diff --git a/Classification/include/CGAL/Classification/Feature_set.h b/Classification/include/CGAL/Classification/Feature_set.h
index b234f380fdb..e225649cf84 100644
--- a/Classification/include/CGAL/Classification/Feature_set.h
+++ b/Classification/include/CGAL/Classification/Feature_set.h
@@ -19,7 +19,7 @@
 #include <boost/make_shared.hpp>
 
 #ifdef CGAL_LINKED_WITH_TBB
-#include <tbb/mutex.h>
+#include <mutex>
 #include <tbb/task_group.h>
 #endif // CGAL_LINKED_WITH_TBB
 
diff --git a/Classification/include/CGAL/Classification/Local_eigen_analysis.h b/Classification/include/CGAL/Classification/Local_eigen_analysis.h
index b3849220d54..ad3b83c0eef 100644
--- a/Classification/include/CGAL/Classification/Local_eigen_analysis.h
+++ b/Classification/include/CGAL/Classification/Local_eigen_analysis.h
@@ -34,7 +34,8 @@
 #include <tbb/parallel_for.h>
 #include <tbb/blocked_range.h>
 #include <tbb/scalable_allocator.h>
-#include <tbb/mutex.h>
+#include <mutex>
+
 #endif // CGAL_LINKED_WITH_TBB
 
 namespace CGAL {
@@ -69,7 +70,7 @@ private:
     PointMap m_point_map;
     const NeighborQuery& m_neighbor_query;
     float& m_mean_range;
-    tbb::mutex& m_mutex;
+    std::mutex& m_mutex;
     
   public:
     
@@ -78,7 +79,7 @@ private:
                           PointMap point_map,
                           const NeighborQuery& neighbor_query,
                           float& mean_range,
-                          tbb::mutex& mutex)
+                          std::mutex& mutex)
       : m_eigen (eigen), m_input (input), m_point_map (point_map),
         m_neighbor_query (neighbor_query), m_mean_range (mean_range), m_mutex (mutex)
     { }
@@ -120,7 +121,7 @@ private:
     const FaceListGraph& m_input;
     const NeighborQuery& m_neighbor_query;
     float& m_mean_range;
-    tbb::mutex& m_mutex;
+    std::mutex& m_mutex;
     
   public:
     
@@ -128,7 +129,7 @@ private:
                                 const FaceListGraph& input,
                                 const NeighborQuery& neighbor_query,
                                 float& mean_range,
-                                tbb::mutex& mutex)
+                                std::mutex& mutex)
       : m_eigen (eigen), m_input (input),
         m_neighbor_query (neighbor_query), m_mean_range (mean_range), m_mutex (mutex)
     { }
@@ -271,7 +272,7 @@ public:
 #else
     if (boost::is_convertible<ConcurrencyTag,Parallel_tag>::value)
     {
-      tbb::mutex mutex;
+      std::mutex mutex;
       Compute_eigen_values<PointRange, PointMap, NeighborQuery, DiagonalizeTraits>
         f(out, input, point_map, neighbor_query, out.m_content->mean_range, mutex);
       tbb::parallel_for(tbb::blocked_range<size_t>(0, input.size ()), f);
@@ -361,7 +362,7 @@ public:
 #else
     if (boost::is_convertible<ConcurrencyTag,Parallel_tag>::value)
     {
-      tbb::mutex mutex;
+      std::mutex mutex;
       Compute_eigen_values_graph<FaceListGraph, NeighborQuery, DiagonalizeTraits>
         f(out, input, neighbor_query, out.m_content->mean_range, mutex);
 
diff --git a/Classification/include/CGAL/Classification/Mesh_feature_generator.h b/Classification/include/CGAL/Classification/Mesh_feature_generator.h
index 5b90b74b379..d9baf699327 100644
--- a/Classification/include/CGAL/Classification/Mesh_feature_generator.h
+++ b/Classification/include/CGAL/Classification/Mesh_feature_generator.h
@@ -45,7 +45,7 @@
 
 #ifdef CGAL_LINKED_WITH_TBB
 #include <tbb/task_group.h>
-#include <tbb/mutex.h>
+#include <mutex>
 #endif // CGAL_LINKED_WITH_TBB
 
 namespace CGAL {
diff --git a/Classification/include/CGAL/Classification/Sum_of_weighted_features_classifier.h b/Classification/include/CGAL/Classification/Sum_of_weighted_features_classifier.h
index 5b7dc929185..4f17b7da445 100644
--- a/Classification/include/CGAL/Classification/Sum_of_weighted_features_classifier.h
+++ b/Classification/include/CGAL/Classification/Sum_of_weighted_features_classifier.h
@@ -33,7 +33,7 @@
 #include <tbb/parallel_for.h>
 #include <tbb/blocked_range.h>
 #include <tbb/scalable_allocator.h>
-#include <tbb/mutex.h>
+#include <mutex>
 #endif // CGAL_LINKED_WITH_TBB
 
