
+// RUN: %libomp-cxx-compile-and-run
+
+/*
+ * This test aims to check whether hidden helper thread has right gtid. We also
+ * test if there is mixed dependences between regular tasks and hidden helper
+ * tasks, the tasks are executed by right set of threads. It is equivalent to
+ * the following code:
+ *
+ * #pragma omp parallel for
+ * for (int i = 0; i < N; ++i) {
+ *   int data1 = -1, data2 = -1, data3 = -1;
+ *   int depvar;
+ * #pragma omp task shared(data1) depend(inout: depvar)
+ *   {
+ *     data1 = omp_get_global_thread_id();
+ *   }
+ * #pragma omp task hidden helper shared(data2) depend(inout: depvar)
+ *   {
+ *     data2 = omp_get_global_thread_id();
+ *   }
+ * #pragma omp task shared(data3) depend(inout: depvar)
+ *   {
+ *     data3 = omp_get_global_thread_id();
+ *   }
+ * #pragma omp taskwait
+ *   assert(data1 == 0 || data1 > __kmp_num_hidden_helper_threads);
+ *   assert(data2 > 0 && data2 <= __kmp_num_hidden_helper_threads);
+ *   assert(data3 == 0 || data3 > __kmp_num_hidden_helper_threads);
+ * }
+ */
+
+#include "common.h"
+
+extern "C" {
+struct kmp_task_t_with_privates {
+  kmp_task_t task;
+};
+
+struct anon {
+  int32_t *data;
+};
+}
+
+kmp_int32 __kmp_hidden_helper_threads_num;
+
+kmp_int32 omp_task_entry(kmp_int32 gtid, kmp_task_t_with_privates *task) {
+  auto shareds = reinterpret_cast<anon *>(task->task.shareds);
+  auto p = shareds->data;
+  *p = __kmpc_global_thread_num(nullptr);
+  return 0;
+}
+
+template <bool hidden_helper_task> void assert_gtid(int v) {
+  if (__kmp_hidden_helper_threads_num) {
+    if (hidden_helper_task) {
+      assert(v > 0 && v <= __kmp_hidden_helper_threads_num);
+    } else {
+      assert(v == 0 || v > __kmp_hidden_helper_threads_num);
+    }
+  } else {
+    assert(v >= 0);
+  }
+}
+
+int main(int argc, char *argv[]) {
+  __kmp_hidden_helper_threads_num = get_num_hidden_helper_threads();
+
+  constexpr const int N = 1024;
+#pragma omp parallel for
+  for (int i = 0; i < N; ++i) {
+    int32_t data1 = -1, data2 = -1, data3 = -1;
+    int depvar;
+    int32_t gtid = __kmpc_global_thread_num(nullptr);
+
+    // Task 1, regular task
+    auto task1 = __kmpc_omp_task_alloc(
+        nullptr, gtid, 1, sizeof(kmp_task_t_with_privates), sizeof(anon),
+        reinterpret_cast<kmp_routine_entry_t>(omp_task_entry));
+    auto shareds = reinterpret_cast<anon *>(task1->shareds);
+    shareds->data = &data1;
+
+    kmp_depend_info_t depinfo1;
+    depinfo1.base_addr = reinterpret_cast<intptr_t>(&depvar);
+    depinfo1.flags.in = 1;
+    depinfo1.flags.out = 1;
+    depinfo1.len = 4;
+
+    __kmpc_omp_task_with_deps(nullptr, gtid, task1, 1, &depinfo1, 0, nullptr);
+
+    // Task 2, hidden helper task
+    auto task2 = __kmpc_omp_target_task_alloc(
+        nullptr, gtid, 1, sizeof(kmp_task_t_with_privates), sizeof(anon),
+        reinterpret_cast<kmp_routine_entry_t>(omp_task_entry), -1);
+    shareds = reinterpret_cast<anon *>(task2->shareds);
+    shareds->data = &data2;
+
+    kmp_depend_info_t depinfo2;
+    depinfo2.base_addr = reinterpret_cast<intptr_t>(&depvar);
+    depinfo2.flags.in = 1;
+    depinfo2.flags.out = 1;
+    depinfo2.len = 4;
+
+    __kmpc_omp_task_with_deps(nullptr, gtid, task2, 1, &depinfo2, 0, nullptr);
+
+    // Task 3, regular task
+    auto task3 = __kmpc_omp_task_alloc(
+        nullptr, gtid, 1, sizeof(kmp_task_t_with_privates), sizeof(anon),
+        reinterpret_cast<kmp_routine_entry_t>(omp_task_entry));
+    shareds = reinterpret_cast<anon *>(task3->shareds);
+    shareds->data = &data3;
+
+    kmp_depend_info_t depinfo3;
+    depinfo3.base_addr = reinterpret_cast<intptr_t>(&depvar);
+    depinfo3.flags.in = 1;
+    depinfo3.flags.out = 1;
+    depinfo3.len = 4;
+
+    __kmpc_omp_task_with_deps(nullptr, gtid, task3, 1, &depinfo3, 0, nullptr);
+
+    __kmpc_omp_taskwait(nullptr, gtid);
+
+    // FIXME: 8 here is not accurate
+    assert_gtid<false>(data1);
+    assert_gtid<true>(data2);
+    assert_gtid<false>(data3);
+  }
+
+  std::cout << "PASS\n";
+  return 0;
+}
+
