Working recursive invocations.

Recursion works by using a function node, calling the graph again. We separated an graph invocation form an function invocation within an graph, making the graph only handle one concern.

app/playground/main.cpp

@@ -4,6 +4,7 @@
 #include <tuple>
 #include <array>
 
+#include <pls/pls.h>
 #include <pls/dataflow/internal/inputs.h>
 #include <pls/dataflow/internal/outputs.h>
 #include <pls/dataflow/internal/function_node.h>
@@ -17,31 +18,53 @@ int main() {
   // Define
   graph<inputs<int, int>, outputs<int>> graph;
 
-  auto func1 = [](const int &i1, const int &i2, int &o1) {
-    std::cout << "Add up " << i1 << " and " << i2 << "..." << std::endl;
-    o1 = i1 + i2;
+  auto triple = [](const int &i1, int &o1) {
+    o1 = i1 * 3;
   };
-  function_node<inputs<int, int>, outputs<int>, decltype(func1)> node1{func1};
+  function_node<inputs<int>, outputs<int>, decltype(triple)> triple_node{triple};
 
-  auto func2 = [](const int &i1, int &o1) {
-    std::cout << "Print Result " << i1 << std::endl;
-    o1 = i1;
+  auto minus_one = [](const int &i1, int &o1) {
+    o1 = i1 - 1;
   };
-  function_node<inputs<int>, outputs<int>, decltype(func2)> node2{func2};
+  function_node<inputs<int>, outputs<int>, decltype(minus_one)> minus_one_node{minus_one};
+
+  auto recursion = [&](const int &i1, const int &i2, int &o1) {
+    if (i1 > 0) {
+      std::tuple<int> out;
+      graph.run({i1, i2}, out);
+      pls::scheduler::wait_for_all();
+      o1 = std::get<0>(out);
+    } else {
+      o1 = i2;
+    }
+  };
+  function_node<inputs<int, int>, outputs<int>, decltype(recursion)> recursion_node{recursion};
 
   // Connect
-  graph.input<0>() >> node1.in_port<0>();
-  graph.input<1>() >> node1.in_port<1>();
+  graph.input<0>() >> minus_one_node.in_port<0>();
+  minus_one_node.out_port<0>() >> recursion_node.in_port<0>();
 
-  node1.out_port<0>() >> node2.in_port<0>();
+  graph.input<1>() >> triple_node.in_port<0>();
+  triple_node.out_port<0>() >> recursion_node.in_port<1>();
 
-  node2.out_port<0>() >> graph.output<0>();
+  recursion_node.out_port<0>() >> graph.output<0>();
 
   // Build
   graph.build();
 
-  // Execute
-  graph.run({1, 2});
-  graph.run({1, 1});
-  graph.run({5, 6});
+  pls::malloc_scheduler_memory my_scheduler_memory{8, 2u << 18u};
+  pls::scheduler scheduler{&my_scheduler_memory, 8};
+  scheduler.perform_work([&] {
+    // Schedule Execution
+    std::tuple<int> out1, out2, out3;
+    graph.run({1, 2}, out1);
+    graph.run({1, 1}, out2);
+    graph.run({5, 6}, out3);
+
+    // Wait for results and print
+    pls::scheduler::wait_for_all();
+    std::cout << std::get<0>(out1) << std::endl;
+    std::cout << std::get<0>(out2) << std::endl;
+    std::cout << std::get<0>(out3) << std::endl;
+  });
 }

lib/pls/CMakeLists.txt

@@ -49,7 +49,7 @@ add_library(pls STATIC
         include/pls/internal/scheduling/scheduler_impl.h
         include/pls/internal/scheduling/task.h src/internal/scheduling/task.cpp
         include/pls/internal/scheduling/scheduler_memory.h src/internal/scheduling/scheduler_memory.cpp
-        include/pls/internal/scheduling/lambda_task.h include/pls/internal/helpers/seqence.h)
+        include/pls/internal/scheduling/lambda_task.h include/pls/internal/helpers/seqence.h include/pls/dataflow/internal/build_state.h)
 # Add everything in `./include` to be in the include path of this project
 target_include_directories(pls
         PUBLIC

lib/pls/include/pls/dataflow/internal/function_node.h

@@ -64,7 +64,7 @@ class function_node<pls::dataflow::inputs<I0, I...>, pls::dataflow::outputs<O0,
 
   template<int POS, typename T>
   void token_pushed(token<T> token) {
-    auto current_memory = get_invocation<invocation_memory>(token);
+    auto current_memory = get_invocation<invocation_memory>(token.invocation());
 
