--User changes

-Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth.
 -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted.
 -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember.
 -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render.
 -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply.
 -Make default temporal samples be 100, whereas before it was 1000 which was overkill.
 -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box.
  -This wasn't otherwise fixable without writing a lot more code.

--Bug fixes
 -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it.
 -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop.
  -These bugs were due to a previous commit. Revert parts of that commit.
 -Prevent a zoom value of less than 0 when reading from xml.
 -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already.
 -Unique file naming was broken because it was looking for _# and the default names ended with -#.
 -Disallow renaming of an ember in the library tree to an empty string.
 -Severe bug that prevented some variations from being read correctly from params generated outside this program.
 -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1.
 -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double.
 -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU.
 -Prevent user from saving stylesheet to default.qss, it's a special reserved filename.

--Code changes
 -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post.
 -Allow for array variables in variations where the address of each element is stored in m_Params.
 -Qualify all math functions with std::
 -No longer use our own Clamp() in OpenCL, instead use the standard clamp().
 -Redesign how functions are used in the variations OpenCL code.
 -Add tests to EmberTester to verify some of the new functionality.
 -Place more const and override qualifiers on functions where appropriate.
 -Add a global rand with a lock to be used very sparingly.
 -Use a map instead of a vector for bad param names in Xml parsing.
 -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear.
 -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread.
 -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember.
 -Add Contains() function to Utils.h.
 -EmberRender: print names of kernels being printed with --dump_kernel option.
 -Clean up EmberTester to handle some of the recent changes.
 -Fix various casts.
 -Replace % 2 with & 1, even though the compiler was likely doing this already.
 -Add new file Variations06.h to accommodate new variations.
 -General cleanup.

Builds/MSVC/Installer/Product.wxs

@@ -13,7 +13,7 @@
   <!--
   Change this for every release.
   -->
-  <?define ProductCode="{F59721C4-C46D-4FFF-BEA4-803FA3BD0BFE}"?>
+  <?define ProductCode="{96EEC105-DB40-4453-AD22-ABA8BC03267D}"?>
   
   <Product Id="$(var.ProductCode)" Name="$(var.ProductName)" Language="1033" Version="$(var.ProductVersion)" Manufacturer="$(var.Manufacturer)" UpgradeCode="$(var.UpgradeCode)">
     <Package

Builds/MSVC/VS2013/Ember.vcxproj

@@ -157,7 +157,8 @@
       <StructMemberAlignment>Default</StructMemberAlignment>
       <PrecompiledHeaderFile>EmberPch.h</PrecompiledHeaderFile>
       <FloatingPointModel>Precise</FloatingPointModel>
-      <AdditionalOptions>/bigobj %(AdditionalOptions)</AdditionalOptions>
+      <AdditionalOptions>/bigobj
+ %(AdditionalOptions)</AdditionalOptions>
       <StringPooling>true</StringPooling>
       <FloatingPointExceptions>false</FloatingPointExceptions>
       <MultiProcessorCompilation>true</MultiProcessorCompilation>
@@ -228,7 +229,8 @@
       <FavorSizeOrSpeed>Speed</FavorSizeOrSpeed>
       <PrecompiledHeaderFile>EmberPch.h</PrecompiledHeaderFile>
       <StringPooling>true</StringPooling>
-      <AdditionalOptions>/bigobj %(AdditionalOptions)</AdditionalOptions>
+      <AdditionalOptions>/bigobj
+ %(AdditionalOptions)</AdditionalOptions>
       <FloatingPointModel>Precise</FloatingPointModel>
       <MultiProcessorCompilation>true</MultiProcessorCompilation>
       <FloatingPointExceptions>false</FloatingPointExceptions>
@@ -295,12 +297,12 @@
     <ClInclude Include="..\..\..\Source\Ember\Variations03.h" />
     <ClInclude Include="..\..\..\Source\Ember\Variations04.h" />
     <ClInclude Include="..\..\..\Source\Ember\Variations05.h" />
+    <ClInclude Include="..\..\..\Source\Ember\Variations06.h" />
     <ClInclude Include="..\..\..\Source\Ember\VariationsDC.h" />
     <ClInclude Include="..\..\..\Source\Ember\Xform.h" />
     <ClInclude Include="..\..\..\Source\Ember\Isaac.h" />
     <ClInclude Include="..\..\..\Source\Ember\Timing.h" />
     <ClInclude Include="..\..\..\Source\Ember\XmlToEmber.h" />
-    <ClInclude Include="resource1.h" />
   </ItemGroup>
   <ItemGroup>
     <ClCompile Include="..\..\..\Source\Ember\Affine2D.cpp" />

Builds/MSVC/VS2013/Ember.vcxproj.filters

@@ -116,8 +116,8 @@
     <ClInclude Include="..\..\..\Source\Ember\EmberMotion.h">
       <Filter>Header Files</Filter>
     </ClInclude>
-    <ClInclude Include="resource1.h">
+    <ClInclude Include="..\..\..\Source\Ember\Variations06.h">
-      <Filter>Header Files</Filter>
+      <Filter>Header Files\Variations</Filter>
     </ClInclude>
   </ItemGroup>
   <ItemGroup>

Builds/MSVC/VS2013/EmberCL.vcxproj

@@ -298,6 +298,7 @@
     </ClCompile>
     <ClCompile Include="..\..\..\Source\EmberCL\FinalAccumOpenCLKernelCreator.cpp" />
     <ClCompile Include="..\..\..\Source\EmberCL\DEOpenCLKernelCreator.cpp" />
+    <ClCompile Include="..\..\..\Source\EmberCL\FunctionMapper.cpp" />
     <ClCompile Include="..\..\..\Source\EmberCL\IterOpenCLKernelCreator.cpp" />
     <ClCompile Include="..\..\..\Source\EmberCL\OpenCLInfo.cpp" />
     <ClCompile Include="..\..\..\Source\EmberCL\OpenCLWrapper.cpp" />
@@ -309,6 +310,7 @@
     <ClInclude Include="..\..\..\Source\EmberCL\EmberCLStructs.h" />
     <ClInclude Include="..\..\..\Source\EmberCL\DEOpenCLKernelCreator.h" />
     <ClInclude Include="..\..\..\Source\EmberCL\FinalAccumOpenCLKernelCreator.h" />
+    <ClInclude Include="..\..\..\Source\EmberCL\FunctionMapper.h" />
     <ClInclude Include="..\..\..\Source\EmberCL\IterOpenCLKernelCreator.h" />
     <ClInclude Include="..\..\..\Source\EmberCL\OpenCLInfo.h" />
     <ClInclude Include="..\..\..\Source\EmberCL\OpenCLWrapper.h" />

Builds/MSVC/VS2013/EmberCL.vcxproj.filters

@@ -42,6 +42,9 @@
     <ClCompile Include="..\..\..\Source\EmberCL\OpenCLInfo.cpp">
       <Filter>Source Files</Filter>
     </ClCompile>
+    <ClCompile Include="..\..\..\Source\EmberCL\FunctionMapper.cpp">
+      <Filter>Source Files</Filter>
+    </ClCompile>
   </ItemGroup>
   <ItemGroup>
     <ClInclude Include="..\..\..\Source\EmberCL\OpenCLWrapper.h">
@@ -74,6 +77,9 @@
     <ClInclude Include="..\..\..\Source\EmberCL\OpenCLInfo.h">
       <Filter>Header Files</Filter>
     </ClInclude>
+    <ClInclude Include="..\..\..\Source\EmberCL\FunctionMapper.h">
+      <Filter>Header Files</Filter>
+    </ClInclude>
   </ItemGroup>
   <ItemGroup>
     <ResourceCompile Include="EmberCL.rc">

Builds/MSVC/VS2013/Fractorium.vcxproj

@@ -338,7 +338,6 @@ xcopy /F /Y /R /D "$(SolutionDir)..\..\..\Data\flam3-palettes.xml" "$(OutDir)"</
     <ClCompile Include="..\..\..\Source\Fractorium\OptionsDialog.cpp" />
     <ClCompile Include="..\..\..\Source\Fractorium\qcssparser.cpp" />
     <ClCompile Include="..\..\..\Source\Fractorium\qcssscanner.cpp" />
-    <ClCompile Include="..\..\..\Source\Fractorium\qcssscanner.h" />
     <ClCompile Include="..\..\..\Source\Fractorium\QssDialog.cpp" />
     <ClCompile Include="..\..\..\Source\Fractorium\QssTextEdit.cpp" />
     <ClCompile Include="..\..\..\Source\Fractorium\SpinBox.cpp" />

Data/dark.qss

@@ -315,6 +315,14 @@ QTreeView::item
 	background-color: rgb(53, 53, 53);
 	color: darkgray;
 	outline: none;
+	margin-right: 1px;
+}
+
+QTreeView::item:hover
+{
+	border: 0px none black;
+	background-color: gray;
+	outline: none;
 }
 
 QTreeView::item:selected
@@ -609,3 +617,4 @@ QTableView#FinalRenderParamsTable QPushButton
 	margin-bottom: 2px;
 	padding: 0px;
 }
+

Source/Ember/DensityFilter.h

@@ -123,7 +123,7 @@ public:
 		//
 		//    num filters = (de_max_width / de_min_width)^(1 / estimator_curve)
 		//
-		decFilterCount = pow(finalMaxRad / finalMinRad, T(1.0) / m_Curve);
+		decFilterCount = std::pow(finalMaxRad / finalMinRad, T(1.0) / m_Curve);
 
 		if (decFilterCount > 1e7)//Too many filters.
 			return false;
@@ -133,8 +133,8 @@ public:
 		//Condense the smaller kernels to save space.
 		if (intFilterCount > keepThresh)
 		{
-			maxIndex = static_cast<int>(ceil(DE_THRESH + pow(T(intFilterCount - DE_THRESH), m_Curve))) + 1;
+			maxIndex = static_cast<int>(ceil(DE_THRESH + std::pow(T(intFilterCount - DE_THRESH), m_Curve))) + 1;
-			m_MaxFilteredCounts = static_cast<int>(pow(T(maxIndex - DE_THRESH), T(1.0) / m_Curve)) + DE_THRESH;
+			m_MaxFilteredCounts = static_cast<int>(std::pow(T(maxIndex - DE_THRESH), T(1.0) / m_Curve)) + DE_THRESH;
 		}
 		else
 		{
@@ -163,12 +163,12 @@ public:
 			//Calculate the filter width for this number of hits in a bin.
 			if (filterLoop < keepThresh)
 			{
-				filterHeight = (finalMaxRad / pow(T(filterLoop + 1), m_Curve));
+				filterHeight = (finalMaxRad / std::pow(T(filterLoop + 1), m_Curve));
 			}
 			else
 			{
-				loopAdjust = pow(T(filterLoop - keepThresh), (T(1.0) / m_Curve)) + keepThresh;
+				loopAdjust = std::pow(T(filterLoop - keepThresh), (T(1.0) / m_Curve)) + keepThresh;
-				filterHeight = (finalMaxRad / pow(loopAdjust + 1, m_Curve));
+				filterHeight = (finalMaxRad / std::pow(loopAdjust + 1, m_Curve));
 			}
 
 			//Once we've reached the min radius, don't populate any more.
@@ -262,7 +262,7 @@ public:
 		T finalMaxRad = m_MaxRad * m_Supersample + 1;
 		T finalMinRad = m_MinRad * m_Supersample + 1;
 
-		return pow(finalMaxRad / finalMinRad, T(1.0) / m_Curve) <= 1e7;
+		return std::pow(finalMaxRad / finalMinRad, T(1.0) / m_Curve) <= 1e7;
 	}
 
 	/// <summary>

Source/Ember/Ember.cpp

@@ -5,6 +5,7 @@
 namespace EmberNs
 {
 template<> unique_ptr<QTIsaac<ISAAC_SIZE, ISAAC_INT>> QTIsaac<ISAAC_SIZE, ISAAC_INT>::GlobalRand = unique_ptr<QTIsaac<ISAAC_SIZE, ISAAC_INT>>(new QTIsaac<ISAAC_SIZE, ISAAC_INT>());
+CriticalSection QTIsaac<ISAAC_SIZE, ISAAC_INT>::m_CS;
 }
 
 #include "Curves.h"
@@ -20,6 +21,7 @@ template<> unique_ptr<QTIsaac<ISAAC_SIZE, ISAAC_INT>> QTIsaac<ISAAC_SIZE, ISAAC_
 #include "Variations03.h"
 #include "Variations04.h"
 #include "Variations05.h"
+#include "Variations06.h"
 #include "VariationsDC.h"
 #include "VariationList.h"
 #include "Affine2D.h"
@@ -56,7 +58,7 @@ uint Timing::m_ProcessorCount;
 	template<> map<string, vector<Palette<T>>> PaletteList<T>::m_Palettes = map<string, vector<Palette<T>>>(); \
 	template<> bool XmlToEmber<T>::m_Init = false; \
 	template<> vector<string> XmlToEmber<T>::m_FlattenNames = vector<string>(); \
-	template<> vector<pair<string, string>> XmlToEmber<T>::m_BadParamNames = vector<pair<string, string>>(); \
+	template<> unordered_map<string, string> XmlToEmber<T>::m_BadParamNames = unordered_map<string, string>(); \
 	template<> vector<pair<pair<string, string>, vector<string>>> XmlToEmber<T>::m_BadVariationNames = vector<pair<pair<string, string>, vector<string>>>(); \
 	template EMBER_API class Point<T>; \
 	template EMBER_API struct Color<T>; \
@@ -350,6 +352,8 @@ uint Timing::m_ProcessorCount;
 	EXPORTPREPOSTREGVAR(Ho, T) \
 	EXPORTPREPOSTREGVAR(Julia3Dq, T) \
 	EXPORTPREPOSTREGVAR(Line, T) \
+	EXPORTPREPOSTREGVAR(Loonie2, T) \
+	EXPORTPREPOSTREGVAR(Loonie3, T) \
 	EXPORTPREPOSTREGVAR(Loonie3D, T) \
 	EXPORTPREPOSTREGVAR(Mcarpet, T) \
 	EXPORTPREPOSTREGVAR(Waves23D, T) \
@@ -364,12 +368,21 @@ uint Timing::m_ProcessorCount;
 	EXPORTPREPOSTREGVAR(Falloff2, T) \
 	EXPORTPREPOSTREGVAR(Falloff3, T) \
 	EXPORTPREPOSTREGVAR(Xtrb, T) \
-	template EMBER_API class DCBubbleVariation<T>; \
+	EXPORTPREPOSTREGVAR(Hexaplay3D, T) \
+	EXPORTPREPOSTREGVAR(Hexnix3D, T) \
+	EXPORTPREPOSTREGVAR(Hexcrop, T) \
+	EXPORTPREPOSTREGVAR(Hexes, T) \
+	EXPORTPREPOSTREGVAR(Nblur, T) \
+	EXPORTPREPOSTREGVAR(Octapol, T) \
+	EXPORTPREPOSTREGVAR(Crob, T) \
+	EXPORTPREPOSTREGVAR(BubbleT3D, T) \
+	EXPORTPREPOSTREGVAR(Synth, T) \
+	EXPORTPREPOSTREGVAR(DCBubble, T) \
 	EXPORTPREPOSTREGVAR(DCCarpet, T) \
 	EXPORTPREPOSTREGVAR(DCCube, T) \
-	template EMBER_API class DCCylinderVariation<T>; \
+	EXPORTPREPOSTREGVAR(DCCylinder, T) \
 	EXPORTPREPOSTREGVAR(DCGridOut, T) \
-	template EMBER_API class DCLinearVariation<T>; \
+	EXPORTPREPOSTREGVAR(DCLinear, T) \
 	EXPORTPREPOSTREGVAR(DCZTransl, T) \
 	EXPORTPREPOSTREGVAR(DCTriangle, T) \
 	template EMBER_API class VariationList<T>; \

Source/Ember/Ember.h

@@ -201,7 +201,7 @@ public:
 		m_SubBatchSize = DEFAULT_SBS;
 		m_FuseCount = 15;
 		m_Supersample = 1;
-		m_TemporalSamples = 1000;
+		m_TemporalSamples = 100;
 		m_Symmetry = 0;
 		m_Quality = 100;
 		m_PixelsPerUnit = 240;
@@ -226,7 +226,7 @@ public:
 		m_Time = 0;
 		m_Background.Reset();
 		m_Interp = EMBER_INTERP_LINEAR;
-		m_AffineInterp = INTERP_LOG;
+		m_AffineInterp = AFFINE_INTERP_LOG;
 
 		//DE filter.
 		m_MinRadDE = 0;
@@ -505,15 +505,15 @@ public:
 		}
 		else
 		{
-			m_CamMat[0][0] =  cos(-m_CamYaw);
+			m_CamMat[0][0] =  std::cos(-m_CamYaw);
-			m_CamMat[1][0] = -sin(-m_CamYaw);
+			m_CamMat[1][0] = -std::sin(-m_CamYaw);
 			m_CamMat[2][0] = 0;
-			m_CamMat[0][1] =  cos(m_CamPitch) * sin(-m_CamYaw);
+			m_CamMat[0][1] =  std::cos(m_CamPitch) * std::sin(-m_CamYaw);
-			m_CamMat[1][1] =  cos(m_CamPitch) * cos(-m_CamYaw);
+			m_CamMat[1][1] =  std::cos(m_CamPitch) * std::cos(-m_CamYaw);
-			m_CamMat[2][1] = -sin(m_CamPitch);
+			m_CamMat[2][1] = -std::sin(m_CamPitch);
-			m_CamMat[0][2] =  sin(m_CamPitch) * sin(-m_CamYaw);
+			m_CamMat[0][2] =  std::sin(m_CamPitch) * std::sin(-m_CamYaw);
-			m_CamMat[1][2] =  sin(m_CamPitch) * cos(-m_CamYaw);
+			m_CamMat[1][2] =  std::sin(m_CamPitch) * std::cos(-m_CamYaw);
-			m_CamMat[2][2] =  cos(m_CamPitch);
+			m_CamMat[2][2] =  std::cos(m_CamPitch);
 
 			if (projBits & PROJBITS_BLUR)
 			{
@@ -667,10 +667,10 @@ public:
 	}
 
 	/// <summary>
-	/// Flatten all xforms by adding a flatten variation if none is present, and if any of the
+	/// Flatten all xforms by adding a flatten variation if none is present, and if none of the
-	/// variations or parameters in the vector are present.
+	/// variations or parameters in the vector are not present.
 	/// </summary>
-	/// <param name="names">Vector of variation and parameter names</param>
+	/// <param name="names">Vector of variation and parameter names that inhibit flattening</param>
 	/// <returns>True if flatten was added to any of the xforms, false if it already was present or if none of the specified variations or parameters were present.</returns>
 	bool Flatten(vector<string>& names)
 	{
@@ -907,7 +907,7 @@ public:
 			ClampRef<T>(thisXform->m_ColorSpeed, -1, 1);
 
 			//Interp affine and post.
-			if (m_AffineInterp == INTERP_LOG)
+			if (m_AffineInterp == AFFINE_INTERP_LOG)
 			{
 				vector<v2T> cxMag(size);
 				vector<v2T> cxAng(size);
@@ -944,7 +944,7 @@ public:
 					Interpolater<T>::InterpAndConvertBack(coefs, cxAng, cxMag, cxTrn, thisXform->m_Post);
 				}
 			}
-			else if (m_AffineInterp == INTERP_LINEAR)
+			else if (m_AffineInterp == AFFINE_INTERP_LINEAR)
 			{
 				//Interpolate pre and post affine using coefs.
 				allID = true;
@@ -1057,7 +1057,7 @@ public:
 				continue;
 
 			//Assume that if there are no variations, then it's a padding xform.
-			if (m_Xforms[i].Empty() && m_AffineInterp != INTERP_LOG)
+			if (m_Xforms[i].Empty() && m_AffineInterp != AFFINE_INTERP_LOG)
 				continue;
 
 			m_Xforms[i].m_Affine.Rotate(angle);
@@ -1140,8 +1140,8 @@ public:
 			m_Xforms[i].m_ColorSpeed = 0;
 			m_Xforms[i].m_Animate = 0;
 			m_Xforms[i].m_ColorX = m_Xforms[i].m_ColorY = (sym < 3) ? 0 : (T(k - 1) / T(sym - 2));//Added Y.
-			m_Xforms[i].m_Affine.A(Round6(cos(k * a)));
+			m_Xforms[i].m_Affine.A(Round6(std::cos(k * a)));
-			m_Xforms[i].m_Affine.D(Round6(sin(k * a)));
+			m_Xforms[i].m_Affine.D(Round6(std::sin(k * a)));
 			m_Xforms[i].m_Affine.B(Round6(-m_Xforms[i].m_Affine.D()));
 			m_Xforms[i].m_Affine.E(m_Xforms[i].m_Affine.A());
 			m_Xforms[i].m_Affine.C(0);
@@ -1159,7 +1159,7 @@ public:
 	/// Return a uint with bits set to indicate which kind of projection should be done.
 	/// </summary>
 	/// <param name="onlyScaleIfNewIsSmaller">A uint with bits set for each kind of projection that is needed</param>
-	size_t ProjBits()
+	size_t ProjBits() const
 	{
 		size_t val = 0;
 
@@ -1410,9 +1410,9 @@ public:
 			m_MinRadDE = 0;
 			m_CurveDE = T(0.4);
 			m_GammaThresh = T(0.01);
-			m_TemporalSamples = 1000;
+			m_TemporalSamples = 100;
 			m_SpatialFilterType = GAUSSIAN_SPATIAL_FILTER;
-			m_AffineInterp = INTERP_LOG;
+			m_AffineInterp = AFFINE_INTERP_LOG;
 			m_TemporalFilterType = BOX_TEMPORAL_FILTER;
 			m_TemporalFilterWidth = 1;
 			m_TemporalFilterExp = 0;
@@ -1443,7 +1443,7 @@ public:
 			m_GammaThresh = -1;
 			m_TemporalSamples = 0;
 			m_SpatialFilterType = GAUSSIAN_SPATIAL_FILTER;
-			m_AffineInterp = INTERP_LOG;
+			m_AffineInterp = AFFINE_INTERP_LOG;
 			m_TemporalFilterType = BOX_TEMPORAL_FILTER;
 			m_TemporalFilterWidth = -1;
 			m_TemporalFilterExp = -999;

Source/Ember/EmberDefines.h

@@ -29,15 +29,13 @@
 
 //Wrap the sincos function for Macs and PC.
 #if defined(__APPLE__) || defined(_MSC_VER)
-	#define sincos(x, s, c) *(s)=sin(x); *(c)=cos(x);
+	#define sincos(x, s, c) *(s)=std::sin(x); *(c)=std::cos(x);
 #else
-	//extern void sincos(double x, double *s, double *c);
+	/*static void sincos(float x, float* s, float* c)
-	//extern void sincos(float x, float *s, float *c);
-	static void sincos(float x, float* s, float* c)
 	{
-		*s = sin(x);
+		*s = std::sin(x);
-		*c = cos(x);
+		*c = std::cos(x);
-	}
+	}*/
 #endif
 
 namespace EmberNs
@@ -92,6 +90,15 @@ static inline size_t NowMs()
 #define DO_DOUBLE 1//Comment this out for shorter build times during development. Always uncomment for release.
 //#define ISAAC_FLAM3_DEBUG 1//This is almost never needed, but is very useful when troubleshooting difficult bugs. Enable it to do a side by side comparison with flam3.
 
