yse BLI_math for the compositor in more places.

26 files changed, 157 insertions, 272 deletions

diff --git a/source/blender/blenkernel/intern/curve.c b/source/blender/blenkernel/intern/curve.c
index ce62b9c10dc..623d4b8a931 100644
--- a/source/blender/blenkernel/intern/curve.c
+++ b/source/blender/blenkernel/intern/curve.c
@@ -1002,9 +1002,7 @@ void BKE_nurb_makeFaces(Nurb *nu, float *coord_array, int rowstride, int resolu,
 						bp++;
 
 					if (*fp != 0.0f) {
-						in[0] += (*fp) * bp->vec[0];
-						in[1] += (*fp) * bp->vec[1];
-						in[2] += (*fp) * bp->vec[2];
+						madd_v3_v3fl(in, bp->vec, *fp);
 					}
 				}
 			}
@@ -1106,9 +1104,7 @@ void BKE_nurb_makeCurve(Nurb *nu, float *coord_array, float *tilt_array, float *
 				bp++;
 
 			if (*fp != 0.0f) {
-				coord_fp[0] += (*fp) * bp->vec[0];
-				coord_fp[1] += (*fp) * bp->vec[1];
-				coord_fp[2] += (*fp) * bp->vec[2];
+				madd_v3_v3fl(coord_fp, bp->vec, *fp);
 
 				if (tilt_fp)
 					(*tilt_fp) += (*fp) * bp->alfa;
diff --git a/source/blender/blenkernel/intern/tracking.c b/source/blender/blenkernel/intern/tracking.c
index 57983062e3f..b61a39a2cd2 100644
--- a/source/blender/blenkernel/intern/tracking.c
+++ b/source/blender/blenkernel/intern/tracking.c
@@ -1301,6 +1301,8 @@ ImBuf *BKE_tracking_sample_pattern_imbuf(int frame_width, int frame_height, ImBu
 	(void) frame_height;
 	(void) search_ibuf;
 	(void) marker;
+	(void) track;
+	(void) use_mask;
 
 	pattern_ibuf = IMB_allocImBuf(num_samples_x, num_samples_y, 32, IB_rectfloat);
 
diff --git a/source/blender/blenlib/BLI_math_vector.h b/source/blender/blenlib/BLI_math_vector.h
index 7ff52a5824f..be492fb6fdd 100644
--- a/source/blender/blenlib/BLI_math_vector.h
+++ b/source/blender/blenlib/BLI_math_vector.h
@@ -87,6 +87,8 @@ MINLINE void add_v2_v2(float r[2], const float a[2]);
 MINLINE void add_v2_v2v2(float r[2], const float a[2], const float b[2]);
 MINLINE void add_v3_v3(float r[3], const float a[3]);
 MINLINE void add_v3_v3v3(float r[3], const float a[3], const float b[3]);
+MINLINE void add_v4_v4(float r[4], const float a[4]);
+MINLINE void add_v4_v4v4(float r[4], const float a[4], const float b[4]);
 
 MINLINE void sub_v2_v2(float r[2], const float a[2]);
 MINLINE void sub_v2_v2v2(float r[2], const float a[2], const float b[2]);
@@ -103,6 +105,7 @@ MINLINE void mul_v2_v2(float r[2], const float a[2]);
 MINLINE void mul_v3_v3(float r[3], const float a[3]);
 MINLINE void mul_v3_v3v3(float r[3], const float a[3], const float b[3]);
 MINLINE void mul_v4_fl(float r[4], float f);
+MINLINE void mul_v4_v4fl(float r[3], const float a[3], float f);
 
 MINLINE void madd_v3_v3fl(float r[3], const float a[3], float f);
 MINLINE void madd_v3_v3v3(float r[3], const float a[3], const float b[3]);
@@ -110,6 +113,7 @@ MINLINE void madd_v2_v2v2fl(float r[2], const float a[2], const float b[2], floa
 MINLINE void madd_v3_v3v3fl(float r[3], const float a[3], const float b[3], float f);
 MINLINE void madd_v3_v3v3v3(float r[3], const float a[3], const float b[3], const float c[3]);
 MINLINE void madd_v4_v4fl(float r[4], const float a[4], float f);
+MINLINE void madd_v4_v4v4(float r[4], const float a[4], const float b[4]);
 
 MINLINE void negate_v2(float r[2]);
 MINLINE void negate_v2_v2(float r[2], const float a[2]);
diff --git a/source/blender/blenlib/intern/math_vector_inline.c b/source/blender/blenlib/intern/math_vector_inline.c
index 56188048c02..e89b2ece467 100644
--- a/source/blender/blenlib/intern/math_vector_inline.c
+++ b/source/blender/blenlib/intern/math_vector_inline.c
@@ -272,6 +272,22 @@ MINLINE void add_v3_v3v3(float r[3], const float a[3], const float b[3])
 	r[2] = a[2] + b[2];
 }
 
+MINLINE void add_v4_v4(float r[4], const float a[4])
+{
+	r[0] += a[0];
+	r[1] += a[1];
+	r[2] += a[2];
+	r[3] += a[3];
+}
+
+MINLINE void add_v4_v4v4(float r[4], const float a[4], const float b[4])
+{
+	r[0] = a[0] + b[0];
+	r[1] = a[1] + b[1];
+	r[2] = a[2] + b[2];
+	r[3] = a[3] + b[3];
+}
+
 MINLINE void sub_v2_v2(float r[2], const float a[2])
 {
 	r[0] -= a[0];
@@ -361,6 +377,14 @@ MINLINE void mul_v4_fl(float r[4], float f)
 	r[3] *= f;
 }
 
+MINLINE void mul_v4_v4fl(float r[4], const float a[4], float f)
+{
+	r[0] = a[0] * f;
+	r[1] = a[1] * f;
+	r[2] = a[2] * f;
+	r[3] = a[3] * f;
+}
+
 MINLINE void madd_v2_v2fl(float r[2], const float a[2], float f)
 {
 	r[0] += a[0] * f;
@@ -409,6 +433,14 @@ MINLINE void madd_v4_v4fl(float r[4], const float a[4], float f)
 	r[3] += a[3] * f;
 }
 
