Cycles: code refactor to bake using regular render session and tiles

There should be no user visible change from this, except that tile size
now affects performance. The goal here is to simplify bake denoising in
D3099, letting it reuse more denoising tiles and pass code.

A lot of code is now shared with regular rendering, with the two main
differences being that we read some render result passes from the bake API
when starting to render a tile, and call the bake kernel instead of the
path trace kernel.

With this kind of design where Cycles asks for tiles from the bake API,
it should eventually be easier to reduce memory usage, show tiles as
they are baked, or bake multiple passes at once, though there's still
quite some work needed for that.

Reviewers: #cycles

Subscribers: monio, wmatyjewicz, lukasstockner97, michaelknubben

Differential Revision: https://developer.blender.org/D3108

5 files changed, 100 insertions, 36 deletions

diff --git a/intern/cycles/device/cuda/device_cuda.h b/intern/cycles/device/cuda/device_cuda.h
index 3e397da895b..3f23f0fe4c5 100644
--- a/intern/cycles/device/cuda/device_cuda.h
+++ b/intern/cycles/device/cuda/device_cuda.h
@@ -223,7 +223,7 @@ class CUDADevice : public Device {
                               CUdeviceptr d_wtile,
                               CUstream stream = 0);
 
-  void path_trace(DeviceTask &task, RenderTile &rtile, device_vector<WorkTile> &work_tiles);
+  void render(DeviceTask &task, RenderTile &rtile, device_vector<WorkTile> &work_tiles);
 
   void film_convert(DeviceTask &task,
                     device_ptr buffer,
diff --git a/intern/cycles/device/cuda/device_cuda_impl.cpp b/intern/cycles/device/cuda/device_cuda_impl.cpp
index ba5d479e0e7..acf53c3eb1b 100644
--- a/intern/cycles/device/cuda/device_cuda_impl.cpp
+++ b/intern/cycles/device/cuda/device_cuda_impl.cpp
@@ -586,20 +586,23 @@ void CUDADevice::reserve_local_memory(const DeviceRequestedFeatures &requested_f
   cuMemGetInfo(&free_before, &total);
 
   /* Get kernel function. */
-  CUfunction cuPathTrace;
+  CUfunction cuRender;
 
-  if (requested_features.use_integrator_branched) {
-    cuda_assert(cuModuleGetFunction(&cuPathTrace, cuModule, "kernel_cuda_branched_path_trace"));
+  if (requested_features.use_baking) {
+    cuda_assert(cuModuleGetFunction(&cuRender, cuModule, "kernel_cuda_bake"));
+  }
+  else if (requested_features.use_integrator_branched) {
+    cuda_assert(cuModuleGetFunction(&cuRender, cuModule, "kernel_cuda_branched_path_trace"));
   }
   else {
-    cuda_assert(cuModuleGetFunction(&cuPathTrace, cuModule, "kernel_cuda_path_trace"));
+    cuda_assert(cuModuleGetFunction(&cuRender, cuModule, "kernel_cuda_path_trace"));
   }
 
-  cuda_assert(cuFuncSetCacheConfig(cuPathTrace, CU_FUNC_CACHE_PREFER_L1));
+  cuda_assert(cuFuncSetCacheConfig(cuRender, CU_FUNC_CACHE_PREFER_L1));
 
   int min_blocks, num_threads_per_block;
-  cuda_assert(cuOccupancyMaxPotentialBlockSize(
-      &min_blocks, &num_threads_per_block, cuPathTrace, NULL, 0, 0));
+  cuda_assert(
+      cuOccupancyMaxPotentialBlockSize(&min_blocks, &num_threads_per_block, cuRender, NULL, 0, 0));
 
   /* Launch kernel, using just 1 block appears sufficient to reserve
    * memory for all multiprocessors. It would be good to do this in
@@ -609,7 +612,7 @@ void CUDADevice::reserve_local_memory(const DeviceRequestedFeatures &requested_f
 
   void *args[] = {&d_work_tiles, &total_work_size};
 