+//These two must always match.
+#ifdef WIN32
+#define ALIGN __declspec(align(16))
+#else
+#define ALIGN __attribute__ ((aligned (16)))
+#endif
+
+#define ALIGN_CL "((aligned (16)))"//The extra parens are necessary.
+
 #if GLM_VERSION >= 96
 	#define v2T  glm::tvec2<T, glm::defaultp>
 	#define v3T  glm::tvec3<T, glm::defaultp>
@@ -113,7 +120,7 @@ static inline size_t NowMs()
 #endif
 
 enum eInterp : uint { EMBER_INTERP_LINEAR = 0, EMBER_INTERP_SMOOTH = 1 };
-enum eAffineInterp : uint { INTERP_LINEAR = 0, INTERP_LOG = 1, INTERP_COMPAT = 2, INTERP_OLDER = 3 };
+enum eAffineInterp : uint { AFFINE_INTERP_LINEAR = 0, AFFINE_INTERP_LOG = 1, AFFINE_INTERP_COMPAT = 2, AFFINE_INTERP_OLDER = 3 };
 enum ePaletteMode : uint { PALETTE_STEP = 0, PALETTE_LINEAR = 1 };
 enum ePaletteInterp : uint { INTERP_HSV = 0, INTERP_SWEEP = 1 };
 enum eMotion : uint { MOTION_SIN = 1, MOTION_TRIANGLE = 2, MOTION_HILL = 3, MOTION_SAW = 4 };

Source/Ember/EmberPch.h

@@ -53,6 +53,7 @@
 #include <thread>
 #include <time.h>
 #include <vector>
+#include <unordered_map>
 
 //Third party headers.
 #ifdef _WIN32

Source/Ember/EmberToXml.h

@@ -64,6 +64,8 @@ public:
 	bool Save(const string& filename, vector<Ember<T>>& embers, size_t printEditDepth, bool doEdits, bool intPalette, bool hexPalette, bool append = false, bool start = false, bool finish = false)
 	{
 		bool b = false;
+		bool hasTimes = false;
+		T t = 0;
 		string temp;
 		ofstream f;
 
@@ -76,6 +78,21 @@ public:
 
 			if (f.is_open())
 			{
+				//Check to see if there are valid times by checking if any differed.
+				//If so, assume they were intentionally entered times.
+				for (size_t i = 1; i < embers.size(); i++)
+				{
+					if (embers[i].m_Time != embers[i - 1].m_Time)
+					{
+						hasTimes = true;
+						break;
+					}
+				}
+
+				if (!hasTimes)
+					for (auto& ember : embers)
+						ember.m_Time = t++;
+
 				if ((append && start) || !append)
 				{
 					temp = "<flames>\n";
@@ -170,7 +187,6 @@ public:
 		os << " gamma=\"" << ember.m_Gamma << "\"";
 		os << " highlight_power=\"" << ember.m_HighlightPower << "\"";
 		os << " vibrancy=\"" << ember.m_Vibrancy << "\"";
-		//os << " hue=\"" << ember.m_Hue << "\"";//Oddly enough, flam3 never wrote this value out.//ORIG
 		os << " estimator_radius=\"" << ember.m_MaxRadDE << "\"";
 		os << " estimator_minimum=\"" << ember.m_MinRadDE << "\"";
 		os << " estimator_curve=\"" << ember.m_CurveDE << "\"";
@@ -186,16 +202,18 @@ public:
 		else if (ember.m_PaletteMode == PALETTE_LINEAR)
 			os << " palette_mode=\"linear\"";
 
-		if (ember.m_Interp == EMBER_INTERP_SMOOTH)
+		if (ember.m_Interp == EMBER_INTERP_LINEAR)
+			os << " interpolation=\"linear\"";
+		else if (ember.m_Interp == EMBER_INTERP_SMOOTH)
 			os << " interpolation=\"smooth\"";
 
-		if (ember.m_AffineInterp == INTERP_LINEAR)
+		if (ember.m_AffineInterp == AFFINE_INTERP_LINEAR)
 			os << " interpolation_type=\"linear\"";
-		else if (ember.m_AffineInterp == INTERP_LOG)
+		else if (ember.m_AffineInterp == AFFINE_INTERP_LOG)
 			os << " interpolation_type=\"log\"";
-		else if (ember.m_AffineInterp == INTERP_COMPAT)
+		else if (ember.m_AffineInterp == AFFINE_INTERP_COMPAT)
 			os << " interpolation_type=\"old\"";
-		else if (ember.m_AffineInterp == INTERP_OLDER)
+		else if (ember.m_AffineInterp == AFFINE_INTERP_OLDER)
 			os << " interpolation_type=\"older\"";
 
 		if (ember.m_PaletteInterp == INTERP_SWEEP)

Source/Ember/Interpolate.h

@@ -110,8 +110,8 @@ public:
 					destXform->DeleteVariationById(VAR_LINEAR);
 
 					//Only do the next substitution for log interpolation.
-					if ((i == 0 && destEmbers[i].m_AffineInterp     == INTERP_LOG) ||
+					if ((i == 0 && destEmbers[i].m_AffineInterp     == AFFINE_INTERP_LOG) ||
-						(i > 0  && destEmbers[i - 1].m_AffineInterp == INTERP_LOG))
+						(i > 0  && destEmbers[i - 1].m_AffineInterp == AFFINE_INTERP_LOG))
 					{
 						for (ii = -1; ii <= 1; ii += 2)
 						{
@@ -790,7 +790,7 @@ public:
 				accang[col] += coefs[i] * cxAng[i][col];
 
 				if (accmode[col] == 0)
-					accmag[col] += coefs[i] * log(cxMag[i][col]);
+					accmag[col] += coefs[i] * std::log(cxMag[i][col]);
 				else
 					accmag[col] += coefs[i] * (cxMag[i][col]);
 
@@ -803,12 +803,12 @@ public:
 		for (col = 0; col < 2; col++)
 		{
 			if (accmode[col] == 0)
-				expmag = exp(accmag[col]);
+				expmag = std::exp(accmag[col]);
 			else
 				expmag = accmag[col];
 
-			store.m_Mat[0][col] = expmag * cos(accang[col]);
+			store.m_Mat[0][col] = expmag * std::cos(accang[col]);
-			store.m_Mat[1][col] = expmag * sin(accang[col]);
+			store.m_Mat[1][col] = expmag * std::sin(accang[col]);
 		}
 	}
 
@@ -865,7 +865,7 @@ public:
 		//abs peak values should be not be greater than 1.
 		if (funcNum == MOTION_SIN)
 		{
-			return sin(T(2.0) * T(M_PI) * timeVal);
+			return std::sin(T(2.0) * T(M_PI) * timeVal);
 		}
 		else if (funcNum == MOTION_TRIANGLE)
 		{
@@ -885,7 +885,7 @@ public:
 		}
 		else if (funcNum == MOTION_HILL)
 		{
-			return ((1 - cos(T(2.0) * T(M_PI) * timeVal)) * T(0.5));
+			return ((1 - std::cos(T(2.0) * T(M_PI) * timeVal)) * T(0.5));
 		}
 		else //MOTION_SAW
 		{

Source/Ember/Isaac.h

@@ -1,6 +1,6 @@
 #pragma once
 
-#include "EmberDefines.h"
+#include "Timing.h"
 
 /// <summary>
 /// C++ TEMPLATE VERSION OF Robert J. Jenkins Jr.'s
@@ -65,6 +65,8 @@ public:
 	/// </summary>
 	static unique_ptr<QTIsaac<ALPHA, ISAAC_INT> > GlobalRand;
 
+	static CriticalSection m_CS;
+
 	/// <summary>
 	/// The structure which holds all of the random information.
 	/// </summary>
@@ -116,6 +118,18 @@ public:
 		return ret;
 	}
 
+	/// <summary>
+	/// Locked version of RandByte().
+	/// </summary>
+	/// <returns>The next random integer in the range of 0-255</returns>
+	static inline T LockedRandByte()
+	{
+		m_CS.Enter();
+		T t = GlobalRand->RandByte();
+		m_CS.Leave();
+		return t;
+	}
+
 	/// <summary>
 	/// Return the next random integer.
 	/// </summary>
@@ -129,6 +143,18 @@ public:
 #endif
 	}
 
+	/// <summary>
+	/// Locked version of Rand().
+	/// </summary>
+	/// <returns>The next random integer</returns>
+	static inline T LockedRand()
+	{
+		m_CS.Enter();
+		T t = GlobalRand->Rand();
+		m_CS.Leave();
+		return t;
+	}
+
 	/// <summary>
 	/// Return the next random integer between 0 and the value passed in minus 1.
 	/// </summary>
@@ -139,6 +165,19 @@ public:
 		return (upper == 0) ? Rand() : Rand() % upper;
 	}
 
+	/// <summary>
+	/// Locked version of Rand().
+	/// </summary>
+	/// <param name="upper">A value one greater than the maximum value that will be returned</param>
+	/// <returns>A value between 0 and the value passed in minus 1</returns>
+	static inline T LockedRand(T upper)
+	{
+		m_CS.Enter();
+		T t = GlobalRand->Rand(upper);
+		m_CS.Leave();
+		return t;
+	}
+
 	/// <summary>
 	/// Returns a random floating point value between the specified minimum and maximum.
 	/// Template argument expected to be float or double.
@@ -153,6 +192,21 @@ public:
 		return fMin + (f * (fMax - fMin));
 	}
 
+	/// <summary>
+	/// Locked version of Frand().
+	/// </summary>
+	/// <param name="fMin">The minimum value allowed, inclusive.</param>
+	/// <param name="fMax">The maximum value allowed, inclusive.</param>
+	/// <returns>A new random floating point value within the specified range, inclusive.</returns>
+	template<typename floatType>
+	static inline floatType LockedFrand(floatType fMin, floatType fMax)
+	{
+		m_CS.Enter();
+		floatType t = GlobalRand->Frand<floatType>(fMin, fMax);
+		m_CS.Leave();
+		return t;
+	}
+
 	/// <summary>
 	/// Thin wrapper around a call to Frand() with a range of 0-1.
 	/// Template argument expected to be float or double.
@@ -168,6 +222,19 @@ public:
 #endif
 	}
 
+	/// <summary>
+	/// Locked version of Frand01().
+	/// </summary>
+	/// <returns>A new random number in the range of 0-1, inclusive.</returns>
+	template<typename floatType>
+	static inline floatType LockedFrand01()
+	{
+		m_CS.Enter();
+		floatType t = GlobalRand->Frand01<floatType>();
+		m_CS.Leave();
+		return t;
+	}
+
 	/// <summary>
 	/// Thin wrapper around a call to Frand() with a range of -1-1.
 	/// Template argument expected to be float or double.
@@ -183,6 +250,19 @@ public:
 #endif
 	}
 
+	/// <summary>
+	/// Locked version of Frand11().
+	/// </summary>
+	/// <returns>A new random number in the range of -1-1, inclusive.</returns>
+	template<typename floatType>
+	static inline floatType LockedFrand11()
+	{
+		m_CS.Enter();
+		floatType t = GlobalRand->Frand11<floatType>();
+		m_CS.Leave();
+		return t;
+	}
+
 	/// <summary>
 	/// Not sure what this does.
 	/// </summary>
@@ -193,6 +273,19 @@ public:
 		return RandBit() ? floatType(0.38196) : floatType(0.61804);
 	}
 
+	/// <summary>
+	/// Locked version of GoldenBit().
+	/// </summary>
+	/// <returns>Something that is golden</returns>
+	template<typename floatType>
+	static inline floatType LockedGoldenBit()
+	{
+		m_CS.Enter();
+		floatType t = GlobalRand->GoldenBit<floatType>();
+		m_CS.Leave();
+		return t;
+	}
+
 	/// <summary>
 	/// Returns a random 0 or 1.
 	/// </summary>
@@ -202,6 +295,18 @@ public:
 		return RandByte() & 1;
 	}
 
+	/// <summary>
+	/// Locked version of RandBit().
+	/// </summary>
+	/// <returns>A random 0 or 1</returns>
+	static inline uint LockedRandBit()
+	{
+		m_CS.Enter();
+		uint t = GlobalRand->RandBit();
+		m_CS.Leave();
+		return t;
+	}
+
 	/// <summary>
 	/// A different way of getting a floating point rand in the range -1-1.
 	/// Flam3 used this but it seems unnecessary now, keep around if it's ever needed.

Source/Ember/Iterator.h

@@ -108,12 +108,12 @@ public:
 
 		for (size_t distrib = 0; distrib < distribCount; distrib++)
 		{
-			T totalDensity = 0;
+			double totalDensity = 0;
 
 			//First find the total densities of all xforms.
 			for (i = 0; i < ember.XformCount(); i++)
 			{
-				T d = xforms[i].m_Weight;
+				double d = xforms[i].m_Weight;
 
 				if (distrib > 0)
 					d *= xforms[distrib - 1].Xaos(i);
@@ -127,14 +127,15 @@ public:
 
 			//Calculate how much of a fraction of a the total density each element represents.
 			size_t j = 0;
-			T tempDensity = 0, currentDensityLimit = 0, densityPerElement = totalDensity / CHOOSE_XFORM_GRAIN;
+			//These must be double, else roundoff error will prevent the last element of m_XformDistributions from being set.
+			double tempDensity = 0, currentDensityLimit = 0, densityPerElement = totalDensity / CHOOSE_XFORM_GRAIN;
 
 			//Assign xform indices in order to each element of m_XformDistributions.
 			//The number of elements assigned a given index is proportional to that xform's
 			//density relative to the sum of all densities.
 			for (i = 0; i < ember.XformCount(); i++)
 			{
-				T temp = xforms[i].m_Weight;
+				double temp = xforms[i].m_Weight;
 
 				if (distrib > 0)
 					temp *= xforms[distrib - 1].Xaos(i);
@@ -145,6 +146,11 @@ public:
 				//Also check that j is within the bounds of the distribution array just to be safe in the case of a rounding error.
 				while (tempDensity < currentDensityLimit && j < CHOOSE_XFORM_GRAIN)
 				{
+#ifdef _DEBUG
+					//Ensure distribution contains no out of bounds indices.
+					if (byte(i) >= ember.XformCount())
+						throw "Out of bounds xform index in selection distribution.";
+#endif
 					//printf("offset = %d, xform = %d, running sum = %f\n", j, i, tempDensity);
 					m_XformDistributions[(distrib * CHOOSE_XFORM_GRAIN) + j] = byte(i);
 					tempDensity += densityPerElement;
@@ -152,6 +158,11 @@ public:
 				}
 			}
 
+#ifdef _DEBUG
+			//Ensure every element of the distribution was populated.
+			if (j < CHOOSE_XFORM_GRAIN)
+				throw "Not all distribution elements set, undefined behavior.";
+#endif
 			//Flam3 did this, which gives the same result.
 			//T t = xforms[0].m_Weight;
 			//

Source/Ember/Palette.h

@@ -492,17 +492,17 @@ public:
 	static T CalcAlpha(T density, T gamma, T linrange)
 	{
 		T frac, alpha;
-		T funcval = pow(linrange, gamma);
+		T funcval = std::pow(linrange, gamma);
 
 		if (density > 0)
 		{
 			if (density < linrange)
 			{
 				frac = density / linrange;
-				alpha = (T(1.0) - frac) * density * (funcval / linrange) + frac * pow(density, gamma);
+				alpha = (T(1.0) - frac) * density * (funcval / linrange) + frac * std::pow(density, gamma);
 			}
 			else
-				alpha = pow(density, gamma);
+				alpha = std::pow(density, gamma);
 		}
 		else
 			alpha = 0;
@@ -545,7 +545,7 @@ public:
 		if (maxa > 255 && highPow >= 0)
 		{
 			newls = T(255.0) / maxc;
-			lsratio = pow(newls / ls, highPow);
+			lsratio = std::pow(newls / ls, highPow);
 
 			//Calculate the max-value color (ranged 0 - 1).
 			for (rgbi = 0; rgbi < 3; rgbi++)

Source/Ember/Renderer.cpp

@@ -113,7 +113,7 @@ void Renderer<T, bucketT>::ComputeBounds()
 template <typename T, typename bucketT>
 void Renderer<T, bucketT>::ComputeQuality()
 {
-	m_Scale = pow(T(2.0), Zoom());
+	m_Scale = std::pow(T(2.0), Zoom());
 	m_ScaledQuality = Quality() * m_Scale * m_Scale;
 }
 
@@ -470,7 +470,7 @@ eRenderStatus Renderer<T, bucketT>::Run(vector<byte>& finalImage, double time, s
 			goto Finish;
 		}
 
-		//Don't need to do this every time through for a single image.
+		//Do this every iteration for an animation, or else do it once for a single image.
 		if (TemporalSamples() > 1 || !resume)
 		{
 			ComputeQuality();
@@ -573,7 +573,7 @@ FilterAndAccum:
 		//Apply appropriate filter if iterating is complete.
 		if (filterAndAccumOnly || temporalSample >= TemporalSamples())
 		{
-			fullRun = m_DensityFilter.get() ? GaussianDensityFilter() : LogScaleDensityFilter();
+			fullRun = m_DensityFilter.get() ? GaussianDensityFilter() : LogScaleDensityFilter(forceOutput);
 		}
 		else
 		{
@@ -581,7 +581,7 @@ FilterAndAccum:
 			if (m_DensityFilter.get() && m_InteractiveFilter == FILTER_DE)
 				fullRun = GaussianDensityFilter();
 			else if (!m_DensityFilter.get() || m_InteractiveFilter == FILTER_LOG)
-				fullRun = LogScaleDensityFilter();
+				fullRun = LogScaleDensityFilter(forceOutput);
 		}
 
 		//Only update state if iterating and filtering finished completely (didn't arrive here via forceOutput).
@@ -801,14 +801,42 @@ bool Renderer<T, bucketT>::ResetBuckets(bool resetHist, bool resetAccum)
 	return resetHist || resetAccum;
 }
 
+/// <summary>
+/// Log scales a single row with a specially structured loop that will be vectorized by the compiler.
+/// Note this adds an epsilon to the denomiator used to compute the logScale
+/// value because the conditional check for zero would have prevented the loop from
+/// being vectorized.
+/// </summary>
+/// <param name="row">The absolute element index in the histogram this row starts on</param>
+/// <param name="rowEnd">The absolute element index in the histogram this row ends on</param>
+template <typename T, typename bucketT>
+void Renderer<T, bucketT>::VectorizedLogScale(size_t row, size_t rowEnd)
+{
+	float k1 = float(m_K1);//All types must be float.
+	float k2 = float(m_K2);
+	auto* __restrict hist = m_HistBuckets.data();//Vectorizer can't tell these point to different locations.
+	auto* __restrict acc = m_AccumulatorBuckets.data();
+
+	for (size_t i = row; i < rowEnd; i++)
+	{
+		float logScale = (k1 * std::log(1.0f + hist[i].a * k2)) / (hist[i].a + std::numeric_limits<float>::epsilon());
+
+		acc[i].r = hist[i].r * logScale;//Must break these out individually. Vectorizer can't reason about vec4's overloaded * operator.
+		acc[i].g = hist[i].g * logScale;
+		acc[i].b = hist[i].b * logScale;
+		acc[i].a = hist[i].a * logScale;
+	}
+}
+
 /// <summary>
 /// Perform log scale density filtering.
 /// Base case for simple log scale density estimation as discussed (mostly) in the paper
 /// in section 4, p. 6-9.
 /// </summary>
+/// <param name="forceOutput">Whether this output was forced due to an interactive render</param>
 /// <returns>True if not prematurely aborted, else false.</returns>
 template <typename T, typename bucketT>
-eRenderStatus Renderer<T, bucketT>::LogScaleDensityFilter()
+eRenderStatus Renderer<T, bucketT>::LogScaleDensityFilter(bool forceOutput)
 {
 	size_t startRow = 0;
 	size_t endRow = m_SuperRasH;
@@ -816,30 +844,46 @@ eRenderStatus Renderer<T, bucketT>::LogScaleDensityFilter()
 	size_t endCol = m_SuperRasW;
 	//Timing t(4);
 
-	//Original didn't parallelize this, doing so gives a 50-75% speedup.
+	//if (forceOutput)//Assume interactive render, so speed up at the expense of slight quality.
-	//The value can be directly assigned, which is quicker than summing.
+	//{
-	parallel_for(startRow, endRow, [&] (size_t j)
+	//	parallel_for(startRow, endRow, [&](size_t j)
+	//	{
+	//		if (!m_Abort)
+	//		{
+	//			size_t row = j * m_SuperRasW;
+	//			size_t rowEnd = row + endCol;
+
+	//			VectorizedLogScale(row, rowEnd);
+	//		}
+	//	});
+	//}
+	//else
 	{
-		size_t row = j * m_SuperRasW;
+		//Original didn't parallelize this, doing so gives a 50-75% speedup.
-		//__m128 logm128;//Figure out SSE at some point.
+		//The value can be directly assigned, which is quicker than summing.
-		//__m128 bucketm128;
+		parallel_for(startRow, endRow, [&](size_t j)
-		//__m128 scaledBucket128;
-
-		for (size_t i = startCol; (i < endCol) && !m_Abort; i++)
 		{
-			size_t index = row + i;
+			size_t row = j * m_SuperRasW;
+			size_t rowEnd = row + endCol;
 
-			//Check for visibility first before doing anything else to avoid all possible unnecessary calculations.
+			if (!m_Abort)
-			if (m_HistBuckets[index].a != 0)
 			{
-				bucketT logScale = (m_K1 * log(1 + m_HistBuckets[index].a * m_K2)) / m_HistBuckets[index].a;
+				for (size_t i = row; i < rowEnd; i++)
+				{
+					//Check for visibility first before doing anything else to avoid all possible unnecessary calculations.
+					if (m_HistBuckets[i].a != 0)
+					{
+						bucketT logScale = (m_K1 * std::log(1 + m_HistBuckets[i].a * m_K2)) / m_HistBuckets[i].a;
 
