Cycles: more accurately count main paths for adding work tiles

Easy now thanks to the main and shadow path decoupling. Doesn't help
in an benchmark scene except Spring, where it reduces render time by
maybe 2-3%.

Ref T87836

1 files changed, 4 insertions, 2 deletions

diff --git a/intern/cycles/integrator/path_trace_work_gpu.cpp b/intern/cycles/integrator/path_trace_work_gpu.cpp
index f898f3fce81..2c71b1cf876 100644
--- a/intern/cycles/integrator/path_trace_work_gpu.cpp
+++ b/intern/cycles/integrator/path_trace_work_gpu.cpp
@@ -731,7 +731,6 @@ void PathTraceWorkGPU::enqueue_work_tiles(DeviceKernel kernel,
 
 int PathTraceWorkGPU::num_active_main_paths_paths()
 {
-  /* TODO: this is wrong, does not account for duplicates with shadow! */
   IntegratorQueueCounter *queue_counter = integrator_queue_counter_.data();
 
   int num_paths = 0;
@@ -739,7 +738,10 @@ int PathTraceWorkGPU::num_active_main_paths_paths()
     DCHECK_GE(queue_counter->num_queued[i], 0)
         << "Invalid number of queued states for kernel "
         << device_kernel_as_string(static_cast<DeviceKernel>(i));
-    num_paths += queue_counter->num_queued[i];
+
+    if (!kernel_is_shadow_path((DeviceKernel)i)) {
+      num_paths += queue_counter->num_queued[i];
+    }
   }
 
   return num_paths;