     std::get<POS>(current_memory->input_buffer_) = token;
     auto remaining_inputs = --(current_memory->inputs_missing_);
@@ -141,7 +141,6 @@ class function_node<pls::dataflow::inputs<I0, I...>, pls::dataflow::outputs<O0,
   };
 
  private:
-  // Input/Output ports
   multi_in_port_type in_port_;
   multi_out_port_type out_port_;
 

lib/pls/include/pls/dataflow/internal/graph.h

@@ -6,12 +6,16 @@
 #include <tuple>
 #include <iostream>
 
+#include "build_state.h"
 #include "node.h"
+#include "function_node.h"
 #include "in_port.h"
 #include "out_port.h"
 #include "inputs.h"
 #include "outputs.h"
 
+#include "pls/pls.h"
+
 namespace pls {
 namespace dataflow {
 namespace internal {
@@ -25,126 +29,70 @@ class graph<pls::dataflow::inputs<I0, I...>, pls::dataflow::outputs<O0, O...>> :
   // Our own type
   using self_type = graph<pls::dataflow::inputs<I0, I...>, pls::dataflow::outputs<O0, O...>>;
 
-  // Callbacks for input/output ports
-  struct in_port_cb {
-    self_type *self_;
-    template<int POS, typename T>
-    void token_pushed(token<T> token) {
-      self_->in_port_token_pushed<POS, T>(token);
-    }
-  };
-  struct out_port_cb {
-    self_type *self_;
-    template<int POS, typename T>
-    void token_pushed(token<T> token) {
-      self_->out_port_token_pushed<POS, T>(token);
-    }
-  };
-
-  // Input-Output port types (external)
-  using multi_in_port_type = multi_in_port<in_port_cb, 0, I0, I...>;
-  using multi_out_port_type = multi_out_port<O0, O...>;
-
   // Input-Output port types (internal)
-  using internal_in_port_type = multi_out_port<I0, I...>;
-  using internal_out_port_type = multi_in_port<out_port_cb, 0, O0, O...>;
+  using inputs_type = multi_out_port<I0, I...>;
+  using outputs_type = multi_in_port<self_type, 0, O0, O...>;
 
   // Input-Output value tuples
   using value_input_tuple = std::tuple<I0, I...>;
   using input_tuple = std::tuple<token < I0>, token<I>...>;
+
+  using value_output_tuple = std::tuple<O0, O...>;
   using output_tuple = std::tuple<token < O0>, token<O>...>;
+
   static constexpr int num_in_ports = std::tuple_size<input_tuple>();
   static constexpr int num_out_ports = std::tuple_size<output_tuple>();
 
   // Memory used for ONE active invocation of the dataflow-graph
   struct invocation_memory {
-    std::atomic<int> inputs_missing_;
-    input_tuple input_buffer_;
+    value_output_tuple *output_buffer_;
   };
 
  public:
   template<int POS>
-  using in_port_at = typename multi_in_port_type::template in_port_type_at<POS>;
-  template<int POS>
-  using out_port_at = typename multi_out_port_type::template out_port_type_at<POS>;
-
-  template<int POS>
-  in_port_at<POS> &in_port() {
-    return in_port_.template get<POS>();
-  }
-  template<int POS>
-  out_port_at<POS> &out_port() {
-    return out_port_.template get<POS>();
-  }
-
-  template<int POS, typename T>
-  void in_port_token_pushed(token<T> token) {
-    auto invocation = get_invocation<invocation_memory>(token);
-
-    // TODO: Invoke a new instance of our graph (recursive call) when all inputs are here
-    std::cout << "Token pushed at " << POS << " with value " << token.value() << std::endl;
-    auto remaining_inputs = --(invocation->inputs_missing_);
-    if (remaining_inputs == 0) {
-      std::cout << "All inputs there!" << std::endl;
-    }
-  }
-
-  template<int POS>
-  using internal_in_port_at = typename internal_in_port_type::template out_port_type_at<POS>;
+  using input_at = typename inputs_type::template out_port_type_at<POS>;
   template<int POS>
-  using internal_out_port_at = typename internal_out_port_type::template in_port_type_at<POS>;
+  using output_at = typename outputs_type::template in_port_type_at<POS>;
 
   template<int POS>
-  internal_in_port_at<POS> &input() {
-    return internal_in_port_.template get<POS>();
+  input_at<POS> &input() {
+    return inputs_.template get<POS>();
   }
   template<int POS>
-  internal_out_port_at<POS> &output() {
-    return internal_out_port_.template get<POS>();
+  output_at<POS> &output() {
+    return outputs_.template get<POS>();
   }
 
   template<int POS, typename T>
-  void out_port_token_pushed(token<T> token) {
-    auto invocation = get_invocation<invocation_memory>(token);
-
-    std::cout << "Output produced at port " << POS << " with value " << token.value() << std::endl;
-    auto remaining_inputs = --(invocation->inputs_missing_);
-    if (remaining_inputs == 0) {
-      std::cout << "All inputs there!" << std::endl;
-    }
+  void token_pushed(token<T> token) {
+    auto invocation = get_invocation<invocation_memory>(token.invocation());
+    std::get<POS>(*invocation->output_buffer_) = token.value();
   }
 