+MINLINE void madd_v4_v4v4(float r[4], const float a[4], const float b[4])
+{
+	r[0] += a[0] * b[0];
+	r[1] += a[1] * b[1];
+	r[2] += a[2] * b[2];
+	r[3] += a[3] * b[3];
+}
+
 MINLINE void mul_v3_v3v3(float r[3], const float v1[3], const float v2[3])
 {
 	r[0] = v1[0] * v2[0];
diff --git a/source/blender/compositor/intern/COM_MemoryBuffer.cpp b/source/blender/compositor/intern/COM_MemoryBuffer.cpp
index 3ebf8398c02..b88d42ea8d0 100644
--- a/source/blender/compositor/intern/COM_MemoryBuffer.cpp
+++ b/source/blender/compositor/intern/COM_MemoryBuffer.cpp
@@ -329,9 +329,7 @@ void MemoryBuffer::readEWA(float result[4], float fx, float fy, float dx, float
 				float tc[4];
 				const float wt = EWA_WTS[(Q < 0.f) ? 0 : (unsigned int)Q];
 				read(tc, clipuv(u, width), clipuv(v, height));
-				result[0] += tc[0]*wt;
-				result[1] += tc[1]*wt;
-				result[2] += tc[2]*wt;
+				madd_v3_v3fl(result, tc, wt);
 				result[3] += result[3] ? tc[3]*wt : 0.f;
 				d += wt;
 			}
diff --git a/source/blender/compositor/intern/COM_NodeOperation.h b/source/blender/compositor/intern/COM_NodeOperation.h
index db1fdda0bcf..3f536fb3f2d 100644
--- a/source/blender/compositor/intern/COM_NodeOperation.h
+++ b/source/blender/compositor/intern/COM_NodeOperation.h
@@ -35,6 +35,9 @@ class NodeOperation;
 #include "list"
 #include "BLI_threads.h"
 
+#include "BLI_math_color.h"
+#include "BLI_math_vector.h"
+
 class ReadBufferOperation;
 
 /**
diff --git a/source/blender/compositor/operations/COM_AlphaOverKeyOperation.cpp b/source/blender/compositor/operations/COM_AlphaOverKeyOperation.cpp
index c891142b808..0c9f9b97031 100644
--- a/source/blender/compositor/operations/COM_AlphaOverKeyOperation.cpp
+++ b/source/blender/compositor/operations/COM_AlphaOverKeyOperation.cpp
@@ -37,16 +37,10 @@ void AlphaOverKeyOperation::executePixel(float *outputValue, float x, float y, P
 	inputColor2Operation->read(inputOverColor, x, y, sampler, inputBuffers);
 	
 	if (inputOverColor[3] <= 0.0f) {
-		outputValue[0] = inputColor1[0];
-		outputValue[1] = inputColor1[1];
-		outputValue[2] = inputColor1[2];
-		outputValue[3] = inputColor1[3];
+		copy_v4_v4(outputValue, inputColor1);
 	}
 	else if (value[0] == 1.0f && inputOverColor[3] >= 1.0f) {
-		outputValue[0] = inputOverColor[0];
-		outputValue[1] = inputOverColor[1];
-		outputValue[2] = inputOverColor[2];
-		outputValue[3] = inputOverColor[3];
+		copy_v4_v4(outputValue, inputOverColor);
 	}
 	else {
 		float premul = value[0]*inputOverColor[3];
diff --git a/source/blender/compositor/operations/COM_AlphaOverMixedOperation.cpp b/source/blender/compositor/operations/COM_AlphaOverMixedOperation.cpp
index 7d489ce856e..850bbd5cc00 100644
--- a/source/blender/compositor/operations/COM_AlphaOverMixedOperation.cpp
+++ b/source/blender/compositor/operations/COM_AlphaOverMixedOperation.cpp
@@ -38,16 +38,10 @@ void AlphaOverMixedOperation::executePixel(float outputValue[4], float x, float
 	inputColor2Operation->read(inputOverColor, x, y, sampler, inputBuffers);
 	
 	if (inputOverColor[3] <= 0.0f) {
-		outputValue[0] = inputColor1[0];
-		outputValue[1] = inputColor1[1];
-		outputValue[2] = inputColor1[2];
-		outputValue[3] = inputColor1[3];
+		copy_v4_v4(outputValue, inputColor1);
 	}
 	else if (value[0] == 1.0f && inputOverColor[3] >= 1.0f) {
-		outputValue[0] = inputOverColor[0];
-		outputValue[1] = inputOverColor[1];
-		outputValue[2] = inputOverColor[2];
-		outputValue[3] = inputOverColor[3];
+		copy_v4_v4(outputValue, inputOverColor);
 	}
 	else {
 		float addfac = 1.0f - this->x + inputOverColor[3]*this->x;
diff --git a/source/blender/compositor/operations/COM_AlphaOverPremultiplyOperation.cpp b/source/blender/compositor/operations/COM_AlphaOverPremultiplyOperation.cpp
index 6cc33387917..db67f2e0406 100644
--- a/source/blender/compositor/operations/COM_AlphaOverPremultiplyOperation.cpp
+++ b/source/blender/compositor/operations/COM_AlphaOverPremultiplyOperation.cpp
@@ -38,16 +38,10 @@ void AlphaOverPremultiplyOperation::executePixel(float *outputValue, float x, fl
 	
 	/* Zero alpha values should still permit an add of RGB data */
 	if (inputOverColor[3]<0.0f) {
-		outputValue[0] = inputColor1[0];
-		outputValue[1] = inputColor1[1];
-		outputValue[2] = inputColor1[2];
-		outputValue[3] = inputColor1[3];
+		copy_v4_v4(outputValue, inputColor1);
 	}
 	else if (value[0] == 1.0f && inputOverColor[3] >= 1.0f) {
-		outputValue[0] = inputOverColor[0];
-		outputValue[1] = inputOverColor[1];
-		outputValue[2] = inputOverColor[2];
-		outputValue[3] = inputOverColor[3];
+		copy_v4_v4(outputValue, inputOverColor);
 	}
 	else {
 		float mul = 1.0f - value[0]*inputOverColor[3];
diff --git a/source/blender/compositor/operations/COM_BilateralBlurOperation.cpp b/source/blender/compositor/operations/COM_BilateralBlurOperation.cpp
index 88fe17f633e..7129e7c140a 100644
--- a/source/blender/compositor/operations/COM_BilateralBlurOperation.cpp
+++ b/source/blender/compositor/operations/COM_BilateralBlurOperation.cpp
@@ -62,36 +62,27 @@ void BilateralBlurOperation::executePixel(float *color, int x, int y, MemoryBuff
 	int maxy = ceil(y + space);
 	float deltaColor;
 	this->inputDeterminatorProgram->read(determinatorReferenceColor, x, y, inputBuffers, data);
-	
-	blurColor[0] = 0.0f;
-	blurColor[1] = 0.0f;
-	blurColor[2] = 0.0f;
-	blurColor[3] = 0.0f;
+
+	zero_v4(blurColor);
 	blurDivider = 0.0f;
 	for (int yi = miny ; yi < maxy ; yi+=QualityStepHelper::getStep()) {
 		for (int xi = minx ; xi < maxx ; xi+=QualityStepHelper::getStep()) {
 			// read determinator
 			this->inputDeterminatorProgram->read(determinator, xi, yi, inputBuffers, data);
-			deltaColor = fabsf(determinatorReferenceColor[0] - determinator[0])+
-				fabsf(determinatorReferenceColor[1] - determinator[1])+
-				fabsf(determinatorReferenceColor[2] - determinator[2]); // do not take the alpha channel into account
+			deltaColor = (fabsf(determinatorReferenceColor[0] - determinator[0]) +
+			              fabsf(determinatorReferenceColor[1] - determinator[1]) +
+			              fabsf(determinatorReferenceColor[2] - determinator[2])); // do not take the alpha channel into account
 			if (deltaColor< sigmacolor) {
 				// add this to the blur
 				this->inputColorProgram->read(tempColor, xi, yi, inputBuffers, data);
-				blurColor[0]+=tempColor[0];
-				blurColor[1]+=tempColor[1];
-				blurColor[2]+=tempColor[2];
-				blurColor[3]+=tempColor[3];
+				add_v4_v4(blurColor, tempColor);
 				blurDivider += 1.0f;
 			}
 		}
 	}
 	