 #define CLASSIFICATION_TRAINING_QUICK_ESTIMATION
@@ -72,9 +72,9 @@ private:
     std::vector<std::size_t>& m_true_positives;
     std::vector<std::size_t>& m_false_positives;
     std::vector<std::size_t>& m_false_negatives;
-    std::vector<tbb::mutex>& m_tp_mutex;
-    std::vector<tbb::mutex>& m_fp_mutex;
-    std::vector<tbb::mutex>& m_fn_mutex;
+    std::vector<std::mutex>& m_tp_mutex;
+    std::vector<std::mutex>& m_fp_mutex;
+    std::vector<std::mutex>& m_fn_mutex;
 
     
   public:
@@ -85,9 +85,9 @@ private:
                  std::vector<std::size_t>& true_positives,
                  std::vector<std::size_t>& false_positives,
                  std::vector<std::size_t>& false_negatives,
-                 std::vector<tbb::mutex>& tp_mutex,
-                 std::vector<tbb::mutex>& fp_mutex,
-                 std::vector<tbb::mutex>& fn_mutex)
+                 std::vector<std::mutex>& tp_mutex,
+                 std::vector<std::mutex>& fp_mutex,
+                 std::vector<std::mutex>& fn_mutex)
       : m_training_set (training_set)
       , m_classifier (classifier)
       , m_label (label)
@@ -901,9 +901,9 @@ private:
 #else
       if (boost::is_convertible<ConcurrencyTag,Parallel_tag>::value)
       {
-        std::vector<tbb::mutex> tp_mutex (m_labels.size());
-        std::vector<tbb::mutex> fp_mutex (m_labels.size());
-        std::vector<tbb::mutex> fn_mutex (m_labels.size());
+        std::vector<std::mutex> tp_mutex (m_labels.size());
+        std::vector<std::mutex> fp_mutex (m_labels.size());
+        std::vector<std::mutex> fn_mutex (m_labels.size());
         Compute_iou f(training_sets[j], *this, j,
                       true_positives, false_positives, false_negatives,
                       tp_mutex, fp_mutex, fn_mutex);
diff --git a/Classification/include/CGAL/Classification/classify.h b/Classification/include/CGAL/Classification/classify.h
index 6fb787a43df..28ec87f5fdb 100644
--- a/Classification/include/CGAL/Classification/classify.h
+++ b/Classification/include/CGAL/Classification/classify.h
@@ -24,7 +24,7 @@
 #include <tbb/parallel_for.h>
 #include <tbb/blocked_range.h>
 #include <tbb/scalable_allocator.h>
-#include <tbb/mutex.h>
+#include <mutex>
 #endif // CGAL_LINKED_WITH_TBB
 
 namespace CGAL {
diff --git a/Mesh_3/include/CGAL/Compact_mesh_cell_base_3.h b/Mesh_3/include/CGAL/Compact_mesh_cell_base_3.h
index 077de31abcf..af4041ed3df 100644
--- a/Mesh_3/include/CGAL/Compact_mesh_cell_base_3.h
+++ b/Mesh_3/include/CGAL/Compact_mesh_cell_base_3.h
@@ -34,7 +34,7 @@
 
 
 #ifdef CGAL_LINKED_WITH_TBB
-# include <tbb/atomic.h>
+# include <atomic>
 #endif
 
 namespace CGAL {
@@ -152,7 +152,8 @@ public:
   {
     CGAL_precondition(facet>=0 && facet<4);
     char current_bits = bits_;
-    while (bits_.compare_and_swap(current_bits | char(1 << facet), current_bits) != current_bits)
+   
+    while (!bits_.compare_exchange_weak(current_bits, current_bits | char(1 << facet)))
     {
       current_bits = bits_;
     }
@@ -165,7 +166,7 @@ public:
     char current_bits = bits_;
     char mask = char(15 & ~(1 << facet));
     char wanted_value = current_bits & mask;
-    while (bits_.compare_and_swap(wanted_value, current_bits) != current_bits)
+    while (!bits_.compare_exchange_weak(current_bits, wanted_value))
     {
       current_bits = bits_;
     }
@@ -182,18 +183,19 @@ public:
   /// this function "deletes" cc
   void try_to_set_circumcenter(Point_3 *cc) const
   {
-    if (weighted_circumcenter_.compare_and_swap(cc, nullptr) != nullptr)
+    Point_3* base_test = nullptr;
+    if (!weighted_circumcenter_.compare_exchange_strong(base_test, cc))
       delete cc;
   }
 
 private:
-  typedef tbb::atomic<unsigned int> Erase_counter_type;
+  typedef std::atomic<unsigned int> Erase_counter_type;
   Erase_counter_type                m_erase_counter;
   /// Stores visited facets (4 first bits)
-  tbb::atomic<char> bits_;
+  std::atomic<char> bits_;
 
 protected:
-  mutable tbb::atomic<Point_3*> weighted_circumcenter_;
+  mutable std::atomic<Point_3*> weighted_circumcenter_;
 };
 
 #endif // CGAL_LINKED_WITH_TBB
diff --git a/Mesh_3/include/CGAL/Mesh_3/C3T3_helpers.h b/Mesh_3/include/CGAL/Mesh_3/C3T3_helpers.h
index 3f4421ed3e9..b899ad8eeff 100644
--- a/Mesh_3/include/CGAL/Mesh_3/C3T3_helpers.h
+++ b/Mesh_3/include/CGAL/Mesh_3/C3T3_helpers.h
@@ -47,7 +47,7 @@
 
 #ifdef CGAL_LINKED_WITH_TBB
 # include <tbb/parallel_do.h>
-# include <tbb/mutex.h>
+# include <mutex>
 #endif
 
 #include <functional>
@@ -595,7 +595,7 @@ public:
 protected:
   Lock_data_structure *m_lock_ds;
 
-  typedef tbb::mutex  Mutex_type;
+  typedef std::mutex  Mutex_type;
   mutable Mutex_type  m_mut_outdated_cells;
   mutable Mutex_type  m_mut_moving_vertices;
   mutable Mutex_type  m_mut_vertex_to_proj;
diff --git a/Mesh_3/include/CGAL/Mesh_3/Mesh_complex_3_in_triangulation_3_base.h b/Mesh_3/include/CGAL/Mesh_3/Mesh_complex_3_in_triangulation_3_base.h
index e621c87d4b1..fc1e3e0763d 100644
--- a/Mesh_3/include/CGAL/Mesh_3/Mesh_complex_3_in_triangulation_3_base.h
+++ b/Mesh_3/include/CGAL/Mesh_3/Mesh_complex_3_in_triangulation_3_base.h
@@ -41,7 +41,7 @@
 
