Compositor: Implement per-node execution time report

Visually it is the same as the execution time implemented for the
geometry nodes, and it is to be enabled in the overlay popover.

The implementation is separate from the geometry nodes, as it is
not easy or practical to re-use the geometry nodes implementation.

The execution time is stored in a run-time hash, indexed by a node
instance key. This is similar to the storage of the mode preview
images.

The time is only implemented for full-frame compositor, as for the
tiled compositor it could be tricky to calculate reliable time for
pixel processing nodes which process one pixel at a time.

@@ -0,0 +17,4 @@
+
+class Profiler {
+  /* Per-node accumulated execution time. Includes execution time of all operations the node was
+   * break down into. */

@@ -2459,12 +2463,89 @@ static std::optional<std::chrono::nanoseconds> node_get_execution_time(
   return std::nullopt;
 }
 
+/* Cerate node key instance, assuming the node comes from the currently editing node tree. */

@@ -2626,2 +2740,3 @@
+
   if (!(snode.edittree->type == NTREE_GEOMETRY)) {
-    /* Currently geometry nodes are the only nodes to have extra infos per nodes. */
+    /* Currently geometry and compositor nodes are the only nodes to have extra infos per nodes. */

@@ -321,2 +321,4 @@
   BKE_callback_exec_id(bmain, &scene->id, BKE_CB_EVT_COMPOSITE_CANCEL);
   cj->cancelled = true;
+
+  cj->ntree->runtime->compositor.per_node_execution_time =

@@ -0,0 +26,4 @@
+/* Profiler implementation which is used by the node execution system. */
+class Profiler {
+  /* Local copy of the profiling data, which is known to not cause threading conflicts with the
+   * interface thread while the compositing tree is evaluated in background. */

@@ -0,0 +39,4 @@
+
+void Profiler::finalize(const bNodeTree &node_tree)
+{
+  accumulate_node_group_times(node_tree);

@@ -0,0 +51,4 @@
+    const bNodeInstanceKey key = BKE_node_instance_key(parent_key, &node_tree, node);
+
+    if (node->type != NODE_GROUP) {
+      /* Non-group node, no need to recurs into. Simply accumulate the node's execution time to the

@@ -2459,12 +2467,87 @@ static std::optional<std::chrono::nanoseconds> node_get_execution_time(
   return std::nullopt;
 }
 
+/* Create node key instance, assuming the node comes from the currently editing node tree. */