Cycles: use TBB for task pools and task scheduler

No significant performance improvement is expected, but it means we have a
single thread pool throughout Blender. And it should make adding more
parallellization in the future easier.

After previous refactoring commits this is basically a drop-in replacement.
One difference is that the task pool had a mechanism for scheduling tasks to
the front of the queue to minimize memory usage. TBB has a smarter algorithm
to balance depth-first and breadth-first scheduling of tasks and we assume that
removes the need to manually provide hints to the scheduler.

Fixes T77533

2 files changed, 9 insertions, 13 deletions

diff --git a/intern/cycles/bvh/bvh_build.cpp b/intern/cycles/bvh/bvh_build.cpp
index 116576b101d..0235ac33c77 100644
--- a/intern/cycles/bvh/bvh_build.cpp
+++ b/intern/cycles/bvh/bvh_build.cpp
@@ -626,8 +626,8 @@ BVHNode *BVHBuild::build_node(const BVHObjectBinning &range, int level)
     /* Threaded build */
     inner = new InnerNode(bounds);
 
-    task_pool.push([=] { thread_build_node(inner, 0, left, level + 1); }, true);
-    task_pool.push([=] { thread_build_node(inner, 1, right, level + 1); }, true);
+    task_pool.push([=] { thread_build_node(inner, 0, left, level + 1); });
+    task_pool.push([=] { thread_build_node(inner, 1, right, level + 1); });
   }
 
   if (do_unalinged_split) {
@@ -742,16 +742,12 @@ BVHNode *BVHBuild::build_node(const BVHRange &range,
 
     /* Create tasks for left and right nodes, using copy for most arguments and
      * move for reference to avoid memory copies. */
-    task_pool.push(
-        [=, refs = std::move(left_references)]() mutable {
-          thread_build_spatial_split_node(inner, 0, left, refs, level + 1);
-        },
-        true);
-    task_pool.push(
-        [=, refs = std::move(right_references)]() mutable {
-          thread_build_spatial_split_node(inner, 1, right, refs, level + 1);
-        },
-        true);
+    task_pool.push([=, refs = std::move(left_references)]() mutable {
+      thread_build_spatial_split_node(inner, 0, left, refs, level + 1);
+    });
+    task_pool.push([=, refs = std::move(right_references)]() mutable {
+      thread_build_spatial_split_node(inner, 1, right, refs, level + 1);
+    });
   }
 
   if (do_unalinged_split) {
diff --git a/intern/cycles/bvh/bvh_sort.cpp b/intern/cycles/bvh/bvh_sort.cpp
index 5bdded354bc..b01785b547a 100644
--- a/intern/cycles/bvh/bvh_sort.cpp
+++ b/intern/cycles/bvh/bvh_sort.cpp
@@ -147,7 +147,7 @@ static void bvh_reference_sort_threaded(TaskPool *task_pool,
     if (left < end) {
       if (start < right) {
         task_pool->push(
-            function_bind(bvh_reference_sort_threaded, task_pool, data, left, end, compare), true);
+            function_bind(bvh_reference_sort_threaded, task_pool, data, left, end, compare));
       }
       else {
         start = left;