-				//Original did a temporary assignment, then *= logScale, then passed the result to bump_no_overflow().
+						//Original did a temporary assignment, then *= logScale, then passed the result to bump_no_overflow().
-				//Combine here into one operation for a slight speedup.
+						//Combine here into one operation for a slight speedup.
-				m_AccumulatorBuckets[index] = m_HistBuckets[index] * logScale;
+						m_AccumulatorBuckets[i] = m_HistBuckets[i] * logScale;
+					}
+				}
 			}
-		}
+		});
-	});
+	}
+
 	//t.Toc(__FUNCTION__);
 
 	return m_Abort ? RENDER_ABORT : RENDER_OK;
@@ -858,7 +902,7 @@ eRenderStatus Renderer<T, bucketT>::GaussianDensityFilter()
 	Timing totalTime, localTime;
 	bool scf = !(Supersample() & 1);
 	intmax_t ss = Floor<T>(Supersample() / T(2));
-	T scfact = pow(Supersample() / (Supersample() + T(1)), T(2));
+	T scfact = std::pow(Supersample() / (Supersample() + T(1)), T(2));
 
 	size_t threads = m_ThreadsToUse;
 	size_t startRow = Supersample() - 1;
@@ -896,7 +940,7 @@ eRenderStatus Renderer<T, bucketT>::GaussianDensityFilter()
 				if (bucket->a == 0)
 					continue;
 
-				bucketT cacheLog = (m_K1 * log(1 + bucket->a * m_K2)) / bucket->a;//Caching this calculation gives a 30% speedup.
+				bucketT cacheLog = (m_K1 * std::log(1 + bucket->a * m_K2)) / bucket->a;//Caching this calculation gives a 30% speedup.
 
 				if (ss == 0)
 				{
@@ -928,7 +972,7 @@ eRenderStatus Renderer<T, bucketT>::GaussianDensityFilter()
 				else if (filterSelect <= DE_THRESH)
 					filterSelectInt = size_t(ceil(filterSelect)) - 1;
 				else
-					filterSelectInt = DE_THRESH + size_t(Floor<T>(pow(filterSelect - DE_THRESH, m_DensityFilter->Curve())));
+					filterSelectInt = DE_THRESH + size_t(Floor<T>(std::pow(filterSelect - DE_THRESH, m_DensityFilter->Curve())));
 
 				//If the filter selected below the min specified clamp it to the min.
 				if (filterSelectInt > m_DensityFilter->MaxFilterIndex())
@@ -1214,6 +1258,13 @@ EmberStats Renderer<T, bucketT>::Iterate(size_t iterCount, size_t temporalSample
 	double percent, etaMs;
 	EmberStats stats;
 
+	//Do this every iteration for an animation, or else do it once for a single image. CPU only.
+	if (!m_LastIter)
+	{
+		m_ThreadEmbers.clear();
+		m_ThreadEmbers.insert(m_ThreadEmbers.begin(), m_ThreadsToUse, m_Ember);
+	}
+	
 #ifdef TG
 	size_t threadIndex;
 
@@ -1261,7 +1312,8 @@ EmberStats Renderer<T, bucketT>::Iterate(size_t iterCount, size_t temporalSample
 			//Finally, iterate.
 			//t.Tic();
 			//Iterating, loop 3.
-			m_BadVals[threadIndex] += m_Iterator->Iterate(m_Ember, params, m_Samples[threadIndex].data(), m_Rand[threadIndex]);
+			m_BadVals[threadIndex] += m_Iterator->Iterate(m_ThreadEmbers[threadIndex], params, m_Samples[threadIndex].data(), m_Rand[threadIndex]);
+			//m_BadVals[threadIndex] += m_Iterator->Iterate(m_Ember, params, m_Samples[threadIndex].data(), m_Rand[threadIndex]);
 			//iterationTime += t.Toc();
 
 			if (m_LockAccum)
@@ -1459,8 +1511,6 @@ void Renderer<T, bucketT>::Accumulate(QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand, Poin
 	size_t histIndex, intColorIndex, histSize = m_HistBuckets.size();
 	bucketT colorIndex, colorIndexFrac;
 	auto dmap = palette->m_Entries.data();
-	//T oneColDiv2 = m_CarToRas.OneCol() / 2;
-	//T oneRowDiv2 = m_CarToRas.OneRow() / 2;
 	
 	//It's critical to understand what's going on here as it's one of the most important parts of the algorithm.
 	//A color value gets retrieved from the palette and
@@ -1484,17 +1534,6 @@ void Renderer<T, bucketT>::Accumulate(QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand, Poin
 			p.m_Y = (p00 * m_RotMat.D()) + (p11 * m_RotMat.E()) + m_Ember.m_RotCenterY;
 		}
 
-		//T angle = rand.Frand01<T>() * M_2PI;
-		//T r = exp(T(0.5) * sqrt(-log(rand.Frand01<T>()))) - 1;
-
-		//T r = (rand.Frand01<T>() + rand.Frand01<T>() - 1);
-		//T r = (rand.Frand01<T>() + rand.Frand01<T>() + rand.Frand01<T>() + rand.Frand01<T>() - 2);
-
-		//p.m_X += (r * oneColDiv2) * cos(angle);
-		//p.m_Y += (r * oneRowDiv2) * sin(angle);
-		//p.m_X += r * cos(angle);
-		//p.m_Y += r * sin(angle);
-
 		//Checking this first before converting gives better performance than converting and checking a single value, which the original did.
 		//Second, an interesting optimization observation is that when keeping the bounds vars within m_CarToRas and calling its InBounds() member function,
 		//rather than here as members, about a 7% speedup is achieved. This is possibly due to the fact that data from m_CarToRas is accessed
@@ -1610,7 +1649,7 @@ void Renderer<T, bucketT>::GammaCorrection(tvec4<bucketT, glm::defaultp>& bucket
 
 	for (glm::length_t rgbi = 0; rgbi < 3; rgbi++)
 	{
-		a = newRgb[rgbi] + ((1 - vibrancy) * 255 * pow(bucket[rgbi], g));
+		a = newRgb[rgbi] + ((1 - vibrancy) * 255 * std::pow(bucket[rgbi], g));
 
 		if (NumChannels() <= 3 || !Transparency())
 		{

Source/Ember/Renderer.h

@@ -70,7 +70,7 @@ protected:
 	virtual void MakeDmap(T colorScalar);
 	virtual bool Alloc(bool histOnly = false);
 	virtual bool ResetBuckets(bool resetHist = true, bool resetAccum = true);
-	virtual eRenderStatus LogScaleDensityFilter();
+	virtual eRenderStatus LogScaleDensityFilter(bool forceOutput = false);
 	virtual eRenderStatus GaussianDensityFilter();
 	virtual eRenderStatus AccumulatorToFinalImage(vector<byte>& pixels, size_t finalOffset);
 	virtual eRenderStatus AccumulatorToFinalImage(byte* pixels, size_t finalOffset);
@@ -152,6 +152,7 @@ protected:
 	/*inline*/ void AddToAccum(const tvec4<bucketT, glm::defaultp>& bucket, intmax_t i, intmax_t ii, intmax_t j, intmax_t jj);
 	template <typename accumT> void GammaCorrection(tvec4<bucketT, glm::defaultp>& bucket, Color<bucketT>& background, bucketT g, bucketT linRange, bucketT vibrancy, bool doAlpha, bool scale, accumT* correctedChannels);
 	void CurveAdjust(bucketT& a, const glm::length_t& index);
+	void VectorizedLogScale(size_t row, size_t rowEnd);
 
 protected:
 	T m_Scale;
@@ -173,6 +174,7 @@ protected:
 	Ember<T> m_TempEmber;
 	Ember<T> m_LastEmber;
 	vector<Ember<T>> m_Embers;
+	vector<Ember<T>> m_ThreadEmbers;
 	CarToRas<T> m_CarToRas;
 	Iterator<T>* m_Iterator;
 	unique_ptr<StandardIterator<T>> m_StandardIterator;

Source/Ember/SheepTools.h

@@ -185,8 +185,9 @@ public:
 	/// <param name="useVars">The variations to use if the mutation mode is random</param>
 	/// <param name="sym">The type of symmetry to add if random specified. If 0, it will be added randomly.</param>
 	/// <param name="speed">The speed to multiply the pre affine transforms by if the mutate mode is MUTATE_ALL_COEFS, else ignored.</param>
+	/// <param name="maxVars">The maximum number of variations to allow in any single xform in the ember.</param>
 	/// <returns>A string describing what was done</returns>
-	string Mutate(Ember<T>& ember, eMutateMode mode, vector<eVariationId>& useVars, intmax_t sym, T speed)
+	string Mutate(Ember<T>& ember, eMutateMode mode, vector<eVariationId>& useVars, intmax_t sym, T speed, size_t maxVars)
 	{
 		bool done = false;
 		size_t modXform;
@@ -224,7 +225,7 @@ public:
 			do
 			{
 				//Create a random flame, and use the variations to replace those in the original.
-				Random(mutation, useVars, sym, ember.TotalXformCount());
+				Random(mutation, useVars, sym, ember.TotalXformCount(), maxVars);
 
 				for (size_t i = 0; i < ember.TotalXformCount(); i++)
 				{
@@ -255,7 +256,7 @@ public:
 		else if (mode == MUTATE_ONE_XFORM_COEFS)
 		{
 			//Generate a 2-xform random.
-			Random(mutation, useVars, sym, 2);
+			Random(mutation, useVars, sym, 2, maxVars);
 
 			//Which xform to mutate?
 			modXform = m_Rand.Rand() % ember.TotalXformCount();
@@ -305,10 +306,10 @@ public:
 						T f = T(M_PI) * m_Rand.Frand11<T>();
 						T ra, rb, rd, re;
 
-						ra = (xform->m_Affine.A() * cos(f) + xform->m_Affine.B() * -sin(f));
+						ra = (xform->m_Affine.A() * std::cos(f) + xform->m_Affine.B() * -std::sin(f));
-						rd = (xform->m_Affine.A() * sin(f) + xform->m_Affine.D() *  cos(f));
+						rd = (xform->m_Affine.A() * std::sin(f) + xform->m_Affine.D() *  std::cos(f));
-						rb = (xform->m_Affine.B() * cos(f) + xform->m_Affine.E() * -sin(f));
+						rb = (xform->m_Affine.B() * std::cos(f) + xform->m_Affine.E() * -std::sin(f));
-						re = (xform->m_Affine.B() * sin(f) + xform->m_Affine.E() *  cos(f));
+						re = (xform->m_Affine.B() * std::sin(f) + xform->m_Affine.E() *  std::cos(f));
 
 						xform->m_Affine.A(ra);
 						xform->m_Affine.B(rb);
@@ -317,10 +318,10 @@ public:
 
 						f *= -1;
 
-						ra = (xform->m_Post.A() * cos(f) + xform->m_Post.B() * -sin(f));
+						ra = (xform->m_Post.A() * std::cos(f) + xform->m_Post.B() * -std::sin(f));
-						rd = (xform->m_Post.A() * sin(f) + xform->m_Post.D() *  cos(f));
+						rd = (xform->m_Post.A() * std::sin(f) + xform->m_Post.D() *  std::cos(f));
-						rb = (xform->m_Post.B() * cos(f) + xform->m_Post.E() * -sin(f));
+						rb = (xform->m_Post.B() * std::cos(f) + xform->m_Post.E() * -std::sin(f));
-						re = (xform->m_Post.B() * sin(f) + xform->m_Post.E() *  cos(f));
+						re = (xform->m_Post.B() * std::sin(f) + xform->m_Post.E() *  std::cos(f));
 
 						xform->m_Post.A(ra);
 						xform->m_Post.B(rb);
@@ -408,7 +409,7 @@ public:
 		else if (mode == MUTATE_ALL_COEFS)
 		{
 			os << "mutate all coefs";
-			Random(mutation, useVars, sym, ember.TotalXformCount());
+			Random(mutation, useVars, sym, ember.TotalXformCount(), maxVars);
 
 			//Change all the coefs by a fraction of the random.
 			for (size_t x = 0; x < ember.TotalXformCount(); x++)
@@ -599,11 +600,12 @@ public:
 	/// Thin wrapper around Random() that passes an empty vector for useVars, a random value for symmetry and 0 for max xforms.
 	/// </summary>
 	/// <param name="ember">The newly created random ember</param>
-	void Random(Ember<T>& ember)
+	/// <param name="maxVars">The maximum number of variations to allow in any single xform in the ember.</param>
+	void Random(Ember<T>& ember, size_t maxVars)
 	{
 		vector<eVariationId> useVars;
 
-		Random(ember, useVars, static_cast<intmax_t>(m_Rand.Frand<T>(-2, 2)), 0);
+		Random(ember, useVars, static_cast<intmax_t>(m_Rand.Frand<T>(-2, 2)), 0, maxVars);
 	}
 
 	/// <summary>
@@ -613,7 +615,8 @@ public:
 	/// <param name="useVars">A list of variations to use. If empty, any variation can be used.</param>
 	/// <param name="sym">The symmetry type to use from -2 to 2</param>
 	/// <param name="specXforms">The number of xforms to use. If 0, a quasi random count is used.</param>
-	void Random(Ember<T>& ember, vector<eVariationId>& useVars, intmax_t sym, size_t specXforms)
+	/// <param name="maxVars">The maximum number of variations to allow in any single xform in the ember.</param>
+	void Random(Ember<T>& ember, vector<eVariationId>& useVars, intmax_t sym, size_t specXforms, size_t maxVars)
 	{
 		bool postid, addfinal = false;
 		int var, samed, multid, samepost;
@@ -705,11 +708,13 @@ public:
 
 				if (var > -1)
 				{
-					xform->AddVariation(m_VariationList.GetVariation(var)->Copy());//Use only one variation specified for all xforms.
+					if (xform->TotalVariationCount() < maxVars)
+						xform->AddVariation(m_VariationList.GetVariation(var)->Copy());//Use only one variation specified for all xforms.
 				}
 				else if (multid && var == -1)
 				{
-					xform->AddVariation(m_VariationList.GetVariation(m_Rand.Rand() % varCount)->Copy());//Choose a random var for this xform.
+					if (xform->TotalVariationCount() < maxVars)
+						xform->AddVariation(m_VariationList.GetVariation(m_Rand.Rand() % varCount)->Copy());//Choose a random var for this xform.
 				}
 				else
 				{
@@ -719,7 +724,8 @@ public:
 						Xform<T>* prevXform = ember.GetXform(i - 1);
 
 						for (j = 0; j < prevXform->TotalVariationCount(); j++)
-							xform->AddVariation(prevXform->GetVariation(j)->Copy());
+							if (xform->TotalVariationCount() < maxVars)
+								xform->AddVariation(prevXform->GetVariation(j)->Copy());
 					}
 					else
 					{
@@ -735,21 +741,24 @@ public:
 						//the probability that multiple vars are used.
 						for (j = 0; j < n; j++)
 						{
-							if (var != -2)
+							if (xform->TotalVariationCount() < maxVars)
 							{
-								//Pick a random variation and use a random weight from 0-1.
+								if (var != -2)
-								Variation<T>* v = m_VariationList.GetVariationCopy(static_cast<size_t>(m_Rand.Rand() % varCount), m_Rand.Frand<T>(T(0.001), 1));
+								{
+									//Pick a random variation and use a random weight from 0-1.
+									Variation<T>* v = m_VariationList.GetVariationCopy(static_cast<size_t>(m_Rand.Rand() % varCount), m_Rand.Frand<T>(T(0.001), 1));
 
-								if (v && !xform->AddVariation(v))
+									if (v && !xform->AddVariation(v))
-									delete v;//It already existed and therefore was not added.
+										delete v;//It already existed and therefore was not added.
-							}
+								}
-							else
+								else
-							{
+								{
-								//Pick a random variation from the suppled IDs and use a random weight from 0-1.
+									//Pick a random variation from the suppled IDs and use a random weight from 0-1.
-								Variation<T>* v = m_VariationList.GetVariationCopy(useVars[m_Rand.Rand() % useVars.size()], m_Rand.Frand<T>(T(0.001), 1));
+									Variation<T>* v = m_VariationList.GetVariationCopy(useVars[m_Rand.Rand() % useVars.size()], m_Rand.Frand<T>(T(0.001), 1));
 
-								if (v && !xform->AddVariation(v))
+									if (v && !xform->AddVariation(v))
-									delete v;
+										delete v;
+								}
 							}
 						}
 
@@ -770,15 +779,18 @@ public:
 				//the probability that multiple vars are used.
 				for (j = 0; j < n; j++)
 				{
-					if (var != -2)
+					if (xform->TotalVariationCount() < maxVars)
 					{
-						//Pick a random variation and use a random weight from 0-1.
+						if (var != -2)
-						xform->AddVariation(m_VariationList.GetVariationCopy(static_cast<size_t>(m_Rand.Rand() % varCount), m_Rand.Frand<T>(T(0.001), 1)));
+						{
-					}
+							//Pick a random variation and use a random weight from 0-1.
-					else
+							xform->AddVariation(m_VariationList.GetVariationCopy(static_cast<size_t>(m_Rand.Rand() % varCount), m_Rand.Frand<T>(T(0.001), 1)));
-					{
+						}
-						//Pick a random variation from the suppled IDs and use a random weight from 0-1.
+						else
-						xform->AddVariation(m_VariationList.GetVariationCopy(useVars[m_Rand.Rand() % useVars.size()], m_Rand.Frand<T>(T(0.001), 1)));
+						{
+							//Pick a random variation from the suppled IDs and use a random weight from 0-1.
+							xform->AddVariation(m_VariationList.GetVariationCopy(useVars[m_Rand.Rand() % useVars.size()], m_Rand.Frand<T>(T(0.001), 1)));
+						}
 					}
 				}
 
@@ -1240,8 +1252,8 @@ public:
 	{
 		T r[2];
 		T th = by * 2 * T(M_PI) / 360;
-		T c = cos(th);
+		T c = std::cos(th);
-		T s = -sin(th);
+		T s = -std::sin(th);
 
 		newCenterX -= oldCenterX;
 		newCenterY -= oldCenterY;

Source/Ember/SpatialFilter.h

@@ -207,7 +207,7 @@ protected:
 		x *= T(M_PI);
 
 		if (x != 0)
-			return sin(x) / x;
+			return std::sin(x) / x;
 
 		return 1.0;
 	}
@@ -269,7 +269,7 @@ public:
 	/// <returns>The filtered value</returns>
 	virtual T Filter(T t) const
 	{
-		return exp(-2 * t * t) * std::sqrt(2 / T(M_PI));
+		return std::exp(-2 * t * t) * std::sqrt(2 / T(M_PI));
 	}
 };
 
@@ -604,7 +604,7 @@ public:
 	/// <returns>The filtered value</returns>
 	virtual T Filter(T t) const
 	{
-		return (T(0.42) + T(0.5) * cos(T(M_PI) * t) + T(0.08) * cos(2 * T(M_PI) * t));
+		return (T(0.42) + T(0.5) * std::cos(T(M_PI) * t) + T(0.08) * std::cos(2 * T(M_PI) * t));
 	}
 };
 
@@ -675,7 +675,7 @@ public:
 	/// <returns>The filtered value</returns>
 	virtual T Filter(T t) const
 	{
-		return T(0.54) + T(0.46) * cos(T(M_PI) * t);
+		return T(0.54) + T(0.46) * std::cos(T(M_PI) * t);
 	}
 };
 
@@ -703,7 +703,7 @@ public:
 	/// <returns>The filtered value</returns>
 	virtual T Filter(T t) const
 	{
-		return T(0.5) + T(0.5) * cos(T(M_PI) * t);
+		return T(0.5) + T(0.5) * std::cos(T(M_PI) * t);
 	}
 };
 

Source/Ember/TemporalFilter.h

@@ -207,7 +207,7 @@ public:
 					slpx = T(Size() - i) / Size();
 
 				//Scale the color based on these values.
-				m_Filter[i] = pow(slpx, fabs(filterExp));
+				m_Filter[i] = std::pow(slpx, fabs(filterExp));
 
 				//Keep the max.
 				if (m_Filter[i] > maxFilt)

Source/Ember/Utils.h

@@ -41,6 +41,18 @@ static inline bool FindIf(c& container, pr pred)
 	return std::find_if(container.begin(), container.end(), pred) != container.end();
 }
 
+/// <summary>
+/// Thin wrapper around std::find_if() determine if a value exists at least once.
+/// </summary>
+/// <param name="container">The container to call find_if() on</param>
+/// <param name="val">The value to search for</param>
+/// <returns>True if the value was contained at least once, else false.</returns>
+template<class c, class T>
+static inline bool Contains(c& container, const T& val)
+{
+	return std::find_if(container.begin(), container.end(), [&](const T& t) -> bool { return t == val; }) != container.end();
+}
+
 /// <summary>
 /// Thin wrapper around computing the total size of a vector.
 /// </summary>
@@ -574,7 +586,7 @@ static
 #endif
 float SafeTan<float>(float x)
 {
-	return tan(Clamp<float>(x, FLOAT_MIN_TAN, FLOAT_MAX_TAN));
+	return std::tan(Clamp<float>(x, FLOAT_MIN_TAN, FLOAT_MAX_TAN));
 }
 
 template <>
@@ -583,7 +595,7 @@ static
 #endif
 double SafeTan<double>(double x)
 {
-	return tan(x);
+	return std::tan(x);
 }
 
 /// <summary>
@@ -632,7 +644,7 @@ static inline T Spread(T x, T y)
 template <typename T>
 static inline T Powq4(T x, T y)
 {
-	return pow(fabs(x), y) * SignNz(x);
+	return std::pow(std::fabs(x), y) * SignNz(x);
 }
 
 /// <summary>
@@ -696,7 +708,7 @@ static inline T Fabsmod(T v)
 template <typename T>
 static inline T Fosc(T p, T amp, T ph)
 {
-	return T(0.5) - cos(p * amp + ph) * T(0.5);
+	return T(0.5) - std::cos(p * amp + ph) * T(0.5);
 }
 
 /// <summary>
@@ -708,7 +720,7 @@ static inline T Fosc(T p, T amp, T ph)
 template <typename T>
 static inline T Foscn(T p, T ph)
 {
-	return T(0.5) - cos(p + ph) * T(0.5);
+	return T(0.5) - std::cos(p + ph) * T(0.5);
 }
 
 /// <summary>
@@ -719,7 +731,7 @@ static inline T Foscn(T p, T ph)
 template <typename T>
 static inline T LogScale(T x)
 {
-	return x == 0 ? 0 : log((fabs(x) + 1) * T(M_E)) * SignNz(x) / T(M_E);
+	return x == 0 ? 0 : std::log((fabs(x) + 1) * T(M_E)) * SignNz(x) / T(M_E);
 }
 
