Code refactor: split displace/background into separate kernels, remove luma.

8 files changed, 22 insertions, 30 deletions

diff --git a/intern/cycles/device/device_cpu.cpp b/intern/cycles/device/device_cpu.cpp
index ff34f4f9ce4..19e3c0a9075 100644
--- a/intern/cycles/device/device_cpu.cpp
+++ b/intern/cycles/device/device_cpu.cpp
@@ -171,10 +171,10 @@ public:
 
 	DeviceRequestedFeatures requested_features;
 
-	KernelFunctions<void(*)(KernelGlobals *, float *, int, int, int, int, int)>   path_trace_kernel;
-	KernelFunctions<void(*)(KernelGlobals *, uchar4 *, float *, float, int, int, int, int)>       convert_to_half_float_kernel;
-	KernelFunctions<void(*)(KernelGlobals *, uchar4 *, float *, float, int, int, int, int)>       convert_to_byte_kernel;
-	KernelFunctions<void(*)(KernelGlobals *, uint4 *, float4 *, float*, int, int, int, int, int)> shader_kernel;
+	KernelFunctions<void(*)(KernelGlobals *, float *, int, int, int, int, int)>             path_trace_kernel;
+	KernelFunctions<void(*)(KernelGlobals *, uchar4 *, float *, float, int, int, int, int)> convert_to_half_float_kernel;
+	KernelFunctions<void(*)(KernelGlobals *, uchar4 *, float *, float, int, int, int, int)> convert_to_byte_kernel;
+	KernelFunctions<void(*)(KernelGlobals *, uint4 *, float4 *, int, int, int, int, int)>   shader_kernel;
 
 	KernelFunctions<void(*)(int, TilesInfo*, int, int, float*, float*, float*, float*, float*, int*, int, int)> filter_divide_shadow_kernel;
 	KernelFunctions<void(*)(int, TilesInfo*, int, int, int, int, float*, float*, int*, int, int)>               filter_get_feature_kernel;
@@ -756,7 +756,6 @@ public:
 				shader_kernel()(&kg,
 				                (uint4*)task.shader_input,
 				                (float4*)task.shader_output,
-				                (float*)task.shader_output_luma,
 				                task.shader_eval_type,
 				                task.shader_filter,
 				                x,
diff --git a/intern/cycles/device/device_cuda.cpp b/intern/cycles/device/device_cuda.cpp
index 8cfc5332e94..734edcff503 100644
--- a/intern/cycles/device/device_cuda.cpp
+++ b/intern/cycles/device/device_cuda.cpp
@@ -1424,14 +1424,16 @@ public:
 		CUfunction cuShader;
 		CUdeviceptr d_input = cuda_device_ptr(task.shader_input);
 		CUdeviceptr d_output = cuda_device_ptr(task.shader_output);
-		CUdeviceptr d_output_luma = cuda_device_ptr(task.shader_output_luma);
 
 		/* 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) {
+			cuda_assert(cuModuleGetFunction(&cuShader, cuModule, "kernel_cuda_displace"));
+		}
 		else {
-			cuda_assert(cuModuleGetFunction(&cuShader, cuModule, "kernel_cuda_shader"));
+			cuda_assert(cuModuleGetFunction(&cuShader, cuModule, "kernel_cuda_background"));
 		}
 
 		/* do tasks in smaller chunks, so we can cancel it */
@@ -1450,9 +1452,6 @@ public:
 				int arg = 0;
 				args[arg++] = &d_input;
 				args[arg++] = &d_output;
-				if(task.shader_eval_type < SHADER_EVAL_BAKE) {
-					args[arg++] = &d_output_luma;
-				}
 				args[arg++] = &task.shader_eval_type;
 				if(task.shader_eval_type >= SHADER_EVAL_BAKE) {
 					args[arg++] = &task.shader_filter;
diff --git a/intern/cycles/device/device_multi.cpp b/intern/cycles/device/device_multi.cpp
index 164ed50bdf6..b17b972b06f 100644
--- a/intern/cycles/device/device_multi.cpp
+++ b/intern/cycles/device/device_multi.cpp
@@ -383,7 +383,6 @@ public:
 				if(task.rgba_half) subtask.rgba_half = sub.ptr_map[task.rgba_half];
 				if(task.shader_input) subtask.shader_input = sub.ptr_map[task.shader_input];
 				if(task.shader_output) subtask.shader_output = sub.ptr_map[task.shader_output];
-				if(task.shader_output_luma) subtask.shader_output_luma = sub.ptr_map[task.shader_output_luma];
 
 				sub.device->task_add(subtask);
 			}
diff --git a/intern/cycles/device/device_network.cpp b/intern/cycles/device/device_network.cpp
index 4ff8647f66b..deea59f1d23 100644
--- a/intern/cycles/device/device_network.cpp
+++ b/intern/cycles/device/device_network.cpp
@@ -660,10 +660,6 @@ protected:
 			if(task.shader_output)
 				task.shader_output = device_ptr_from_client_pointer(task.shader_output);
 