+// CHECK: PASS
diff --git a/openmp/runtime/test/tasking/hidden_helper_task/taskgroup.cpp b/openmp/runtime/test/tasking/hidden_helper_task/taskgroup.cpp
new file mode 100644
--- /dev/null
+++ b/openmp/runtime/test/tasking/hidden_helper_task/taskgroup.cpp
@@ -0,0 +1,78 @@
+// RUN: %libomp-cxx-compile-and-run
+
+/*
+ * This test aims to check whether hidden helper task can work with regular task
+ * in terms of dependences. It is equivalent to the following code:
+ *
+ * #pragma omp parallel
+ * for (int i = 0; i < N; ++i) {
+ *   int data1 = 0, data2 = 0;
+ * #pragma omp taskgroup
+ *   {
+ * #pragma omp hidden helper task shared(data1)
+ *    {
+ *      data1 = 1;
+ *    }
+ * #pragma omp hidden helper task shared(data2)
+ *    {
+ *      data2 = 2;
+ *    }
+ *   }
+ *   assert(data1 == 1);
+ *   assert(data2 == 2);
+ * }
+ */
+
+#include "common.h"
+
+extern "C" {
+struct kmp_task_t_with_privates {
+  kmp_task_t task;
+};
+
+struct anon {
+  int32_t *data;
+};
+}
+
+template <int I>
+kmp_int32 omp_task_entry(kmp_int32 gtid, kmp_task_t_with_privates *task) {
+  auto shareds = reinterpret_cast<anon *>(task->task.shareds);
+  auto p = shareds->data;
+  *p = I;
+  return 0;
+}
+
+int main(int argc, char *argv[]) {
+  constexpr const int N = 1024;
+#pragma omp parallel for
+  for (int i = 0; i < N; ++i) {
+    int32_t gtid = __kmpc_global_thread_num(nullptr);
+    int32_t data1 = 0, data2 = 0;
+    __kmpc_taskgroup(nullptr, gtid);
+
+    auto task1 = __kmpc_omp_target_task_alloc(
+        nullptr, gtid, 1, sizeof(kmp_task_t_with_privates), sizeof(anon),
+        reinterpret_cast<kmp_routine_entry_t>(omp_task_entry<1>), -1);
+    auto shareds = reinterpret_cast<anon *>(task1->shareds);
+    shareds->data = &data1;
+    __kmpc_omp_task(nullptr, gtid, task1);
+
+    auto task2 = __kmpc_omp_target_task_alloc(
+        nullptr, gtid, 1, sizeof(kmp_task_t_with_privates), sizeof(anon),
+        reinterpret_cast<kmp_routine_entry_t>(omp_task_entry<2>), -1);
+    shareds = reinterpret_cast<anon *>(task2->shareds);
+    shareds->data = &data2;
+    __kmpc_omp_task(nullptr, gtid, task2);
+
+    __kmpc_end_taskgroup(nullptr, gtid);
+
+    assert(data1 == 1);
+    assert(data2 == 2);
+  }
+
+  std::cout << "PASS\n";
+  return 0;
+}
+
+// CHECK: PASS
diff --git a/openmp/runtime/test/worksharing/for/kmp_sch_simd_guided.c b/openmp/runtime/test/worksharing/for/kmp_sch_simd_guided.c
--- a/openmp/runtime/test/worksharing/for/kmp_sch_simd_guided.c
+++ b/openmp/runtime/test/worksharing/for/kmp_sch_simd_guided.c
@@ -5,6 +5,7 @@
   value 46 to the OpenMP RTL. Test uses numerous loop parameter combinations.
 */
 #include <stdio.h>
+#include <stdlib.h>
 #include <omp.h>
 
 #if defined(WIN32) || defined(_WIN32)
@@ -47,6 +48,9 @@
 // End of definitions copied from OpenMP RTL.
 // ---------------------------------------------------------------------------
 static id loc = {0, 2, 0, 0, ";file;func;0;0;;"};
+// This variable is defined in OpenMP RTL but we can't have it exposed so we
+// need to redefine it here.
+static int __kmp_hidden_helper_threads_num = 8;
 
 // ---------------------------------------------------------------------------
 int run_loop_64(i64 loop_lb, i64 loop_ub, i64 loop_st, int loop_chunk) {
@@ -58,6 +62,9 @@
   int rc;
   int tid = omp_get_thread_num();
   int gtid = tid;
+  if (gtid) {
+    gtid += __kmp_hidden_helper_threads_num;
+  }
   int last;
 #if DEBUG
   printf("run_loop_<%d>(lb=%d, ub=%d, st=%d, ch=%d)\n",
@@ -210,6 +217,9 @@
   int rc;
   int tid = omp_get_thread_num();
   int gtid = tid;
+  if (gtid) {
+    gtid += __kmp_hidden_helper_threads_num;
+  }
   int last;
 #if DEBUG
   printf("run_loop_<%d>(lb=%d, ub=%d, st=%d, ch=%d)\n",
@@ -397,6 +407,13 @@
 // ---------------------------------------------------------------------------
 int main()
 {
+  {
+    const char *env = getenv("LIBOMP_NUM_HIDDEN_HELPER_THREADS");
+    if (env) {
+      __kmp_hidden_helper_threads_num = atoi(env);
+    }
+  }
+
   int n, err = 0;
   for (n = 1; n <= 4; ++ n) {
     err += run_32(n);