 /// <summary>
@@ -730,7 +742,7 @@ static inline T LogScale(T x)
 template <typename T>
 static inline T LogMap(T x)
 {
-	return x == 0 ? 0 : (T(M_E) + log(x * T(M_E))) * T(0.25) * SignNz(x);
+	return x == 0 ? 0 : (T(M_E) + std::log(x * T(M_E))) * T(0.25) * SignNz(x);
 }
 
 /// <summary>

Source/Ember/Variation.h

@@ -73,6 +73,7 @@ enum eVariationId
 	VAR_BTRANSFORM	   ,
 	VAR_BUBBLE		   ,
 	VAR_BUBBLE2		   ,
+	VAR_BUBBLET3D	   ,
 	VAR_BUTTERFLY	   ,
 	VAR_BWRAPS		   ,
 	VAR_CARDIOID	   ,
@@ -101,6 +102,7 @@ enum eVariationId
 	VAR_CPOW		   ,
 	VAR_CPOW2		   ,
 	VAR_CRESCENTS	   ,
+	VAR_CROB		   ,
 	VAR_CROP		   ,
 	VAR_CROPN		   ,
 	VAR_CROSS		   ,
@@ -173,6 +175,10 @@ enum eVariationId
 	VAR_HEART		   ,
 	VAR_HEAT           ,
 	VAR_HEMISPHERE	   ,
+	VAR_HEXAPLAY3D	   ,
+	VAR_HEXCROP		   ,
+	VAR_HEXES		   ,
+	VAR_HEXNIX3D	   ,
 	VAR_HO  		   ,
 	VAR_HOLE		   ,
 	VAR_HORSESHOE	   ,
@@ -210,6 +216,8 @@ enum eVariationId
 	VAR_LOG			   ,
 	VAR_LOQ			   ,
 	VAR_LOONIE		   ,
+	VAR_LOONIE2		   ,
+	VAR_LOONIE3		   ,
 	VAR_LOONIE3D	   ,
 	VAR_MASK		   ,
 	VAR_MCARPET		   ,
@@ -223,10 +231,12 @@ enum eVariationId
 	VAR_MODULUS		   ,
 	VAR_MURL		   ,
 	VAR_MURL2		   ,
+	VAR_NBLUR		   ,
 	VAR_NGON		   ,
 	VAR_NOISE		   ,
 	VAR_NPOLAR		   ,
 	VAR_OCTAGON		   ,
+	VAR_OCTAPOL		   ,
 	VAR_ORTHO		   ,
 	VAR_OSCILLOSCOPE   ,
 	VAR_OVOID		   ,
@@ -284,6 +294,7 @@ enum eVariationId
 	VAR_SINUS_GRID	   ,
 	VAR_SINUSOIDAL	   ,
 	VAR_SINUSOIDAL3D   ,
+	//VAR_SMARTCROP	   ,
 	VAR_SPHERICAL	   ,
 	VAR_SPHERICAL3D	   ,
 	VAR_SPHERICALN	   ,
@@ -308,6 +319,7 @@ enum eVariationId
 	VAR_SUPER_SHAPE3D  ,
 	VAR_SVF			   ,
 	VAR_SWIRL		   ,
+	VAR_SYNTH		   ,
 	VAR_TAN			   ,
 	VAR_TANCOS         ,
 	VAR_TANGENT		   ,
@@ -369,6 +381,7 @@ enum eVariationId
 	VAR_PRE_BTRANSFORM,
 	VAR_PRE_BUBBLE,
 	VAR_PRE_BUBBLE2,
+	VAR_PRE_BUBBLET3D,
 	VAR_PRE_BUTTERFLY,
 	VAR_PRE_BWRAPS,
 	VAR_PRE_CARDIOID,
@@ -397,6 +410,7 @@ enum eVariationId
 	VAR_PRE_CPOW,
 	VAR_PRE_CPOW2,
 	VAR_PRE_CRESCENTS,
+	VAR_PRE_CROB,
 	VAR_PRE_CROP,
 	VAR_PRE_CROPN,
 	VAR_PRE_CROSS,
@@ -469,6 +483,10 @@ enum eVariationId
 	VAR_PRE_HEART,
 	VAR_PRE_HEAT,
 	VAR_PRE_HEMISPHERE,
+	VAR_PRE_HEXAPLAY3D,
+	VAR_PRE_HEXCROP,
+	VAR_PRE_HEXES,
+	VAR_PRE_HEXNIX3D,
 	VAR_PRE_HO,
 	VAR_PRE_HOLE,
 	VAR_PRE_HORSESHOE,
@@ -506,6 +524,8 @@ enum eVariationId
 	VAR_PRE_LOG,
 	VAR_PRE_LOQ,
 	VAR_PRE_LOONIE,
+	VAR_PRE_LOONIE2,
+	VAR_PRE_LOONIE3,
 	VAR_PRE_LOONIE3D,
 	VAR_PRE_MASK,
 	VAR_PRE_MCARPET,
@@ -519,10 +539,12 @@ enum eVariationId
 	VAR_PRE_MODULUS,
 	VAR_PRE_MURL,
 	VAR_PRE_MURL2,
+	VAR_PRE_NBLUR,
 	VAR_PRE_NGON,
 	VAR_PRE_NOISE,
 	VAR_PRE_NPOLAR,
 	VAR_PRE_OCTAGON,
+	VAR_PRE_OCTAPOL,
 	VAR_PRE_ORTHO,
 	VAR_PRE_OSCILLOSCOPE,
 	VAR_PRE_OVOID,
@@ -580,6 +602,7 @@ enum eVariationId
 	VAR_PRE_SINUS_GRID,
 	VAR_PRE_SINUSOIDAL,
 	VAR_PRE_SINUSOIDAL3D,
+	//VAR_PRE_SMARTCROP,
 	VAR_PRE_SPHERICAL,
 	VAR_PRE_SPHERICAL3D,
 	VAR_PRE_SPHERICALN,
@@ -604,6 +627,7 @@ enum eVariationId
 	VAR_PRE_SUPER_SHAPE3D,
 	VAR_PRE_SVF,
 	VAR_PRE_SWIRL,
+	VAR_PRE_SYNTH,
 	VAR_PRE_TAN,
 	VAR_PRE_TANCOS,
 	VAR_PRE_TANGENT,
@@ -665,6 +689,7 @@ enum eVariationId
 	VAR_POST_BTRANSFORM,
 	VAR_POST_BUBBLE,
 	VAR_POST_BUBBLE2,
+	VAR_POST_BUBBLET3D,
 	VAR_POST_BUTTERFLY,
 	VAR_POST_BWRAPS,
 	VAR_POST_CARDIOID,
@@ -693,6 +718,7 @@ enum eVariationId
 	VAR_POST_CPOW,
 	VAR_POST_CPOW2,
 	VAR_POST_CRESCENTS,
+	VAR_POST_CROB,
 	VAR_POST_CROP,
 	VAR_POST_CROPN,
 	VAR_POST_CROSS,
@@ -765,6 +791,10 @@ enum eVariationId
 	VAR_POST_HEART,
 	VAR_POST_HEAT,
 	VAR_POST_HEMISPHERE,
+	VAR_POST_HEXAPLAY3D,
+	VAR_POST_HEXCROP,
+	VAR_POST_HEXES,
+	VAR_POST_HEXNIX3D,
 	VAR_POST_HO,
 	VAR_POST_HOLE,
 	VAR_POST_HORSESHOE,
@@ -802,6 +832,8 @@ enum eVariationId
 	VAR_POST_LOG,
 	VAR_POST_LOQ,
 	VAR_POST_LOONIE,
+	VAR_POST_LOONIE2,
+	VAR_POST_LOONIE3,
 	VAR_POST_LOONIE3D,
 	VAR_POST_MASK,
 	VAR_POST_MCARPET,
@@ -815,10 +847,12 @@ enum eVariationId
 	VAR_POST_MODULUS,
 	VAR_POST_MURL,
 	VAR_POST_MURL2,
+	VAR_POST_NBLUR,
 	VAR_POST_NGON,
 	VAR_POST_NOISE,
 	VAR_POST_NPOLAR,
 	VAR_POST_OCTAGON,
+	VAR_POST_OCTAPOL,
 	VAR_POST_ORTHO,
 	VAR_POST_OSCILLOSCOPE,
 	VAR_POST_OVOID,
@@ -876,6 +910,7 @@ enum eVariationId
 	VAR_POST_SINUS_GRID,
 	VAR_POST_SINUSOIDAL,
 	VAR_POST_SINUSOIDAL3D,
+	//VAR_POST_SMARTCROP,
 	VAR_POST_SPHERICAL,
 	VAR_POST_SPHERICAL3D,
 	VAR_POST_SPHERICALN,
@@ -900,6 +935,7 @@ enum eVariationId
 	VAR_POST_SUPER_SHAPE3D,
 	VAR_POST_SVF,
 	VAR_POST_SWIRL,
+	VAR_POST_SYNTH,
 	VAR_POST_TAN,
 	VAR_POST_TANCOS,
 	VAR_POST_TANGENT,
@@ -931,7 +967,7 @@ enum eVariationId
 	VAR_POST_ZSCALE,
 	VAR_POST_ZTRANSLATE,
 
-	//Direct color are special and only some have pre/post counterparts.
+	//Direct color variations are special.
 	VAR_DC_BUBBLE,
 	VAR_DC_CARPET,
 	VAR_DC_CUBE,
@@ -941,15 +977,21 @@ enum eVariationId
 	VAR_DC_TRIANGLE,
 	VAR_DC_ZTRANSL,
 
+	VAR_PRE_DC_BUBBLE,
 	VAR_PRE_DC_CARPET,
 	VAR_PRE_DC_CUBE,
+	VAR_PRE_DC_CYLINDER,
 	VAR_PRE_DC_GRIDOUT,
+	VAR_PRE_DC_LINEAR,
 	VAR_PRE_DC_TRIANGLE,
 	VAR_PRE_DC_ZTRANSL,
 
+	VAR_POST_DC_BUBBLE,
 	VAR_POST_DC_CARPET,
 	VAR_POST_DC_CUBE,
+	VAR_POST_DC_CYLINDER,
 	VAR_POST_DC_GRIDOUT,
+	VAR_POST_DC_LINEAR,
 	VAR_POST_DC_TRIANGLE,
 	VAR_POST_DC_ZTRANSL,
 
@@ -1215,6 +1257,30 @@ public:
 		return ss.str();
 	}
 
+	/// <summary>
+	/// Returns an OpenCL string for the fields in this variation
+	/// that change during iterations.
+	/// Note these are different than regular variation parameters,
+	/// and thus require a completely different solution.
+	/// </summary>
+	/// <returns></returns>
+	virtual string StateOpenCLString() const
+	{
+		return "";
+	}
+
+	/// <summary>
+	/// Returns an OpenCL string for the initialization of the fields in this variation
+	/// that change during iterations.
+	/// Note these are different than regular variation parameters,
+	/// and thus require a completely different solution.
+	/// </summary>
+	/// <returns></returns>
+	virtual string StateInitOpenCLString() const
+	{
+		return "";
+	}
+
 	/// <summary>
 	/// Return the name and weight of the variation as a string.
 	/// </summary>
@@ -1274,6 +1340,12 @@ public:
 	/// <returns>The OpenCL string to perform the equivalent calculation on the GPU in OpenCL</returns>
 	virtual string OpenCLString() const { return ""; }
 
+	/// <summary>
+	/// If the OpenCL string depends on any global functions specific to this variation, return their names.
+	/// </summary>
+	/// <returns>The names for global OpenCL functions specific to this variation</returns>
+	virtual vector<string> OpenCLGlobalFuncNames() const { return vector<string>(); }
+
 	/// <summary>
 	/// If the OpenCL string depends on any functions specific to this variation, return them.
 	/// </summary>
@@ -1389,6 +1461,7 @@ template <typename T> class ParametricVariation;
 /// Each of these takes the form of a name string and a pointer to a value.
 /// Also, some of them can be considered precalculated values, rather than
 /// formal parameters.
+/// Further, some can change state between iterations.
 /// This class encapsulates a single parameter.
 /// Template argument expected to be float or double.
 /// </summary>
@@ -1409,7 +1482,7 @@ public:
 	/// <summary>
 	/// Constructor for a precalc param that takes arguments.
 	/// </summary>
-	/// <param name="isPrecalc">Whether the parameter is actually a precalculated value. Default: false.</param>
+	/// <param name="isPrecalc">Whether the parameter is actually a precalculated value. Always true.</param>
 	/// <param name="param">A pointer to the parameter</param>
 	/// <param name="name">The name of the parameter</param>
 	ParamWithName(bool isPrecalc,
@@ -1419,6 +1492,21 @@ public:
 		Init(param, name, 0, REAL, TLOW, TMAX, true);
 	}
 
+	/// <summary>
+	/// Constructor for a state param that takes arguments.
+	/// </summary>
+	/// <param name="isPrecalc">Whether the parameter is actually a precalculated value. Always true.</param>
+	/// <param name="isState">Whether the parameter changes state between iterations. Always true.</param>
+	/// <param name="param">A pointer to the parameter</param>
+	/// <param name="name">The name of the parameter</param>
+	ParamWithName(bool isPrecalc,
+		bool isState,
+		T* param,
+		string name)
+	{
+		Init(param, name, 0, REAL, TLOW, TMAX, true, true);
+	}
+
 	/// <summary>
 	/// Constructor for a non-precalc param that takes arguments.
 	/// </summary>
@@ -1466,6 +1554,7 @@ public:
 			m_Type = paramWithName.m_Type;
 			m_Name = paramWithName.m_Name;
 			m_IsPrecalc = paramWithName.m_IsPrecalc;
+			m_IsState = paramWithName.m_IsState;
 		}
 
 		return *this;
@@ -1481,7 +1570,8 @@ public:
 	/// <param name="min">The minimum value the parameter can be</param>
 	/// <param name="max">The maximum value the parameter can be</param>
 	/// <param name="isPrecalc">Whether the parameter is actually a precalculated value. Default: false.</param>
-	void Init(T* param, const string& name, T def = 0, eParamType type = REAL, T min = TLOW, T max = TMAX, bool isPrecalc = false)
+	/// <param name="isState">Whether the parameter changes state between iterations. Default: false.</param>
+	void Init(T* param, const string& name, T def = 0, eParamType type = REAL, T min = TLOW, T max = TMAX, bool isPrecalc = false, bool isState = false)
 	{
 		m_Param = param;
 		m_Def = def;
@@ -1490,6 +1580,7 @@ public:
 		m_Type = type;
 		m_Name = name;
 		m_IsPrecalc = isPrecalc;
+		m_IsState = isState;
 
 		Set(m_Def);//Initial value.
 	}
@@ -1569,7 +1660,8 @@ public:
 		   << "Param Min: " << m_Min << endl
 		   << "Param Max: " << m_Max << endl
 		   << "Param Type: " << m_Type << endl
-		   << "Is Precalc: " << m_IsPrecalc << endl;
+		   << "Is Precalc: " << m_IsPrecalc << endl
+		   << "Is State: " << m_IsState << endl;
 
 		return ss.str();
 	}
@@ -1585,6 +1677,7 @@ public:
 	eParamType Type() const { return m_Type; }
 	string Name() const { return m_Name; }
 	bool IsPrecalc() const { return m_IsPrecalc; }
+	bool IsState() const { return m_IsState; }
 
 private:
 	T* m_Param;//Pointer to the parameter value.
@@ -1594,6 +1687,7 @@ private:
 	eParamType m_Type;//The type of the parameter.
 	string m_Name;//Name of the parameter.
 	bool m_IsPrecalc;//Whether the parameter is actually a precalculated value.
+	bool m_IsState;//Whether the parameter changes state between iterations. This is also considered precalc.
 };
 
 /// <summary>
@@ -1694,7 +1788,7 @@ public:
 	/// Get a pointer to a parameter value with the specified name.
 	/// </summary>
 	/// <param name="name">The name to search for</param>
-	/// <returns>A pointer to the parameter value if the name matched, else false.</returns>
+	/// <returns>A pointer to the parameter value if the name matched, else null.</returns>
 	T* GetParam(const char* name) const
 	{
 		for (auto& param : m_Params)
@@ -1808,11 +1902,35 @@ public:
 		return vec;
 	}
 
+	/// <summary>
+	/// Returns an OpenCL string for the fields in this variation
+	/// that change during iterations.
+	/// Note these are different than regular variation parameters,
+	/// and thus require a completely different solution.
+	/// </summary>
+	/// <returns></returns>
+	virtual string StateOpenCLString() const override
+	{
+		ostringstream os, os2;
+		os2 << "_" << XformIndexInEmber() << ";";
+		string index = os2.str();
+
+		for (auto& param : m_Params)
+		{
+			if (param.IsState())
+			{
+				os << "\n\treal_t " << param.Name() << index;
+			}
+		}
+
+		return os.str();
+	}
+
 	/// <summary>
 	/// Return the name, weight and parameters of the variation as a string.
 	/// </summary>
 	/// <returns>The name, weight and parameters of the variation</returns>
-	virtual string ToString() const
+	virtual string ToString() const override
 	{
 		ostringstream ss;
 
@@ -1849,6 +1967,20 @@ protected:
 		}
 	}
 
+	/// <summary>
+	/// Get a pointer to the underlying ParamWithName object with the specified name.
+	/// </summary>
+	/// <param name="name">The name to search for</param>
+	/// <returns>A pointer to the underlying ParamWithName object if the name matched, else null.</returns>
+	const ParamWithName<T>* GetUnderlyingParam(const char* name) const
+	{
+		for (auto& param : m_Params)
+			if (!_stricmp(param.Name().c_str(), name))
+				return &param;
+
+		return nullptr;
+	}
+
 	vector<ParamWithName<T>> m_Params;//The params pointer vector which stores pointer to parameter members of derived classes.
 };
 

Source/Ember/VariationList.h

@@ -5,6 +5,7 @@
 #include "Variations03.h"
 #include "Variations04.h"
 #include "Variations05.h"
+#include "Variations06.h"
 #include "VariationsDC.h"
 
 /// <summary>
@@ -35,7 +36,7 @@ public:
 	/// </summary>
 	VariationList()
 	{
-		m_Variations.reserve(900);//Change this as the list grows.
+		m_Variations.reserve(eVariationId::LAST_VAR);
 		ADDPREPOSTREGVAR(Linear)
 		ADDPREPOSTREGVAR(Sinusoidal)
 		ADDPREPOSTREGVAR(Spherical)
@@ -315,6 +316,8 @@ public:
 		ADDPREPOSTREGVAR(Ho)
 		ADDPREPOSTREGVAR(Julia3Dq)
 		ADDPREPOSTREGVAR(Line)
+		ADDPREPOSTREGVAR(Loonie2)
+		ADDPREPOSTREGVAR(Loonie3)
 		ADDPREPOSTREGVAR(Loonie3D)
 		ADDPREPOSTREGVAR(Mcarpet)
 		ADDPREPOSTREGVAR(Waves23D)
@@ -329,16 +332,25 @@ public:
 		ADDPREPOSTREGVAR(Falloff2)
 		ADDPREPOSTREGVAR(Falloff3)
 		ADDPREPOSTREGVAR(Xtrb)
+		ADDPREPOSTREGVAR(Hexaplay3D)
+		ADDPREPOSTREGVAR(Hexnix3D)
+		ADDPREPOSTREGVAR(Hexcrop)
+		ADDPREPOSTREGVAR(Hexes)
+		ADDPREPOSTREGVAR(Nblur)
+		ADDPREPOSTREGVAR(Octapol)
+		ADDPREPOSTREGVAR(Crob)
+		ADDPREPOSTREGVAR(BubbleT3D)
+		ADDPREPOSTREGVAR(Synth)
 		//ADDPREPOSTREGVAR(LinearXZ)
 		//ADDPREPOSTREGVAR(LinearYZ)
 
 		//DC are special.
-		m_Variations.push_back(new DCBubbleVariation<T>());
+		ADDPREPOSTREGVAR(DCBubble)
 		ADDPREPOSTREGVAR(DCCarpet)
 		ADDPREPOSTREGVAR(DCCube)
-		m_Variations.push_back(new DCCylinderVariation<T>());
+		ADDPREPOSTREGVAR(DCCylinder)
 		ADDPREPOSTREGVAR(DCGridOut)
-		m_Variations.push_back(new DCLinearVariation<T>());
+		ADDPREPOSTREGVAR(DCLinear)
 		ADDPREPOSTREGVAR(DCTriangle)
 		ADDPREPOSTREGVAR(DCZTransl)
 
@@ -348,6 +360,7 @@ public:
 		m_RegVariations.reserve(m_Variations.size()  / 3);
 		m_PreVariations.reserve(m_Variations.size()  / 3);
 		m_PostVariations.reserve(m_Variations.size() / 3);
+		m_ParametricVariations.reserve(size_t(m_Variations.size() * .90));//This is a rough guess at how many are parametric.
 
 		for (auto var : m_Variations) if (var->VarType() == VARTYPE_REG)  m_RegVariations.push_back(var);
 		for (auto var : m_Variations) if (var->VarType() == VARTYPE_PRE)  m_PreVariations.push_back(var);

Source/Ember/Variations01.h

@@ -58,8 +58,8 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		helper.Out.x = m_Weight * sin(helper.In.x);
+		helper.Out.x = m_Weight * std::sin(helper.In.x);
-		helper.Out.y = m_Weight * sin(helper.In.y);
+		helper.Out.y = m_Weight * std::sin(helper.In.y);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -118,6 +118,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
 };
 
 /// <summary>
@@ -208,6 +213,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
 };
 
 /// <summary>
@@ -264,8 +274,8 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		helper.Out.x = m_Weight * helper.m_PrecalcSqrtSumSquares * sin(helper.m_PrecalcAtanxy + helper.m_PrecalcSqrtSumSquares);
+		helper.Out.x = m_Weight * helper.m_PrecalcSqrtSumSquares * std::sin(helper.m_PrecalcAtanxy + helper.m_PrecalcSqrtSumSquares);
-		helper.Out.y = m_Weight * helper.m_PrecalcSqrtSumSquares * cos(helper.m_PrecalcAtanxy - helper.m_PrecalcSqrtSumSquares);
+		helper.Out.y = m_Weight * helper.m_PrecalcSqrtSumSquares * std::cos(helper.m_PrecalcAtanxy - helper.m_PrecalcSqrtSumSquares);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -305,8 +315,8 @@ public:
 		T a = helper.m_PrecalcSqrtSumSquares * helper.m_PrecalcAtanxy;
 		T r = m_Weight * helper.m_PrecalcSqrtSumSquares;
 
-		helper.Out.x = r * sin(a);
+		helper.Out.x = r * std::sin(a);
-		helper.Out.y = (-r) * cos(a);
+		helper.Out.y = (-r) * std::cos(a);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -353,8 +363,8 @@ public:
 		T val = T(M_PI) * helper.m_PrecalcSqrtSumSquares;
 		T r = m_WeightByPI * helper.m_PrecalcAtanxy;
 