-  cuda_assert(cuLaunchKernel(cuPathTrace, 1, 1, 1, num_threads_per_block, 1, 1, 0, 0, args, 0));
+  cuda_assert(cuLaunchKernel(cuRender, 1, 1, 1, num_threads_per_block, 1, 1, 0, 0, args, 0));
 
   cuda_assert(cuCtxSynchronize());
 
@@ -1780,9 +1783,7 @@ void CUDADevice::adaptive_sampling_post(RenderTile &rtile,
                              0));
 }
 
-void CUDADevice::path_trace(DeviceTask &task,
-                            RenderTile &rtile,
-                            device_vector<WorkTile> &work_tiles)
+void CUDADevice::render(DeviceTask &task, RenderTile &rtile, device_vector<WorkTile> &work_tiles)
 {
   scoped_timer timer(&rtile.buffers->render_time);
 
@@ -1790,21 +1791,24 @@ void CUDADevice::path_trace(DeviceTask &task,
     return;
 
   CUDAContextScope scope(this);
-  CUfunction cuPathTrace;
+  CUfunction cuRender;
 
   /* Get kernel function. */
-  if (task.integrator_branched) {
-    cuda_assert(cuModuleGetFunction(&cuPathTrace, cuModule, "kernel_cuda_branched_path_trace"));
+  if (rtile.task == RenderTile::BAKE) {
+    cuda_assert(cuModuleGetFunction(&cuRender, cuModule, "kernel_cuda_bake"));
+  }
+  else if (task.integrator_branched) {
+    cuda_assert(cuModuleGetFunction(&cuRender, cuModule, "kernel_cuda_branched_path_trace"));
   }
   else {
-    cuda_assert(cuModuleGetFunction(&cuPathTrace, cuModule, "kernel_cuda_path_trace"));
+    cuda_assert(cuModuleGetFunction(&cuRender, cuModule, "kernel_cuda_path_trace"));
   }
 
   if (have_error()) {
     return;
   }
 
-  cuda_assert(cuFuncSetCacheConfig(cuPathTrace, CU_FUNC_CACHE_PREFER_L1));
+  cuda_assert(cuFuncSetCacheConfig(cuRender, CU_FUNC_CACHE_PREFER_L1));
 
   /* Allocate work tile. */
   work_tiles.alloc(1);
@@ -1822,8 +1826,8 @@ void CUDADevice::path_trace(DeviceTask &task,
    * remain conservative for GPUs connected to a display to avoid driver
    * timeouts and display freezing. */
   int min_blocks, num_threads_per_block;
-  cuda_assert(cuOccupancyMaxPotentialBlockSize(
-      &min_blocks, &num_threads_per_block, cuPathTrace, NULL, 0, 0));
+  cuda_assert(
+      cuOccupancyMaxPotentialBlockSize(&min_blocks, &num_threads_per_block, cuRender, NULL, 0, 0));
   if (!info.display_device) {
     min_blocks *= 8;
   }
@@ -1851,7 +1855,7 @@ void CUDADevice::path_trace(DeviceTask &task,
     void *args[] = {&d_work_tiles, &total_work_size};
 
     cuda_assert(
-        cuLaunchKernel(cuPathTrace, num_blocks, 1, 1, num_threads_per_block, 1, 1, 0, 0, args, 0));
+        cuLaunchKernel(cuRender, num_blocks, 1, 1, num_threads_per_block, 1, 1, 0, 0, args, 0));
 
     /* Run the adaptive sampling kernels at selected samples aligned to step samples. */
     uint filter_sample = sample + wtile->num_samples - 1;
@@ -1957,10 +1961,7 @@ void CUDADevice::shader(DeviceTask &task)
   CUdeviceptr d_output = (CUdeviceptr)task.shader_output;
 
