Merge branch 'master' into cycles_disney_brdf

5 files changed, 75 insertions, 28 deletions

diff --git a/intern/cycles/kernel/kernels/cpu/kernel.cpp b/intern/cycles/kernel/kernels/cpu/kernel.cpp
index 365ce891354..f11c85d5f6a 100644
--- a/intern/cycles/kernel/kernels/cpu/kernel.cpp
+++ b/intern/cycles/kernel/kernels/cpu/kernel.cpp
@@ -95,7 +95,7 @@ void kernel_tex_copy(KernelGlobals *kg,
 		int id = atoi(name + strlen("__tex_image_float4_"));
 		int array_index = id;
 
-		if(array_index >= 0 && array_index < TEX_NUM_FLOAT4_IMAGES_CPU) {
+		if(array_index >= 0 && array_index < TEX_NUM_FLOAT4_CPU) {
 			tex = &kg->texture_float4_images[array_index];
 		}
 
@@ -109,9 +109,9 @@ void kernel_tex_copy(KernelGlobals *kg,
 	else if(strstr(name, "__tex_image_float")) {
 		texture_image_float *tex = NULL;
 		int id = atoi(name + strlen("__tex_image_float_"));
-		int array_index = id - TEX_IMAGE_FLOAT_START_CPU;
+		int array_index = id - TEX_START_FLOAT_CPU;
 
-		if(array_index >= 0 && array_index < TEX_NUM_FLOAT_IMAGES_CPU) {
+		if(array_index >= 0 && array_index < TEX_NUM_FLOAT_CPU) {
 			tex = &kg->texture_float_images[array_index];
 		}
 
@@ -125,9 +125,9 @@ void kernel_tex_copy(KernelGlobals *kg,
 	else if(strstr(name, "__tex_image_byte4")) {
 		texture_image_uchar4 *tex = NULL;
 		int id = atoi(name + strlen("__tex_image_byte4_"));
-		int array_index = id - TEX_IMAGE_BYTE4_START_CPU;
+		int array_index = id - TEX_START_BYTE4_CPU;
 
-		if(array_index >= 0 && array_index < TEX_NUM_BYTE4_IMAGES_CPU) {
+		if(array_index >= 0 && array_index < TEX_NUM_BYTE4_CPU) {
 			tex = &kg->texture_byte4_images[array_index];
 		}
 
@@ -141,9 +141,9 @@ void kernel_tex_copy(KernelGlobals *kg,
 	else if(strstr(name, "__tex_image_byte")) {
 		texture_image_uchar *tex = NULL;
 		int id = atoi(name + strlen("__tex_image_byte_"));
-		int array_index = id - TEX_IMAGE_BYTE_START_CPU;
+		int array_index = id - TEX_START_BYTE_CPU;
 
-		if(array_index >= 0 && array_index < TEX_NUM_BYTE_IMAGES_CPU) {
+		if(array_index >= 0 && array_index < TEX_NUM_BYTE_CPU) {
 			tex = &kg->texture_byte_images[array_index];
 		}
 
@@ -154,6 +154,38 @@ void kernel_tex_copy(KernelGlobals *kg,
 			tex->extension = extension;
 		}
 	}
+	else if(strstr(name, "__tex_image_half4")) {
+		texture_image_half4 *tex = NULL;
+		int id = atoi(name + strlen("__tex_image_half4_"));
+		int array_index = id - TEX_START_HALF4_CPU;
+
+		if(array_index >= 0 && array_index < TEX_NUM_HALF4_CPU) {
+			tex = &kg->texture_half4_images[array_index];
+		}
+
+		if(tex) {
+			tex->data = (half4*)mem;
+			tex->dimensions_set(width, height, depth);
+			tex->interpolation = interpolation;
+			tex->extension = extension;
+		}
+	}
+	else if(strstr(name, "__tex_image_half")) {
+		texture_image_half *tex = NULL;
+		int id = atoi(name + strlen("__tex_image_half_"));
+		int array_index = id - TEX_START_HALF_CPU;
+
+		if(array_index >= 0 && array_index < TEX_NUM_HALF_CPU) {
+			tex = &kg->texture_half_images[array_index];
+		}
+
+		if(tex) {
+			tex->data = (half*)mem;
+			tex->dimensions_set(width, height, depth);
+			tex->interpolation = interpolation;
+			tex->extension = extension;
+		}
+	}
 	else
 		assert(0);
 }
diff --git a/intern/cycles/kernel/kernels/cpu/kernel_cpu_image.h b/intern/cycles/kernel/kernels/cpu/kernel_cpu_image.h
index 4807f96a0df..47383140170 100644
--- a/intern/cycles/kernel/kernels/cpu/kernel_cpu_image.h
+++ b/intern/cycles/kernel/kernels/cpu/kernel_cpu_image.h
@@ -23,24 +23,32 @@ CCL_NAMESPACE_BEGIN
 