 	if (blurDivider > 0.0f) {
-		color[0] = blurColor[0]/blurDivider;
-		color[1] = blurColor[1]/blurDivider;
-		color[2] = blurColor[2]/blurDivider;
-		color[3] = blurColor[3]/blurDivider;
+		mul_v4_v4fl(color, blurColor, 1.0f / blurDivider);
 	}
 	else {
 		color[0] = 0.0f;
diff --git a/source/blender/compositor/operations/COM_BokehBlurOperation.cpp b/source/blender/compositor/operations/COM_BokehBlurOperation.cpp
index b4811c89dc5..67e24496cbe 100644
--- a/source/blender/compositor/operations/COM_BokehBlurOperation.cpp
+++ b/source/blender/compositor/operations/COM_BokehBlurOperation.cpp
@@ -79,14 +79,7 @@ void BokehBlurOperation::executePixel(float *color, int x, int y, MemoryBuffer *
 
 	inputBoundingBoxReader->read(tempBoundingBox, x, y, COM_PS_NEAREST, inputBuffers);
 	if (tempBoundingBox[0] >0.0f) {
-		tempColor[0] = 0;
-		tempColor[1] = 0;
-		tempColor[2] = 0;
-		tempColor[3] = 0;
-		float overallmultiplyerr = 0;
-		float overallmultiplyerg = 0;
-		float overallmultiplyerb = 0;
-		float overallmultiplyera = 0;
+		float overallmultiplyer[4] = {0.0f, 0.0f, 0.0f, 0.0f};
 		MemoryBuffer *inputBuffer = (MemoryBuffer*)data;
 		float *buffer = inputBuffer->getBuffer();
 		int bufferwidth = inputBuffer->getWidth();
@@ -103,6 +96,8 @@ void BokehBlurOperation::executePixel(float *color, int x, int y, MemoryBuffer *
 		maxy = min(maxy, inputBuffer->getRect()->ymax);
 		maxx = min(maxx, inputBuffer->getRect()->xmax);
 
+		zero_v4(tempColor);
+
 		int step = getStep();
 		int offsetadd = getOffsetAdd();
 
@@ -113,21 +108,15 @@ void BokehBlurOperation::executePixel(float *color, int x, int y, MemoryBuffer *
 				float u = this->bokehMidX - (nx-x) *m;
 				float v = this->bokehMidY - (ny-y) *m;
 				inputBokehProgram->read(bokeh, u, v, COM_PS_NEAREST, inputBuffers);
-				tempColor[0] += bokeh[0] * buffer[bufferindex];
-				tempColor[1] += bokeh[1] * buffer[bufferindex+1];
-				tempColor[2] += bokeh[2]* buffer[bufferindex+2];
-				tempColor[3] += bokeh[3]* buffer[bufferindex+3];
-				overallmultiplyerr += bokeh[0];
-				overallmultiplyerg += bokeh[1];
-				overallmultiplyerb += bokeh[2];
-				overallmultiplyera += bokeh[3];
+				madd_v4_v4v4(tempColor, bokeh, &buffer[bufferindex]);
+				add_v4_v4(overallmultiplyer, bokeh);
 				bufferindex +=offsetadd;
 			}
 		}
-		color[0] = tempColor[0] * (1.0f / overallmultiplyerr);
-		color[1] = tempColor[1] * (1.0f / overallmultiplyerg);
-		color[2] = tempColor[2] * (1.0f / overallmultiplyerb);
-		color[3] = tempColor[3] * (1.0f / overallmultiplyera);
+		color[0] = tempColor[0] * (1.0f / overallmultiplyer[0]);
+		color[1] = tempColor[1] * (1.0f / overallmultiplyer[1]);
+		color[2] = tempColor[2] * (1.0f / overallmultiplyer[2]);
+		color[3] = tempColor[3] * (1.0f / overallmultiplyer[3]);
 	}
 	else {
 		inputProgram->read(color, x, y, COM_PS_NEAREST, inputBuffers);
diff --git a/source/blender/compositor/operations/COM_ColorCurveOperation.cpp b/source/blender/compositor/operations/COM_ColorCurveOperation.cpp
index a38012271f1..c4336ed5e06 100644
--- a/source/blender/compositor/operations/COM_ColorCurveOperation.cpp
+++ b/source/blender/compositor/operations/COM_ColorCurveOperation.cpp
@@ -79,9 +79,7 @@ void ColorCurveOperation::executePixel(float *color, float x, float y, PixelSamp
 	if (*fac >= 1.0f)
 		curvemapping_evaluate_premulRGBF(workingCopy, color, image);
 	else if (*fac <= 0.0f) {
-		color[0] = image[0];
-		color[1] = image[1];
-		color[2] = image[2];
+		copy_v3_v3(color, image);
 	}
 	else {
 		float col[4], mfac = 1.0f - *fac;
@@ -140,9 +138,7 @@ void ConstantLevelColorCurveOperation::executePixel(float *color, float x, float
 	if (*fac >= 1.0f)
 		curvemapping_evaluate_premulRGBF(this->curveMapping, color, image);
 	else if (*fac <= 0.0f) {
-		color[0] = image[0];
-		color[1] = image[1];
-		color[2] = image[2];
+		copy_v3_v3(color, image);
 	}
 	else {
 		float col[4], mfac = 1.0f - *fac;
diff --git a/source/blender/compositor/operations/COM_ConvertKeyToPremulOperation.cpp b/source/blender/compositor/operations/COM_ConvertKeyToPremulOperation.cpp
index 547915f58c9..db27e07d52f 100644
--- a/source/blender/compositor/operations/COM_ConvertKeyToPremulOperation.cpp
+++ b/source/blender/compositor/operations/COM_ConvertKeyToPremulOperation.cpp
@@ -43,9 +43,7 @@ void ConvertKeyToPremulOperation::executePixel(float *outputValue, float x, floa
 	this->inputColor->read(inputValue, x, y, sampler, inputBuffers);
 	alpha = inputValue[3];
 