   /* get kernel function */
-  if (task.shader_eval_type >= SHADER_EVAL_BAKE) {
-    cuda_assert(cuModuleGetFunction(&cuShader, cuModule, "kernel_cuda_bake"));
-  }
-  else if (task.shader_eval_type == SHADER_EVAL_DISPLACE) {
+  if (task.shader_eval_type == SHADER_EVAL_DISPLACE) {
     cuda_assert(cuModuleGetFunction(&cuShader, cuModule, "kernel_cuda_displace"));
   }
   else {
@@ -2297,9 +2298,12 @@ void CUDADevice::thread_run(DeviceTask *task)
           split_kernel->path_trace(task, tile, void_buffer, void_buffer);
         }
         else {
-          path_trace(*task, tile, work_tiles);
+          render(*task, tile, work_tiles);
         }
       }
+      else if (tile.task == RenderTile::BAKE) {
+        render(*task, tile, work_tiles);
+      }
       else if (tile.task == RenderTile::DENOISE) {
         tile.sample = tile.start_sample + tile.num_samples;
 
diff --git a/intern/cycles/device/device_cpu.cpp b/intern/cycles/device/device_cpu.cpp
index c701c14318f..fc6febd8cee 100644
--- a/intern/cycles/device/device_cpu.cpp
+++ b/intern/cycles/device/device_cpu.cpp
@@ -188,6 +188,7 @@ class CPUDevice : public Device {
       convert_to_byte_kernel;
   KernelFunctions<void (*)(KernelGlobals *, uint4 *, float4 *, int, int, int, int, int)>
       shader_kernel;
+  KernelFunctions<void (*)(KernelGlobals *, float *, int, int, int, int, int)> bake_kernel;
 
   KernelFunctions<void (*)(
       int, TileInfo *, int, int, float *, float *, float *, float *, float *, int *, int, int)>
@@ -270,6 +271,7 @@ class CPUDevice : public Device {
         REGISTER_KERNEL(convert_to_half_float),
         REGISTER_KERNEL(convert_to_byte),
         REGISTER_KERNEL(shader),
+        REGISTER_KERNEL(bake),
         REGISTER_KERNEL(filter_divide_shadow),
         REGISTER_KERNEL(filter_get_feature),
         REGISTER_KERNEL(filter_write_feature),
@@ -895,7 +897,7 @@ class CPUDevice : public Device {
     }
   }
 
-  void path_trace(DeviceTask &task, RenderTile &tile, KernelGlobals *kg)
+  void render(DeviceTask &task, RenderTile &tile, KernelGlobals *kg)
   {
     const bool use_coverage = kernel_data.film.cryptomatte_passes & CRYPT_ACCURATE;
 
@@ -919,12 +921,21 @@ class CPUDevice : public Device {
           break;
       }
 
-      for (int y = tile.y; y < tile.y + tile.h; y++) {
-        for (int x = tile.x; x < tile.x + tile.w; x++) {
-          if (use_coverage) {
-            coverage.init_pixel(x, y);
+      if (tile.task == RenderTile::PATH_TRACE) {
+        for (int y = tile.y; y < tile.y + tile.h; y++) {
+          for (int x = tile.x; x < tile.x + tile.w; x++) {
+            if (use_coverage) {
+              coverage.init_pixel(x, y);
+            }
+            path_trace_kernel()(kg, render_buffer, sample, x, y, tile.offset, tile.stride);
+          }
+        }
+      }
+      else {
+        for (int y = tile.y; y < tile.y + tile.h; y++) {
+          for (int x = tile.x; x < tile.x + tile.w; x++) {
+            bake_kernel()(kg, render_buffer, sample, x, y, tile.offset, tile.stride);
           }
-          path_trace_kernel()(kg, render_buffer, sample, x, y, tile.offset, tile.stride);
         }
       }
       tile.sample = sample + 1;
@@ -1019,9 +1030,12 @@ class CPUDevice : public Device {
           split_kernel->path_trace(&task, tile, kgbuffer, void_buffer);
         }
         else {
-          path_trace(task, tile, kg);
+          render(task, tile, kg);
         }
       }
+      else if (tile.task == RenderTile::BAKE) {
+        render(task, tile, kg);
+      }
       else if (tile.task == RenderTile::DENOISE) {
         denoise(denoising, tile);
         task.update_progress(&tile, tile.w * tile.h);
diff --git a/intern/cycles/device/opencl/device_opencl.h b/intern/cycles/device/opencl/device_opencl.h
index d6f4fb43061..389268e1c2a 100644
--- a/intern/cycles/device/opencl/device_opencl.h
+++ b/intern/cycles/device/opencl/device_opencl.h
@@ -451,6 +451,7 @@ class OpenCLDevice : public Device {
                     device_ptr rgba_half);
   void shader(DeviceTask &task);
   void update_adaptive(DeviceTask &task, RenderTile &tile, int sample);
+  void bake(DeviceTask &task, RenderTile &tile);
 