 
 #ifdef CGAL_LINKED_WITH_TBB
-  #include <tbb/atomic.h>
+  #include <atomic>
   #include <tbb/concurrent_hash_map.h>
 
 namespace CGAL {
@@ -174,7 +174,7 @@ public:
     , manifold_info_initialized_(false) //TODO: parallel!
   {
     // We don't put it in the initialization list because
-    // tbb::atomic has no contructors
+    // std::atomic has no contructors
     number_of_facets_ = 0;
     number_of_cells_ = 0;
   }
@@ -186,8 +186,9 @@ public:
     , edge_facet_counter_(rhs.edge_facet_counter_)
     , manifold_info_initialized_(rhs.manifold_info_initialized_)
   {
-    number_of_facets_ = rhs.number_of_facets_;
-    number_of_cells_ = rhs.number_of_cells_;
+    Init_number_of_elements<Concurrency_tag> init;
+    init(number_of_facets_, rhs.number_of_facets_);
+    init(number_of_cells_, rhs.number_of_cells_);
   }
 
   /// Destructor
@@ -505,9 +506,10 @@ public:
   /// Swaps this & rhs
   void swap(Self& rhs)
   {
-    std::swap(rhs.number_of_facets_, number_of_facets_);
+    Swap_elements<Concurrency_tag> swapper;
+    swapper(rhs.number_of_facets_, number_of_facets_);
     tr_.swap(rhs.tr_);
-    std::swap(rhs.number_of_cells_, number_of_cells_);
+    swapper(rhs.number_of_cells_, number_of_cells_);
   }
 
   /// Returns bbox
@@ -850,12 +852,55 @@ private:
   {
     typedef size_type type;
   };
+  
+  template<typename Concurrency_tag2, typename dummy = void>
+  struct Init_number_of_elements
+  {
+    template<typename T>
+    void operator()(T& a, const T& b)
+    {
+      a = b;
+    }
+  };
+  
+  template<typename Concurrency_tag2, typename dummy = void>
+  struct Swap_elements
+  {
+    template<typename T>
+    void operator()(T& a, T& b)
+    {
+      std::swap(a, b);
+    }
+  };
 #ifdef CGAL_LINKED_WITH_TBB
   // Parallel: atomic
   template<typename dummy>
   struct Number_of_elements<Parallel_tag, dummy>
   {
-    typedef tbb::atomic<size_type> type;
+    typedef std::atomic<size_type> type;
+  };
+  
+  template<typename dummy>
+  struct Init_number_of_elements<Parallel_tag, dummy>
+  {
+    template<typename T>
+    void operator()(T& a, const T& b)
+    {
+      a = b.load();
+    }
+  };
+  
+  template<typename dummy>
+  struct Swap_elements<Parallel_tag, dummy>
+  {
+    template<typename T>
+    void operator()(T& a, T& b)
+    {
+      T tmp;
+      tmp.exchange(a);
+      a.exchange(b);
+      b.exchange(tmp);
+    }
   };
 #endif // CGAL_LINKED_WITH_TBB
 
diff --git a/Mesh_3/include/CGAL/Mesh_3/Mesh_global_optimizer.h b/Mesh_3/include/CGAL/Mesh_3/Mesh_global_optimizer.h
index 0e4145b26f7..a058ff1dd90 100644
--- a/Mesh_3/include/CGAL/Mesh_3/Mesh_global_optimizer.h
+++ b/Mesh_3/include/CGAL/Mesh_3/Mesh_global_optimizer.h
@@ -47,7 +47,8 @@
 #include <boost/type_traits/is_convertible.hpp>
 
 #ifdef CGAL_LINKED_WITH_TBB
-# include <tbb/atomic.h>
+# include <atomic>
+# include <mutex>
 # include <tbb/parallel_do.h>
 # include <tbb/concurrent_vector.h>
 #endif
@@ -122,7 +123,7 @@ protected:
   typedef typename Tr::Lock_data_structure                  Lock_data_structure;
   typedef tbb::concurrent_vector<std::tuple<typename Tr::Vertex_handle, Vector_3, FT> >
                                                             Moves_vector;
-  typedef tbb::atomic<unsigned int>                         Nb_frozen_points_type ;
+  typedef std::atomic<unsigned int>                         Nb_frozen_points_type ;
 
   Mesh_global_optimizer_base(const Bbox_3 &bbox, int num_grid_cells_per_axis)
     : big_moves_size_(0)
@@ -136,7 +137,7 @@ protected:
   {
     if (++big_moves_current_size_ <= big_moves_size_ )
     {
-      tbb::mutex::scoped_lock lock(m_big_moves_mutex);
+      std::lock_guard<std::mutex> lock(m_big_moves_mutex);
       typename std::multiset<FT>::const_iterator it = big_moves_.insert(new_sq_move);
 
       // New smallest move of all big moves?
@@ -149,7 +150,7 @@ protected:
 
       if( new_sq_move > big_moves_smallest_ )
       {
-        tbb::mutex::scoped_lock lock(m_big_moves_mutex);
+        std::lock_guard<std::mutex> lock(m_big_moves_mutex);
         // Test it again since it may have been modified by another
         // thread in the meantime
         if( new_sq_move > big_moves_smallest_ )
@@ -185,11 +186,11 @@ protected:
 public:
 
 protected:
-  tbb::atomic<std::size_t>  big_moves_current_size_;
-  tbb::atomic<FT>           big_moves_smallest_;
+  std::atomic<std::size_t>  big_moves_current_size_;
+  std::atomic<FT>           big_moves_smallest_;
   std::size_t               big_moves_size_;
   std::multiset<FT>         big_moves_;
-  tbb::mutex                m_big_moves_mutex;
+  std::mutex                m_big_moves_mutex;
 
