Cycles Bake: show progress bar during bake

Baking progress preview is not possible, in parts due to the way the API
was designed. But at least you get to see the progress bar while baking.

Reviewers: sergey

Differential Revision: https://developer.blender.org/D656

intern/cycles/blender/blender_session.cpp

@@ -554,6 +554,8 @@ void BlenderSession::bake(BL::Object b_object, const string& pass_type, BL::Bake
 	session->reset(buffer_params, session_params.samples);
 	session->update_scene();
 
+	session->progress.set_update_callback(function_bind(&BlenderSession::update_bake_progress, this));
+
 	scene->bake_manager->bake(scene->device, &scene->dscene, scene, session->progress, shader_type, bake_data, result);
 
 	/* free all memory used (host and device), so we wouldn't leave render
@@ -765,6 +767,26 @@ void BlenderSession::get_progress(float& progress, double& total_time)
 		progress = 0.0;
 }
 
+void BlenderSession::update_bake_progress()
+{
+	float progress;
+	int sample, samples_per_task, parts_total;
+
+	sample = session->progress.get_sample();
+	samples_per_task = scene->bake_manager->num_samples;
+	parts_total = scene->bake_manager->num_parts;
+
+	if(samples_per_task)
+		progress = ((float)sample / (float)(parts_total * samples_per_task));
+	else
+		progress = 0.0;
+
+	if(progress != last_progress) {
+		b_engine.update_progress(progress);
+		last_progress = progress;
+	}
+}
+
 void BlenderSession::update_status_progress()
 {
 	string timestatus, status, substatus;

intern/cycles/blender/blender_session.h

@@ -73,6 +73,7 @@ public:
 	void get_progress(float& progress, double& total_time);
 	void test_cancel();
 	void update_status_progress();
+	void update_bake_progress();
 
 	bool background;
 	Session *session;

intern/cycles/device/device.h

@@ -122,6 +122,7 @@ public:
 	virtual bool load_kernels(bool experimental) { return true; }
 
 	/* tasks */
+	virtual int get_split_task_count(DeviceTask& task) = 0;
 	virtual void task_add(DeviceTask& task) = 0;
 	virtual void task_wait() = 0;
 	virtual void task_cancel() = 0;

intern/cycles/device/device_cpu.cpp

@@ -185,7 +185,7 @@ public:
 
 					tile.sample = sample + 1;
 
-					task.update_progress(tile);
+					task.update_progress(&tile);
 				}
 			}
 			else
@@ -207,7 +207,7 @@ public:
 
 					tile.sample = sample + 1;
 
-					task.update_progress(tile);
+					task.update_progress(&tile);
 				}
 			}
 			else
@@ -229,7 +229,7 @@ public:
 
 					tile.sample = sample + 1;
 
-					task.update_progress(tile);
+					task.update_progress(&tile);
 				}
 			}
 			else
@@ -251,7 +251,7 @@ public:
 
 					tile.sample = sample + 1;
 
-					task.update_progress(tile);
+					task.update_progress(&tile);
 				}
 			}
 			else
@@ -273,7 +273,7 @@ public:
 
 					tile.sample = sample + 1;
 
-					task.update_progress(tile);
+					task.update_progress(&tile);
 				}
 			}
 			else
@@ -294,7 +294,7 @@ public:
 
 					tile.sample = sample + 1;
 
-					task.update_progress(tile);
+					task.update_progress(&tile);
 				}
 			}
 
@@ -433,71 +433,83 @@ public:
 
