D627: Memory usage optimization for the compositor.

The compostor used a fixed size of 4 floats to hold pixel data. this
patch will select size of a pixel based on its type.
It uses 1 float for Value, 3 float for vector and 4 floats for color
data types.

When benchmarking on shots (opening shot of caminandes) we get a
reduction of memory of 30% and a tiny speedup as less data
transformations needs to take place (but these are negligable.

More information of the patch can be found on
https://developer.blender.org/D627 and
http://wiki.blender.org/index.php/Dev:Ref/Proposals/Compositor2014_p1.1_TD

Developers: jbakker & mdewanchand
Thanks for Sergey for his indept review.

1 files changed, 9 insertions, 11 deletions

diff --git a/source/blender/compositor/operations/COM_ConvertOperation.cpp b/source/blender/compositor/operations/COM_ConvertOperation.cpp
index 1445b78552f..977586acb60 100644
--- a/source/blender/compositor/operations/COM_ConvertOperation.cpp
+++ b/source/blender/compositor/operations/COM_ConvertOperation.cpp
@@ -49,9 +49,9 @@ ConvertValueToColorOperation::ConvertValueToColorOperation() : ConvertBaseOperat
 
 void ConvertValueToColorOperation::executePixelSampled(float output[4], float x, float y, PixelSampler sampler)
 {
-	float inputValue[4];
-	this->m_inputOperation->readSampled(inputValue, x, y, sampler);
-	output[0] = output[1] = output[2] = inputValue[0];
+	float value;
+	this->m_inputOperation->readSampled(&value, x, y, sampler);
+	output[0] = output[1] = output[2] = value;
 	output[3] = 1.0f;
 }
 
@@ -98,8 +98,9 @@ ConvertColorToVectorOperation::ConvertColorToVectorOperation() : ConvertBaseOper
 
 void ConvertColorToVectorOperation::executePixelSampled(float output[4], float x, float y, PixelSampler sampler)
 {
-	this->m_inputOperation->readSampled(output, x, y, sampler);
-}
+	float color[4];
+	this->m_inputOperation->readSampled(color, x, y, sampler);
+	copy_v3_v3(output, color);}
 
 
 /* ******** Value to Vector ******** */
@@ -112,12 +113,9 @@ ConvertValueToVectorOperation::ConvertValueToVectorOperation() : ConvertBaseOper
 
 void ConvertValueToVectorOperation::executePixelSampled(float output[4], float x, float y, PixelSampler sampler)
 {
-	float input[4];
-	this->m_inputOperation->readSampled(input, x, y, sampler);
-	output[0] = input[0];
-	output[1] = input[0];
-	output[2] = input[0];
-	output[3] = 0.0f;
+	float value;
+	this->m_inputOperation->readSampled(&value, x, y, sampler);
+	output[0] = output[1] = output[2] = value;
 }