   /// Lock data structure
   Lock_data_structure m_lock_ds;
diff --git a/Mesh_3/include/CGAL/Mesh_3/Mesh_surface_cell_base_3.h b/Mesh_3/include/CGAL/Mesh_3/Mesh_surface_cell_base_3.h
index e95510e101c..40b492c1467 100644
--- a/Mesh_3/include/CGAL/Mesh_3/Mesh_surface_cell_base_3.h
+++ b/Mesh_3/include/CGAL/Mesh_3/Mesh_surface_cell_base_3.h
@@ -27,7 +27,7 @@
 #include <CGAL/Mesh_3/io_signature.h>
 
 #ifdef CGAL_LINKED_WITH_TBB
-# include <tbb/atomic.h>
+# include <atomic>
 #endif
 
 #ifdef _MSC_VER
@@ -123,7 +123,7 @@ public:
 
 protected:
   /// Stores visited facets (4 first bits)
-  tbb::atomic<char> bits_;
+  std::atomic<char> bits_;
 };
 #endif // CGAL_LINKED_WITH_TBB
 
diff --git a/Mesh_3/include/CGAL/Mesh_3/Mesher_3.h b/Mesh_3/include/CGAL/Mesh_3/Mesher_3.h
index e58a21ff397..479a1a51023 100644
--- a/Mesh_3/include/CGAL/Mesh_3/Mesher_3.h
+++ b/Mesh_3/include/CGAL/Mesh_3/Mesher_3.h
@@ -50,7 +50,7 @@
 #endif
 
 #ifdef CGAL_LINKED_WITH_TBB
-#  include <tbb/task_scheduler_init.h>
+#  include <thread>
 #endif
 
 #include <boost/format.hpp>
@@ -659,7 +659,7 @@ initialize()
 #  endif
       Random_points_on_sphere_3<Bare_point> random_point(radius);
       const int NUM_PSEUDO_INFINITE_VERTICES = static_cast<int>(
-        float(tbb::task_scheduler_init::default_num_threads())
+        float(std::thread::hardware_concurrency())
         * Concurrent_mesher_config::get().num_pseudo_infinite_vertices_per_core);
       for (int i = 0 ; i < NUM_PSEUDO_INFINITE_VERTICES ; ++i, ++random_point)
         r_c3t3_.add_far_point(r_c3t3_.triangulation().geom_traits().construct_weighted_point_3_object()
diff --git a/Mesh_3/include/CGAL/Mesh_3/Worksharing_data_structures.h b/Mesh_3/include/CGAL/Mesh_3/Worksharing_data_structures.h
index 8abc76f5a03..32b5816c83c 100644
--- a/Mesh_3/include/CGAL/Mesh_3/Worksharing_data_structures.h
+++ b/Mesh_3/include/CGAL/Mesh_3/Worksharing_data_structures.h
@@ -28,7 +28,7 @@
 #include <tbb/concurrent_vector.h>
 #include <tbb/scalable_allocator.h>
 
-#include <tbb/atomic.h>
+#include <atomic>
 
 #include <vector>
 
@@ -65,8 +65,8 @@ public:
 
     m_num_cells =
       num_grid_cells_per_axis*num_grid_cells_per_axis*num_grid_cells_per_axis;
-    m_occupation_grid = new tbb::atomic<int>[m_num_cells];
-    m_num_batches_grid = new tbb::atomic<int>[m_num_cells];
+    m_occupation_grid = new std::atomic<int>[m_num_cells];
+    m_num_batches_grid = new std::atomic<int>[m_num_cells];
     // Initialize grid
     for (int i = 0 ; i < m_num_cells ; ++i)
     {
@@ -106,15 +106,15 @@ public:
 
   void add_batch(int cell_index, int to_add)
   {
-    m_num_batches_grid[cell_index].fetch_and_add(to_add);
+    m_num_batches_grid[cell_index].fetch_add(to_add);
   }
 
   void add_occupation(int cell_index, int to_add, int)
   {
-    m_occupation_grid[cell_index].fetch_and_add(to_add);
+    m_occupation_grid[cell_index].fetch_add(to_add);
 
     /*int new_occupation =
-      (m_occupation_grid[cell_index].fetch_and_add(to_add))
+      (m_occupation_grid[cell_index].fetch_add(to_add))
       + to_add;
     //m_num_batches_grid[cell_index] = num_items_in_work_queue;
 
@@ -221,7 +221,7 @@ public:
 
 
     // Rotate
-    static tbb::atomic<int> last_cell_index;
+    static std::atomic<int> last_cell_index;
     //std::cerr << "last=" << last_cell_index << std::endl;
     int i = (last_cell_index + 1) % m_num_cells;
     for ( ; i != last_cell_index ; i = (i + 1) % m_num_cells)
@@ -246,11 +246,11 @@ protected:
 
   int                                             m_num_grid_cells_per_axis;
   int                                             m_num_cells;
-  tbb::atomic<int> *                              m_occupation_grid;
-  tbb::atomic<int> *                              m_num_batches_grid;
+  std::atomic<int> *                              m_occupation_grid;
+  std::atomic<int> *                              m_num_batches_grid;
 
-  tbb::atomic<int>                                m_laziest_cell_index;
-  tbb::atomic<int>                                m_laziest_cell_occupation;
+  std::atomic<int>                                m_laziest_cell_index;
+  std::atomic<int>                                m_laziest_cell_occupation;
 };
 