-			if(task.shader_output_luma)
-				task.shader_output_luma = device_ptr_from_client_pointer(task.shader_output_luma);
-
-
 			task.acquire_tile = function_bind(&DeviceServer::task_acquire_tile, this, _1, _2);
 			task.release_tile = function_bind(&DeviceServer::task_release_tile, this, _1);
 			task.update_progress_sample = function_bind(&DeviceServer::task_update_progress_sample, this);
diff --git a/intern/cycles/device/device_network.h b/intern/cycles/device/device_network.h
index 7bfebaf5aec..3d3bd99dfe7 100644
--- a/intern/cycles/device/device_network.h
+++ b/intern/cycles/device/device_network.h
@@ -132,7 +132,7 @@ public:
 		archive & type & task.x & task.y & task.w & task.h;
 		archive & task.rgba_byte & task.rgba_half & task.buffer & task.sample & task.num_samples;
 		archive & task.offset & task.stride;
-		archive & task.shader_input & task.shader_output & task.shader_output_luma & task.shader_eval_type;
+		archive & task.shader_input & task.shader_output & task.shader_eval_type;
 		archive & task.shader_x & task.shader_w;
 		archive & task.need_finish_queue;
 	}
@@ -291,7 +291,7 @@ public:
 		*archive & type & task.x & task.y & task.w & task.h;
 		*archive & task.rgba_byte & task.rgba_half & task.buffer & task.sample & task.num_samples;
 		*archive & task.offset & task.stride;
-		*archive & task.shader_input & task.shader_output & task.shader_output_luma & task.shader_eval_type;
+		*archive & task.shader_input & task.shader_output & task.shader_eval_type;
 		*archive & task.shader_x & task.shader_w;
 		*archive & task.need_finish_queue;
 
diff --git a/intern/cycles/device/device_task.cpp b/intern/cycles/device/device_task.cpp
index 3bc4c310283..3c7d24fb5b7 100644
--- a/intern/cycles/device/device_task.cpp
+++ b/intern/cycles/device/device_task.cpp
@@ -31,7 +31,7 @@ CCL_NAMESPACE_BEGIN
 DeviceTask::DeviceTask(Type type_)
 : type(type_), x(0), y(0), w(0), h(0), rgba_byte(0), rgba_half(0), buffer(0),
   sample(0), num_samples(1),
-  shader_input(0), shader_output(0), shader_output_luma(0),
+  shader_input(0), shader_output(0),
   shader_eval_type(0), shader_filter(0), shader_x(0), shader_w(0)
 {
 	last_update_time = time_dt();
diff --git a/intern/cycles/device/device_task.h b/intern/cycles/device/device_task.h
index 44a1efff1f5..b9658eb978f 100644
--- a/intern/cycles/device/device_task.h
+++ b/intern/cycles/device/device_task.h
@@ -46,7 +46,7 @@ public:
 	int offset, stride;
 
 	device_ptr shader_input;
-	device_ptr shader_output, shader_output_luma;
+	device_ptr shader_output;
 	int shader_eval_type;
 	int shader_filter;
 	int shader_x, shader_w;
diff --git a/intern/cycles/device/opencl/opencl_base.cpp b/intern/cycles/device/opencl/opencl_base.cpp
index 8095611f099..3db3efd1103 100644
--- a/intern/cycles/device/opencl/opencl_base.cpp
+++ b/intern/cycles/device/opencl/opencl_base.cpp
@@ -228,7 +228,8 @@ bool OpenCLDeviceBase::load_kernels(const DeviceRequestedFeatures& requested_fea
 	base_program = OpenCLProgram(this, "base", "kernel.cl", build_options_for_base_program(requested_features));
 	base_program.add_kernel(ustring("convert_to_byte"));
 	base_program.add_kernel(ustring("convert_to_half_float"));
-	base_program.add_kernel(ustring("shader"));
+	base_program.add_kernel(ustring("displace"));
+	base_program.add_kernel(ustring("background"));
 	base_program.add_kernel(ustring("bake"));
 	base_program.add_kernel(ustring("zero_buffer"));
 
@@ -1112,7 +1113,6 @@ void OpenCLDeviceBase::shader(DeviceTask& task)
 	cl_mem d_data = CL_MEM_PTR(const_mem_map["__data"]->device_pointer);
 	cl_mem d_input = CL_MEM_PTR(task.shader_input);
 	cl_mem d_output = CL_MEM_PTR(task.shader_output);
-	cl_mem d_output_luma = CL_MEM_PTR(task.shader_output_luma);
 	cl_int d_shader_eval_type = task.shader_eval_type;
 	cl_int d_shader_filter = task.shader_filter;
 	cl_int d_shader_x = task.shader_x;
@@ -1121,10 +1121,15 @@ void OpenCLDeviceBase::shader(DeviceTask& task)
 
 	cl_kernel kernel;
 
-	if(task.shader_eval_type >= SHADER_EVAL_BAKE)
+	if(task.shader_eval_type >= SHADER_EVAL_BAKE) {
 		kernel = base_program(ustring("bake"));
-	else
-		kernel = base_program(ustring("shader"));
+	}
+	else if(task.shader_eval_type >= SHADER_EVAL_DISPLACE) {
+		kernel = base_program(ustring("displace"));
+	}
+	else {
+		kernel = base_program(ustring("background"));
+	}
 
 	cl_uint start_arg_index =
 		kernel_set_args(kernel,
@@ -1133,12 +1138,6 @@ void OpenCLDeviceBase::shader(DeviceTask& task)
 		                d_input,
 		                d_output);
 
-	if(task.shader_eval_type < SHADER_EVAL_BAKE) {
-		start_arg_index += kernel_set_args(kernel,
-		                                   start_arg_index,
-		                                   d_output_luma);
-	}
-
 	set_kernel_arg_buffers(kernel, &start_arg_index);
 
 	start_arg_index += kernel_set_args(kernel,