-		helper.Out.x = sin(val) * r;
+		helper.Out.x = std::sin(val) * r;
-		helper.Out.y = cos(val) * r;
+		helper.Out.y = std::cos(val) * r;
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -416,8 +426,8 @@ public:
 		T r = Zeps(helper.m_PrecalcSqrtSumSquares);
 		T r1 = m_Weight / r;
 
-		helper.Out.x = r1 * (helper.m_PrecalcCosa + sin(r));
+		helper.Out.x = r1 * (helper.m_PrecalcCosa + std::sin(r));
-		helper.Out.y = r1 * (helper.m_PrecalcSina - cos(r));
+		helper.Out.y = r1 * (helper.m_PrecalcSina - std::cos(r));
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -437,6 +447,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
 };
 
 /// <summary>
@@ -478,6 +493,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
 };
 
 /// <summary>
@@ -497,8 +517,8 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		helper.Out.x = m_Weight * helper.m_PrecalcSina * cos(helper.m_PrecalcSqrtSumSquares);
+		helper.Out.x = m_Weight * helper.m_PrecalcSina * std::cos(helper.m_PrecalcSqrtSumSquares);
-		helper.Out.y = m_Weight * helper.m_PrecalcCosa * sin(helper.m_PrecalcSqrtSumSquares);
+		helper.Out.y = m_Weight * helper.m_PrecalcCosa * std::sin(helper.m_PrecalcSqrtSumSquares);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -540,8 +560,8 @@ public:
 	{
 		T a = helper.m_PrecalcAtanxy;
 		T r = helper.m_PrecalcSqrtSumSquares;
-		T n0 = sin(a + r);
+		T n0 = std::sin(a + r);
-		T n1 = cos(a - r);
+		T n1 = std::cos(a - r);
 		T m0 = n0 * n0 * n0 * r;
 		T m1 = n1 * n1 * n1 * r;
 
@@ -598,8 +618,8 @@ public:
 		if (rand.RandBit())
 			a += T(M_PI);
 
-		helper.Out.x = r * cos(a);
+		helper.Out.x = r * std::cos(a);
-		helper.Out.y = r * sin(a);
+		helper.Out.y = r * std::sin(a);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -694,8 +714,8 @@ public:
 	{
 		T c10 = m_Xform->m_Affine.B();
 		T c11 = m_Xform->m_Affine.E();
-		T nx = helper.In.x + c10 * sin(helper.In.y * m_Dx2);
+		T nx = helper.In.x + c10 * std::sin(helper.In.y * m_Dx2);
-		T ny = helper.In.y + c11 * sin(helper.In.x * m_Dy2);
+		T ny = helper.In.y + c11 * std::sin(helper.In.x * m_Dy2);
 
 		helper.Out.x = m_Weight * nx;
 		helper.Out.y = m_Weight * ny;
@@ -817,8 +837,8 @@ public:
 	{
 		T dx = SafeTan<T>(3 * helper.In.y);
 		T dy = SafeTan<T>(3 * helper.In.x);
-		T nx = helper.In.x + m_Xform->m_Affine.C() * sin(dx);
+		T nx = helper.In.x + m_Xform->m_Affine.C() * std::sin(dx);
-		T ny = helper.In.y + m_Xform->m_Affine.F() * sin(dy);
+		T ny = helper.In.y + m_Xform->m_Affine.F() * std::sin(dy);
 
 		helper.Out.x = m_Weight * nx;
 		helper.Out.y = m_Weight * ny;
@@ -864,11 +884,11 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		T dx = m_Weight * exp(helper.In.x - 1);
+		T dx = m_Weight * std::exp(helper.In.x - 1);
 		T dy = T(M_PI) * helper.In.y;
 
-		helper.Out.x = dx * cos(dy);
+		helper.Out.x = dx * std::cos(dy);
-		helper.Out.y = dx * sin(dy);
+		helper.Out.y = dx * std::sin(dy);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -911,7 +931,7 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		T r = m_Weight * pow(helper.m_PrecalcSqrtSumSquares, helper.m_PrecalcSina);
+		T r = m_Weight * std::pow(helper.m_PrecalcSqrtSumSquares, helper.m_PrecalcSina);
 
 		helper.Out.x = r * helper.m_PrecalcCosa;
 		helper.Out.y = r * helper.m_PrecalcSina;
@@ -953,8 +973,8 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T a = helper.In.x * T(M_PI);
-		T nx =  cos(a) * cosh(helper.In.y);
+		T nx = std::cos(a) * std::cosh(helper.In.y);
-		T ny = -sin(a) * sinh(helper.In.y);
+		T ny = -std::sin(a) * std::sinh(helper.In.y);
 
 		helper.Out.x = m_Weight * nx;
 		helper.Out.y = m_Weight * ny;
@@ -1028,6 +1048,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
 };
 
 /// <summary>
@@ -1061,8 +1086,8 @@ public:
 		T r = m_Weight * helper.m_PrecalcSqrtSumSquares;
 
 		a += (fmod(a + dy, dx) > dx2) ? -dx2 : dx2;
-		helper.Out.x = r * cos(a);
+		helper.Out.x = r * std::cos(a);
-		helper.Out.y = r * sin(a);
+		helper.Out.y = r * std::sin(a);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -1085,6 +1110,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
 };
 
 /// <summary>
@@ -1111,7 +1141,7 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		T r = helper.m_PrecalcSqrtSumSquares * (m_BlobLow + m_BlobDiff * (T(0.5) + T(0.5) * sin(m_BlobWaves * helper.m_PrecalcAtanxy)));
+		T r = helper.m_PrecalcSqrtSumSquares * (m_BlobLow + m_BlobDiff * (T(0.5) + T(0.5) * std::sin(m_BlobWaves * helper.m_PrecalcAtanxy)));
 
 		helper.Out.x = m_Weight * helper.m_PrecalcSina * r;
 		helper.Out.y = m_Weight * helper.m_PrecalcCosa * r;
@@ -1194,10 +1224,10 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		T nx1 = cos(m_PdjB * helper.In.x);
+		T nx1 = std::cos(m_PdjB * helper.In.x);
-		T nx2 = sin(m_PdjC * helper.In.x);
+		T nx2 = std::sin(m_PdjC * helper.In.x);
-		T ny1 = sin(m_PdjA * helper.In.y);
+		T ny1 = std::sin(m_PdjA * helper.In.y);
-		T ny2 = cos(m_PdjD * helper.In.y);
+		T ny2 = std::cos(m_PdjD * helper.In.y);
 
 		helper.Out.x = m_Weight * (ny1 - nx1);
 		helper.Out.y = m_Weight * (nx2 - ny2);
@@ -1300,8 +1330,8 @@ public:
 		else
 			a = a + m_Fan2Dx2;
 
-		helper.Out.x = r * sin(a);
+		helper.Out.x = r * std::sin(a);
-		helper.Out.y = r * cos(a);
+		helper.Out.y = r * std::cos(a);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -1534,9 +1564,9 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		helper.Out.x = m_Weight * sin(helper.In.x);
+		helper.Out.x = m_Weight * std::sin(helper.In.x);
 		helper.Out.y = m_Weight * helper.In.y;
-		helper.Out.z = m_Weight * cos(helper.In.x);
+		helper.Out.z = m_Weight * std::cos(helper.In.x);
 	}
 
 	virtual string OpenCLString() const override
@@ -1601,12 +1631,17 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
+
 	virtual void Precalc() override
 	{
 		T angle = m_Angle * T(M_PI) / 2;
 
-		m_Vsin = sin(angle);
+		m_Vsin = std::sin(angle);
-		m_VfCos = m_Dist * cos(angle);
+		m_VfCos = m_Dist * std::cos(angle);
 	}
 
 	virtual void Random(QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
@@ -1650,8 +1685,8 @@ public:
 		T tempr = rand.Frand01<T>() * M_2PI;
 		T r = m_Weight * rand.Frand01<T>();
 
-		helper.Out.x = helper.In.x * r * cos(tempr);
+		helper.Out.x = helper.In.x * r * std::cos(tempr);
-		helper.Out.y = helper.In.y * r * sin(tempr);
+		helper.Out.y = helper.In.y * r * std::sin(tempr);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -1690,10 +1725,10 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T tempr = (helper.m_PrecalcAtanyx + M_2PI * rand.Rand(ISAAC_INT(m_Rn))) / m_Power;
-		T r = m_Weight * pow(helper.m_PrecalcSumSquares, m_Cn);
+		T r = m_Weight * std::pow(helper.m_PrecalcSumSquares, m_Cn);
 
-		helper.Out.x = r * cos(tempr);
+		helper.Out.x = r * std::cos(tempr);
-		helper.Out.y = r * sin(tempr);
+		helper.Out.y = r * std::sin(tempr);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -1770,15 +1805,15 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		int rnd = int(m_Rn * rand.Frand01<T>());
-		T tempr, r = m_Weight * pow(helper.m_PrecalcSumSquares, m_Cn);
+		T tempr, r = m_Weight * std::pow(helper.m_PrecalcSumSquares, m_Cn);
 
 		if ((rnd & 1) == 0)
 			tempr = (M_2PI * rnd + helper.m_PrecalcAtanyx) / m_Power;
 		else
 			tempr = (M_2PI * rnd - helper.m_PrecalcAtanyx) / m_Power;
 
-		helper.Out.x = r * cos(tempr);
+		helper.Out.x = r * std::cos(tempr);
-		helper.Out.y = r * sin(tempr);
+		helper.Out.y = r * std::sin(tempr);
 		helper.Out.z = m_Weight * helper.In.z;
 
 		//int rnd = (int)(m_Rn * rand.Frand01<T>());
@@ -1892,8 +1927,8 @@ public:
 		T tempr = rand.Frand01<T>() * M_2PI;
 		T r = m_Weight * rand.Frand01<T>();
 
-		helper.Out.x = r * cos(tempr);
+		helper.Out.x = r * std::cos(tempr);
-		helper.Out.y = r * sin(tempr);
+		helper.Out.y = r * std::sin(tempr);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -1931,8 +1966,8 @@ public:
 		T angle = rand.Frand01<T>() * M_2PI;
 		T r = m_Weight * (rand.Frand01<T>() + rand.Frand01<T>() + rand.Frand01<T>() + rand.Frand01<T>() - 2);
 
-		helper.Out.x = r * cos(angle);
+		helper.Out.x = r * std::cos(angle);
-		helper.Out.y = r * sin(angle);
+		helper.Out.y = r * std::sin(angle);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -1979,8 +2014,8 @@ public:
 		T tempa = helper.m_PrecalcAtanyx + m_Spin * rndG;
 		T rz = m_Zoom * rndG - 1;
 
-		helper.Out.x = ra * cos(tempa) + rz * helper.In.x;
+		helper.Out.x = ra * std::cos(tempa) + rz * helper.In.x;
-		helper.Out.y = ra * sin(tempa) + rz * helper.In.y;
+		helper.Out.y = ra * std::sin(tempa) + rz * helper.In.y;
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -2055,8 +2090,8 @@ public:
 		T a = m_Rotation + M_2PI * (sl + rand.Frand01<T>() * m_Thickness) / m_Slices;
 		T r = m_Weight * rand.Frand01<T>();
 
-		helper.Out.x = r * cos(a);
+		helper.Out.x = r * std::cos(a);
-		helper.Out.y = r * sin(a);
+		helper.Out.y = r * std::sin(a);
 	}
 
 	virtual string OpenCLString() const override
@@ -2127,14 +2162,14 @@ public:
 		if ((helper.In.x == 0) && (helper.In.y == 0))
 			rFactor = 0;
 		else
-			rFactor = pow(helper.m_PrecalcSumSquares, m_CPower);
+			rFactor = std::pow(helper.m_PrecalcSumSquares, m_CPower);
 
 		T phi = helper.m_PrecalcAtanyx - m_CSides * Floor<T>(helper.m_PrecalcAtanyx * m_CSidesInv);
 
 		if (phi > T(0.5) * m_CSides)
 			phi -= m_CSides;
 
-		T amp = (m_Corners * (1 / cos(phi) - 1) + m_Circle) * m_Weight * rFactor;
+		T amp = (m_Corners * (1 / std::cos(phi) - 1) + m_Circle) * m_Weight * rFactor;
 
 		helper.Out.x = amp * helper.In.x;
 		helper.Out.y = amp * helper.In.y;
@@ -2269,6 +2304,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
+
 	virtual void Precalc() override
 	{
 		m_C22 = 2 * m_C2;
@@ -2427,7 +2467,7 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		helper.Out.x = m_Weight * sin(helper.In.x) / cos(helper.In.y);
+		helper.Out.x = m_Weight * std::sin(helper.In.x) / std::cos(helper.In.y);
 		helper.Out.y = m_Weight * SafeTan<T>(helper.In.y);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
@@ -2497,8 +2537,8 @@ public:
 		T r = m_Weight / Zeps(helper.m_PrecalcSumSquares);
 		T tanr = m_Weight * SafeTan<T>(ang) * r;
 
-		helper.Out.x = tanr * cos(helper.In.x);
+		helper.Out.x = tanr * std::cos(helper.In.x);
-		helper.Out.y = tanr * sin(helper.In.y);
+		helper.Out.y = tanr * std::sin(helper.In.y);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -2519,6 +2559,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
 };
 
 /// <summary>
@@ -2576,7 +2621,7 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T r = m_Weight * helper.m_PrecalcSqrtSumSquares;
-		T cr = cos(r);
+		T cr = std::cos(r);
 		T icr = 1 / cr;
 
 		helper.Out.x = m_Weight * helper.In.x;
@@ -2630,7 +2675,7 @@ public:
 		T sinr, cosr, diff;
 
 		sincos(r, &sinr, &cosr);
-		diff = log10(sinr * sinr) + cosr;
+		diff = std::log10(sinr * sinr) + cosr;
 
 		if (BadVal(diff))
 			diff = -30.0;
@@ -2698,6 +2743,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
 };
 
 /// <summary>
@@ -2820,14 +2870,14 @@ public:
 	{
 		T theta = m_Pm4 * helper.m_PrecalcAtanyx + T(M_PI_4);
 
-		T t1 = fabs(cos(theta));
+		T t1 = fabs(std::cos(theta));
-		t1 = pow(t1, m_N2);
+		t1 = std::pow(t1, m_N2);
 
-		T t2 = fabs(sin(theta));
+		T t2 = fabs(std::sin(theta));
-		t2 = pow(t2, m_N3);
+		t2 = std::pow(t2, m_N3);
 
 		T r = m_Weight * ((m_Rnd * rand.Frand01<T>() + (1 - m_Rnd) * helper.m_PrecalcSqrtSumSquares) - m_Holes)
-			* pow(t1 + t2, m_PNeg1N1) / helper.m_PrecalcSqrtSumSquares;
+			* std::pow(t1 + t2, m_PNeg1N1) / helper.m_PrecalcSqrtSumSquares;
 
 		helper.Out.x = r * helper.In.x;
 		helper.Out.y = r * helper.In.y;
@@ -2925,7 +2975,7 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T theta = helper.m_PrecalcAtanyx;
-		T r = m_Weight * (rand.Frand01<T>() - m_Holes) * cos(m_Petals * theta) / helper.m_PrecalcSqrtSumSquares;
+		T r = m_Weight * (rand.Frand01<T>() - m_Holes) * std::cos(m_Petals * theta) / helper.m_PrecalcSqrtSumSquares;
 
 		helper.Out.x = r * helper.In.x;
 		helper.Out.y = r * helper.In.y;
@@ -3241,7 +3291,7 @@ public:
 		}
 		else
 		{
-			helper.Out.x = m_V4 * log((t + x2) / (t - x2));
+			helper.Out.x = m_V4 * std::log((t + x2) / (t - x2));
 			helper.Out.y = m_V * y;
 			helper.Out.z = m_Weight * helper.In.z;
 		}
@@ -3497,6 +3547,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
 };
 
 /// <summary>
@@ -3651,12 +3706,12 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T a = helper.m_PrecalcAtanyx;
-		T lnr = T(0.5) * log(helper.m_PrecalcSumSquares);
+		T lnr = T(0.5) * std::log(helper.m_PrecalcSumSquares);
 		T angle = m_C * a + m_D * lnr + m_Ang * Floor<T>(m_Power * rand.Frand01<T>());
-		T m = m_Weight * exp(m_C * lnr - m_D * a);
+		T m = m_Weight * std::exp(m_C * lnr - m_D * a);
 
-		helper.Out.x = m * cos(angle);
+		helper.Out.x = m * std::cos(angle);
-		helper.Out.y = m * sin(angle);
+		helper.Out.y = m * std::sin(angle);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -3740,8 +3795,8 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		helper.Out.x = m_Weight * helper.In.x + m_XAmpV * exp(-helper.In.y * helper.In.y * m_XLengthV);
+		helper.Out.x = m_Weight * helper.In.x + m_XAmpV * std::exp(-helper.In.y * helper.In.y * m_XLengthV);
-		helper.Out.y = m_Weight * helper.In.y + m_YAmpV * exp(-helper.In.x * helper.In.x * m_YLengthV);
+		helper.Out.y = m_Weight * helper.In.y + m_YAmpV * std::exp(-helper.In.x * helper.In.x * m_YLengthV);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -3830,8 +3885,8 @@ public:
 		T r1 = std::sqrt(tmp + tmp2);
 		T r2 = std::sqrt(tmp - tmp2);
 		T xmax = (r1 + r2) * T(0.5);
-		T a1 = log(xmax + std::sqrt(xmax - 1));
+		T a1 = std::log(xmax + std::sqrt(xmax - 1));
-		T a2 = -acos(Clamp<T>(helper.In.x / xmax, -1, 1));
+		T a2 = -std::acos(Clamp<T>(helper.In.x / xmax, -1, 1));
 		T w = m_Weight / T(11.57034632);//This is an interesting magic number.
 		T snv, csv, snhu, cshu;
 
@@ -3860,7 +3915,7 @@ public:
 		   << "\t\treal_t r2 = sqrt(tmp - tmp2);\n"
 		   << "\t\treal_t xmax = (r1 + r2) * (real_t)(0.5);\n"
 		   << "\t\treal_t a1 = log(xmax + sqrt(xmax - (real_t)(1.0)));\n"
-		   << "\t\treal_t a2 = -acos(Clamp(vIn.x / xmax, -(real_t)(1.0), (real_t)(1.0)));\n"
+		   << "\t\treal_t a2 = -acos(clamp(vIn.x / xmax, -(real_t)(1.0), (real_t)(1.0)));\n"
 		   << "\t\treal_t w = xform->m_VariationWeights[" << varIndex << "] / (real_t)(11.57034632);\n"
 		   << "\t\treal_t snv = sin(a1);\n"
 		   << "\t\treal_t csv = cos(a1);\n"
@@ -3916,9 +3971,9 @@ public:
 		helper.Out.x = w * atan2(a, b);
 
 		if (helper.In.y > 0)
-			helper.Out.y = w * log(xmax + ssx);
+			helper.Out.y = w * std::log(xmax + ssx);
 		else
-			helper.Out.y = -(w * log(xmax + ssx));
+			helper.Out.y = -(w * std::log(xmax + ssx));
 
 		helper.Out.z = m_Weight * helper.In.z;
 	}
@@ -3998,12 +4053,12 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T a = helper.m_PrecalcAtanyx;
-		T lnr = T(0.5) * log(helper.m_PrecalcSumSquares);
+		T lnr = T(0.5) * std::log(helper.m_PrecalcSumSquares);
-		T m = m_Weight * exp(m_C * lnr - m_D * a);
+		T m = m_Weight * std::exp(m_C * lnr - m_D * a);
 		T n = m_C * a + m_D * lnr;
 
-		helper.Out.x = m * cos(n);
+		helper.Out.x = m * std::cos(n);
-		helper.Out.y = m * sin(n);
+		helper.Out.y = m * std::sin(n);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -4085,7 +4140,7 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		T expx = exp(helper.In.x) * T(0.5);
+		T expx = std::exp(helper.In.x) * T(0.5);
 		T expnx = T(0.25) / expx;
 		T sn, cn, tmp;
 
@@ -4119,6 +4174,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
 };
 
 /// <summary>
@@ -4145,8 +4205,8 @@ public:
 		{
 			T a = atan2(y, x) + m_Spin + m_Twist * (m_Weight - r);
 
-			helper.Out.x = m_Weight * (r * cos(a) + m_X);//Fix to make it colapse to 0 when weight is 0.//SMOULDER
+			helper.Out.x = m_Weight * (r * std::cos(a) + m_X);//Fix to make it colapse to 0 when weight is 0.//SMOULDER
-			helper.Out.y = m_Weight * (r * sin(a) - m_Y);
+			helper.Out.y = m_Weight * (r * std::sin(a) - m_Y);
 		}
 		else
 		{
@@ -4197,6 +4257,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
+
 	virtual bool SetParamVal(const char* name, T val) override
 	{
 		if (!_stricmp(name, "lazysusan_spin"))
@@ -4431,9 +4496,9 @@ public:
 		T t;
 
 		if (m_Damping == 0.0)
-			t = m_Amplitude * cos(m_2PiFreq * helper.In.x) + m_Separation;
+			t = m_Amplitude * std::cos(m_2PiFreq * helper.In.x) + m_Separation;
 		else
-			t = m_Amplitude * exp(-fabs(helper.In.x) * m_Damping) * cos(m_2PiFreq * helper.In.x) + m_Separation;
+			t = m_Amplitude * std::exp(-fabs(helper.In.x) * m_Damping) * std::cos(m_2PiFreq * helper.In.x) + m_Separation;
 
 		if (fabs(helper.In.y) <= t)
 		{
@@ -4537,7 +4602,7 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		helper.Out.x = m_Vvar * helper.m_PrecalcAtanxy;
-		helper.Out.y = m_Vvar2 * log(helper.m_PrecalcSumSquares);
+		helper.Out.y = m_Vvar2 * std::log(helper.m_PrecalcSumSquares);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -4596,8 +4661,8 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		helper.Out.x = m_Weight * (helper.In.x + m_X * sin(SafeTan<T>(helper.In.y * m_C)));
+		helper.Out.x = m_Weight * (helper.In.x + m_X * std::sin(SafeTan<T>(helper.In.y * m_C)));
-		helper.Out.y = m_Weight * (helper.In.y + m_Y * sin(SafeTan<T>(helper.In.x * m_C)));
+		helper.Out.y = m_Weight * (helper.In.y + m_Y * std::sin(SafeTan<T>(helper.In.x * m_C)));
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -4691,6 +4756,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
+
 	virtual void Precalc() override
 	{
 		m_InvWeight = 1 / Zeps(m_Weight);
@@ -5057,8 +5127,8 @@ public:
 
 		a = a * m_CompFac + c * m_Angle;
 		r = m_Weight * (r + m_Hole);
-		helper.Out.x = r * cos(a);
+		helper.Out.x = r * std::cos(a);
-		helper.Out.y = r * sin(a);
+		helper.Out.y = r * std::sin(a);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -5139,14 +5209,14 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		T r = m_Weight * pow(helper.m_PrecalcSumSquares, m_Cn);
+		T r = m_Weight * std::pow(helper.m_PrecalcSumSquares, m_Cn);
 		int tRand = int(m_Rn * rand.Frand01<T>());
 		T a = (helper.m_PrecalcAtanyx + M_2PI * tRand) / m_Power;
 		T c = T(Floor<T>((m_Count * a + T(M_PI)) * T(M_1_PI) * T(0.5)));
 
 		a = a * m_Cf + c * m_Angle;
-		helper.Out.x = r * cos(a);
+		helper.Out.x = r * std::cos(a);
-		helper.Out.y = r * sin(a);
+		helper.Out.y = r * std::sin(a);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -5242,8 +5312,8 @@ public:
 
 		a = a * compFac + c * m_Angle;
 		r = m_Weight * (r + m_Hole);
-		helper.Out.x = r * cos(a);
+		helper.Out.x = r * std::cos(a);
-		helper.Out.y = r * sin(a);
+		helper.Out.y = r * std::sin(a);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -5274,6 +5344,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
+
 	virtual void Random(QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		m_Angle = T(M_PI) * rand.Frand01<T>();
@@ -5324,8 +5399,8 @@ public:
 		else
 			a = helper.m_PrecalcAtanyx + m_Outside / (m_Weight - r);
 