 
@@ -504,7 +504,7 @@ public:
   {
     m_tls_work_buffers = new TLS_WorkBuffer[m_num_cells];
     m_work_batches = new tbb::concurrent_queue<WorkBatch>[m_num_cells];
-    m_num_batches = new tbb::atomic<int>[m_num_cells];
+    m_num_batches = new std::atomic<int>[m_num_cells];
 
     for (int i = 0 ; i < m_num_cells ; ++i)
       m_num_batches[i] = 0;
@@ -673,7 +673,7 @@ protected:
   Work_statistics                   m_stats;
   TLS_WorkBuffer                   *m_tls_work_buffers;
   tbb::concurrent_queue<WorkBatch> *m_work_batches;
-  tbb::atomic<int>                 *m_num_batches;
+  std::atomic<int>                 *m_num_batches;
 };
 
 
@@ -822,7 +822,7 @@ protected:
   }
 
   const size_t                      NUM_WORK_ITEMS_PER_BATCH;
-  tbb::atomic<int>                  m_cache_number_of_tasks;
+  std::atomic<int>                  m_cache_number_of_tasks;
   TLS_WorkBuffer                    m_tls_work_buffers;
 };
 
diff --git a/Mesh_3/include/CGAL/Mesh_3/vertex_perturbation.h b/Mesh_3/include/CGAL/Mesh_3/vertex_perturbation.h
index 82bafa3f43b..603fd7626b1 100644
--- a/Mesh_3/include/CGAL/Mesh_3/vertex_perturbation.h
+++ b/Mesh_3/include/CGAL/Mesh_3/vertex_perturbation.h
@@ -29,7 +29,7 @@
   #include <CGAL/Timer.h>
   #ifdef CGAL_LINKED_WITH_TBB
     #include <tbb/enumerable_thread_specific.h>
-    #include <tbb/atomic.h>
+    #include <atomic>
   #endif
 #endif
 
@@ -192,7 +192,7 @@ public:
 #endif
   {
 #ifdef CGAL_MESH_3_PERTURBER_VERBOSE
-    // Initialized here in case it's some tbb::atomic
+    // Initialized here in case it's some std::atomic
     total_counter_ = 0;
     total_time_ = 0;
 #endif
@@ -353,8 +353,8 @@ private:
   mutable int counter_;
 #ifdef CGAL_LINKED_WITH_TBB
   mutable tbb::enumerable_thread_specific<CGAL::Timer> timer_;
-  tbb::atomic<int> total_counter_;
-  tbb::atomic<std::size_t> total_time_;
+  std::atomic<int> total_counter_;
+  std::atomic<std::size_t> total_time_;
 #else
   mutable CGAL::Timer timer_;
   int total_counter_;
diff --git a/Mesh_3/include/CGAL/Mesh_cell_base_3.h b/Mesh_3/include/CGAL/Mesh_cell_base_3.h
index 75e85ea869d..379bd4c6c54 100644
--- a/Mesh_3/include/CGAL/Mesh_cell_base_3.h
+++ b/Mesh_3/include/CGAL/Mesh_cell_base_3.h
@@ -33,7 +33,7 @@
 #include <boost/type_traits/is_convertible.hpp>
 
 #ifdef CGAL_LINKED_WITH_TBB
-# include <tbb/atomic.h>
+# include <atomic>
 #endif
 
 namespace CGAL {
@@ -87,7 +87,7 @@ public:
   }
   
 protected:
-  typedef tbb::atomic<unsigned int> Erase_counter_type;
+  typedef std::atomic<unsigned int> Erase_counter_type;
   Erase_counter_type                m_erase_counter;
 };
 #endif // CGAL_LINKED_WITH_TBB
diff --git a/Mesh_3/test/Mesh_3/CMakeLists.txt b/Mesh_3/test/Mesh_3/CMakeLists.txt
index 03f021a7172..90d9e5e949e 100644
--- a/Mesh_3/test/Mesh_3/CMakeLists.txt
+++ b/Mesh_3/test/Mesh_3/CMakeLists.txt
@@ -73,6 +73,7 @@ if ( CGAL_FOUND )
       test_backward_compatibility
       test_meshing_polyhedron
       test_meshing_polyhedral_complex
+      
       )
     if(TBB_FOUND AND TARGET ${target})
       CGAL_target_use_TBB(${target})
diff --git a/Mesh_3/test/Mesh_3/test_meshing_3D_gray_image.cpp b/Mesh_3/test/Mesh_3/test_meshing_3D_gray_image.cpp
index 0055e50b125..4b232d42d20 100644
--- a/Mesh_3/test/Mesh_3/test_meshing_3D_gray_image.cpp
+++ b/Mesh_3/test/Mesh_3/test_meshing_3D_gray_image.cpp
@@ -13,7 +13,6 @@
 //******************************************************************************
 // File Description :
 //***************************************************************************
-
 #include "test_meshing_utilities.h"
 #include <CGAL/Image_3.h>
 #include <CGAL/Labeled_mesh_domain_3.h>
diff --git a/Mesh_3/test/Mesh_3/test_meshing_determinism.cpp b/Mesh_3/test/Mesh_3/test_meshing_determinism.cpp
index e71c56853bd..e184d689e99 100644
--- a/Mesh_3/test/Mesh_3/test_meshing_determinism.cpp
+++ b/Mesh_3/test/Mesh_3/test_meshing_determinism.cpp
@@ -13,9 +13,9 @@
 #include <fstream>
 #include <sstream>
 #include <cstring>
-
 #ifdef CGAL_LINKED_WITH_TBB
-#  include <tbb/task_scheduler_init.h>
+#define TBB_PREVIEW_GLOBAL_CONTROL 1
+#  include <tbb/global_control.h>
 #endif
 