   void denoise(RenderTile &tile, DenoisingTask &denoising);
 
diff --git a/intern/cycles/device/opencl/device_opencl_impl.cpp b/intern/cycles/device/opencl/device_opencl_impl.cpp
index 2766f85d17c..beb3174b111 100644
--- a/intern/cycles/device/opencl/device_opencl_impl.cpp
+++ b/intern/cycles/device/opencl/device_opencl_impl.cpp
@@ -1367,6 +1367,9 @@ void OpenCLDevice::thread_run(DeviceTask *task)
          */
         clFinish(cqCommandQueue);
       }
+      else if (tile.task == RenderTile::BAKE) {
+        bake(*task, tile);
+      }
       else if (tile.task == RenderTile::DENOISE) {
         tile.sample = tile.start_sample + tile.num_samples;
         denoise(tile, denoising);
@@ -1858,10 +1861,7 @@ void OpenCLDevice::shader(DeviceTask &task)
   cl_int d_offset = task.offset;
 
   OpenCLDevice::OpenCLProgram *program = &background_program;
-  if (task.shader_eval_type >= SHADER_EVAL_BAKE) {
-    program = &bake_program;
-  }
-  else if (task.shader_eval_type == SHADER_EVAL_DISPLACE) {
+  if (task.shader_eval_type == SHADER_EVAL_DISPLACE) {
     program = &displace_program;
   }
   program->wait_for_availability();
@@ -1892,6 +1892,51 @@ void OpenCLDevice::shader(DeviceTask &task)
   }
 }
 
+void OpenCLDevice::bake(DeviceTask &task, RenderTile &rtile)
+{
+  scoped_timer timer(&rtile.buffers->render_time);
+
+  /* Cast arguments to cl types. */
+  cl_mem d_data = CL_MEM_PTR(const_mem_map["__data"]->device_pointer);
+  cl_mem d_buffer = CL_MEM_PTR(rtile.buffer);
+  cl_int d_x = rtile.x;
+  cl_int d_y = rtile.y;
+  cl_int d_w = rtile.w;
+  cl_int d_h = rtile.h;
+  cl_int d_offset = rtile.offset;
+  cl_int d_stride = rtile.stride;
+
+  bake_program.wait_for_availability();
+  cl_kernel kernel = bake_program();
+
+  cl_uint start_arg_index = kernel_set_args(kernel, 0, d_data, d_buffer);
+
+  set_kernel_arg_buffers(kernel, &start_arg_index);
+
+  start_arg_index += kernel_set_args(
+      kernel, start_arg_index, d_x, d_y, d_w, d_h, d_offset, d_stride);
+
+  int start_sample = rtile.start_sample;
+  int end_sample = rtile.start_sample + rtile.num_samples;
+
+  for (int sample = start_sample; sample < end_sample; sample++) {
+    if (task.get_cancel()) {
+      if (task.need_finish_queue == false)
+        break;
+    }
+
+    kernel_set_args(kernel, start_arg_index, sample);
+
+    enqueue_kernel(kernel, d_w, d_h);
+
+    rtile.sample = sample + 1;
+
+    task.update_progress(&rtile, rtile.w * rtile.h);
+  }
+
+  clFinish(cqCommandQueue);
+}
+
 string OpenCLDevice::kernel_build_options(const string *debug_src)
 {
   string build_options = "-cl-no-signed-zeros -cl-mad-enable ";