-  graph() : in_port_cb_{this},
-            out_port_cb_{this},
-            in_port_{this, &in_port_cb_},
-            out_port_{},
-            internal_in_port_{},
-            internal_out_port_{this, &out_port_cb_} {
-  }
+  graph() : inputs_{}, outputs_{this, this} {}
 
   /////////////////////////////////////////////////////////////////
   // Graph building
   /////////////////////////////////////////////////////////////////
   void build() {
-    state_ = building;
-    PLS_ASSERT(is_internal_fully_connected(), "Must fully connect all inputs/outputs inside a dataflow graph!")
+    PLS_ASSERT(build_state_ == build_state::fresh, "Must only build a dataflow graph once!")
+    PLS_ASSERT(is_fully_connected(), "Must fully connect all inputs/outputs inside a dataflow graph!")
 
-    num_nodes_ = 0;
+    build_state_ = build_state::building;
+    node_list_start_ = node_list_current_ = this;
+    memory_index_ = 0;
+    num_nodes_ = 1;
     for (int i = 0; i < num_in_ports; i++) {
-      build_recursive(internal_in_port_.next_node_at(i));
+      build_recursive(inputs_.next_node_at(i));
     }
-
-    state_ = built;
+    build_state_ = build_state::built;
   }
 
-  node *start_{nullptr};
-  node *current_{nullptr};
-
   void build_recursive(node *node) {
-    if (node->state_ != fresh) {
+    if (node->build_state_ != build_state::fresh) {
       return; // Already visited
     }
-    node->state_ = building;
+    node->build_state_ = build_state::building;
     PLS_ASSERT(node->is_fully_connected(), "Must fully connect dataflow graph nodes!")
 
     add_node(node);
@@ -152,34 +100,33 @@ class graph<pls::dataflow::inputs<I0, I...>, pls::dataflow::outputs<O0, O...>> :
       build_recursive(node->successor_at(i));
     }
 
-    node->state_ = built;
+    node->build_state_ = build_state::built;
   }
 
   void add_node(node *new_node) {
     new_node->memory_index_ = num_nodes_++;
 
-    if (start_ == nullptr) {
-      start_ = new_node;
-      current_ = new_node;
+    if (node_list_current_ == nullptr) {
+      node_list_current_ = new_node;
+      node_list_start_ = new_node;
     } else {
-      current_->direct_successor_ = new_node;
-      current_ = new_node;
+      node_list_current_->direct_successor_ = new_node;
+      node_list_current_ = new_node;
     }
   }
 
-  // TODO: Clean up + move to async task
   template<int N, typename ...IT>
   struct feed_inputs {
-    feed_inputs(internal_in_port_type &, value_input_tuple &, invocation_info &) {}
+    feed_inputs(inputs_type &, value_input_tuple &, invocation_info &) {}
     void run() {}
   };
   template<int N, typename IT1, typename ...IT>
   struct feed_inputs<N, IT1, IT...> {
-    internal_in_port_type &inputs_;
+    inputs_type &inputs_;
     value_input_tuple &input_values_;
     invocation_info &invocation_;
 
-    feed_inputs(internal_in_port_type &inputs,
+    feed_inputs(inputs_type &inputs,
                 value_input_tuple &input_values,
                 invocation_info &invocation) : inputs_{inputs},
                                                input_values_{input_values},
@@ -191,38 +138,56 @@ class graph<pls::dataflow::inputs<I0, I...>, pls::dataflow::outputs<O0, O...>> :
     }
   };
 