-	outputValue[0] = inputValue[0] * alpha;
-	outputValue[1] = inputValue[1] * alpha;
-	outputValue[2] = inputValue[2] * alpha;
+	mul_v3_v3fl(outputValue, inputValue, alpha);
 
 	/* never touches the alpha */
 	outputValue[3] = alpha;
diff --git a/source/blender/compositor/operations/COM_ConvertPremulToKeyOperation.cpp b/source/blender/compositor/operations/COM_ConvertPremulToKeyOperation.cpp
index 16636ee2afc..920b5f8a775 100644
--- a/source/blender/compositor/operations/COM_ConvertPremulToKeyOperation.cpp
+++ b/source/blender/compositor/operations/COM_ConvertPremulToKeyOperation.cpp
@@ -44,14 +44,10 @@ void ConvertPremulToKeyOperation::executePixel(float *outputValue, float x, floa
 	alpha = inputValue[3];
 
 	if (fabsf(alpha) < 1e-5f) {
-		outputValue[0] = 0.f;
-		outputValue[1] = 0.f;
-		outputValue[2] = 0.f;
+		zero_v3(outputValue);
 	}
 	else {
-		outputValue[0] = inputValue[0] / alpha;
-		outputValue[1] = inputValue[1] / alpha;
-		outputValue[2] = inputValue[2] / alpha;
+		mul_v3_v3fl(outputValue, inputValue, 1.0f / alpha);
 	}
 
 	/* never touches the alpha */
diff --git a/source/blender/compositor/operations/COM_ConvertRGBToYCCOperation.cpp b/source/blender/compositor/operations/COM_ConvertRGBToYCCOperation.cpp
index c626dc03000..ce62cf0ae49 100644
--- a/source/blender/compositor/operations/COM_ConvertRGBToYCCOperation.cpp
+++ b/source/blender/compositor/operations/COM_ConvertRGBToYCCOperation.cpp
@@ -60,9 +60,8 @@ void ConvertRGBToYCCOperation::executePixel(float *outputValue, float x, float y
 	rgb_to_ycc(inputColor[0], inputColor[1], inputColor[2], &color[0], &color[1], &color[2], this->mode);
 
 	/* divided by 255 to normalize for viewing in */
-	outputValue[0] = color[0]/255.f; /* Y */
-	outputValue[1] = color[1]/255.f; /* Cb*/
-	outputValue[2] = color[2]/255.f; /* Cr*/
+	/* R,G,B --> Y,Cb,Cr */
+	mul_v3_v3fl(outputValue, color, 1.0f / 255.0f);
 	outputValue[3] = inputColor[3];
 }
 
diff --git a/source/blender/compositor/operations/COM_ConvertYCCToRGBOperation.cpp b/source/blender/compositor/operations/COM_ConvertYCCToRGBOperation.cpp
index d3048c131e4..dbfe4847c78 100644
--- a/source/blender/compositor/operations/COM_ConvertYCCToRGBOperation.cpp
+++ b/source/blender/compositor/operations/COM_ConvertYCCToRGBOperation.cpp
@@ -57,9 +57,8 @@ void ConvertYCCToRGBOperation::executePixel(float *outputValue, float x, float y
 	inputOperation->read(inputColor, x, y, sampler, inputBuffers);
 
 	/* need to un-normalize the data */
-	inputColor[0] *= 255.f; /* Y */
-	inputColor[1] *= 255.f; /* Cb*/
-	inputColor[2] *= 255.f; /* Cr*/
+	/* R,G,B --> Y,Cb,Cr */
+	mul_v3_fl(inputColor, 255.0f);
 
 	ycc_to_rgb(inputColor[0], inputColor[1], inputColor[2], &outputValue[0], &outputValue[1], &outputValue[2], this->mode);
 	outputValue[3] = inputColor[3];
diff --git a/source/blender/compositor/operations/COM_ConvolutionEdgeFilterOperation.cpp b/source/blender/compositor/operations/COM_ConvolutionEdgeFilterOperation.cpp
index db67412f3e7..6edb046a070 100644
--- a/source/blender/compositor/operations/COM_ConvolutionEdgeFilterOperation.cpp
+++ b/source/blender/compositor/operations/COM_ConvolutionEdgeFilterOperation.cpp
@@ -26,12 +26,6 @@
 ConvolutionEdgeFilterOperation::ConvolutionEdgeFilterOperation() : ConvolutionFilterOperation()
 {
 }
-inline void addFilter(float *result, float*input, float value)
-{
-	result[0] += input[0] * value;
-	result[1] += input[1] * value;
-	result[2] += input[2] * value;
-}
 
 void ConvolutionEdgeFilterOperation::executePixel(float *color,int x, int y, MemoryBuffer *inputBuffers[], void *data)
 {
@@ -64,48 +58,48 @@ void ConvolutionEdgeFilterOperation::executePixel(float *color,int x, int y, Mem
 	res2[3] = 0.0f;
 	
 	this->inputOperation->read(in1, x1, y1, inputBuffers, NULL);
-	addFilter(res1, in1, this->filter[0]);
-	addFilter(res2, in1, this->filter[0]);
+	madd_v3_v3fl(res1, in1, this->filter[0]);
+	madd_v3_v3fl(res2, in1, this->filter[0]);
 	
 	this->inputOperation->read(in1, x2, y1, inputBuffers, NULL);
-	addFilter(res1, in1, this->filter[1]);
-	addFilter(res2, in1, this->filter[3]);
+	madd_v3_v3fl(res1, in1, this->filter[1]);
+	madd_v3_v3fl(res2, in1, this->filter[3]);
 	
