Cycles: Move triangle intersection functions into own file

This way extending intersection routines with some pre-calculation step wouldn't
explode the single file size, hopefully keeping them all in a nice maintainable
state.

4 files changed, 233 insertions, 208 deletions

diff --git a/intern/cycles/kernel/CMakeLists.txt b/intern/cycles/kernel/CMakeLists.txt
index 0ff227938ae..ca1065f114a 100644
--- a/intern/cycles/kernel/CMakeLists.txt
+++ b/intern/cycles/kernel/CMakeLists.txt
@@ -124,6 +124,7 @@ set(SRC_GEOM_HEADERS
 	geom/geom_object.h
 	geom/geom_primitive.h
 	geom/geom_triangle.h
+	geom/geom_triangle_intersect.h
 	geom/geom_volume.h
 )
 
diff --git a/intern/cycles/kernel/geom/geom.h b/intern/cycles/kernel/geom/geom.h
index 9495a2541f9..3a768f37dd9 100644
--- a/intern/cycles/kernel/geom/geom.h
+++ b/intern/cycles/kernel/geom/geom.h
@@ -35,6 +35,7 @@
 #include "geom_attribute.h"
 #include "geom_object.h"
 #include "geom_triangle.h"
+#include "geom_triangle_intersect.h"
 #include "geom_motion_triangle.h"
 #include "geom_motion_curve.h"
 #include "geom_curve.h"
diff --git a/intern/cycles/kernel/geom/geom_triangle.h b/intern/cycles/kernel/geom/geom_triangle.h
index c08a82ee038..0390804d131 100644
--- a/intern/cycles/kernel/geom/geom_triangle.h
+++ b/intern/cycles/kernel/geom/geom_triangle.h
@@ -17,105 +17,10 @@
 
 /* Triangle Primitive
  *
- * Basic triangle with 3 vertices is used to represent mesh surfaces. For BVH
- * ray intersection we use a precomputed triangle storage to accelerate
- * intersection at the cost of more memory usage */
+ * Basic triangle with 3 vertices is used to represent mesh surfaces. */
 
 CCL_NAMESPACE_BEGIN
 
-/* Refine triangle intersection to more precise hit point. For rays that travel
- * far the precision is often not so good, this reintersects the primitive from
- * a closer distance. */
-
-ccl_device_inline float3 triangle_refine(KernelGlobals *kg, ShaderData *sd, const Intersection *isect, const Ray *ray)
-{
-	float3 P = ray->P;
-	float3 D = ray->D;
-	float t = isect->t;
-
-#ifdef __INTERSECTION_REFINE__
-	if(isect->object != OBJECT_NONE) {
-#ifdef __OBJECT_MOTION__
-		Transform tfm = sd->ob_itfm;
-#else
-		Transform tfm = object_fetch_transform(kg, isect->object, OBJECT_INVERSE_TRANSFORM);
-#endif
-
-		P = transform_point(&tfm, P);
-		D = transform_direction(&tfm, D*t);
-		D = normalize_len(D, &t);
-	}
-
-	P = P + D*t;
-
-	float4 v00 = kernel_tex_fetch(__tri_woop, isect->prim*TRI_NODE_SIZE+0);
-	float Oz = v00.w - P.x*v00.x - P.y*v00.y - P.z*v00.z;
-	float invDz = 1.0f/(D.x*v00.x + D.y*v00.y + D.z*v00.z);
-	float rt = Oz * invDz;
-
-	P = P + D*rt;
-
-	if(isect->object != OBJECT_NONE) {
-#ifdef __OBJECT_MOTION__
-		Transform tfm = sd->ob_tfm;
-#else
-		Transform tfm = object_fetch_transform(kg, isect->object, OBJECT_TRANSFORM);
-#endif
-
-		P = transform_point(&tfm, P);
-	}
-
-	return P;
-#else
-	return P + D*t;