-		helper.Out.x = m_Weight * r * cos(a);
+		helper.Out.x = m_Weight * r * std::cos(a);
-		helper.Out.y = m_Weight * r * sin(a);
+		helper.Out.y = m_Weight * r * std::sin(a);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -5392,9 +5467,9 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		helper.Out.x = m_Weight * (helper.In.x + m_ScaleX * sin(helper.In.y * m_FreqX));
+		helper.Out.x = m_Weight * (helper.In.x + m_ScaleX * std::sin(helper.In.y * m_FreqX));
-		helper.Out.y = m_Weight * (helper.In.y + m_ScaleY * sin(helper.In.x * m_FreqY));
+		helper.Out.y = m_Weight * (helper.In.y + m_ScaleY * std::sin(helper.In.x * m_FreqY));
-		helper.Out.z = m_Weight * (helper.In.z + m_ScaleZ * sin(helper.m_PrecalcSqrtSumSquares * m_FreqZ));
+		helper.Out.z = m_Weight * (helper.In.z + m_ScaleZ * std::sin(helper.m_PrecalcSqrtSumSquares * m_FreqZ));
 	}
 
 	virtual string OpenCLString() const override
@@ -5465,10 +5540,10 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		T expe = m_Weight * exp(helper.In.x);
+		T expe = m_Weight * std::exp(helper.In.x);
 
-		helper.Out.x = expe * cos(helper.In.y);
+		helper.Out.x = expe * std::cos(helper.In.y);
-		helper.Out.y = expe * sin(helper.In.y);
+		helper.Out.y = expe * std::sin(helper.In.y);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -5505,7 +5580,7 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		helper.Out.x = m_Weight * log(helper.m_PrecalcSumSquares) * m_Denom;
+		helper.Out.x = m_Weight * std::log(helper.m_PrecalcSumSquares) * m_Denom;
 		helper.Out.y = m_Weight * helper.m_PrecalcAtanyx;
 		helper.Out.z = m_Weight * helper.In.z;
 	}
@@ -5530,7 +5605,7 @@ public:
 
 	virtual void Precalc() override
 	{
-		m_Denom = T(0.5) / log(m_Base);
+		m_Denom = T(0.5) / std::log(m_Base);
 	}
 
 protected:
@@ -5561,8 +5636,8 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		helper.Out.x = m_Weight * sin(helper.In.x) * cosh(helper.In.y);
+		helper.Out.x = m_Weight * std::sin(helper.In.x) * cosh(helper.In.y);
-		helper.Out.y = m_Weight * cos(helper.In.x) * sinh(helper.In.y);
+		helper.Out.y = m_Weight * std::cos(helper.In.x) * sinh(helper.In.y);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -5595,8 +5670,8 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		//clamp fabs x and y to 7.104760e+002 for cosh, and |x| 7.104760e+002 for sinh
-		helper.Out.x = m_Weight * cos(helper.In.x) * cosh(helper.In.y);
+		helper.Out.x = m_Weight * std::cos(helper.In.x) * cosh(helper.In.y);
-		helper.Out.y = -(m_Weight * sin(helper.In.x) * sinh(helper.In.y));
+		helper.Out.y = -(m_Weight * std::sin(helper.In.x) * sinh(helper.In.y));
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -5721,9 +5796,9 @@ public:
 		T cscsin, csccos, cscsinh, csccosh, cscden;
 
 		sincos(helper.In.x, &cscsin, &csccos);
-		cscsinh = sinh(helper.In.y);
+		cscsinh = std::sinh(helper.In.y);
-		csccosh = cosh(helper.In.y);
+		csccosh = std::cosh(helper.In.y);
-		cscden = 2 / (cosh(2 * helper.In.y) - cos(2 * helper.In.x));
+		cscden = 2 / (std::cosh(2 * helper.In.y) - std::cos(2 * helper.In.x));
 		helper.Out.x = m_Weight * cscden * cscsin * csccosh;
 		helper.Out.y = -(m_Weight * cscden * csccos * cscsinh);
 		helper.Out.z = m_Weight * helper.In.z;
@@ -5987,9 +6062,9 @@ public:
 		T cschsin, cschcos, cschsinh, cschcosh, cschden;
 
 		sincos(helper.In.y, &cschsin, &cschcos);
-		cschsinh = sinh(helper.In.x);
+		cschsinh = std::sinh(helper.In.x);
-		cschcosh = cosh(helper.In.x);
+		cschcosh = std::cosh(helper.In.x);
-		cschden = 2 / (cosh(2 * helper.In.x) - cos(2 * helper.In.y));
+		cschden = 2 / (std::cosh(2 * helper.In.x) - std::cos(2 * helper.In.y));
 		helper.Out.x = m_Weight * cschden * cschsinh * cschcos;
 		helper.Out.y = -(m_Weight * cschden * cschcosh * cschsin);
 		helper.Out.z = m_Weight * helper.In.z;
@@ -6077,8 +6152,8 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		T s = sin(m_Freq * helper.In.x);
+		T s = std::sin(m_Freq * helper.In.x);
-		T t = sin(m_Freq * helper.In.y);
+		T t = std::sin(m_Freq * helper.In.y);
 		T dy = helper.In.y + m_AugerWeight * (m_Scale * s / 2 + fabs(helper.In.y) * s);
 		T dx = helper.In.x + m_AugerWeight * (m_Scale * t / 2 + fabs(helper.In.x) * t);
 
@@ -6166,10 +6241,9 @@ public:
 		T avgr = m_Weight * (m_Spr * std::sqrt(std::sqrt(yy + SQR(xpw)) / frac));
 		T avga = (atan2(helper.In.y, xmw) - atan2(helper.In.y, xpw)) * T(0.5);
 
-		helper.Out.x = avgr * cos(avga);
+		helper.Out.x = avgr * std::cos(avga);
-		helper.Out.y = avgr * sin(avga);
+		helper.Out.y = avgr * std::sin(avga);
 		helper.Out.z = helper.In.z;
-		//helper.Out.z = m_Weight * helper.In.z;
 	}
 
 	virtual string OpenCLString() const override

Source/Ember/Variations02.h

@@ -58,12 +58,12 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T theta = helper.m_PrecalcAtanyx;
-		T t = (rand.Frand01<T>() * m_Thickness) * (1 / cos(m_N * theta)) - m_Holes;
+		T t = (rand.Frand01<T>() * m_Thickness) * (1 / std::cos(m_N * theta)) - m_Holes;
 
 		if (fabs(t) != 0)
 		{
-			helper.Out.x = m_Weight * t * cos(theta);
+			helper.Out.x = m_Weight * t * std::cos(theta);
-			helper.Out.y = m_Weight * t * sin(theta);
+			helper.Out.y = m_Weight * t * std::sin(theta);
 			helper.Out.z = 0;
 		}
 		else
@@ -170,8 +170,8 @@ public:
 				r = (SQR(lx) + SQR(ly)) / m_R2;
 
 				T theta = m_BwrapsInnerTwist * (1 - r) + m_BwrapsOuterTwist * r;
-				T s = sin(theta);
+				T s = std::sin(theta);
-				T c = cos(theta);
+				T c = std::cos(theta);
 
 				vx = cx + c * lx + s * ly;
 				vy = cy - s * lx + c * ly;
@@ -336,8 +336,8 @@ public:
 
 		T r = m_Weight * side;
 		T val = T(M_PI_4) * perimeter / side - T(M_PI_4);
-		T sina = sin(val);
+		T sina = std::sin(val);
-		T cosa = cos(val);
+		T cosa = std::cos(val);
 
 		helper.Out.x = r * cosa;
 		helper.Out.y = r * sina;
@@ -716,10 +716,10 @@ public:
 		{
 			if (m_Bcbw != 0)
 			{
-				T ang = atan2(y, x);
+				T ang = std::atan2(y, x);
 				T omega = (T(0.2) * m_Bcbw * rand.Frand01<T>()) + 1;
-				T px = omega * cos(ang);
+				T px = omega * std::cos(ang);
-				T py = omega * sin(ang);
+				T py = omega * std::sin(ang);
 				helper.Out.x = m_Weight * px;
 				helper.Out.y = m_Weight * py;
 			}
@@ -1134,10 +1134,10 @@ public:
 		T angle = rand.Frand01<T>() * M_2PI;
 		T r = m_Weight * (rand.Frand01<T>() + rand.Frand01<T>() + rand.Frand01<T>() + rand.Frand01<T>() - 2);
 		T angle2 = rand.Frand01<T>() * T(M_PI);
-		T sina = sin(angle);
+		T sina = std::sin(angle);
-		T cosa = cos(angle);
+		T cosa = std::cos(angle);
-		T sinb = sin(angle2);
+		T sinb = std::sin(angle2);
-		T cosb = cos(angle2);
+		T cosb = std::cos(angle2);
 
 		helper.Out.x = r * sinb * cosa;
 		helper.Out.y = r * sinb * sina;
@@ -1202,6 +1202,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
 };
 
 /// <summary>
@@ -1254,6 +1259,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
+
 	virtual void Precalc() override
 	{
 		m_C2x = 2 * m_Cx;
@@ -1305,13 +1315,13 @@ public:
 	{
 		T r = helper.m_PrecalcSqrtSumSquares;
 		T temp = r * m_Pi;
-		T sr = sin(temp);
+		T sr = std::sin(temp);
-		T cr = cos(temp);
+		T cr = std::cos(temp);
 		T vv = m_Weight * helper.m_PrecalcAtanxy / Zeps(m_Pi);
 
 		helper.Out.x = vv * sr;
 		helper.Out.y = vv * cr;
-		helper.Out.z = vv * (r * cos(helper.In.z));
+		helper.Out.z = vv * (r * std::cos(helper.In.z));
 	}
 
 	virtual string OpenCLString() const override
@@ -1337,6 +1347,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
+
 protected:
 	void Init()
 	{
@@ -1519,7 +1534,7 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		T r = m_Weight * std::sqrt(helper.m_PrecalcSumSquares + sin(helper.m_PrecalcAtanyx * m_A) + 1);
+		T r = m_Weight * std::sqrt(helper.m_PrecalcSumSquares + std::sin(helper.m_PrecalcAtanyx * m_A) + 1);
 
 		helper.Out.x = r * helper.m_PrecalcCosa;
 		helper.Out.y = r * helper.m_PrecalcSina;
@@ -1580,7 +1595,7 @@ public:
 
 		int isXY = int(LRint(helper.In.x * m_Cs) + LRint(helper.In.y * m_Cs));
 
-		if (isXY % 2)
+		if (isXY & 1)
 		{
 			dx = m_Ncx + rnx;
 			dy = m_Ncy;
@@ -1619,7 +1634,7 @@ public:
 		   << "\n"
 		   << "\t\tint isXY = (int)(LRint(vIn.x * " << cs << ") + LRint(vIn.y * " << cs << "));\n"
 		   << "\n"
-		   << "\t\tif (isXY % 2)\n"
+		   << "\t\tif (isXY & 1)\n"
 		   << "\t\t{\n"
 		   << "\t\t	dx = " << ncx << " + rnx;\n"
 		   << "\t\t	dy = " << ncy << ";\n"
@@ -1638,6 +1653,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "LRint" };
+	}
+
 	virtual void Precalc() override
 	{
 		m_Cs = 1 / Zeps(m_Size);
@@ -1719,8 +1739,8 @@ public:
 
 		r = m_Vvar4Pi * side + m_Hole;
 		val = T(M_PI_4) * perimeter / side - T(M_PI_4);
-		helper.Out.x = r * cos(val);
+		helper.Out.x = r * std::cos(val);
-		helper.Out.y = r * sin(val);
+		helper.Out.y = r * std::sin(val);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -1832,8 +1852,8 @@ public:
 		T r = m_Weight * (side + m_Hole);
 		T val = T(M_PI_4) * perimeter / side - T(M_PI_4);
 
-		helper.Out.x = r * cos(val);
+		helper.Out.x = r * std::cos(val);
-		helper.Out.y = r * sin(val);
+		helper.Out.y = r * std::sin(val);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -1956,6 +1976,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Fabsmod", "Fosc", "Foscn", "Lerp" };
+	}
+
 	virtual void Precalc() override
 	{
 		m_Ax  = M_2PI * fabs(m_AmountX);
@@ -2061,10 +2086,10 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		T expor = exp(helper.In.x * m_K - helper.In.y * m_T);
+		T expor = std::exp(helper.In.x * m_K - helper.In.y * m_T);
 		T temp = helper.In.x * m_T + helper.In.y * m_K;
-		T snv = sin(temp);
+		T snv = std::sin(temp);
-		T csv = cos(temp);
+		T csv = std::cos(temp);
 
 		helper.Out.x = m_Weight * expor * csv;
 		helper.Out.y = m_Weight * expor * snv;
@@ -2098,7 +2123,7 @@ public:
 
 	virtual void Precalc() override
 	{
-		m_K = T(0.5) * log(Zeps(SQR(m_Real) + SQR(m_Imag)));//Original used 1e-300, which isn't representable with a float.
+		m_K = T(0.5) * std::log(Zeps(SQR(m_Real) + SQR(m_Imag)));//Original used 1e-300, which isn't representable with a float.
 		m_T = atan2(m_Imag, m_Real);
 	}
 
@@ -2272,8 +2297,8 @@ public:
 		temp = (helper.In.x * T(M_PI) + helper.In.y * m_NatLog) * -1;
 		sincos(temp, &snum2, &cnum2);
 
-		T eradius1 = exp(helper.In.x * m_NatLog);
+		T eradius1 = std::exp(helper.In.x * m_NatLog);
-		T eradius2 = exp((helper.In.x * m_NatLog - helper.In.y * T(M_PI)) * -1);
+		T eradius2 = std::exp((helper.In.x * m_NatLog - helper.In.y * T(M_PI)) * -1);
 
 		helper.Out.x = m_Weight * (eradius1 * cnum1 - eradius2 * cnum2) * m_Five;
 		helper.Out.y = m_Weight * (eradius1 * snum1 - eradius2 * snum2) * m_Five;
@@ -2310,7 +2335,7 @@ public:
 	virtual void Precalc() override
 	{
 		m_Five = 1 / SQRT5;
-		m_NatLog = log(M_PHI);
+		m_NatLog = std::log(M_PHI);
 	}
 
 protected:
@@ -2351,8 +2376,8 @@ public:
 		temp = (helper.In.x * T(M_PI) + helper.In.y * m_NatLog) * -1;
 		sincos(temp, &snum2, &cnum2);
 
-		T eradius1 = m_Sc * exp(m_Sc2 * (helper.In.x * m_NatLog));
+		T eradius1 = m_Sc * std::exp(m_Sc2 * (helper.In.x * m_NatLog));
-		T eradius2 = m_Sc * exp(m_Sc2 * ((helper.In.x * m_NatLog - helper.In.y * T(M_PI)) * -1));
+		T eradius2 = m_Sc * std::exp(m_Sc2 * ((helper.In.x * m_NatLog - helper.In.y * T(M_PI)) * -1));
 
 		helper.Out.x = m_Weight * (eradius1 * cnum1 - eradius2 * cnum2) * m_Five;
 		helper.Out.y = m_Weight * (eradius1 * snum1 - eradius2 * snum2) * m_Five;
@@ -2391,7 +2416,7 @@ public:
 	virtual void Precalc() override
 	{
 		m_Five = 1 / SQRT5;
-		m_NatLog = log(M_PHI);
+		m_NatLog = std::log(M_PHI);
 	}
 
 protected:
@@ -2686,6 +2711,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "LRint" };
+	}
 };
 
 /// <summary>
@@ -2704,7 +2734,7 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		T r, delta = pow(helper.m_PrecalcAtanyx / T(M_PI) + 1, m_A);
+		T r, delta = std::pow(helper.m_PrecalcAtanyx / T(M_PI) + 1, m_A);
 
 		if (m_Inside != 0)
 			r = m_Weight * delta / (helper.m_PrecalcSqrtSumSquares + delta);
@@ -2814,7 +2844,7 @@ public:
 	{
 		T pa = 2 * T(M_PI) / m_P;
 		T qa = 2 * T(M_PI) / m_Q;
-		T r = (1 - cos(pa)) / (cos(pa) + cos(qa)) + 1;
+		T r = (1 - std::cos(pa)) / (std::cos(pa) + std::cos(qa)) + 1;
 		T a = m_N * pa;
 
 		if (r > 0)
@@ -2822,8 +2852,8 @@ public:
 		else
 			r = 1;
 
-		m_Real = r * cos(a);
+		m_Real = r * std::cos(a);
-		m_Imag = r * sin(a);
+		m_Imag = r * std::sin(a);
 	}
 
 protected:
@@ -2864,8 +2894,8 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T temp = rand.Rand() * m_Pa;
-		T sina = sin(temp);
+		T sina = std::sin(temp);
-		T cosa = cos(temp);
+		T cosa = std::cos(temp);
 		T re = m_R * cosa;
 		T im = m_R * sina;
 		T a = helper.In.x + re;
@@ -2912,8 +2942,8 @@ public:
 
 	virtual void Precalc() override
 	{
-		T r2 = 1 - (cos(2 * T(M_PI) / m_P) - 1) /
+		T r2 = 1 - (std::cos(2 * T(M_PI) / m_P) - 1) /
-			(cos(2 * T(M_PI) / m_P) + cos(2 * T(M_PI) / m_Q));
+			(std::cos(2 * T(M_PI) / m_P) + std::cos(2 * T(M_PI) / m_Q));
 
 		if (r2 > 0)
 			m_R = 1 / std::sqrt(r2);
@@ -2966,8 +2996,8 @@ public:
 		T y = (b * c - a * d);
 		T vr = m_Weight / (SQR(c) + SQR(d));
 		T temp = rand.Rand() * m_Pa;
-		T sina = sin(temp);
+		T sina = std::sin(temp);
-		T cosa = cos(temp);
+		T cosa = std::cos(temp);
 
 		helper.Out.x = vr * (x * cosa + y * sina);
 		helper.Out.y = vr * (y * cosa - x * sina);
@@ -3007,8 +3037,8 @@ public:
 
 	virtual void Precalc() override
 	{
-		T r2 = 1 - (cos(2 * T(M_PI) / m_P) - 1) /
+		T r2 = 1 - (std::cos(2 * T(M_PI) / m_P) - 1) /
-			(cos(2 * T(M_PI) / m_P) + cos(2 * T(M_PI) / m_Q));
+			(std::cos(2 * T(M_PI) / m_P) + std::cos(2 * T(M_PI) / m_Q));
 
 		if (r2 > 0)
 			m_R = 1 / std::sqrt(r2);
@@ -3101,7 +3131,7 @@ public:
 	{
 		T pa = 2 * T(M_PI) / m_P;
 		T qa = 2 * T(M_PI) / m_Q;
-		T r  = -(cos(pa) - 1) / (cos(pa) + cos(qa));
+		T r = -(std::cos(pa) - 1) / (std::cos(pa) + std::cos(qa));
 		T na = m_N * pa;
 
 		if (r > 0)
@@ -3109,8 +3139,8 @@ public:
 		else
 			r = 1;
 