 	this->inputOperation->read(in1, x3, y1, inputBuffers, NULL);
-	addFilter(res1, in1, this->filter[2]);
-	addFilter(res2, in1, this->filter[6]);
+	madd_v3_v3fl(res1, in1, this->filter[2]);
+	madd_v3_v3fl(res2, in1, this->filter[6]);
 	
 	this->inputOperation->read(in1, x1, y2, inputBuffers, NULL);
-	addFilter(res1, in1, this->filter[3]);
-	addFilter(res2, in1, this->filter[1]);
+	madd_v3_v3fl(res1, in1, this->filter[3]);
+	madd_v3_v3fl(res2, in1, this->filter[1]);
 	
 	this->inputOperation->read(in2, x2, y2, inputBuffers, NULL);
-	addFilter(res1, in2, this->filter[4]);
-	addFilter(res2, in2, this->filter[4]);
+	madd_v3_v3fl(res1, in2, this->filter[4]);
+	madd_v3_v3fl(res2, in2, this->filter[4]);
 	
 	this->inputOperation->read(in1, x3, y2, inputBuffers, NULL);
-	addFilter(res1, in1, this->filter[5]);
-	addFilter(res2, in1, this->filter[7]);
+	madd_v3_v3fl(res1, in1, this->filter[5]);
+	madd_v3_v3fl(res2, in1, this->filter[7]);
 	
 	this->inputOperation->read(in1, x1, y3, inputBuffers, NULL);
-	addFilter(res1, in1, this->filter[6]);
-	addFilter(res2, in1, this->filter[2]);
+	madd_v3_v3fl(res1, in1, this->filter[6]);
+	madd_v3_v3fl(res2, in1, this->filter[2]);
 	
 	this->inputOperation->read(in1, x2, y3, inputBuffers, NULL);
-	addFilter(res1, in1, this->filter[7]);
-	addFilter(res2, in1, this->filter[5]);
+	madd_v3_v3fl(res1, in1, this->filter[7]);
+	madd_v3_v3fl(res2, in1, this->filter[5]);
 	
 	this->inputOperation->read(in1, x3, y3, inputBuffers, NULL);
-	addFilter(res1, in1, this->filter[8]);
-	addFilter(res2, in1, this->filter[8]);
+	madd_v3_v3fl(res1, in1, this->filter[8]);
+	madd_v3_v3fl(res2, in1, this->filter[8]);
 	
-	color[0] = sqrt(res1[0]*res1[0]+res2[0]*res2[0]);
-	color[1] = sqrt(res1[1]*res1[1]+res2[1]*res2[1]);
-	color[2] = sqrt(res1[2]*res1[2]+res2[2]*res2[2]);
+	color[0] = sqrt(res1[0] * res1[0] + res2[0] * res2[0]);
+	color[1] = sqrt(res1[1] * res1[1] + res2[1] * res2[1]);
+	color[2] = sqrt(res1[2] * res1[2] + res2[2] * res2[2]);
 	
-	color[0] = color[0]*value[0] + in2[0] * mval;
-	color[1] = color[1]*value[0] + in2[1] * mval;
-	color[2] = color[2]*value[0] + in2[2] * mval;
+	color[0] = color[0] * value[0] + in2[0] * mval;
+	color[1] = color[1] * value[0] + in2[1] * mval;
+	color[2] = color[2] * value[0] + in2[2] * mval;
 	
 	color[3] = in2[3];
 }
diff --git a/source/blender/compositor/operations/COM_ConvolutionFilterOperation.cpp b/source/blender/compositor/operations/COM_ConvolutionFilterOperation.cpp
index 2720a0a4146..3c9cde92e2e 100644
--- a/source/blender/compositor/operations/COM_ConvolutionFilterOperation.cpp
+++ b/source/blender/compositor/operations/COM_ConvolutionFilterOperation.cpp
@@ -69,10 +69,6 @@ void ConvolutionFilterOperation::deinitExecution()
 
 void ConvolutionFilterOperation::executePixel(float *color,int x, int y, MemoryBuffer *inputBuffers[], void *data)
 {
-	color[0] = 0.0;
-	color[1] = 0.0;
-	color[2] = 0.0;
-	color[3] = 0.0;
 	float in1[4];
 	float in2[4];
 	int x1 = x - 1;
@@ -89,57 +85,32 @@ void ConvolutionFilterOperation::executePixel(float *color,int x, int y, MemoryB
 	CLAMP(y3, 0, getHeight()-1);
 	float value[4];
 	this->inputValueOperation->read(value, x2, y2, inputBuffers, NULL);
-	float mval = 1.0f - value[0];
+	const float mval = 1.0f - value[0];
+
+	zero_v4(color);
 	this->inputOperation->read(in1, x1, y1, inputBuffers, NULL);
-	color[0] += in1[0] * this->filter[0];
-	color[1] += in1[1] * this->filter[0];
-	color[2] += in1[2] * this->filter[0];
-	color[3] += in1[3] * this->filter[0];
+	madd_v4_v4fl(color, in1, this->filter[0]);
 	this->inputOperation->read(in1, x2, y1, inputBuffers, NULL);
-	color[0] += in1[0] * this->filter[1];
-	color[1] += in1[1] * this->filter[1];
-	color[2] += in1[2] * this->filter[1];
-	color[3] += in1[3] * this->filter[1];
+	madd_v4_v4fl(color, in1, this->filter[1]);
 	this->inputOperation->read(in1, x3, y1, inputBuffers, NULL);
-	color[0] += in1[0] * this->filter[2];
-	color[1] += in1[1] * this->filter[2];
-	color[2] += in1[2] * this->filter[2];
-	color[3] += in1[3] * this->filter[2];
+	madd_v4_v4fl(color, in1, this->filter[2]);
 	this->inputOperation->read(in1, x1, y2, inputBuffers, NULL);
-	color[0] += in1[0] * this->filter[3];
-	color[1] += in1[1] * this->filter[3];
-	color[2] += in1[2] * this->filter[3];
-	color[3] += in1[3] * this->filter[3];
+	madd_v4_v4fl(color, in1, this->filter[3]);
 	this->inputOperation->read(in2, x2, y2, inputBuffers, NULL);
-	color[0] += in2[0] * this->filter[4];
-	color[1] += in2[1] * this->filter[4];
-	color[2] += in2[2] * this->filter[4];
-	color[3] += in2[3] * this->filter[4];
+	madd_v4_v4fl(color, in2, this->filter[4]);
 	this->inputOperation->read(in1, x3, y2, inputBuffers, NULL);
-	color[0] += in1[0] * this->filter[5];
-	color[1] += in1[1] * this->filter[5];
-	color[2] += in1[2] * this->filter[5];
-	color[3] += in1[3] * this->filter[5];
+	madd_v4_v4fl(color, in1, this->filter[5]);
 	this->inputOperation->read(in1, x1, y3, inputBuffers, NULL);
-	color[0] += in1[0] * this->filter[6];
-	color[1] += in1[1] * this->filter[6];
-	color[2] += in1[2] * this->filter[6];
-	color[3] += in1[3] * this->filter[6];
+	madd_v4_v4fl(color, in1, this->filter[6]);
 	this->inputOperation->read(in1, x2, y3, inputBuffers, NULL);
-	color[0] += in1[0] * this->filter[7];
-	color[1] += in1[1] * this->filter[7];
-	color[2] += in1[2] * this->filter[7];
-	color[3] += in1[3] * this->filter[7];
+	madd_v4_v4fl(color, in1, this->filter[7]);
 	this->inputOperation->read(in1, x3, y3, inputBuffers, NULL);
-	color[0] += in1[0] * this->filter[8];
-	color[1] += in1[1] * this->filter[8];
-	color[2] += in1[2] * this->filter[8];
-	color[3] += in1[3] * this->filter[8];
+	madd_v4_v4fl(color, in1, this->filter[8]);
 	