-#endif
-}
-
-/* same as above, except that isect->t is assumed to be in object space for instancing */
-ccl_device_inline float3 triangle_refine_subsurface(KernelGlobals *kg, ShaderData *sd, const Intersection *isect, const Ray *ray)
-{
-	float3 P = ray->P;
-	float3 D = ray->D;
-	float t = isect->t;
-
-#ifdef __INTERSECTION_REFINE__
-	if(isect->object != OBJECT_NONE) {
-#ifdef __OBJECT_MOTION__
-		Transform tfm = sd->ob_itfm;
-#else
-		Transform tfm = object_fetch_transform(kg, isect->object, OBJECT_INVERSE_TRANSFORM);
-#endif
-
-		P = transform_point(&tfm, P);
-		D = transform_direction(&tfm, D);
-		D = normalize(D);
-	}
-
-	P = P + D*t;
-
-	float4 v00 = kernel_tex_fetch(__tri_woop, isect->prim*TRI_NODE_SIZE+0);
-	float Oz = v00.w - P.x*v00.x - P.y*v00.y - P.z*v00.z;
-	float invDz = 1.0f/(D.x*v00.x + D.y*v00.y + D.z*v00.z);
-	float rt = Oz * invDz;
-
-	P = P + D*rt;
-
-	if(isect->object != OBJECT_NONE) {
-#ifdef __OBJECT_MOTION__
-		Transform tfm = sd->ob_tfm;
-#else
-		Transform tfm = object_fetch_transform(kg, isect->object, OBJECT_TRANSFORM);
-#endif
-
-		P = transform_point(&tfm, P);
-	}
-
-	return P;
-#else
-	return P + D*t;
-#endif
-}
-
 /* normal on triangle  */
 ccl_device_inline float3 triangle_normal(KernelGlobals *kg, ShaderData *sd)
 {
@@ -298,116 +203,4 @@ ccl_device float3 triangle_attribute_float3(KernelGlobals *kg, const ShaderData
 	}
 }
 
-/* Ray-Triangle intersection for BVH traversal
- *
- * Based on Sven Woop's algorithm with precomputed triangle storage */
-
-ccl_device_inline bool triangle_intersect(KernelGlobals *kg, Intersection *isect,
-	float3 P, float3 dir, uint visibility, int object, int triAddr)
-{
-	/* compute and check intersection t-value */
-	float4 v00 = kernel_tex_fetch(__tri_woop, triAddr*TRI_NODE_SIZE+0);
-	float4 v11 = kernel_tex_fetch(__tri_woop, triAddr*TRI_NODE_SIZE+1);
-
-	float Oz = v00.w - P.x*v00.x - P.y*v00.y - P.z*v00.z;
-	float invDz = 1.0f/(dir.x*v00.x + dir.y*v00.y + dir.z*v00.z);
-	float t = Oz * invDz;
-
-	if(t > 0.0f && t < isect->t) {
-		/* compute and check barycentric u */
-		float Ox = v11.w + P.x*v11.x + P.y*v11.y + P.z*v11.z;
-		float Dx = dir.x*v11.x + dir.y*v11.y + dir.z*v11.z;
-		float u = Ox + t*Dx;
-
-		if(u >= 0.0f) {
-			/* compute and check barycentric v */
-			float4 v22 = kernel_tex_fetch(__tri_woop, triAddr*TRI_NODE_SIZE+2);
-			float Oy = v22.w + P.x*v22.x + P.y*v22.y + P.z*v22.z;
-			float Dy = dir.x*v22.x + dir.y*v22.y + dir.z*v22.z;
-			float v = Oy + t*Dy;
-
-			if(v >= 0.0f && u + v <= 1.0f) {
-#ifdef __VISIBILITY_FLAG__
-				/* visibility flag test. we do it here under the assumption
-				 * that most triangles are culled by node flags */
-				if(kernel_tex_fetch(__prim_visibility, triAddr) & visibility)
-#endif
-				{
-					/* record intersection */
-					isect->t = t;
-					isect->u = u;
-					isect->v = v;
-					isect->prim = triAddr;
-					isect->object = object;