 // To avoid verbose function and named parameters call
@@ -119,7 +119,7 @@ int main(int, char*[])
 {
   test<CGAL::Sequential_tag>();
 #ifdef CGAL_LINKED_WITH_TBB
-  tbb::task_scheduler_init init(1);
+  tbb::global_control c(tbb::global_control::max_allowed_parallelism, 1);
   test<CGAL::Parallel_tag>();
 #endif
 }
diff --git a/Point_set_processing_3/include/CGAL/bilateral_smooth_point_set.h b/Point_set_processing_3/include/CGAL/bilateral_smooth_point_set.h
index e80d6a9be52..d69d85d9bf4 100644
--- a/Point_set_processing_3/include/CGAL/bilateral_smooth_point_set.h
+++ b/Point_set_processing_3/include/CGAL/bilateral_smooth_point_set.h
@@ -38,11 +38,12 @@
 #include <CGAL/property_map.h>
 
 #ifdef CGAL_LINKED_WITH_TBB
+
 #include <CGAL/Point_set_processing_3/internal/Parallel_callback.h>
 #include <tbb/parallel_for.h>
 #include <tbb/blocked_range.h>
 #include <tbb/scalable_allocator.h>  
-#include <tbb/atomic.h>
+#include <atomic>
 #endif // CGAL_LINKED_WITH_TBB
 
 // Default allocator: use TBB allocators if available
diff --git a/Polygon_mesh_processing/include/CGAL/Polygon_mesh_processing/distance.h b/Polygon_mesh_processing/include/CGAL/Polygon_mesh_processing/distance.h
index de52cfac3f2..e6681cd3d9d 100644
--- a/Polygon_mesh_processing/include/CGAL/Polygon_mesh_processing/distance.h
+++ b/Polygon_mesh_processing/include/CGAL/Polygon_mesh_processing/distance.h
@@ -34,7 +34,7 @@
 #ifdef CGAL_LINKED_WITH_TBB
 #include <tbb/parallel_for.h>
 #include <tbb/blocked_range.h>
-#include <tbb/atomic.h>
+#include <atomic>
 #endif // CGAL_LINKED_WITH_TBB
 
 #include <boost/unordered_set.hpp>
@@ -78,13 +78,13 @@ struct Distance_computation{
   const AABB_tree& tree;
   const PointRange& sample_points;
   Point_3 initial_hint;
-  tbb::atomic<double>* distance;
+  std::atomic<double>* distance;
 
   Distance_computation(
           const AABB_tree& tree,
           const Point_3& p,
           const PointRange& sample_points,
-          tbb::atomic<double>* d)
+          std::atomic<double>* d)
     : tree(tree)
     , sample_points(sample_points)
     , initial_hint(p)
@@ -108,7 +108,8 @@ struct Distance_computation{
     double current_value = *distance;
     while( current_value < hdist )
     {
-      current_value = distance->compare_and_swap(hdist, current_value);
+      if(distance->compare_exchange_weak(current_value, hdist))
+        current_value = hdist;
     }
   }
 };
@@ -129,7 +130,7 @@ double approximate_Hausdorff_distance_impl(
 #else
   if (boost::is_convertible<Concurrency_tag,Parallel_tag>::value)
   {
-    tbb::atomic<double> distance;
+    std::atomic<double> distance;
     distance=0;
     Distance_computation<AABBTree, PointRange> f(tree, hint, sample_points, &distance);
     tbb::parallel_for(tbb::blocked_range<std::size_t>(0, sample_points.size()), f);
diff --git a/Polyhedron/demo/Polyhedron/Plugins/Point_set/Point_set_selection_plugin.cpp b/Polyhedron/demo/Polyhedron/Plugins/Point_set/Point_set_selection_plugin.cpp
index ebff326b116..ff0ef0d3f45 100644
--- a/Polyhedron/demo/Polyhedron/Plugins/Point_set/Point_set_selection_plugin.cpp
+++ b/Polyhedron/demo/Polyhedron/Plugins/Point_set/Point_set_selection_plugin.cpp
@@ -42,7 +42,7 @@
 //#undef  CGAL_LINKED_WITH_TBB
 #ifdef CGAL_LINKED_WITH_TBB
 #include <tbb/parallel_for.h>
-#include <tbb/mutex.h>
+#include <mutex>
 #include <tbb/blocked_range.h>
 #include <tbb/scalable_allocator.h>  
 #endif // CGAL_LINKED_WITH_TBB
diff --git a/Polyhedron/demo/Polyhedron/Plugins/Point_set/Point_set_to_mesh_distance_plugin.cpp b/Polyhedron/demo/Polyhedron/Plugins/Point_set/Point_set_to_mesh_distance_plugin.cpp
index 797a5ba644a..cc05ec26642 100644
--- a/Polyhedron/demo/Polyhedron/Plugins/Point_set/Point_set_to_mesh_distance_plugin.cpp
+++ b/Polyhedron/demo/Polyhedron/Plugins/Point_set/Point_set_to_mesh_distance_plugin.cpp
@@ -22,7 +22,7 @@
 #ifdef CGAL_LINKED_WITH_TBB
 #include <tbb/parallel_for.h>
 #include <tbb/blocked_range.h>
-#include <tbb/atomic.h>
+#include <atomic>
 #endif // CGAL_LINKED_WITH_TBB
 
 #if defined(CGAL_LINKED_WITH_TBB)
@@ -31,13 +31,13 @@ struct Distance_computation{
   const AABB_tree& tree;
   const Point_set & point_set;
   Point_3 initial_hint;
-  tbb::atomic<double>* distance;
+  std::atomic<double>* distance;
   std::vector<double>& output;
 