-  void run(value_input_tuple input) {
-    // TODO: clearly move this onto the task stack instead fo malloc (without any free...)
-    void **buffers = reinterpret_cast<void **>(malloc(num_nodes_ * sizeof(void *)));
-    node *iterator = start_;
-    for (int i = 0; i < num_nodes_; i++) {
-      auto required_size = iterator->instance_buffer_size();
-      buffers[i] = malloc(required_size);
-      iterator->init_instance_buffer(buffers[i]);
+  class run_graph_task : public pls::task {
+    graph *self_;
+    value_input_tuple input_;
+    value_output_tuple *output_;
+
+    // Buffers for actual execution
+    invocation_info invocation_;
+
+   public:
+    run_graph_task(graph *self, value_input_tuple &input, value_output_tuple *output) : self_{self},
+                                                                                        input_{input},
+                                                                                        output_{output},
+                                                                                        invocation_{nullptr} {
+      void **buffers;
+      buffers = reinterpret_cast<void **>(allocate_memory(self_->num_nodes_ * sizeof(void *)));
+
+      node *iterator = self_->node_list_start_;
+      for (int i = 0; i < self_->num_nodes_; i++) {
+        auto required_size = iterator->instance_buffer_size();
+        buffers[i] = allocate_memory(required_size);
+        iterator->init_instance_buffer(buffers[i]);
+
+        iterator = iterator->direct_successor_;
+      }
+      invocation_ = invocation_info{buffers};
+
+      self_->get_invocation<invocation_memory>(invocation_)->output_buffer_ = output;
+    }
 
-      iterator = iterator->direct_successor_;
+    void execute_internal() override {
+      feed_inputs<0, I0, I...>{self_->inputs_, input_, invocation_}.run();
     }
+  };
 
-    invocation_info invocation{buffers};
-    feed_inputs<0, I0, I...>{internal_in_port_, input, invocation}.run();
+  void run(value_input_tuple input, value_output_tuple &output) {
+    pls::scheduler::spawn_child<run_graph_task>(this, input, &output);
   }
 
   //////////////////////////////////////////////////////////////////
   // Overrides for generic node functionality (building the graph)
   //////////////////////////////////////////////////////////////////
   int num_successors() const override {
-    return num_out_ports;
+    return 0;
   }
   node *successor_at(int pos) const override {
-    return out_port_.next_node_at(pos);
+    PLS_ERROR("A graph instance has no direct successor!")
   }
 
   bool is_fully_connected() const override {
-    return in_port_.fully_connected() && out_port_.fully_connected();
-  }
-
-  bool is_internal_fully_connected() const {
-    return internal_in_port_.fully_connected() && internal_out_port_.fully_connected();
+    return inputs_.fully_connected() && outputs_.fully_connected();
   }
 
   int instance_buffer_size() const override {
@@ -230,20 +195,15 @@ class graph<pls::dataflow::inputs<I0, I...>, pls::dataflow::outputs<O0, O...>> :
   }
   void init_instance_buffer(void *memory) const override {
     auto invocation = new(memory) invocation_memory{};
-    invocation->inputs_missing_ = num_in_ports;
   };
 
  private:
-  // Input/Output ports
-  in_port_cb in_port_cb_;
-  out_port_cb out_port_cb_;
-
-  multi_in_port_type in_port_;
-  multi_out_port_type out_port_;
-
-  internal_in_port_type internal_in_port_;
-  internal_out_port_type internal_out_port_;
+  inputs_type inputs_;
+  outputs_type outputs_;
 
+  // Information about building the graph
+  node *node_list_start_{nullptr};
+  node *node_list_current_{nullptr};
   int num_nodes_{0};
 };
 

lib/pls/include/pls/dataflow/internal/node.h

@@ -2,17 +2,22 @@
 #ifndef PLS_DATAFLOW_INTERNAL_NODE_H_
 #define PLS_DATAFLOW_INTERNAL_NODE_H_
 
+#include "build_state.h"
+
 namespace pls {
 namespace dataflow {
 namespace internal {
 
+/**
+ * Represents a single piece included in a dataflow graph.
+ * This can be anything connected to form the dataflow that requires
+ * memory during invocation and is present in the connected flow graph.
+ */
 class node {
   template<typename INS, typename OUTS>
   friend
   class graph;
 
-  enum state { fresh, building, built, teardown };
-
  public:
   virtual int num_successors() const = 0;
   virtual node *successor_at(int pos) const = 0;
@@ -22,18 +27,15 @@ class node {
 
   virtual bool is_fully_connected() const = 0;
 
-  template<typename M, typename T>
-  M *get_invocation(token<T> token) {
-    return token.invocation().template get_instance_buffer<M>(memory_index_);
+  template<typename M>
+  M *get_invocation(invocation_info invocation) {
+    return invocation.template get_instance_buffer<M>(memory_index_);
   }
 
-  // TODO: Remove
-  void set_memory_index(int i) { memory_index_ = i; }
-
  private:
   int memory_index_{0};
   node *direct_successor_{nullptr};
-  state state_{fresh};
+  build_state build_state_{build_state::fresh};
 };
 
 }