-					isect->type = PRIMITIVE_TRIANGLE;
-					return true;
-				}
-			}
-		}
-	}
-
-	return false;
-}
-
-/* Special ray intersection routines for subsurface scattering. In that case we
- * only want to intersect with primitives in the same object, and if case of
- * multiple hits we pick a single random primitive as the intersection point. */
-
-#ifdef __SUBSURFACE__
-ccl_device_inline void triangle_intersect_subsurface(KernelGlobals *kg, Intersection *isect_array,
-	float3 P, float3 dir, int object, int triAddr, float tmax, uint *num_hits, uint *lcg_state, int max_hits)
-{
-	/* compute and check intersection t-value */
-	float4 v00 = kernel_tex_fetch(__tri_woop, triAddr*TRI_NODE_SIZE+0);
-	float4 v11 = kernel_tex_fetch(__tri_woop, triAddr*TRI_NODE_SIZE+1);
-
-	float Oz = v00.w - P.x*v00.x - P.y*v00.y - P.z*v00.z;
-	float invDz = 1.0f/(dir.x*v00.x + dir.y*v00.y + dir.z*v00.z);
-	float t = Oz * invDz;
-
-	if(t > 0.0f && t < tmax) {
-		/* compute and check barycentric u */
-		float Ox = v11.w + P.x*v11.x + P.y*v11.y + P.z*v11.z;
-		float Dx = dir.x*v11.x + dir.y*v11.y + dir.z*v11.z;
-		float u = Ox + t*Dx;
-
-		if(u >= 0.0f) {
-			/* compute and check barycentric v */
-			float4 v22 = kernel_tex_fetch(__tri_woop, triAddr*TRI_NODE_SIZE+2);
-			float Oy = v22.w + P.x*v22.x + P.y*v22.y + P.z*v22.z;
-			float Dy = dir.x*v22.x + dir.y*v22.y + dir.z*v22.z;
-			float v = Oy + t*Dy;
-
-			if(v >= 0.0f && u + v <= 1.0f) {
-				(*num_hits)++;
-
-				int hit;
-
-				if(*num_hits <= max_hits) {
-					hit = *num_hits - 1;
-				}
-				else {
-					/* reservoir sampling: if we are at the maximum number of
-					 * hits, randomly replace element or skip it */
-					hit = lcg_step_uint(lcg_state) % *num_hits;
-
-					if(hit >= max_hits)
-						return;
-				}
-
-				/* record intersection */
-				Intersection *isect = &isect_array[hit];
-				isect->t = t;
-				isect->u = u;
-				isect->v = v;
-				isect->prim = triAddr;
-				isect->object = object;
-				isect->type = PRIMITIVE_TRIANGLE;
-			}
-		}
-	}
-}
-#endif
-
 CCL_NAMESPACE_END
-
diff --git a/intern/cycles/kernel/geom/geom_triangle_intersect.h b/intern/cycles/kernel/geom/geom_triangle_intersect.h
new file mode 100644
index 00000000000..b965bddf330
--- /dev/null
+++ b/intern/cycles/kernel/geom/geom_triangle_intersect.h
@@ -0,0 +1,230 @@
+/*
+ * Adapted from code Copyright 2009-2010 NVIDIA Corporation
+ * Modifications Copyright 2011, Blender Foundation.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+/* Triangle/Ray intersections
+ *
+ * Basic triangle with 3 vertices is used to represent mesh surfaces. For BVH
+ * ray intersection we use a precomputed triangle storage to accelerate
+ * intersection at the cost of more memory usage */
+
+CCL_NAMESPACE_BEGIN
+
+/* Ray-Triangle intersection for BVH traversal
+ *
+ * Based on Sven Woop's algorithm with precomputed triangle storage */
+
