Cycles: support for motion vector and UV passes.

Most of the changes are related to adding support for motion data throughout
the code. There's some code for actual camera/object motion blur raytracing
but it's unfinished (it badly slows down the raytracing kernel even when the
option is turned off), so that code it disabled still.

Motion vector export from Blender tries to avoid computing derived meshes
when the mesh does not have a deforming modifier, and it also won't store
motion vectors for every vertex if only the object or camera is moving.

1 files changed, 43 insertions, 33 deletions

diff --git a/intern/cycles/util/util_math.h b/intern/cycles/util/util_math.h
index 53c1302b4a1..f09803d8b09 100644
--- a/intern/cycles/util/util_math.h
+++ b/intern/cycles/util/util_math.h
@@ -55,6 +55,10 @@ CCL_NAMESPACE_BEGIN
 #ifndef M_2_PI_F
 #define M_2_PI_F	((float)0.636619772367581343075535053490057448)
 #endif
+#ifndef M_SQRT2_F
+#define M_SQRT2_F	((float)1.41421356237309504880)
+#endif
+
 
 /* Scalar */
 
@@ -719,6 +723,45 @@ __device_inline float4 cross(const float4& a, const float4& b)
 #endif
 }
 
+__device_inline bool is_zero(const float4& a)
+{
+#ifdef __KERNEL_SSE__
+	return a == make_float4(0.0f);
+#else
+	return (a.x == 0.0f && a.y == 0.0f && a.z == 0.0f && a.w == 0.0f);
+#endif
+}
+
+__device_inline float reduce_add(const float4& a)
+{
+#ifdef __KERNEL_SSE__
+	float4 h = shuffle<1,0,3,2>(a) + a;
+	return _mm_cvtss_f32(shuffle<2,3,0,1>(h) + h); /* todo: efficiency? */
+#else
+	return ((a.x + a.y) + (a.z + a.w));
+#endif
+}
+
+__device_inline float average(const float4& a)
+{
+	return reduce_add(a) * 0.25f;
+}
+
+__device_inline float dot(const float4& a, const float4& b)
+{
+	return reduce_add(a * b);
+}
+
+__device_inline float len(const float4 a)
+{
+	return sqrtf(dot(a, a));
+}
+
+__device_inline float4 normalize(const float4 a)
+{
+	return a/len(a);
+}
+
 __device_inline float4 min(float4 a, float4 b)
 {
 #ifdef __KERNEL_SSE__
@@ -790,39 +833,6 @@ __device_inline void print_float4(const char *label, const float4& a)
 
 #endif
 
-#ifndef __KERNEL_OPENCL__
-
-__device_inline bool is_zero(const float4& a)
-{
-#ifdef __KERNEL_SSE__
-	return a == make_float4(0.0f);
-#else
-	return (a.x == 0.0f && a.y == 0.0f && a.z == 0.0f && a.w == 0.0f);
-#endif
-}
-
-__device_inline float reduce_add(const float4& a)
-{
-#ifdef __KERNEL_SSE__
-	float4 h = shuffle<1,0,3,2>(a) + a;
-	return _mm_cvtss_f32(shuffle<2,3,0,1>(h) + h); /* todo: efficiency? */
-#else
-	return ((a.x + a.y) + (a.z + a.w));
-#endif
-}
-
-__device_inline float average(const float4& a)
-{
-	return reduce_add(a) * 0.25f;
-}
-
-__device_inline float dot(const float4& a, const float4& b)
-{
-	return reduce_add(a * b);
-}
-
-#endif
-
 /* Int3 */
 
 #ifndef __KERNEL_OPENCL__