 #ifdef WITH_CYCLES_OPTIMIZED_KERNEL_AVX2
 		if(system_cpu_support_avx2()) {
-			for(int x = task.shader_x; x < task.shader_x + task.shader_w; x++) {
+			for(int sample = 0; sample < task.num_samples; sample++) {
-				for(int sample = 0; sample < task.num_samples; sample++)
+				for(int x = task.shader_x; x < task.shader_x + task.shader_w; x++)
 					kernel_cpu_avx2_shader(&kg, (uint4*)task.shader_input, (float4*)task.shader_output, task.shader_eval_type, x, sample);
 
 				if(task.get_cancel() || task_pool.canceled())
 					break;
+
+				task.update_progress(NULL);
 			}
 		}
 		else
 #endif
 #ifdef WITH_CYCLES_OPTIMIZED_KERNEL_AVX
 		if(system_cpu_support_avx()) {
-			for(int x = task.shader_x; x < task.shader_x + task.shader_w; x++) {
+			for(int sample = 0; sample < task.num_samples; sample++) {
-				for(int sample = 0; sample < task.num_samples; sample++)
+				for(int x = task.shader_x; x < task.shader_x + task.shader_w; x++)
 					kernel_cpu_avx_shader(&kg, (uint4*)task.shader_input, (float4*)task.shader_output, task.shader_eval_type, x, sample);
 
 				if(task.get_cancel() || task_pool.canceled())
 					break;
+
+				task.update_progress(NULL);
 			}
 		}
 		else
 #endif
 #ifdef WITH_CYCLES_OPTIMIZED_KERNEL_SSE41			
 		if(system_cpu_support_sse41()) {
-			for(int x = task.shader_x; x < task.shader_x + task.shader_w; x++) {
+			for(int sample = 0; sample < task.num_samples; sample++) {
-				for(int sample = 0; sample < task.num_samples; sample++)
+				for(int x = task.shader_x; x < task.shader_x + task.shader_w; x++)
 					kernel_cpu_sse41_shader(&kg, (uint4*)task.shader_input, (float4*)task.shader_output, task.shader_eval_type, x, sample);
 
 				if(task.get_cancel() || task_pool.canceled())
 					break;
+
+				task.update_progress(NULL);
 			}
 		}
 		else
 #endif
 #ifdef WITH_CYCLES_OPTIMIZED_KERNEL_SSE3
 		if(system_cpu_support_sse3()) {
-			for(int x = task.shader_x; x < task.shader_x + task.shader_w; x++) {
+			for(int sample = 0; sample < task.num_samples; sample++) {
-				for(int sample = 0; sample < task.num_samples; sample++)
+				for(int x = task.shader_x; x < task.shader_x + task.shader_w; x++)
 					kernel_cpu_sse3_shader(&kg, (uint4*)task.shader_input, (float4*)task.shader_output, task.shader_eval_type, x, sample);
 
 				if(task.get_cancel() || task_pool.canceled())
 					break;
+
+				task.update_progress(NULL);
 			}
 		}
 		else
 #endif
 #ifdef WITH_CYCLES_OPTIMIZED_KERNEL_SSE2
 		if(system_cpu_support_sse2()) {
-			for(int x = task.shader_x; x < task.shader_x + task.shader_w; x++) {
+			for(int sample = 0; sample < task.num_samples; sample++) {
-				for(int sample = 0; sample < task.num_samples; sample++)
+				for(int x = task.shader_x; x < task.shader_x + task.shader_w; x++)
 					kernel_cpu_sse2_shader(&kg, (uint4*)task.shader_input, (float4*)task.shader_output, task.shader_eval_type, x, sample);
 
 				if(task.get_cancel() || task_pool.canceled())
 					break;
+
+				task.update_progress(NULL);
 			}
 		}
 		else
 #endif
 		{
-			for(int x = task.shader_x; x < task.shader_x + task.shader_w; x++) {
+			for(int sample = 0; sample < task.num_samples; sample++) {
-				for(int sample = 0; sample < task.num_samples; sample++)
+				for(int x = task.shader_x; x < task.shader_x + task.shader_w; x++)
 					kernel_cpu_shader(&kg, (uint4*)task.shader_input, (float4*)task.shader_output, task.shader_eval_type, x, sample);
 
 				if(task.get_cancel() || task_pool.canceled())
 					break;
+
+				task.update_progress(NULL);
 			}
 		}
 