+ccl_device_inline bool triangle_intersect(KernelGlobals *kg, Intersection *isect,
+	float3 P, float3 dir, uint visibility, int object, int triAddr)
+{
+	/* compute and check intersection t-value */
+	float4 v00 = kernel_tex_fetch(__tri_woop, triAddr*TRI_NODE_SIZE+0);
+	float4 v11 = kernel_tex_fetch(__tri_woop, triAddr*TRI_NODE_SIZE+1);
+
+	float Oz = v00.w - P.x*v00.x - P.y*v00.y - P.z*v00.z;
+	float invDz = 1.0f/(dir.x*v00.x + dir.y*v00.y + dir.z*v00.z);
+	float t = Oz * invDz;
+
+	if(t > 0.0f && t < isect->t) {
+		/* compute and check barycentric u */
+		float Ox = v11.w + P.x*v11.x + P.y*v11.y + P.z*v11.z;
+		float Dx = dir.x*v11.x + dir.y*v11.y + dir.z*v11.z;
+		float u = Ox + t*Dx;
+
+		if(u >= 0.0f) {
+			/* compute and check barycentric v */
+			float4 v22 = kernel_tex_fetch(__tri_woop, triAddr*TRI_NODE_SIZE+2);
+			float Oy = v22.w + P.x*v22.x + P.y*v22.y + P.z*v22.z;
+			float Dy = dir.x*v22.x + dir.y*v22.y + dir.z*v22.z;
+			float v = Oy + t*Dy;
+
+			if(v >= 0.0f && u + v <= 1.0f) {
+#ifdef __VISIBILITY_FLAG__
+				/* visibility flag test. we do it here under the assumption
+				 * that most triangles are culled by node flags */
+				if(kernel_tex_fetch(__prim_visibility, triAddr) & visibility)
+#endif
+				{
+					/* record intersection */
+					isect->t = t;
+					isect->u = u;
+					isect->v = v;
+					isect->prim = triAddr;
+					isect->object = object;
+					isect->type = PRIMITIVE_TRIANGLE;
+					return true;
+				}
+			}
+		}
+	}
+
+	return false;
+}
+
+/* Special ray intersection routines for subsurface scattering. In that case we
+ * only want to intersect with primitives in the same object, and if case of
+ * multiple hits we pick a single random primitive as the intersection point. */
+
+#ifdef __SUBSURFACE__
+ccl_device_inline void triangle_intersect_subsurface(KernelGlobals *kg, Intersection *isect_array,
+	float3 P, float3 dir, int object, int triAddr, float tmax, uint *num_hits, uint *lcg_state, int max_hits)
+{
+	/* compute and check intersection t-value */
+	float4 v00 = kernel_tex_fetch(__tri_woop, triAddr*TRI_NODE_SIZE+0);
+	float4 v11 = kernel_tex_fetch(__tri_woop, triAddr*TRI_NODE_SIZE+1);
+
+	float Oz = v00.w - P.x*v00.x - P.y*v00.y - P.z*v00.z;
+	float invDz = 1.0f/(dir.x*v00.x + dir.y*v00.y + dir.z*v00.z);
+	float t = Oz * invDz;
+
+	if(t > 0.0f && t < tmax) {
+		/* compute and check barycentric u */
+		float Ox = v11.w + P.x*v11.x + P.y*v11.y + P.z*v11.z;
+		float Dx = dir.x*v11.x + dir.y*v11.y + dir.z*v11.z;
+		float u = Ox + t*Dx;
+
+		if(u >= 0.0f) {
+			/* compute and check barycentric v */
+			float4 v22 = kernel_tex_fetch(__tri_woop, triAddr*TRI_NODE_SIZE+2);
+			float Oy = v22.w + P.x*v22.x + P.y*v22.y + P.z*v22.z;
+			float Dy = dir.x*v22.x + dir.y*v22.y + dir.z*v22.z;
+			float v = Oy + t*Dy;
+
+			if(v >= 0.0f && u + v <= 1.0f) {
+				(*num_hits)++;
+
+				int hit;
+
+				if(*num_hits <= max_hits) {
+					hit = *num_hits - 1;
+				}