-		m_Cx = r * cos(na);
+		m_Cx = r * std::cos(na);
-		m_Cy = r * sin(na);
+		m_Cy = r * std::sin(na);
 		m_C2 = SQR(m_Cx) + SQR(m_Cy);
 		m_C2x = 2 * m_Cx;
 		m_C2y = 2 * m_Cy;
@@ -3173,8 +3203,8 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T temp = rand.Rand() * m_Pa;
-		T cx = m_R * cos(temp);
+		T cx = m_R * std::cos(temp);
-		T cy = m_R * sin(temp);
+		T cy = m_R * std::sin(temp);
 		T s2x = 1 + SQR(cx) - SQR(cy);
 		T s2y = 1 + SQR(cy) - SQR(cx);
 		T r2 = helper.m_PrecalcSumSquares + SQR(helper.In.z);
@@ -3223,7 +3253,7 @@ public:
 	{
 		T pa = M_2PI / m_P;
 		T qa = M_2PI / m_Q;
-		T r = -(cos(pa) - 1) / (cos(pa) + cos(qa));
+		T r = -(std::cos(pa) - 1) / (std::cos(pa) + std::cos(qa));
 
 		if (r > 0)
 			r = 1 / std::sqrt(1 + r);
@@ -3281,8 +3311,8 @@ public:
 		T y = helper.In.y * m_S2y;
 		T vr = m_Weight / (m_C2 * r2 + x2cx + 1);
 		T temp = rand.Rand() * m_Pa;
-		T sina = sin(temp);
+		T sina = std::sin(temp);
-		T cosa = cos(temp);
+		T cosa = std::cos(temp);
 
 		helper.Out.x = vr * (x * cosa + y * sina);
 		helper.Out.y = vr * (y * cosa - x * sina);
@@ -3327,7 +3357,7 @@ public:
 	{
 		T pa = M_2PI / m_P;
 		T qa = M_2PI / m_Q;
-		T r = -(cos(pa) - 1) / (cos(pa) + cos(qa));
+		T r = -(std::cos(pa) - 1) / (std::cos(pa) + std::cos(qa));
 
 		if (r > 0)
 			r = 1 / std::sqrt(1 + r);
@@ -3389,8 +3419,8 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T a = T(M_PI) / (helper.m_PrecalcSqrtSumSquares + 1);
-		T s = sin(a);
+		T s = std::sin(a);
-		T c = cos(a);
+		T c = std::cos(a);
 		T r = helper.m_PrecalcAtanyx * m_V;
 
 		helper.Out.x = r * c;
@@ -3457,9 +3487,9 @@ public:
 		T x = m_A * helper.In.x + m_B * helper.In.y + m_E;
 		T y = m_C * helper.In.x + m_D * helper.In.y + m_F;
 		T angle = (atan2(y, x) + M_2PI * rand.Rand(int(m_AbsN))) / m_Power;
-		T sina = sin(angle);
+		T sina = std::sin(angle);
-		T cosa = cos(angle);
+		T cosa = std::cos(angle);
-		T r = m_Weight * pow(SQR(x) + SQR(y), m_Cn);
+		T r = m_Weight * std::pow(SQR(x) + SQR(y), m_Cn);
 
 		helper.Out.x = r * cosa;
 		helper.Out.y = r * sina;
@@ -3556,9 +3586,9 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T a = helper.m_PrecalcAtanyx * m_InvPower + rand.Rand() * m_InvPower2pi;
-		T sina = sin(a);
+		T sina = std::sin(a);
-		T cosa = cos(a);
+		T cosa = std::cos(a);
-		T r = m_Weight * pow(helper.m_PrecalcSumSquares, m_HalfInvPower);
+		T r = m_Weight * std::pow(helper.m_PrecalcSumSquares, m_HalfInvPower);
 
 		helper.Out.x = r * cosa;
 		helper.Out.y = r * sina;
@@ -3636,9 +3666,9 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T angle = helper.m_PrecalcAtanyx * m_Power;
-		T sina = sin(angle);
+		T sina = std::sin(angle);
-		T cosa = cos(angle);
+		T cosa = std::cos(angle);
-		T r = m_Cp * pow(helper.m_PrecalcSumSquares, m_P2);
+		T r = m_Cp * std::pow(helper.m_PrecalcSumSquares, m_P2);
 		T re = r * cosa + 1;
 		T im = r * sina;
 		T r1 = m_Vp / (SQR(re) + SQR(im));
@@ -3726,16 +3756,16 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T angle = helper.m_PrecalcAtanyx * m_Power;
-		T sina = sin(angle);
+		T sina = std::sin(angle);
-		T cosa = cos(angle);
+		T cosa = std::cos(angle);
-		T r = m_C * pow(helper.m_PrecalcSumSquares, m_P2);
+		T r = m_C * std::pow(helper.m_PrecalcSumSquares, m_P2);
 		T re = r * cosa + 1;
 		T im = r * sina;
 
-		r = pow(SQR(re) + SQR(im), m_InvP);
+		r = std::pow(SQR(re) + SQR(im), m_InvP);
-		angle = atan2(im, re) * m_InvP2;
+		angle = std::atan2(im, re) * m_InvP2;
-		sina = sin(angle);
+		sina = std::sin(angle);
-		cosa = cos(angle);
+		cosa = std::cos(angle);
 		re = r * cosa;
 		im = r * sina;
 
@@ -3793,7 +3823,7 @@ public:
 		if (m_C == -1)
 			m_Vp = 0;
 		else
-			m_Vp = m_Weight * pow(m_C + 1, 2 / m_Power);
+			m_Vp = m_Weight * std::pow(m_C + 1, 2 / m_Power);
 	}
 
 protected:
@@ -3836,14 +3866,14 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T x = (m_IsOdd != 0) ? helper.In.x : m_Vvar * helper.m_PrecalcAtanxy;
-		T y = (m_IsOdd != 0) ? helper.In.y : m_Vvar2 * log(helper.m_PrecalcSumSquares);
+		T y = (m_IsOdd != 0) ? helper.In.y : m_Vvar2 * std::log(helper.m_PrecalcSumSquares);
-		T angle = (atan2(y, x) + M_2PI * rand.Rand(int(m_AbsN))) / m_Nnz;
+		T angle = (std::atan2(y, x) + M_2PI * rand.Rand(int(m_AbsN))) / m_Nnz;
-		T r = m_Weight * pow(SQR(x) + SQR(y), m_Cn) * ((m_IsOdd == 0) ? 1 : m_Parity);
+		T r = m_Weight * std::pow(SQR(x) + SQR(y), m_Cn) * ((m_IsOdd == 0) ? 1 : m_Parity);
-		T sina = sin(angle) * r;
+		T sina = std::sin(angle) * r;
-		T cosa = cos(angle) * r;
+		T cosa = std::cos(angle) * r;
 
-		x = (m_IsOdd != 0) ? cosa : (m_Vvar2 * log(SQR(cosa) + SQR(sina)));
+		x = (m_IsOdd != 0) ? cosa : (m_Vvar2 * std::log(SQR(cosa) + SQR(sina)));
-		y = (m_IsOdd != 0) ? sina : (m_Vvar * atan2(cosa, sina));
+		y = (m_IsOdd != 0) ? sina : (m_Vvar * std::atan2(cosa, sina));
 		helper.Out.x = x;
 		helper.Out.y = y;
 		helper.Out.z = m_Weight * helper.In.z;
@@ -3889,7 +3919,7 @@ public:
 		m_Vvar2 = m_Vvar * T(0.5);
 		m_AbsN = abs(m_Nnz);
 		m_Cn = 1 / m_Nnz / 2;
-		m_IsOdd = T(abs(int(m_Parity)) % 2);
+		m_IsOdd = T(abs(int(m_Parity)) & 1);
 	}
 
 protected:
@@ -3972,8 +4002,8 @@ public:
 		else
 		{
 			r = 1 / std::sqrt(r);
-			ts = sin(helper.m_PrecalcAtanyx);
+			ts = std::sin(helper.m_PrecalcAtanyx);
-			tc = cos(helper.m_PrecalcAtanyx);
+			tc = std::cos(helper.m_PrecalcAtanyx);
 			x = r * tc;
 			y = r * ts;
 
@@ -3981,13 +4011,13 @@ public:
 			{
 				xo = (SQR(x) + SQR(y) + 1) / (2 * x);
 				ro = std::sqrt(SQR(x - xo) + SQR(y));
-				theta = atan2(T(1), ro);
+				theta = std::atan2(T(1), ro);
 				a = fmod(m_Out * theta + atan2(y, xo - x) + theta, 2 * theta) - theta;
 				sincos(a, &s, &c);
 
 				x = (xo - c * ro);
 				y =  s * ro;
-				theta = atan2(y, x);
+				theta = std::atan2(y, x);
 				sincos(theta, &ts, &tc);
 				r = 1 / std::sqrt(SQR(x) + SQR(y));
 
@@ -3998,13 +4028,13 @@ public:
 			{
 				xo = - (SQR(x) + SQR(y) + 1) / (2 * x);
 				ro = std::sqrt(SQR(-x - xo) + SQR(y));
-				theta = atan2(T(1), ro);
+				theta = std::atan2(T(1), ro);
-				a = fmod(m_Out * theta + atan2(y, xo + x) + theta, 2 * theta) - theta;
+				a = fmod(m_Out * theta + std::atan2(y, xo + x) + theta, 2 * theta) - theta;
 				sincos(a, &s, &c);
 
 				x = (xo - c * ro);
 				y =  s * ro;
-				theta = atan2(y, x);
+				theta = std::atan2(y, x);
 				sincos(theta, &ts, &tc);
 				r = 1 / std::sqrt(SQR(x) + SQR(y));
 
@@ -4189,10 +4219,10 @@ public:
 		m_C1d = std::sqrt(1 + SQR(m_C1r));
 		m_C2d = std::sqrt(1 + SQR(m_C2r));
 
-		m_C1x = m_C1d * cos(fmod(m_C1a, T(M_PI)));
+		m_C1x = m_C1d * std::cos(fmod(m_C1a, T(M_PI)));
-		m_C1y = m_C1d * sin(fmod(m_C1a, T(M_PI)));
+		m_C1y = m_C1d * std::sin(fmod(m_C1a, T(M_PI)));
-		m_C2x = m_C2d * cos(fmod(m_C2a, T(M_PI)));
+		m_C2x = m_C2d * std::cos(fmod(m_C2a, T(M_PI)));
-		m_C2y = m_C2d * sin(fmod(m_C2a, T(M_PI)));
+		m_C2y = m_C2d * std::sin(fmod(m_C2a, T(M_PI)));
 	}
 
 protected:
@@ -4290,11 +4320,16 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
+
 	virtual void Precalc() override
 	{
-		m_Cx = -m_R * cos(m_A * T(M_PI_2)) * cos(m_B * T(M_PI_2));
+		m_Cx = -m_R * std::cos(m_A * T(M_PI_2)) * std::cos(m_B * T(M_PI_2));
-		m_Cy =  m_R * sin(m_A * T(M_PI_2)) * cos(m_B * T(M_PI_2));
+		m_Cy =  m_R * std::sin(m_A * T(M_PI_2)) * std::cos(m_B * T(M_PI_2));
-		m_Cz = -m_R * sin(m_B * T(M_PI_2));
+		m_Cz = -m_R * std::sin(m_B * T(M_PI_2));
 
 		m_C2 = SQR(m_Cx) + SQR(m_Cy) + SQR(m_Cz);
 
@@ -4360,8 +4395,8 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		T xp = pow(fabs(m_Weight) * fabs(helper.In.x), m_Powx);//Original did not fabs.
+		T xp = std::pow(fabs(m_Weight) * fabs(helper.In.x), m_Powx);//Original did not fabs.
-		T yp = pow(fabs(m_Weight) * fabs(helper.In.y), m_Powy);
+		T yp = std::pow(fabs(m_Weight) * fabs(helper.In.y), m_Powy);
 
 		helper.Out.x = xp * Sign(helper.In.x) + m_Lcx * helper.In.x + m_Scx;
 		helper.Out.y = yp * Sign(helper.In.y) + m_Lcy * helper.In.y + m_Scy;
@@ -4408,6 +4443,11 @@ protected:
 		m_Params.push_back(ParamWithName<T>(&m_Scy,  prefix + "polynomial_scy"));
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Sign" };
+	}
+
 private:
 	T m_Powx;
 	T m_Powy;
@@ -4618,7 +4658,7 @@ public:
 
 		//Calculate cosine wave with given frequency, velocity
 		//and phase based on the distance to center.
-		T wave = cos(m_F * d - m_Vxp);
+		T wave = std::cos(m_F * d - m_Vxp);
 
 		//Calculate the wave offsets
 		T d1 = wave * m_Pxa + d;
@@ -4686,6 +4726,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Lerp" };
+	}
+
 	virtual void Precalc() override
 	{
 		m_F = m_Frequency * 5;
@@ -4757,8 +4802,8 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		T c0 = m_Ax / (1 + exp(m_Sx * helper.In.x));
+		T c0 = m_Ax / (1 + std::exp(m_Sx * helper.In.x));
-		T c1 = m_Ay / (1 + exp(m_Sy * helper.In.y));
+		T c1 = m_Ay / (1 + std::exp(m_Sy * helper.In.y));
 		T x = (2 * (c0 - T(0.5)));
 		T y = (2 * (c1 - T(0.5)));
 
@@ -4879,8 +4924,8 @@ public:
 	{
 		T x = helper.In.x;
 		T y = helper.In.y;
-		T sx = -1 * cos(x * m_Fx);
+		T sx = -1 * std::cos(x * m_Fx);
-		T sy = -1 * cos(y * m_Fy);
+		T sy = -1 * std::cos(y * m_Fy);
 		T tx = Lerp(helper.In.x, sx, m_Ax);
 		T ty = Lerp(helper.In.y, sy, m_Ay);
 
@@ -4921,6 +4966,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Lerp" };
+	}
+
 	virtual void Precalc() override
 	{
 		m_Ax = m_AmpX;
@@ -4980,7 +5030,7 @@ public:
 		T xPlusy = x + y;
 		T x2Minusy2 = x2 - y2;
 		T x2Plusy2 = x2 + y2;
-		T result = x2Minusy2 * sin(M_2PI * m_Distort * xPlusy);
+		T result = x2Minusy2 * std::sin(M_2PI * m_Distort * xPlusy);
 		T divident = 1;
 
 		if (x2Plusy2 != 0)
@@ -5097,9 +5147,9 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		T r = exp(helper.In.y);
+		T r = std::exp(helper.In.y);
-		T s = sin(helper.In.x);
+		T s = std::sin(helper.In.x);
-		T c = cos(helper.In.x);
+		T c = std::cos(helper.In.x);
 
 		helper.Out.x = m_Vvar2 * r * s;
 		helper.Out.y = m_Vvar2 * r * c;
@@ -5162,13 +5212,13 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T angle = (helper.m_PrecalcAtanyx + M_2PI * rand.Rand(int(m_AbsN))) / m_Power;
-		T r = m_Weight * pow(helper.m_PrecalcSumSquares, m_Cn);
+		T r = m_Weight * std::pow(helper.m_PrecalcSumSquares, m_Cn);
-		T sina = sin(angle);
+		T sina = std::sin(angle);
-		T cosa = cos(angle);
+		T cosa = std::cos(angle);
 		T xn = r * cosa;
 		T yn = r * sina;
-		T siny = sin(m_FreqX * yn);
+		T siny = std::sin(m_FreqX * yn);
-		T sinx = sin(m_FreqY * xn);
+		T sinx = std::sin(m_FreqY * xn);
 		T dx = xn + T(0.5) * (m_ScaleX * siny + fabs(xn) * m_IncX * siny);
 		T dy = yn + T(0.5) * (m_ScaleY * sinx + fabs(yn) * m_IncY * sinx);
 
@@ -5432,6 +5482,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Sign" };
+	}
+
 protected:
 	void Init()
 	{
@@ -5467,8 +5522,8 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		helper.Out.x = m_Weight01 / SafeTan<T>(helper.In.x) * cos(helper.In.y);
+		helper.Out.x = m_Weight01 / SafeTan<T>(helper.In.x) * std::cos(helper.In.y);
-		helper.Out.y = m_Weight01 / sin(helper.In.x) * (-helper.In.y);
+		helper.Out.y = m_Weight01 / std::sin(helper.In.x) * (-helper.In.y);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -5520,8 +5575,13 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		helper.Out.x = m_Weight * sin(helper.In.x) * (cosh(helper.In.y) + 1) * Sqr<T>(sin(helper.In.x));
+		T sinx = std::sin(helper.In.x);
-		helper.Out.y = m_Weight * cos(helper.In.x) * (cosh(helper.In.y) + 1) * Sqr<T>(sin(helper.In.x));
+		T sinx2 = SQR(sinx);
+		T cosx = std::cos(helper.In.x);
+		T coshy1 = std::cosh(helper.In.y) + 1;
+
+		helper.Out.x = m_Weight * sinx * coshy1 * sinx2;
+		helper.Out.y = m_Weight * cosx * coshy1 * sinx2;
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -5531,13 +5591,23 @@ public:
 		intmax_t varIndex = IndexInXform();
 
 		ss << "\t{\n"
-		   << "\t\tvOut.x = xform->m_VariationWeights[" << varIndex << "] * sin(vIn.x) * (cosh(vIn.y) + (real_t)(1.0)) * Sqr(sin(vIn.x));\n"
+		   << "\t\treal_t sinx = sin(vIn.x);\n"
-		   << "\t\tvOut.y = xform->m_VariationWeights[" << varIndex << "] * cos(vIn.x) * (cosh(vIn.y) + (real_t)(1.0)) * Sqr(sin(vIn.x));\n"
+		   << "\t\treal_t sinx2 = SQR(sinx);\n"
+		   << "\t\treal_t cosx = cos(vIn.x);\n"
+		   << "\t\treal_t coshy1 = cosh(vIn.y) + 1.0;\n"
+		   << "\n"
+		   << "\t\tvOut.x = xform->m_VariationWeights[" << varIndex << "] * sinx * coshy1 * sinx2;\n"
+		   << "\t\tvOut.y = xform->m_VariationWeights[" << varIndex << "] * cosx * coshy1 * sinx2;\n"
 		   << "\t\tvOut.z = xform->m_VariationWeights[" << varIndex << "] * vIn.z;\n"
 		   << "\t}\n";
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Sqr" };
+	}
 };
 
 /// <summary>
@@ -5554,9 +5624,13 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T d = m_Weight / helper.m_PrecalcSumSquares;
+		T sinx = std::sin(helper.In.x);
+		T sinx2 = SQR(sinx);
+		T cosx = std::cos(helper.In.x);
+		T coshy1 = std::cosh(helper.In.y) + 1;
 
-		helper.Out.x = d * sin(helper.In.x) * (cosh(helper.In.y) + 1) * Sqr(sin(helper.In.x));
+		helper.Out.x = d * sinx * coshy1 * sinx2;
-		helper.Out.y = d * cos(helper.In.x) * (cosh(helper.In.y) + 1) * Sqr(sin(helper.In.x));
+		helper.Out.y = d * cosx * coshy1 * sinx2;
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -5567,14 +5641,23 @@ public:
 
 		ss << "\t{\n"
 		   << "\t\treal_t d = xform->m_VariationWeights[" << varIndex << "] / precalcSumSquares;\n"
+		   << "\t\treal_t sinx = sin(vIn.x);\n"
+		   << "\t\treal_t sinx2 = SQR(sinx);\n"
+		   << "\t\treal_t cosx = cos(vIn.x);\n"
+		   << "\t\treal_t coshy1 = cosh(vIn.y) + 1.0;\n"
 		   << "\n"
-		   << "\t\tvOut.x = d * sin(vIn.x) * (cosh(vIn.y) + (real_t)(1.0)) * Sqr(sin(vIn.x));\n"
+		   << "\t\tvOut.x = d * sinx * coshy1 * sinx2;\n"
-		   << "\t\tvOut.y = d * cos(vIn.x) * (cosh(vIn.y) + (real_t)(1.0)) * Sqr(sin(vIn.x));\n"
+		   << "\t\tvOut.y = d * cosx * coshy1 * sinx2;\n"
 		   << "\t\tvOut.z = xform->m_VariationWeights[" << varIndex << "] * vIn.z;\n"
 		   << "\t}\n";
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Sqr" };
+	}
 };
 
 /// <summary>
@@ -5601,15 +5684,15 @@ public:
 
 		a += M_2PI * n;
 
-		if (cos(a * m_InvSpread) < rand.Rand() * 2 / 0xFFFFFFFF - 1)//Rand max.
+		if (std::cos(a * m_InvSpread) < rand.Rand() * 2 / 0xFFFFFFFF - 1)//Rand max.
 			a -= m_FullSpread;
 
-		T lnr2 = log(helper.m_PrecalcSumSquares);
+		T lnr2 = std::log(helper.m_PrecalcSumSquares);
-		T r = m_Weight * exp(m_HalfC * lnr2 - m_D * a);
+		T r = m_Weight * std::exp(m_HalfC * lnr2 - m_D * a);
 		T temp = m_C * a + m_HalfD * lnr2 + m_Ang * rand.Rand();
 
-		helper.Out.x = r * cos(temp);
+		helper.Out.x = r * std::cos(temp);
-		helper.Out.y = r * sin(temp);
+		helper.Out.y = r * std::sin(temp);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -5658,8 +5741,8 @@ public:
 	virtual void Precalc() override
 	{
 		m_Ang = M_2PI / m_Divisor;
-		m_C = m_R * cos(T(M_PI) / 2 * m_A) / m_Divisor;
+		m_C = m_R * std::cos(T(M_PI) / 2 * m_A) / m_Divisor;
-		m_D = m_R * sin(T(M_PI) / 2 * m_A) / m_Divisor;
+		m_D = m_R * std::sin(T(M_PI) / 2 * m_A) / m_Divisor;
 		m_HalfC = m_C / 2;
 		m_HalfD = m_D / 2;
 		m_InvSpread = T(0.5) / m_Spread;

Source/Ember/Variations04.h

@@ -26,16 +26,16 @@ public:
 
 		ClampGteRef<T>(xmax, -1);
 
-		T mu = acosh(xmax);
+		T mu = std::acosh(xmax);
-		T nu = acos(Clamp<T>(helper.In.x / xmax, -1, 1));//-Pi < nu < Pi.
+		T nu = std::acos(Clamp<T>(helper.In.x / xmax, -1, 1));//-Pi < nu < Pi.
 
 		if (helper.In.y < 0)
 			nu *= -1;
 
 		nu = nu + mu * m_Out + m_In / mu;
 
-		helper.Out.x = m_Weight * cosh(mu) * cos(nu);
+		helper.Out.x = m_Weight * std::cosh(mu) * std::cos(nu);
-		helper.Out.y = m_Weight * sinh(mu) * sin(nu);
+		helper.Out.y = m_Weight * std::sinh(mu) * std::sin(nu);
 		helper.Out.z = m_Weight * helper.In.z;
 	}
 
@@ -57,7 +57,7 @@ public:
 		   << "\t\t	xmax = 1;\n"
 		   << "\n"
 		   << "\t\treal_t mu = acosh(xmax);\n"
-		   << "\t\treal_t nu = acos(Clamp(vIn.x / xmax, -(real_t)(1.0), (real_t)(1.0)));\n"
+		   << "\t\treal_t nu = acos(clamp(vIn.x / xmax, -(real_t)(1.0), (real_t)(1.0)));\n"
 		   << "\n"
 		   << "\t\tif (vIn.y < 0)\n"
 		   << "\t\t	nu *= -1;\n"
@@ -72,6 +72,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "SafeSqrt" };
+	}
+
 protected:
 	void Init()
 	{
@@ -623,6 +628,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
 };
 
 /// <summary>
@@ -660,6 +670,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
 };
 
 /// <summary>
@@ -683,8 +698,8 @@ public:
 
 		if (m_VarType == VARTYPE_REG)
 		{
-			helper.Out.x = 0;
+			helper.Out.x = helper.In.x;
-			outPoint.m_X = helper.In.x;
+			outPoint.m_X = 0;
 		}
 		else
 		{
@@ -729,8 +744,8 @@ public:
 
 	virtual void Precalc() override
 	{
-		m_RxSin = sin(m_Weight * T(M_PI_2));
+		m_RxSin = std::sin(m_Weight * T(M_PI_2));
-		m_RxCos = cos(m_Weight * T(M_PI_2));
+		m_RxCos = std::cos(m_Weight * T(M_PI_2));
 	}
 
 protected:
@@ -812,8 +827,8 @@ public:
 
 	virtual void Precalc() override
 	{
-		m_RySin = sin(m_Weight * T(M_PI_2));
+		m_RySin = std::sin(m_Weight * T(M_PI_2));
-		m_RyCos = cos(m_Weight * T(M_PI_2));
+		m_RyCos = std::cos(m_Weight * T(M_PI_2));
 	}
 
 protected:
@@ -853,8 +868,8 @@ public:
 
 		if (m_VarType == VARTYPE_REG)
 		{
-			helper.Out.z = 0;
+			helper.Out.z = helper.In.z;
-			outPoint.m_Z = helper.In.z;
+			outPoint.m_Z = 0;
 		}
 		else
 		{
@@ -894,8 +909,8 @@ public:
 
 	virtual void Precalc() override
 	{
-		m_RzSin = sin(m_Weight * T(M_PI_2));
+		m_RzSin = std::sin(m_Weight * T(M_PI_2));
-		m_RzCos = cos(m_Weight * T(M_PI_2));
+		m_RzCos = std::cos(m_Weight * T(M_PI_2));
 	}
 
 protected:
@@ -1181,7 +1196,7 @@ public:
 		   << "\n"
 		   << "\t\tvOut.x = xform->m_VariationWeights[" << varIndex << "] * (vIn.x + (MwcNext01(mwc) - (real_t)(0.5)) * r);\n"
 		   << "\t\tvOut.y = xform->m_VariationWeights[" << varIndex << "] * (vIn.y + (MwcNext01(mwc) - (real_t)(0.5)) * r);\n"
-		   << "\t\tvOut.z = xform->m_VariationWeights[" << varIndex << "] * vIn.z;\n"
+		   << "\t\tvOut.z = xform->m_VariationWeights[" << varIndex << "] * vIn.z;\n"//WHY IS THIS DIFFERENT THAN THE CPU ONE?//TODO
 		   << "\t}\n";
 
 		return ss.str();
@@ -1231,11 +1246,11 @@ public:
 	{
 		T jun = Zeps(fabs(m_N));
 
-		T a = (atan2(helper.In.y, pow(fabs(helper.In.x), m_Sep)) + M_2PI * Floor<T>(rand.Frand01<T>() * m_AbsN)) / jun;
+		T a = (std::atan2(helper.In.y, std::pow(fabs(helper.In.x), m_Sep)) + M_2PI * Floor<T>(rand.Frand01<T>() * m_AbsN)) / jun;
-		T r = m_Weight * pow(helper.m_PrecalcSumSquares, m_Cn * m_A);
+		T r = m_Weight * std::pow(helper.m_PrecalcSumSquares, m_Cn * m_A);
 
-		helper.Out.x = r * cos(a) + m_B;
+		helper.Out.x = r * std::cos(a) + m_B;
-		helper.Out.y = r * sin(a) + m_B;
+		helper.Out.y = r * std::sin(a) + m_B;
 		helper.Out.z = m_Weight * helper.In.z;//Original did not multiply by weight. Do it here to be consistent with others.
 	}
 
@@ -1266,6 +1281,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
+
 	virtual void Precalc() override
 	{
 		T jun = Zeps(fabs(m_N));
@@ -1317,8 +1337,8 @@ public:
 		T sqY = SQR(helper.In.y);
 		T v = (sqX + sqY) * m_W;//Do not use precalcSumSquares here because its components are needed below.
 