-	color[0] = color[0]*value[0] + in2[0] * mval;
-	color[1] = color[1]*value[0] + in2[1] * mval;
-	color[2] = color[2]*value[0] + in2[2] * mval;
-	color[3] = color[3]*value[0] + in2[3] * mval;
+	color[0] = color[0] * value[0] + in2[0] * mval;
+	color[1] = color[1] * value[0] + in2[1] * mval;
+	color[2] = color[2] * value[0] + in2[2] * mval;
+	color[3] = color[3] * value[0] + in2[3] * mval;
 }
 
 bool ConvolutionFilterOperation::determineDependingAreaOfInterest(rcti *input, ReadBufferOperation *readOperation, rcti *output)
diff --git a/source/blender/compositor/operations/COM_DirectionalBlurOperation.cpp b/source/blender/compositor/operations/COM_DirectionalBlurOperation.cpp
index 43cba09d16f..60c31400eca 100644
--- a/source/blender/compositor/operations/COM_DirectionalBlurOperation.cpp
+++ b/source/blender/compositor/operations/COM_DirectionalBlurOperation.cpp
@@ -85,10 +85,7 @@ void DirectionalBlurOperation::executePixel(float *color, int x, int y, MemoryBu
 
 		this->inputProgram->read(col, cs * u + ss * v + center_x_pix, cs * v - ss * u + center_y_pix, COM_PS_NEAREST, inputBuffers);
 
-		col2[0] += col[0];
-		col2[1] += col[1];
-		col2[2] += col[2];
-		col2[3] += col[3];
+		add_v4_v4(col2, col);
 
 		/* double transformations */
 		ltx += tx;
@@ -96,10 +93,8 @@ void DirectionalBlurOperation::executePixel(float *color, int x, int y, MemoryBu
 		lrot += rot;
 		lsc += sc;
 	}
-	color[0] = col2[0]/iterations;
-	color[1] = col2[1]/iterations;
-	color[2] = col2[2]/iterations;
-	color[3] = col2[3]/iterations;
+
+	mul_v4_v4fl(color, col2, 1.0f / iterations);
 }
 
 void DirectionalBlurOperation::deinitExecution()
diff --git a/source/blender/compositor/operations/COM_GaussianBokehBlurOperation.cpp b/source/blender/compositor/operations/COM_GaussianBokehBlurOperation.cpp
index 07cc07880e7..80e35f63ac5 100644
--- a/source/blender/compositor/operations/COM_GaussianBokehBlurOperation.cpp
+++ b/source/blender/compositor/operations/COM_GaussianBokehBlurOperation.cpp
@@ -138,21 +138,15 @@ void GaussianBokehBlurOperation::executePixel(float *color, int x, int y, Memory
 		index = ((ny-y)+this->rady) * (this->radx*2+1) + (minx-x+this->radx);
 		int bufferindex = ((minx - bufferstartx)*4)+((ny-bufferstarty)*4*bufferwidth);
 		for (int nx = minx ; nx < maxx ; nx +=step) {
-			float multiplyer = gausstab[index];
-			tempColor[0] += multiplyer * buffer[bufferindex];
-			tempColor[1] += multiplyer * buffer[bufferindex+1];
-			tempColor[2] += multiplyer * buffer[bufferindex+2];
-			tempColor[3] += multiplyer * buffer[bufferindex+3];
+			const float multiplyer = gausstab[index];
+			madd_v4_v4fl(tempColor, &buffer[bufferindex], multiplyer);
 			overallmultiplyer += multiplyer;
 			index += step;
 			bufferindex +=offsetadd;
 		}
 	}
-	float divider = 1.0f / overallmultiplyer;
-	color[0] = tempColor[0] * divider;
-	color[1] = tempColor[1] * divider;
-	color[2] = tempColor[2] * divider;
-	color[3] = tempColor[3] * divider;
+
+	mul_v4_v4fl(color, tempColor, 1.0f / overallmultiplyer);
 }
 
 void GaussianBokehBlurOperation::deinitExecution()
diff --git a/source/blender/compositor/operations/COM_GaussianXBlurOperation.cpp b/source/blender/compositor/operations/COM_GaussianXBlurOperation.cpp
index a7e443838a9..e30cbfeff7e 100644
--- a/source/blender/compositor/operations/COM_GaussianXBlurOperation.cpp
+++ b/source/blender/compositor/operations/COM_GaussianXBlurOperation.cpp
@@ -99,19 +99,12 @@ void GaussianXBlurOperation::executePixel(float *color, int x, int y, MemoryBuff
 	int bufferindex = ((minx - bufferstartx)*4)+((miny-bufferstarty)*4*bufferwidth);
 	for (int nx = minx ; nx < maxx ; nx +=step) {
 		index = (nx-x)+this->rad;
-		float multiplyer = gausstab[index];
-		tempColor[0] += multiplyer * buffer[bufferindex];
-		tempColor[1] += multiplyer * buffer[bufferindex+1];
-		tempColor[2] += multiplyer * buffer[bufferindex+2];
-		tempColor[3] += multiplyer * buffer[bufferindex+3];
+		const float multiplyer = gausstab[index];
+		madd_v4_v4fl(tempColor, &buffer[bufferindex], multiplyer);
 		overallmultiplyer += multiplyer;
-		bufferindex +=offsetadd;
+		bufferindex += offsetadd;
 	}
-	float divider = 1.0f / overallmultiplyer;
-	color[0] = tempColor[0] * divider;
-	color[1] = tempColor[1] * divider;
-	color[2] = tempColor[2] * divider;
-	color[3] = tempColor[3] * divider;
+	mul_v4_v4fl(color, tempColor, 1.0f / overallmultiplyer);
 }
 
