Adaptive Sampling for Cycles.

This feature takes some inspiration from
"RenderMan: An Advanced Path Tracing Architecture for Movie Rendering" and
"A Hierarchical Automatic Stopping Condition for Monte Carlo Global Illumination"

The basic principle is as follows:
While samples are being added to a pixel, the adaptive sampler writes half
of the samples to a separate buffer. This gives it two separate estimates
of the same pixel, and by comparing their difference it estimates convergence.
Once convergence drops below a given threshold, the pixel is considered done.

When a pixel has not converged yet and needs more samples than the minimum,
its immediate neighbors are also set to take more samples. This is done in order
to more reliably detect sharp features such as caustics. A 3x3 box filter that
is run periodically over the tile buffer is used for that purpose.

After a tile has finished rendering, the values of all passes are scaled as if
they were rendered with the full number of samples. This way, any code operating
on these buffers, for example the denoiser, does not need to be changed for
per-pixel sample counts.

Reviewed By: brecht, #cycles

Differential Revision: https://developer.blender.org/D4686

1 files changed, 22 insertions, 0 deletions

diff --git a/intern/cycles/device/cuda/device_cuda.h b/intern/cycles/device/cuda/device_cuda.h
index a825bd6b128..6a0b39434aa 100644
--- a/intern/cycles/device/cuda/device_cuda.h
+++ b/intern/cycles/device/cuda/device_cuda.h
@@ -82,6 +82,17 @@ class CUDADevice : public Device {
   device_vector<TextureInfo> texture_info;
   bool need_texture_info;
 
+  /* Kernels */
+  struct {
+    bool loaded;
+
+    CUfunction adaptive_stopping;
+    CUfunction adaptive_filter_x;
+    CUfunction adaptive_filter_y;
+    CUfunction adaptive_scale_samples;
+    int adaptive_num_threads_per_block;
+  } functions;
+
   static bool have_precompiled_kernels();
 
   virtual bool show_samples() const;
@@ -114,6 +125,8 @@ class CUDADevice : public Device {
 
   virtual bool load_kernels(const DeviceRequestedFeatures &requested_features);
 
+  void load_functions();
+
   void reserve_local_memory(const DeviceRequestedFeatures &requested_features);
 
   void init_host_memory();
@@ -197,6 +210,15 @@ class CUDADevice : public Device {
 
   void denoise(RenderTile &rtile, DenoisingTask &denoising);
 
+  void adaptive_sampling_filter(uint filter_sample,
+                                WorkTile *wtile,
+                                CUdeviceptr d_wtile,
+                                CUstream stream = 0);
+  void adaptive_sampling_post(RenderTile &rtile,
+                              WorkTile *wtile,
+                              CUdeviceptr d_wtile,
+                              CUstream stream = 0);
+
   void path_trace(DeviceTask &task, RenderTile &rtile, device_vector<WorkTile> &work_tiles);
 
   void film_convert(DeviceTask &task,