Cycles: Add optional per-kernel performance statistics

When verbose level 4 is enabled, Blender prints kernel performance
data for Cycles on GPU backends (except Metal that doesn't use
debug_enqueue_* methods) for groups of kernels.
These changes introduce a new CYCLES_DEBUG_PER_KERNEL_PERFORMANCE
environment variable to allow getting timings for each kernels
separately and not grouped with others. This is done by adding
explicit synchronization after each kernel execution.

Differential Revision: https://developer.blender.org/D15971

1 files changed, 3 insertions, 1 deletions

diff --git a/intern/cycles/device/hip/queue.cpp b/intern/cycles/device/hip/queue.cpp
index 8b3d963a32f..3f8b6267100 100644
--- a/intern/cycles/device/hip/queue.cpp
+++ b/intern/cycles/device/hip/queue.cpp
@@ -79,7 +79,7 @@ bool HIPDeviceQueue::enqueue(DeviceKernel kernel,
     return false;
   }
 
-  debug_enqueue(kernel, work_size);
+  debug_enqueue_begin(kernel, work_size);
 
   const HIPContextScope scope(hip_device_);
   const HIPDeviceKernel &hip_kernel = hip_device_->kernels.get(kernel);
@@ -120,6 +120,8 @@ bool HIPDeviceQueue::enqueue(DeviceKernel kernel,
                                        0),
                  "enqueue");
 
+  debug_enqueue_end();
+
   return !(hip_device_->have_error());
 }