 void GaussianXBlurOperation::deinitExecution()
diff --git a/source/blender/compositor/operations/COM_GaussianYBlurOperation.cpp b/source/blender/compositor/operations/COM_GaussianYBlurOperation.cpp
index c7da43fc20f..c0561704596 100644
--- a/source/blender/compositor/operations/COM_GaussianYBlurOperation.cpp
+++ b/source/blender/compositor/operations/COM_GaussianYBlurOperation.cpp
@@ -95,18 +95,11 @@ void GaussianYBlurOperation::executePixel(float *color, int x, int y, MemoryBuff
 	for (int ny = miny ; ny < maxy ; ny +=step) {
 		index = (ny-y)+this->rad;
 		int bufferindex = ((minx - bufferstartx)*4)+((ny-bufferstarty)*4*bufferwidth);
-		float multiplyer = gausstab[index];
-		tempColor[0] += multiplyer * buffer[bufferindex];
-		tempColor[1] += multiplyer * buffer[bufferindex+1];
-		tempColor[2] += multiplyer * buffer[bufferindex+2];
-		tempColor[3] += multiplyer * buffer[bufferindex+3];
+		const float multiplyer = gausstab[index];
+		madd_v4_v4fl(tempColor, &buffer[bufferindex], multiplyer);
 		overallmultiplyer += multiplyer;
 	}
-	float divider = 1.0f / overallmultiplyer;
-	color[0] = tempColor[0] * divider;
-	color[1] = tempColor[1] * divider;
-	color[2] = tempColor[2] * divider;
-	color[3] = tempColor[3] * divider;
+	mul_v4_v4fl(color, tempColor, 1.0f / overallmultiplyer);
 }
 
 void GaussianYBlurOperation::deinitExecution()
diff --git a/source/blender/compositor/operations/COM_GlareSimpleStarOperation.cpp b/source/blender/compositor/operations/COM_GlareSimpleStarOperation.cpp
index 602e18521ee..fba3eca4af9 100644
--- a/source/blender/compositor/operations/COM_GlareSimpleStarOperation.cpp
+++ b/source/blender/compositor/operations/COM_GlareSimpleStarOperation.cpp
@@ -42,28 +42,20 @@ void GlareSimpleStarOperation::generateGlare(float *data, MemoryBuffer *inputTil
 				xm = x - i;
 				xp = x + i;
 				tbuf1->read(c, x, y);
-				c[0]*=f1; c[1]*=f1 ; c[2] *=f1;
+				mul_v3_fl(c, f1);
 				tbuf1->read(tc, (settings->angle ? xm : x), ym);
-				c[0]+=tc[0]*f2;
-				c[1]+=tc[1]*f2;
-				c[2]+=tc[2]*f2;
+				madd_v3_v3fl(c, tc, f2);
 				tbuf1->read(tc, (settings->angle ? xp : x), yp);
-				c[0]+=tc[0]*f2;
-				c[1]+=tc[1]*f2;
-				c[2]+=tc[2]*f2;
+				madd_v3_v3fl(c, tc, f2);
 				c[3] = 1.0f;
 				tbuf1->writePixel(x, y, c);
 
 				tbuf2->read(c, x, y);
-				c[0]*=f1; c[1]*=f1 ; c[2] *=f1;
+				mul_v3_fl(c, f1);
 				tbuf2->read(tc, xm, (settings->angle ? yp : y));
-				c[0]+=tc[0]*f2;
-				c[1]+=tc[1]*f2;
-				c[2]+=tc[2]*f2;
+				madd_v3_v3fl(c, tc, f2);
 				tbuf2->read(tc, xp, (settings->angle ? ym : y));
-				c[0]+=tc[0]*f2;
-				c[1]+=tc[1]*f2;
-				c[2]+=tc[2]*f2;
+				madd_v3_v3fl(c, tc, f2);
 				c[3] = 1.0f;
 				tbuf2->writePixel(x, y, c);
 
@@ -77,28 +69,20 @@ void GlareSimpleStarOperation::generateGlare(float *data, MemoryBuffer *inputTil
 				xm = x - i;
 				xp = x + i;
 				tbuf1->read(c, x, y);
-				c[0]*=f1; c[1]*=f1 ; c[2] *=f1;
+				mul_v3_fl(c, f1);
 				tbuf1->read(tc, (settings->angle ? xm : x), ym);
-				c[0]+=tc[0]*f2;
-				c[1]+=tc[1]*f2;
-				c[2]+=tc[2]*f2;
+				madd_v3_v3fl(c, tc, f2);
 				tbuf1->read(tc, (settings->angle ? xp : x), yp);
-				c[0]+=tc[0]*f2;
-				c[1]+=tc[1]*f2;
-				c[2]+=tc[2]*f2;
+				madd_v3_v3fl(c, tc, f2);
 				c[3] = 1.0f;
 				tbuf1->writePixel(x, y, c);
 
 				tbuf2->read(c, x, y);
-				c[0]*=f1; c[1]*=f1 ; c[2] *=f1;
+				mul_v3_fl(c, f1);
 				tbuf2->read(tc, xm, (settings->angle ? yp : y));
-				c[0]+=tc[0]*f2;
-				c[1]+=tc[1]*f2;
-				c[2]+=tc[2]*f2;
+				madd_v3_v3fl(c, tc, f2);
 				tbuf2->read(tc, xp, (settings->angle ? ym : y));
-				c[0]+=tc[0]*f2;
-				c[1]+=tc[1]*f2;
-				c[2]+=tc[2]*f2;
+				madd_v3_v3fl(c, tc, f2);
 				c[3] = 1.0f;
 				tbuf2->writePixel(x, y, c);
 			}
diff --git a/source/blender/compositor/operations/COM_InvertOperation.cpp b/source/blender/compositor/operations/COM_InvertOperation.cpp
index 982fe1a5450..82158c4adad 100644
--- a/source/blender/compositor/operations/COM_InvertOperation.cpp
+++ b/source/blender/compositor/operations/COM_InvertOperation.cpp
@@ -55,9 +55,7 @@ void InvertOperation::executePixel(float *out, float x, float y, PixelSampler sa
 		out[2] = (1.0f - inputColor[2])*value + inputColor[2]*invertedValue;
 	}
 	else {
-		out[0] = inputColor[0];
-		out[1] = inputColor[1];
-		out[2] = inputColor[2];
+		copy_v3_v3(out, inputColor);
 	}
 	