 ccl_device float4 kernel_tex_image_interp_impl(KernelGlobals *kg, int tex, float x, float y)
 {
-	if(tex >= TEX_IMAGE_BYTE_START_CPU)
-		return kg->texture_byte_images[tex - TEX_IMAGE_BYTE_START_CPU].interp(x, y);
-	else if(tex >= TEX_IMAGE_FLOAT_START_CPU)
-		return kg->texture_float_images[tex - TEX_IMAGE_FLOAT_START_CPU].interp(x, y);
-	else if(tex >= TEX_IMAGE_BYTE4_START_CPU)
-		return kg->texture_byte4_images[tex - TEX_IMAGE_BYTE4_START_CPU].interp(x, y);
+	if(tex >= TEX_START_HALF_CPU)
+		return kg->texture_half_images[tex - TEX_START_HALF_CPU].interp(x, y);
+	else if(tex >= TEX_START_HALF4_CPU)
+		return kg->texture_half4_images[tex - TEX_START_HALF4_CPU].interp(x, y);
+	else if(tex >= TEX_START_BYTE_CPU)
+		return kg->texture_byte_images[tex - TEX_START_BYTE_CPU].interp(x, y);
+	else if(tex >= TEX_START_FLOAT_CPU)
+		return kg->texture_float_images[tex - TEX_START_FLOAT_CPU].interp(x, y);
+	else if(tex >= TEX_START_BYTE4_CPU)
+		return kg->texture_byte4_images[tex - TEX_START_BYTE4_CPU].interp(x, y);
 	else
 		return kg->texture_float4_images[tex].interp(x, y);
 }
 
 ccl_device float4 kernel_tex_image_interp_3d_impl(KernelGlobals *kg, int tex, float x, float y, float z)
 {
-	if(tex >= TEX_IMAGE_BYTE_START_CPU)
-		return kg->texture_byte_images[tex - TEX_IMAGE_BYTE_START_CPU].interp_3d(x, y, z);
-	else if(tex >= TEX_IMAGE_FLOAT_START_CPU)
-		return kg->texture_float_images[tex - TEX_IMAGE_FLOAT_START_CPU].interp_3d(x, y, z);
-	else if(tex >= TEX_IMAGE_BYTE4_START_CPU)
-		return kg->texture_byte4_images[tex - TEX_IMAGE_BYTE4_START_CPU].interp_3d(x, y, z);
+	if(tex >= TEX_START_HALF_CPU)
+		return kg->texture_half_images[tex - TEX_START_HALF_CPU].interp_3d(x, y, z);
+	else if(tex >= TEX_START_HALF4_CPU)
+		return kg->texture_half4_images[tex - TEX_START_HALF4_CPU].interp_3d(x, y, z);
+	else if(tex >= TEX_START_BYTE_CPU)
+		return kg->texture_byte_images[tex - TEX_START_BYTE_CPU].interp_3d(x, y, z);
+	else if(tex >= TEX_START_FLOAT_CPU)
+		return kg->texture_float_images[tex - TEX_START_FLOAT_CPU].interp_3d(x, y, z);
+	else if(tex >= TEX_START_BYTE4_CPU)
+		return kg->texture_byte4_images[tex - TEX_START_BYTE4_CPU].interp_3d(x, y, z);
 	else
 		return kg->texture_float4_images[tex].interp_3d(x, y, z);
 
@@ -48,12 +56,16 @@ ccl_device float4 kernel_tex_image_interp_3d_impl(KernelGlobals *kg, int tex, fl
 