+				else {
+					/* reservoir sampling: if we are at the maximum number of
+					 * hits, randomly replace element or skip it */
+					hit = lcg_step_uint(lcg_state) % *num_hits;
+
+					if(hit >= max_hits)
+						return;
+				}
+
+				/* record intersection */
+				Intersection *isect = &isect_array[hit];
+				isect->t = t;
+				isect->u = u;
+				isect->v = v;
+				isect->prim = triAddr;
+				isect->object = object;
+				isect->type = PRIMITIVE_TRIANGLE;
+			}
+		}
+	}
+}
+#endif
+
+/* Refine triangle intersection to more precise hit point. For rays that travel
+ * far the precision is often not so good, this reintersects the primitive from
+ * a closer distance. */
+
+ccl_device_inline float3 triangle_refine(KernelGlobals *kg, ShaderData *sd, const Intersection *isect, const Ray *ray)
+{
+	float3 P = ray->P;
+	float3 D = ray->D;
+	float t = isect->t;
+
+#ifdef __INTERSECTION_REFINE__
+	if(isect->object != OBJECT_NONE) {
+#ifdef __OBJECT_MOTION__
+		Transform tfm = sd->ob_itfm;
+#else
+		Transform tfm = object_fetch_transform(kg, isect->object, OBJECT_INVERSE_TRANSFORM);
+#endif
+
+		P = transform_point(&tfm, P);
+		D = transform_direction(&tfm, D*t);
+		D = normalize_len(D, &t);
+	}
+
+	P = P + D*t;
+
+	float4 v00 = kernel_tex_fetch(__tri_woop, isect->prim*TRI_NODE_SIZE+0);
+	float Oz = v00.w - P.x*v00.x - P.y*v00.y - P.z*v00.z;
+	float invDz = 1.0f/(D.x*v00.x + D.y*v00.y + D.z*v00.z);
+	float rt = Oz * invDz;
+
+	P = P + D*rt;
+
+	if(isect->object != OBJECT_NONE) {
+#ifdef __OBJECT_MOTION__
+		Transform tfm = sd->ob_tfm;
+#else
+		Transform tfm = object_fetch_transform(kg, isect->object, OBJECT_TRANSFORM);
+#endif
+
+		P = transform_point(&tfm, P);
+	}
+
+	return P;
+#else
+	return P + D*t;
+#endif
+}
+
+/* same as above, except that isect->t is assumed to be in object space for instancing */
+ccl_device_inline float3 triangle_refine_subsurface(KernelGlobals *kg, ShaderData *sd, const Intersection *isect, const Ray *ray)
+{
+	float3 P = ray->P;
+	float3 D = ray->D;
+	float t = isect->t;
+
+#ifdef __INTERSECTION_REFINE__
+	if(isect->object != OBJECT_NONE) {
+#ifdef __OBJECT_MOTION__
+		Transform tfm = sd->ob_itfm;
+#else
+		Transform tfm = object_fetch_transform(kg, isect->object, OBJECT_INVERSE_TRANSFORM);
+#endif
+
+		P = transform_point(&tfm, P);
+		D = transform_direction(&tfm, D);
+		D = normalize(D);
+	}
+
+	P = P + D*t;
+
+	float4 v00 = kernel_tex_fetch(__tri_woop, isect->prim*TRI_NODE_SIZE+0);
+	float Oz = v00.w - P.x*v00.x - P.y*v00.y - P.z*v00.z;
+	float invDz = 1.0f/(D.x*v00.x + D.y*v00.y + D.z*v00.z);
+	float rt = Oz * invDz;
+
+	P = P + D*rt;
+
+	if(isect->object != OBJECT_NONE) {
+#ifdef __OBJECT_MOTION__
+		Transform tfm = sd->ob_tfm;
+#else
+		Transform tfm = object_fetch_transform(kg, isect->object, OBJECT_TRANSFORM);
+#endif
+
+		P = transform_point(&tfm, P);
+	}
+
+	return P;
+#else
+	return P + D*t;
+#endif
+}
+
+CCL_NAMESPACE_END