   Distance_computation(const AABB_tree& tree,
                        const Point_3 p,
                        const Point_set & point_set,
-                       tbb::atomic<double>* d,
+                       std::atomic<double>* d,
                        std::vector<double>& out )
     : tree(tree)
     , point_set(point_set)
@@ -96,7 +96,7 @@ double compute_distances(const Mesh& m,
   }
     return hdist;
 #else
-  tbb::atomic<double> distance;
+  std::atomic<double> distance;
   distance.store(0);
   Distance_computation<Tree, typename Traits::Point_3> f(tree, hint, point_set, &distance, out);
   tbb::parallel_for(tbb::blocked_range<Point_set::const_iterator>(point_set.begin(), point_set.end()), f);
diff --git a/STL_Extension/include/CGAL/Spatial_lock_grid_3.h b/STL_Extension/include/CGAL/Spatial_lock_grid_3.h
index b5bc64efb4f..467a4abac4e 100644
--- a/STL_Extension/include/CGAL/Spatial_lock_grid_3.h
+++ b/STL_Extension/include/CGAL/Spatial_lock_grid_3.h
@@ -18,7 +18,7 @@
 
 #include <boost/bind.hpp>
 
-#include <tbb/atomic.h>
+#include <atomic>
 #if TBB_IMPLEMENT_CPP0X
 # include <tbb/compat/thread>
 #else
@@ -452,12 +452,12 @@ class Spatial_lock_grid_3<Tag_non_blocking>
 public:
   // Constructors
   Spatial_lock_grid_3(const Bbox_3 &bbox, int num_grid_cells_per_axis)
-  : Base(bbox, num_grid_cells_per_axis)
+  : Base(bbox, num_grid_cells_per_axis),
+    m_grid(num_grid_cells_per_axis*num_grid_cells_per_axis*num_grid_cells_per_axis)
   {
     int num_cells =
       num_grid_cells_per_axis*num_grid_cells_per_axis*num_grid_cells_per_axis;
 
-    m_grid.resize(num_cells);
     // Initialize grid (useless?)
     for (int i = 0 ; i < num_cells ; ++i)
       m_grid[i] = false;
@@ -475,7 +475,8 @@ public:
   template <bool no_spin>
   bool try_lock_cell_impl(int cell_index)
   {
-    bool old_value = m_grid[cell_index].compare_and_swap(true, false);
+    bool v1 = true, v2 = false;
+    bool old_value = m_grid[cell_index].compare_exchange_strong(v1,v2);
     if (old_value == false)
     {
       get_thread_local_grid()[cell_index] = true;
@@ -492,7 +493,7 @@ public:
 
 protected:
 
-  std::vector<tbb::atomic<bool> > m_grid;
+  std::vector<std::atomic<bool> > m_grid;
 };
 
 
@@ -512,14 +513,14 @@ public:
 
   Spatial_lock_grid_3(const Bbox_3 &bbox, int num_grid_cells_per_axis)
   : Base(bbox, num_grid_cells_per_axis),
-    m_tls_thread_priorities(init_TLS_thread_priorities)
+    m_tls_thread_priorities(init_TLS_thread_priorities),
+    m_grid(num_grid_cells_per_axis*num_grid_cells_per_axis*num_grid_cells_per_axis)
   {
     int num_cells =
       num_grid_cells_per_axis*num_grid_cells_per_axis*num_grid_cells_per_axis;
-    m_grid.resize(num_cells);
     // Explicitly initialize the atomics
-    std::vector<tbb::atomic<unsigned int> >::iterator it     = m_grid.begin();
-    std::vector<tbb::atomic<unsigned int> >::iterator it_end = m_grid.end();
+    std::vector<std::atomic<unsigned int> >::iterator it     = m_grid.begin();
+    std::vector<std::atomic<unsigned int> >::iterator it_end = m_grid.end();
     for ( ; it != it_end ; ++it)
       *it = 0;
   }
@@ -542,9 +543,8 @@ public:
     // NO SPIN
     if (no_spin)
     {
-      unsigned int old_value =
-        m_grid[cell_index].compare_and_swap(this_thread_priority, 0);
-      if (old_value == 0)
+      unsigned int old_value = 0;
+      if(!m_grid[cell_index].compare_exchange_strong(old_value, this_thread_priority))
       {
         get_thread_local_grid()[cell_index] = true;
         m_tls_locked_cells.local().push_back(cell_index);
@@ -556,9 +556,8 @@ public:
     {
       for(;;)
       {
-        unsigned int old_value =
-          m_grid[cell_index].compare_and_swap(this_thread_priority, 0);
-        if (old_value == 0)
+        unsigned int old_value =0;
+        if(m_grid[cell_index].compare_exchange_weak(old_value, this_thread_priority))
         {
           get_thread_local_grid()[cell_index] = true;
           m_tls_locked_cells.local().push_back(cell_index);
@@ -587,7 +586,7 @@ public:
 private:
   static unsigned int init_TLS_thread_priorities()
   {
-    static tbb::atomic<unsigned int> last_id;
+    static std::atomic<unsigned int> last_id;
     unsigned int id = ++last_id;
     // Ensure it is > 0
     return (1 + id%((std::numeric_limits<unsigned int>::max)()));
@@ -595,7 +594,7 @@ private:
 
 protected:
 
-  std::vector<tbb::atomic<unsigned int> >               m_grid;
+  std::vector<std::atomic<unsigned int> >               m_grid;
 
   typedef tbb::enumerable_thread_specific<unsigned int> TLS_thread_uint_ids;
   TLS_thread_uint_ids                                   m_tls_thread_priorities;
diff --git a/STL_Extension/include/CGAL/thread.h b/STL_Extension/include/CGAL/thread.h
index 148b1b23664..665cc72744c 100644
--- a/STL_Extension/include/CGAL/thread.h
+++ b/STL_Extension/include/CGAL/thread.h
@@ -34,7 +34,7 @@
 #  include <tbb/tbb_config.h>
 #  if TBB_IMPLEMENT_CPP0X
 #    include <tbb/compat/thread>
-#    include <tbb/atomic.h>
+#    include <atomic>
 #    include <tbb/tick_count.h>
 #    define CGAL_USE_TBB_THREADS 1
 #  else
@@ -83,8 +83,8 @@ namespace cpp11 {
 
