dataflow_cpp: preparations for deferred slice initialization

dataflow_cpp/include/embb/dataflow/internal/node.h

@@ -44,6 +44,7 @@ class Node {
   virtual bool HasOutputs() const { return false; }
   virtual void Run(int clock) = 0;
   virtual bool IsFullyConnected() = 0;
+  virtual bool IsSequential() { return true; }
   virtual bool Start(int /*clock*/) {
     EMBB_THROW(embb::base::ErrorException,
       "Nodes are started implicitly.");

dataflow_cpp/include/embb/dataflow/internal/process.h

@@ -98,6 +98,10 @@ class Process< Serial, Inputs<I1, I2, I3, I4, I5>,
     return inputs_.IsFullyConnected() && outputs_.IsFullyConnected();
   }
 
+  virtual bool IsSequential() {
+    return Serial;
+  }
+
   InputsType & GetInputs() {
     return inputs_;
   }

dataflow_cpp/include/embb/dataflow/internal/scheduler.h

@@ -40,6 +40,7 @@ class Scheduler {
   virtual void Spawn(Action & action) = 0;
   virtual void Enqueue(int process_id, Action & action) = 0;
   virtual void WaitForSlice(int slice) = 0;
+  virtual int GetSlices() = 0;
 };
 
 } // namespace internal

dataflow_cpp/include/embb/dataflow/internal/scheduler_mtapi.h

@@ -44,6 +44,13 @@ class SchedulerMTAPI : public Scheduler {
     : slices_(slices) {
     embb::tasks::Node & node = embb::tasks::Node::GetInstance();
 
+    int tl = std::min(
+      static_cast<int>(node.GetTaskLimit()),
+      static_cast<int>(node.GetGroupCount()));
+    if (tl < slices_) {
+      slices_ = tl;
+    }
+
     group_ = reinterpret_cast<embb::tasks::Group**>(
       embb::base::Allocation::Allocate(
       sizeof(embb::tasks::Group*)*slices_));
@@ -93,7 +100,7 @@ class SchedulerMTAPI : public Scheduler {
   virtual void WaitForSlice(int slice) {
     group_[slice]->WaitAll(MTAPI_INFINITE);
   }
-
+  virtual int GetSlices() { return slices_; }
  private:
   embb::tasks::Group ** group_;
   embb::tasks::Queue ** queue_;

dataflow_cpp/include/embb/dataflow/internal/scheduler_sequential.h

@@ -45,6 +45,7 @@ class SchedulerSequential : public Scheduler {
     action.RunSequential();
   }
   virtual void WaitForSlice(int /*slice*/) {}
+  virtual int GetSlices() { return 1; }
 };
 
 } // namespace internal

dataflow_cpp/include/embb/dataflow/network.h

@@ -687,6 +687,9 @@ class Network : public internal::ClockListener {
       slices_ = int(embb_core_count_available())*4;
     }
     sched_ = embb::base::Allocation::New<internal::SchedulerMTAPI>(slices_);
+    if (sched_->GetSlices() != slices_) {
+      slices_ = sched_->GetSlices();
+    }
     sink_counter_ = reinterpret_cast<embb::base::Atomic<int>*>(
       embb::base::Allocation::Allocate(
         sizeof(embb::base::Atomic<int>)*slices_));

tasks_cpp/include/embb/tasks/node.h

@@ -130,6 +130,15 @@ class Node {
   }
 
   /**
+   * Returns the number of available groups.
+   * \return The number of available groups
+   * \waitfree
+   */
+  mtapi_uint_t GetGroupCount() const {
+    return group_count_;
+  }
+
+  /**
     * Returns the number of available tasks.
     * \return The number of available tasks
     * \waitfree
@@ -238,6 +247,7 @@ class Node {
     mtapi_task_context_t * context);
 
   mtapi_uint_t queue_count_;
+  mtapi_uint_t group_count_;
   mtapi_uint_t task_limit_;
   mtapi_uint_t core_count_;
   mtapi_uint_t worker_thread_count_;

tasks_cpp/src/node.cc

@@ -78,6 +78,9 @@ Node::Node(
   mtapi_node_get_attribute(node_id, MTAPI_NODE_MAX_QUEUES, &queue_count_,
     sizeof(queue_count_), &status);
   assert(MTAPI_SUCCESS == status);
+  mtapi_node_get_attribute(node_id, MTAPI_NODE_MAX_GROUPS, &group_count_,
+    sizeof(group_count_), &status);
+  assert(MTAPI_SUCCESS == status);
   mtapi_node_get_attribute(node_id, MTAPI_NODE_MAX_TASKS, &task_limit_,
     sizeof(queue_count_), &status);
   assert(MTAPI_SUCCESS == status);