 	if (alpha)
diff --git a/source/blender/compositor/operations/COM_TonemapOperation.cpp b/source/blender/compositor/operations/COM_TonemapOperation.cpp
index d8089bdf3ea..75adcf524c4 100644
--- a/source/blender/compositor/operations/COM_TonemapOperation.cpp
+++ b/source/blender/compositor/operations/COM_TonemapOperation.cpp
@@ -47,9 +47,7 @@ void TonemapOperation::executePixel(float *color, int x, int y, MemoryBuffer *in
 
 	float output[4];
 	this->imageReader->read(output, x, y, inputBuffers, NULL);
-	output[0] *= avg->al;
-	output[1] *= avg->al;
-	output[2] *= avg->al;
+	mul_v3_fl(output, avg->al);
 	float dr = output[0] + this->data->offset;
 	float dg = output[1] + this->data->offset;
 	float db = output[2] + this->data->offset;
@@ -63,10 +61,7 @@ void TonemapOperation::executePixel(float *color, int x, int y, MemoryBuffer *in
 		output[2] = powf(MAX2(output[2], 0.0f), igm);
 	}
 
-	color[0] = output[0];
-	color[1] = output[1];
-	color[2] = output[2];
-	color[3] = output[3];
+	copy_v4_v4(color, output);
 }
 void PhotoreceptorTonemapOperation::executePixel(float *color, int x, int y, MemoryBuffer *inputBuffers[], void *data)
 {
@@ -94,10 +89,7 @@ void PhotoreceptorTonemapOperation::executePixel(float *color, int x, int y, Mem
 	I_a = I_l + ia * (I_g - I_l);
 	output[2] /= (output[2] + powf(f * I_a, m));
 
-	color[0] = output[0];
-	color[1] = output[1];
-	color[2] = output[2];
-	color[3] = output[3];
+	copy_v4_v4(color, output);
 }
 
 void TonemapOperation::deinitExecution()
@@ -143,18 +135,14 @@ void *TonemapOperation::initializeTileData(rcti *rect, MemoryBuffer **memoryBuff
 		while (p--) {
 			float L = 0.212671f * bc[0] + 0.71516f * bc[1] + 0.072169f * bc[2];
 			Lav += L;
-			cav[0] += bc[0];
-			cav[1] += bc[1];
-			cav[2] += bc[2];
+			add_v3_v3(cav, bc);
 			lsum += logf(MAX2(L, 0.0f) + 1e-5f);
 			maxl = (L > maxl) ? L : maxl;
 			minl = (L < minl) ? L : minl;
 			bc+=4;
 		}
 		data->lav = Lav * sc;
-		data->cav[0] = cav[0]*sc;
-		data->cav[1] = cav[1]*sc;
-		data->cav[2] = cav[2]*sc;
+		mul_v3_v3fl(data->cav, cav, sc);
 		maxl = log((double)maxl + 1e-5); minl = log((double)minl + 1e-5); avl = lsum * sc;
 		data->auto_key = (maxl > minl) ? ((maxl - avl) / (maxl - minl)) : 1.f;
 		float al = exp((double)avl);
diff --git a/source/blender/compositor/operations/COM_VariableSizeBokehBlurOperation.cpp b/source/blender/compositor/operations/COM_VariableSizeBokehBlurOperation.cpp
index 562b0fc2bb5..1544b1c8d06 100644
--- a/source/blender/compositor/operations/COM_VariableSizeBokehBlurOperation.cpp
+++ b/source/blender/compositor/operations/COM_VariableSizeBokehBlurOperation.cpp
@@ -61,10 +61,7 @@ void VariableSizeBokehBlurOperation::executePixel(float *color, int x, int y, Me
 	tempColor[2] = 0;
 	tempColor[3] = 0;
 	float tempSize[4];
-	float overallmultiplyerr = 0;
-	float overallmultiplyerg = 0;
-	float overallmultiplyerb = 0;
-	float overallmultiplyera = 0;
+	float overallmultiplyer[4] = {0.0f, 0.0f, 0.0f, 0.0f};
 
 	int miny = y - maxBlur;
 	int maxy = y + maxBlur;
@@ -76,10 +73,8 @@ void VariableSizeBokehBlurOperation::executePixel(float *color, int x, int y, Me
 		tempColor[1] += readColor[1];
 		tempColor[2] += readColor[2];
 		tempColor[3] += readColor[3];
-		overallmultiplyerr += 1;
-		overallmultiplyerg += 1;
-		overallmultiplyerb += 1;
-		overallmultiplyera += 1;
+		add_v4_v4(tempColor, readColor);
+		add_v3_fl(overallmultiplyer, 1.0f);
 		
 		for (int ny = miny ; ny < maxy ; ny += QualityStepHelper::getStep()) {
 			for (int nx = minx ; nx < maxx ; nx += QualityStepHelper::getStep()) {
@@ -97,22 +92,17 @@ void VariableSizeBokehBlurOperation::executePixel(float *color, int x, int y, Me
 						float v = 256 + dy*256/size;
 						inputBokehProgram->read(bokeh, u, v, COM_PS_NEAREST, inputBuffers);
 						inputProgram->read(readColor, nx, ny, COM_PS_NEAREST, inputBuffers);
-						tempColor[0] += bokeh[1]*readColor[0];
-						tempColor[1] += bokeh[2]*readColor[1];
-						tempColor[2] += bokeh[2]*readColor[2];
-						tempColor[3] += bokeh[3]*readColor[3];
-						overallmultiplyerr += bokeh[0];
-						overallmultiplyerg += bokeh[1];
-						overallmultiplyerb += bokeh[2];
-						overallmultiplyera += bokeh[3];
+						madd_v4_v4v4(tempColor, bokeh, readColor);
+						add_v4_v4(overallmultiplyer, bokeh);
 					}
 				}
 			}
 		}
-		color[0] = tempColor[0] / overallmultiplyerr;
-		color[1] = tempColor[1] / overallmultiplyerg;
-		color[2] = tempColor[2] / overallmultiplyerb;
-		color[3] = tempColor[3] / overallmultiplyera;
+
+		color[0] = tempColor[0] * (1.0f / overallmultiplyer[0]);
+		color[1] = tempColor[1] * (1.0f / overallmultiplyer[1]);
+		color[2] = tempColor[2] * (1.0f / overallmultiplyer[2]);
+		color[3] = tempColor[3] * (1.0f / overallmultiplyer[3]);
 	}
 
 }