@@ -506,6 +518,14 @@ public:
 #endif
 	}
 
+	int get_split_task_count(DeviceTask& task)
+	{
+		if (task.type == DeviceTask::SHADER)
+			return task.get_subtask_count(TaskScheduler::num_threads(), 256);
+		else
+			return task.get_subtask_count(TaskScheduler::num_threads());
+	}
+
 	void task_add(DeviceTask& task)
 	{
 		/* split task into smaller ones */

intern/cycles/device/device_cuda.cpp

@@ -732,14 +732,11 @@ public:
 		const int start = task.shader_x;
 		const int end = task.shader_x + task.shader_w;
 
+		bool cancelled = false;
+		for(int sample = 0; sample < task.num_samples && !cancelled; sample++) {
 			for(int shader_x = start; shader_x < end; shader_x += shader_chunk_size) {
-			if(task.get_cancel())
-				break;
-
 				int shader_w = min(shader_chunk_size, end - shader_x);
 
-			for(int sample = 0; sample < task.num_samples; sample++) {
-
 				/* pass in parameters */
 				void *args[] = {&d_input,
 								 &d_output,
@@ -761,9 +758,16 @@ public:
 										   0, 0, args, 0));
 
 				cuda_assert(cuCtxSynchronize());
+
+				if(task.get_cancel()) {
+					cancelled = false;
+					break;
 				}
 			}
 
+			task.update_progress(NULL);
+		}
+
 		cuda_pop_context();
 	}
 
@@ -991,7 +995,7 @@ public:
 
 					tile.sample = sample + 1;
 
-					task->update_progress(tile);
+					task->update_progress(&tile);
 				}
 
 				task->release_tile(tile);
@@ -1015,6 +1019,11 @@ public:
 		}
 	};
 
+	int get_split_task_count(DeviceTask& task)
+	{
+		return 1;
+	}
+
 	void task_add(DeviceTask& task)
 	{
 		if(task.type == DeviceTask::FILM_CONVERT) {

intern/cycles/device/device_multi.cpp

@@ -278,6 +278,11 @@ public:
 		return -1;
 	}
 
+	int get_split_task_count(DeviceTask& task)
+	{
+		return 1;
+	}
+
 	void task_add(DeviceTask& task)
 	{
 		list<DeviceTask> tasks;

intern/cycles/device/device_opencl.cpp

@@ -1068,7 +1068,11 @@ public:
 			kernel = ckShaderKernel;
 
 		for(int sample = 0; sample < task.num_samples; sample++) {
-			cl_int d_sample = task.sample;
+
+			if(task.get_cancel())
+				break;
+
+			cl_int d_sample = sample;
 
 			opencl_assert(clSetKernelArg(kernel, narg++, sizeof(d_data), (void*)&d_data));
 			opencl_assert(clSetKernelArg(kernel, narg++, sizeof(d_input), (void*)&d_input));
@@ -1084,6 +1088,8 @@ public:
 			opencl_assert(clSetKernelArg(kernel, narg++, sizeof(d_sample), (void*)&d_sample));
 
 			enqueue_kernel(kernel, task.shader_w, 1);
+
+			task.update_progress(NULL);
 		}
 	}
 
@@ -1113,7 +1119,7 @@ public:
 
 					tile.sample = sample + 1;
 
-					task->update_progress(tile);
+					task->update_progress(&tile);
 				}
 
 				task->release_tile(tile);
@@ -1130,6 +1136,11 @@ public:
 		}
 	};
 
+	int get_split_task_count(DeviceTask& task)
+	{
+		return 1;
+	}
+
 	void task_add(DeviceTask& task)
 	{
 		task_pool.push(new OpenCLDeviceTask(this, task));

intern/cycles/device/device_task.cpp

@@ -35,7 +35,7 @@ DeviceTask::DeviceTask(Type type_)
 	last_update_time = time_dt();
 }
 