 ccl_device float4 kernel_tex_image_interp_3d_ex_impl(KernelGlobals *kg, int tex, float x, float y, float z, int interpolation)
 {
-	if(tex >= TEX_IMAGE_BYTE_START_CPU)
-		return kg->texture_byte_images[tex - TEX_IMAGE_BYTE_START_CPU].interp_3d_ex(x, y, z, interpolation);
-	else if(tex >= TEX_IMAGE_FLOAT_START_CPU)
-		return kg->texture_float_images[tex - TEX_IMAGE_FLOAT_START_CPU].interp_3d_ex(x, y, z, interpolation);
-	else if(tex >= TEX_IMAGE_BYTE4_START_CPU)
-		return kg->texture_byte4_images[tex - TEX_IMAGE_BYTE4_START_CPU].interp_3d_ex(x, y, z, interpolation);
+	if(tex >= TEX_START_HALF_CPU)
+		return kg->texture_half4_images[tex - TEX_START_HALF_CPU].interp_3d_ex(x, y, z, interpolation);
+	else if(tex >= TEX_START_HALF4_CPU)
+		return kg->texture_half_images[tex - TEX_START_HALF4_CPU].interp_3d_ex(x, y, z, interpolation);
+	else if(tex >= TEX_START_BYTE_CPU)
+		return kg->texture_byte_images[tex - TEX_START_BYTE_CPU].interp_3d_ex(x, y, z, interpolation);
+	else if(tex >= TEX_START_FLOAT_CPU)
+		return kg->texture_float_images[tex - TEX_START_FLOAT_CPU].interp_3d_ex(x, y, z, interpolation);
+	else if(tex >= TEX_START_BYTE4_CPU)
+		return kg->texture_byte4_images[tex - TEX_START_BYTE4_CPU].interp_3d_ex(x, y, z, interpolation);
 	else
 		return kg->texture_float4_images[tex].interp_3d_ex(x, y, z, interpolation);
 }
diff --git a/intern/cycles/kernel/kernels/cpu/kernel_cpu_impl.h b/intern/cycles/kernel/kernels/cpu/kernel_cpu_impl.h
index 962196ccbdd..ec82d4b4c22 100644
--- a/intern/cycles/kernel/kernels/cpu/kernel_cpu_impl.h
+++ b/intern/cycles/kernel/kernels/cpu/kernel_cpu_impl.h
@@ -109,6 +109,7 @@ void KERNEL_FUNCTION_FULL_NAME(shader)(KernelGlobals *kg,
 {
 	if(type >= SHADER_EVAL_BAKE) {
 		kernel_assert(output_luma == NULL);
+#ifdef __BAKING__
 		kernel_bake_evaluate(kg,
 		                     input,
 		                     output,
@@ -117,6 +118,7 @@ void KERNEL_FUNCTION_FULL_NAME(shader)(KernelGlobals *kg,
 		                     i,
 		                     offset,
 		                     sample);
+#endif
 	}
 	else {
 		kernel_shader_evaluate(kg,
diff --git a/intern/cycles/kernel/kernels/cuda/kernel.cu b/intern/cycles/kernel/kernels/cuda/kernel.cu
index 37fae54faf0..eb2b6ea5414 100644
--- a/intern/cycles/kernel/kernels/cuda/kernel.cu
+++ b/intern/cycles/kernel/kernels/cuda/kernel.cu
@@ -77,8 +77,8 @@
 #  define CUDA_KERNEL_MAX_REGISTERS 63
 #  define CUDA_KERNEL_BRANCHED_MAX_REGISTERS 63
 
-/* 5.0, 5.2 and 5.3 */
-#elif __CUDA_ARCH__ == 500 || __CUDA_ARCH__ == 520 || __CUDA_ARCH__ == 530
+/* 5.0, 5.2, 5.3, 6.0, 6.1 */
+#elif __CUDA_ARCH__ >= 500
 #  define CUDA_MULTIPRESSOR_MAX_REGISTERS 65536
 #  define CUDA_MULTIPROCESSOR_MAX_BLOCKS 32
 #  define CUDA_BLOCK_MAX_THREADS 1024
@@ -86,7 +86,7 @@
 
 /* tunable parameters */
 #  define CUDA_THREADS_BLOCK_WIDTH 16
-#  define CUDA_KERNEL_MAX_REGISTERS 40
+#  define CUDA_KERNEL_MAX_REGISTERS 48
 #  define CUDA_KERNEL_BRANCHED_MAX_REGISTERS 63
 
 /* unknown architecture */
diff --git a/intern/cycles/kernel/kernels/opencl/kernel.cl b/intern/cycles/kernel/kernels/opencl/kernel.cl
index aad06ed5c76..37907cd8fdc 100644
--- a/intern/cycles/kernel/kernels/opencl/kernel.cl
+++ b/intern/cycles/kernel/kernels/opencl/kernel.cl
@@ -35,6 +35,7 @@
 #  include "../../kernel_montecarlo.h"
 #  include "../../kernel_projection.h"
 #  include "../../geom/geom.h"
+#  include "../../bvh/bvh.h"
 
 #  include "../../kernel_accumulate.h"
 #  include "../../kernel_camera.h"