 #if defined(CGAL_NO_ATOMIC) && defined(CGAL_LINKED_WITH_TBB)
   // If <CGAL/atomic.h> did not defined CGAL::cpp11::atomic, then use
-  // tbb::atomic as a fallback.
-  using tbb::atomic;
+  // std::atomic as a fallback.
+  using std::atomic;
 #endif
 
 } // cpp11
diff --git a/STL_Extension/test/STL_Extension/test_Concurrent_compact_container.cpp b/STL_Extension/test/STL_Extension/test_Concurrent_compact_container.cpp
index 30e7f1106bb..8ec883d5f72 100644
--- a/STL_Extension/test/STL_Extension/test_Concurrent_compact_container.cpp
+++ b/STL_Extension/test/STL_Extension/test_Concurrent_compact_container.cpp
@@ -25,7 +25,7 @@ int main()
 
 # include <tbb/task_scheduler_init.h>
 # include <tbb/parallel_for.h>
-# include <tbb/atomic.h>
+# include <atomic>
 
 #include <CGAL/disable_warnings.h>
 
@@ -139,12 +139,12 @@ template <typename Values_vec, typename Cont>
 class Insert_and_erase_in_CCC_functor
 {
   typedef std::vector<typename Cont::iterator>  Iterators_vec;
-  typedef std::vector<tbb::atomic<bool> >       Free_elts_vec;
+  typedef std::vector<std::atomic<bool> >       Free_elts_vec;
 
 public:
   Insert_and_erase_in_CCC_functor(
     const Values_vec &values, Cont &cont, Iterators_vec &iterators,
-    Free_elts_vec &free_elements, tbb::atomic<unsigned int> &num_erasures)
+    Free_elts_vec &free_elements, std::atomic<unsigned int> &num_erasures)
   : m_values(values), m_cont(cont), m_iterators(iterators),
     m_free_elements(free_elements), m_num_erasures(num_erasures)
   {}
@@ -176,7 +176,7 @@ private:
   Cont                      & m_cont;
   Iterators_vec             & m_iterators;
   Free_elts_vec             & m_free_elements;
-  tbb::atomic<unsigned int> & m_num_erasures;
+  std::atomic<unsigned int> & m_num_erasures;
 };
 
 template < class Cont >
@@ -372,14 +372,14 @@ void test(const Cont &)
   {
   Cont c12;
   Vect v12(1000000);
-  std::vector<tbb::atomic<bool> > free_elements(v12.size());
-  for(typename std::vector<tbb::atomic<bool> >::iterator 
+  std::vector<std::atomic<bool> > free_elements(v12.size());
+  for(typename std::vector<std::atomic<bool> >::iterator 
     it = free_elements.begin(), end = free_elements.end(); it != end; ++it) 
   {
     *it = true;
   }
     
-  tbb::atomic<unsigned int> num_erasures; 
+  std::atomic<unsigned int> num_erasures; 
   num_erasures = 0;
   std::vector<typename Cont::iterator> iterators(v12.size());
   tbb::parallel_for(
diff --git a/Triangulation_3/include/CGAL/Delaunay_triangulation_3.h b/Triangulation_3/include/CGAL/Delaunay_triangulation_3.h
index 28d2b3f086f..70764f7bfa0 100644
--- a/Triangulation_3/include/CGAL/Delaunay_triangulation_3.h
+++ b/Triangulation_3/include/CGAL/Delaunay_triangulation_3.h
@@ -45,7 +45,7 @@
 #ifdef CGAL_LINKED_WITH_TBB
 # include <CGAL/point_generators_3.h>
 # include <tbb/parallel_for.h>
-# include <tbb/task_scheduler_init.h>
+# include <thread>
 # include <tbb/enumerable_thread_specific.h>
 # include <tbb/concurrent_vector.h>
 #endif
@@ -289,7 +289,7 @@ private:
                                          bbox.zmin() + 0.5*zdelta);
       Random_points_on_sphere_3<Point> random_point(radius);
       const int NUM_PSEUDO_INFINITE_VERTICES = static_cast<int>(
-                                                 tbb::task_scheduler_init::default_num_threads() * 3.5);
+                                                 std::thread::hardware_concurrency() * 3.5);
       std::vector<Point> points_on_far_sphere;
 
       points_on_far_sphere.reserve(NUM_PSEUDO_INFINITE_VERTICES);
diff --git a/Triangulation_3/include/CGAL/Regular_triangulation_3.h b/Triangulation_3/include/CGAL/Regular_triangulation_3.h
index a4867cf9946..17cabad626a 100644
--- a/Triangulation_3/include/CGAL/Regular_triangulation_3.h
+++ b/Triangulation_3/include/CGAL/Regular_triangulation_3.h
@@ -32,7 +32,7 @@
 #ifdef CGAL_LINKED_WITH_TBB
 # include <CGAL/point_generators_3.h>
 # include <tbb/parallel_for.h>
-# include <tbb/task_scheduler_init.h>
+# include <thread>
 # include <tbb/enumerable_thread_specific.h>
 # include <tbb/concurrent_vector.h>
 #endif
@@ -258,7 +258,7 @@ private:
                                          bbox.zmin() + 0.5*zdelta);
       Random_points_on_sphere_3<Bare_point> random_point(radius);
       const int NUM_PSEUDO_INFINITE_VERTICES = static_cast<int>(
-                                                 tbb::task_scheduler_init::default_num_threads() * 3.5);
+                                                 std::thread::hardware_concurrency() * 3.5);
       typename Gt::Construct_weighted_point_3 cwp =
           geom_traits().construct_weighted_point_3_object();