-void DeviceTask::split(list<DeviceTask>& tasks, int num, int max_size)
+int DeviceTask::get_subtask_count(int num, int max_size)
 {
 	if(max_size != 0) {
 		int max_size_num;
@@ -53,7 +53,21 @@ void DeviceTask::split(list<DeviceTask>& tasks, int num, int max_size)
 
 	if(type == SHADER) {
 		num = min(shader_w, num);
+	}
+	else if(type == PATH_TRACE) {
+	}
+	else {
+		num = min(h, num);
+	}
 
+	return num;
+}
+
+void DeviceTask::split(list<DeviceTask>& tasks, int num, int max_size)
+{
+	num = get_subtask_count(num, max_size);
+
+	if(type == SHADER) {
 		for(int i = 0; i < num; i++) {
 			int tx = shader_x + (shader_w/num)*i;
 			int tw = (i == num-1)? shader_w - i*(shader_w/num): shader_w/num;
@@ -71,8 +85,6 @@ void DeviceTask::split(list<DeviceTask>& tasks, int num, int max_size)
 			tasks.push_back(*this);
 	}
 	else {
-		num = min(h, num);
-
 		for(int i = 0; i < num; i++) {
 			int ty = y + (h/num)*i;
 			int th = (i == num-1)? h - i*(h/num): h/num;
@@ -87,9 +99,10 @@ void DeviceTask::split(list<DeviceTask>& tasks, int num, int max_size)
 	}
 }
 
-void DeviceTask::update_progress(RenderTile &rtile)
+void DeviceTask::update_progress(RenderTile *rtile)
 {
-	if (type != PATH_TRACE)
+	if((type != PATH_TRACE) &&
+	   (type != SHADER))
 		return;
 
 	if(update_progress_sample)
@@ -99,7 +112,7 @@ void DeviceTask::update_progress(RenderTile &rtile)
 		double current_time = time_dt();
 
 		if (current_time - last_update_time >= 1.0) {
-			update_tile_sample(rtile);
+			update_tile_sample(*rtile);
 
 			last_update_time = current_time;
 		}

intern/cycles/device/device_task.h

@@ -52,9 +52,10 @@ public:
 
 	DeviceTask(Type type = PATH_TRACE);
 
+	int get_subtask_count(int num, int max_size = 0);
 	void split(list<DeviceTask>& tasks, int num, int max_size = 0);
 
-	void update_progress(RenderTile &rtile);
+	void update_progress(RenderTile *rtile);
 
 	boost::function<bool(Device *device, RenderTile&)> acquire_tile;
 	boost::function<void(void)> update_progress_sample;

intern/cycles/render/bake.cpp

@@ -155,6 +155,10 @@ bool BakeManager::bake(Device *device, DeviceScene *dscene, Scene *scene, Progre
 	task.shader_w = d_output.size();
 	task.num_samples = is_aa_pass(shader_type)? scene->integrator->aa_samples: 1;
 	task.get_cancel = function_bind(&Progress::get_cancel, &progress);
+	task.update_progress_sample = function_bind(&Progress::increment_sample_update, &progress);
+
+	this->num_parts = device->get_split_task_count(task);
+	this->num_samples = task.num_samples;
 
 	device->task_add(task);
 	device->task_wait();

intern/cycles/render/bake.h

@@ -70,6 +70,9 @@ public:
 
 	bool need_update;
 
+	int num_samples;
+	int num_parts;
+
 private:
 	BakeData *m_bake_data;
 	bool m_is_baking;

intern/cycles/util/util_progress.h

@@ -149,6 +149,12 @@ public:
 		sample++;
 	}
 
+	void increment_sample_update()
+	{
+		increment_sample();
+		set_update();
+	}
+
 	int get_sample()
 	{
 		return sample;