-		helper.Out.x = m_Weight * sin(helper.In.x) * (sqX + m_W - v);
+		helper.Out.x = m_Weight * std::sin(helper.In.x) * (sqX + m_W - v);
-		helper.Out.y = m_Weight * sin(helper.In.y) * (sqY + m_W - v);
+		helper.Out.y = m_Weight * std::sin(helper.In.y) * (sqY + m_W - v);
 		helper.Out.z = (m_VarType == VARTYPE_REG) ? 0 : helper.In.z;
 	}
 
@@ -1467,7 +1487,7 @@ public:
 		return ss.str();
 	}
 
-	virtual string OpenCLFuncsString() const
+	virtual string OpenCLFuncsString() const override
 	{
 		return
 			"real_t VoronDiscreteNoise(int x)\n"
@@ -1611,8 +1631,8 @@ public:
 
 	virtual void Precalc() override
 	{
-		m_SinR = sin(m_Rotation);
+		m_SinR = std::sin(m_Rotation);
-		m_CosR = cos(m_Rotation);
+		m_CosR = std::cos(m_Rotation);
 	}
 
 protected:
@@ -1695,24 +1715,24 @@ public:
 		if (rand.RandBit())
 			phi1 = -phi1;
 
-		sinr = sin(rho1);
+		sinr = std::sin(rho1);
-		cosr = cos(rho1);
+		cosr = std::cos(rho1);
 
-		sinp = sin(phi1);
+		sinp = std::sin(phi1);
-		cosp = cos(phi1);
+		cosp = std::cos(phi1);
 
 		temp = m_M4_1 * rho1;
-		msinr = sin(temp);
+		msinr = std::sin(temp);
-		mcosr = cos(temp);
+		mcosr = std::cos(temp);
 
 		temp = m_M4_2 * phi1;
-		msinp = sin(temp);
+		msinp = std::sin(temp);
-		mcosp = cos(temp);
+		mcosp = std::cos(temp);
 
-		pr1 = m_An2_1 * pow(fabs(mcosr), m_N2_1) + m_Bn3_1 * pow(fabs(msinr), m_N3_1);
+		pr1 = m_An2_1 * std::pow(fabs(mcosr), m_N2_1) + m_Bn3_1 * std::pow(fabs(msinr), m_N3_1);
-		pr2 = m_An2_2 * pow(fabs(mcosp), m_N2_2) + m_Bn3_2 * pow(fabs(msinp), m_N3_2);
+		pr2 = m_An2_2 * std::pow(fabs(mcosp), m_N2_2) + m_Bn3_2 * std::pow(fabs(msinp), m_N3_2);
-		r1 = pow(fabs(pr1), m_N1_1) + m_Spiral * rho1;
+		r1 = std::pow(fabs(pr1), m_N1_1) + m_Spiral * rho1;
-		r2 = pow(fabs(pr2), m_N1_2);
+		r2 = std::pow(fabs(pr2), m_N1_2);
 
 		if (int(m_Toroidmap) == 1)
 		{
@@ -1810,10 +1830,10 @@ public:
 	{
 		m_N1n_1 = (-1 / m_N1_1);
 		m_N1n_2 = (-1 / m_N1_2);
-		m_An2_1 = pow(fabs(1 / m_A1), m_N2_1);
+		m_An2_1 = std::pow(fabs(1 / m_A1), m_N2_1);
-		m_An2_2 = pow(fabs(1 / m_A2), m_N2_2);
+		m_An2_2 = std::pow(fabs(1 / m_A2), m_N2_2);
-		m_Bn3_1 = pow(fabs(1 / m_B1), m_N3_1);
+		m_Bn3_1 = std::pow(fabs(1 / m_B1), m_N3_1);
-		m_Bn3_2 = pow(fabs(1 / m_B2), m_N3_2);
+		m_Bn3_2 = std::pow(fabs(1 / m_B2), m_N3_2);
 		m_M4_1 = m_M1 / 4;
 		m_M4_2 = m_M2 / 4;
 		m_Rho2Pi = m_Rho * T(M_2_PI);
@@ -1902,8 +1922,8 @@ public:
 		T t, rX, rY, rZ;
 
 		t  = Zeps(helper.m_PrecalcSumSquares + SQR(helper.In.z));
-		rX = m_Weight / pow(t, m_StretchX);
+		rX = m_Weight / std::pow(t, m_StretchX);
-		rY = m_Weight / pow(t, m_StretchY);
+		rY = m_Weight / std::pow(t, m_StretchY);
 
 		helper.Out.x = helper.In.x * rX;
 		helper.Out.y = helper.In.y * rY;
@@ -1911,7 +1931,7 @@ public:
 		//Optional 3D calculation.
 		if (int(m_ZOn) == 1)
 		{
-			rZ = m_Weight / pow(t, m_StretchZ);
+			rZ = m_Weight / std::pow(t, m_StretchZ);
 			helper.Out.z = helper.In.z * rZ;
 		}
 	}
@@ -1948,6 +1968,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
+
 protected:
 	void Init()
 	{
@@ -1987,10 +2012,10 @@ public:
 		T yi = helper.In.y - m_Y;
 
 		const T rad = std::sqrt(SQR(xi) + SQR(yi));
-		const T ang = atan2(yi, xi);
+		const T ang = std::atan2(yi, xi);
 		const T rdc = m_Radius + (rand.Frand01<T>() * T(0.5) * m_Ca);
-		const T s = sin(ang);
+		const T s = std::sin(ang);
-		const T c = cos(ang);
+		const T c = std::cos(ang);
 
 		const int esc = rad > m_Radius;
 		const int cr0 = int(m_Zero);
@@ -2131,15 +2156,15 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		const T z = helper.In.z / m_AbsN;
-		const T radiusOut = m_Weight * pow(helper.m_PrecalcSumSquares + z * z, m_Cn);
+		const T radiusOut = m_Weight * std::pow(helper.m_PrecalcSumSquares + z * z, m_Cn);
 		const T x = m_A * helper.In.x + m_B * helper.In.y + m_E;
 		const T y = m_C * helper.In.x + m_D * helper.In.y + m_F;
 		const T tempRand = T(int(rand.Frand01<T>() * m_AbsN));
-		const T alpha = (atan2(y, x) + M_2PI * tempRand) / m_Power;
+		const T alpha = (std::atan2(y, x) + M_2PI * tempRand) / m_Power;
 		const T gamma = radiusOut * helper.m_PrecalcSqrtSumSquares;
 
-		helper.Out.x = gamma * cos(alpha);
+		helper.Out.x = gamma * std::cos(alpha);
-		helper.Out.y = gamma * sin(alpha);
+		helper.Out.y = gamma * std::sin(alpha);
 		helper.Out.z = radiusOut * z;
 	}
 
@@ -2234,8 +2259,8 @@ public:
 		{
 			T r = 1 / helper.m_PrecalcSqrtSumSquares;
 
-			helper.Out.x = m_Weight * r * cos(helper.m_PrecalcAtanyx);
+			helper.Out.x = m_Weight * r * std::cos(helper.m_PrecalcAtanyx);
-			helper.Out.y = m_Weight * r * sin(helper.m_PrecalcAtanyx);
+			helper.Out.y = m_Weight * r * std::sin(helper.m_PrecalcAtanyx);
 		}
 		else if (helper.In.x > 0 && helper.In.y < 0)//Quadrant I: loonie.
 		{
@@ -2262,11 +2287,11 @@ public:
 
 			if (r < m_Weight)
 			{
-				T a = atan2(y, x) + m_Spin + m_Twist * (m_Weight - r);
+				T a = std::atan2(y, x) + m_Spin + m_Twist * (m_Weight - r);
 
 				r *= m_Weight;
-				helper.Out.x = r * cos(a) + m_X;
+				helper.Out.x = r * std::cos(a) + m_X;
-				helper.Out.y = r * sin(a) - m_Y;
+				helper.Out.y = r * std::sin(a) - m_Y;
 			}
 			else
 			{
@@ -2355,6 +2380,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
+
 	virtual void Precalc() override
 	{
 		m_SqrWeight = SQR(m_Weight);
@@ -2556,12 +2586,12 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		const T zr = Hypot<T>(helper.In.z, helper.m_PrecalcSqrtSumSquares);
-		const T phi = acos(Clamp<T>(helper.In.z / zr, -1, 1));
+		const T phi = std::acos(Clamp<T>(helper.In.z / zr, -1, 1));
-		const T ps = sin(phi);
+		const T ps = std::sin(phi);
-		const T pc = cos(phi);
+		const T pc = std::cos(phi);
 
-		helper.Out.x = m_Weight * cos(helper.m_PrecalcAtanyx) * ps * (zr + m_Radius);
+		helper.Out.x = m_Weight * std::cos(helper.m_PrecalcAtanyx) * ps * (zr + m_Radius);
-		helper.Out.y = m_Weight * sin(helper.m_PrecalcAtanyx) * ps * (zr + m_Radius);
+		helper.Out.y = m_Weight * std::sin(helper.m_PrecalcAtanyx) * ps * (zr + m_Radius);
 		helper.Out.z = m_Weight * pc * (zr + m_Radius);
 	}
 
@@ -2575,7 +2605,7 @@ public:
 
 		ss << "\t{\n"
 		   << "\t\tconst real_t zr = Hypot(vIn.z, precalcSqrtSumSquares);\n"
-		   << "\t\tconst real_t phi = acos(Clamp(vIn.z / zr, -(real_t)(1.0), (real_t)(1.0)));\n"
+		   << "\t\tconst real_t phi = acos(clamp(vIn.z / zr, -(real_t)(1.0), (real_t)(1.0)));\n"
 		   << "\t\tconst real_t ps = sin(phi);\n"
 		   << "\t\tconst real_t pc = cos(phi);\n"
 		   << "\n"
@@ -2587,6 +2617,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Hypot" };
+	}
+
 protected:
 	void Init()
 	{
@@ -2621,11 +2656,11 @@ public:
 			Sqr(helper.In.z - m_ZOrigin)) *
 			(rand.Frand01<T>() + rand.Frand01<T>() + rand.Frand01<T>() + rand.Frand01<T>() - 2);
 		T u = rand.Frand01<T>() * M_2PI;
-		T su = sin(u);
+		T su = std::sin(u);
-		T cu = cos(u);
+		T cu = std::cos(u);
 		T v = rand.Frand01<T>() * M_2PI;
-		T sv = sin(v);
+		T sv = std::sin(v);
-		T cv = cos(v);
+		T cv = std::cos(v);
 
 		helper.Out.x = m_X * r * sv * cu;
 		helper.Out.y = m_Y * r * sv * su;
@@ -2665,6 +2700,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Sqr" };
+	}
+
 protected:
 	void Init()
 	{
@@ -2749,12 +2789,17 @@ public:
 		   << "\t\tvOut.x = (x * re + y * im) * r;\n"
 		   << "\t\tvOut.y = (y * re - x * im) * r;\n"
 		   << "\t\tvOut.z = (z * xform->m_VariationWeights[" << varIndex << "]) / c;\n"
-		   << "\t\toutPoint->m_ColorX = Clamp(outPoint->m_ColorX + " << dcAdjust << " * c, (real_t)(0.0), (real_t)(1.0));\n"
+		   << "\t\toutPoint->m_ColorX = clamp(outPoint->m_ColorX + " << dcAdjust << " * c, (real_t)(0.0), (real_t)(1.0));\n"
 		   << "\t}\n";
 
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Hypot", "Spread", "SignNz", "Powq4", "Powq4c", "Zeps" };
+	}
+
 	virtual void Precalc() override
 	{
 		m_C2x2 = 2 * m_C2;
@@ -2810,19 +2855,19 @@ public:
 	{
 		T r = std::sqrt(fabs(helper.m_PrecalcSumSquares + helper.In.z));
 
-		r += m_Ar * sin(fma(m_Br, r, m_Cr));
+		r += m_Ar * std::sin(fma(m_Br, r, m_Cr));
 
 		if (r == 0)
 			r = EPS;
 
-		T temp = fma(m_At, sin(fma(m_Bt, r, m_Ct)), helper.m_PrecalcAtanyx);
+		T temp = fma(m_At, std::sin(fma(m_Bt, r, m_Ct)), helper.m_PrecalcAtanyx);
-		T st = sin(temp);
+		T st = std::sin(temp);
-		T ct = cos(temp);
+		T ct = std::cos(temp);
 
-		temp = fma(m_Ap, sin(fma(m_Bp, r, m_Cp)), acos(Clamp<T>(helper.In.z / r, -1, 1)));
+		temp = fma(m_Ap, std::sin(fma(m_Bp, r, m_Cp)), std::acos(Clamp<T>(helper.In.z / r, -1, 1)));
 
-		T sp = sin(temp);
+		T sp = std::sin(temp);
-		T cp = cos(temp);
+		T cp = std::cos(temp);
 
 		helper.Out.x = r * ct * sp;
 		helper.Out.y = r * st * sp;
@@ -2866,7 +2911,7 @@ public:
 		   << "\t\treal_t st = sin(temp);\n"
 		   << "\t\treal_t ct = cos(temp);\n"
 		   << "\n"
-		   << "\t\ttemp = fma(" << ap << ", sin(fma(" << bp << ", r, " << cp << ")), acos(Clamp(vIn.z / r, -(real_t)(1.0), (real_t)(1.0))));\n"
+		   << "\t\ttemp = fma(" << ap << ", sin(fma(" << bp << ", r, " << cp << ")), acos(clamp(vIn.z / r, -(real_t)(1.0), (real_t)(1.0))));\n"
 		   << "\n"
 		   << "\t\treal_t sp = sin(temp);\n"
 		   << "\t\treal_t cp = cos(temp);\n"
@@ -3080,7 +3125,7 @@ public:
 		return ss.str();
 	}
 
-	virtual string OpenCLFuncsString() const
+	virtual string OpenCLFuncsString() const override
 	{
 		return
 			"real_t Interference2Sine(real_t a, real_t b, real_t c, real_t p, real_t x)\n"
@@ -3121,17 +3166,17 @@ protected:
 private:
 	inline static T Sine(T a, T b, T c, T p, T x)
 	{
-		return a * pow(fabs(sin(b * x + c)), p);//Original did not fabs().
+		return a * std::pow(std::fabs(std::sin(b * x + c)), p);//Original did not fabs().
 	}
 
 	inline static T Tri(T a, T b, T c, T p, T x)
 	{
-		return a * 2 * pow(fabs(asin(cos(b * x + c - T(M_PI_2)))) * T(M_1_PI), p);//Original did not fabs().
+		return a * 2 * std::pow(std::fabs(std::asin(std::cos(b * x + c - T(M_PI_2)))) * T(M_1_PI), p);//Original did not fabs().
 	}
 
 	inline static T Squ(T a, T b, T c, T p, T x)
 	{
-		return a * pow(sin(b * x + c) < 0 ? EPS : T(1), p);//Original passed -1 to pow if sin() was < 0. Doing so will return NaN, so EPS is passed instead.
+		return a * std::pow(std::sin(b * x + c) < 0 ? EPS : T(1), p);//Original passed -1 to pow if sin() was < 0. Doing so will return NaN, so EPS is passed instead.
 	}
 
 	T m_A1;
@@ -3160,10 +3205,10 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T absV = Hypot<T>(helper.In.y, helper.In.z);
-		T s = sin(helper.In.x);
+		T s = std::sin(helper.In.x);
-		T c = cos(helper.In.x);
+		T c = std::cos(helper.In.x);
-		T sh = sinh(absV);
+		T sh = std::sinh(absV);
-		T ch = cosh(absV);
+		T ch = std::cosh(absV);
 		T d = m_Weight * c * sh / absV;
 
 		helper.Out.x = m_Weight * s * ch;
@@ -3191,6 +3236,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Hypot" };
+	}
 };
 
 /// <summary>
@@ -3207,10 +3257,10 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T absV = Hypot<T>(helper.In.y, helper.In.z);
-		T s = sin(absV);
+		T s = std::sin(absV);
-		T c = cos(absV);
+		T c = std::cos(absV);
-		T sh = sinh(helper.In.x);
+		T sh = std::sinh(helper.In.x);
-		T ch = cosh(helper.In.x);
+		T ch = std::cosh(helper.In.x);
 		T d = m_Weight * ch * s / absV;
 
 		helper.Out.x = m_Weight * sh * c;
@@ -3238,6 +3288,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Hypot" };
+	}
 };
 
 /// <summary>
@@ -3255,10 +3310,10 @@ public:
 	{
 		T absV = Hypot<T>(helper.In.y, helper.In.z);
 		T ni = m_Weight / (helper.m_PrecalcSumSquares + SQR(helper.In.z));
-		T s = sin(-helper.In.x);
+		T s = std::sin(-helper.In.x);
-		T c = cos(-helper.In.x);
+		T c = std::cos(-helper.In.x);
-		T sh = sinh(absV);
+		T sh = std::sinh(absV);
-		T ch = cosh(absV);
+		T ch = std::cosh(absV);
 		T d = ni * s * sh / absV;
 
 		helper.Out.x =   c * ch * ni;
@@ -3287,6 +3342,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Hypot" };
+	}
 };
 
 /// <summary>
@@ -3304,10 +3364,10 @@ public:
 	{
 		T absV = Hypot<T>(helper.In.y, helper.In.z);
 		T ni = m_Weight / (helper.m_PrecalcSumSquares + SQR(helper.In.z));
-		T s = sin(absV);
+		T s = std::sin(absV);
-		T c = cos(absV);
+		T c = std::cos(absV);
-		T sh = sinh(helper.In.x);
+		T sh = std::sinh(helper.In.x);
-		T ch = cosh(helper.In.x);
+		T ch = std::cosh(helper.In.x);
 		T d = ni * sh * s / absV;
 
 		helper.Out.x =  ch * c * ni;
@@ -3336,6 +3396,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Hypot" };
+	}
 };
 
 /// <summary>
@@ -3354,10 +3419,10 @@ public:
 		T sysz = SQR(helper.In.y) + SQR(helper.In.z);
 		T absV = std::sqrt(sysz);
 		T ni = m_Weight / (SQR(helper.In.x) + sysz);
-		T s = sin(helper.In.x);
+		T s = std::sin(helper.In.x);
-		T c = cos(helper.In.x);
+		T c = std::cos(helper.In.x);
-		T sh = sinh(absV);
+		T sh = std::sinh(absV);
-		T ch = cosh(absV);
+		T ch = std::cosh(absV);
 		T d = c * sh / absV;
 		T b = -s * sh / absV;
 		T stcv = s * ch;
@@ -3413,10 +3478,10 @@ public:
 		T sysz = SQR(helper.In.y) + SQR(helper.In.z);
 		T absV = std::sqrt(sysz);
 		T ni = m_Weight / (SQR(helper.In.x) + sysz);
-		T s = sin(absV);
+		T s = std::sin(absV);
-		T c = cos(absV);
+		T c = std::cos(absV);
-		T sh = sinh(helper.In.x);
+		T sh = std::sinh(helper.In.x);
-		T ch = cosh(helper.In.x);
+		T ch = std::cosh(helper.In.x);
 		T d = ch * s / absV;
 		T b = sh * s / absV;
 		T stcv = sh * c;
@@ -3470,10 +3535,10 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T absV = Hypot<T>(helper.In.y, helper.In.z);
-		T s = sin(helper.In.x);
+		T s = std::sin(helper.In.x);
-		T c = cos(helper.In.x);
+		T c = std::cos(helper.In.x);
-		T sh = sinh(absV);
+		T sh = std::sinh(absV);
-		T ch = cosh(absV);
+		T ch = std::cosh(absV);
 		T d = -m_Weight * s * sh / absV;
 
 		helper.Out.x = m_Weight * c * ch;
@@ -3501,6 +3566,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Hypot" };
+	}
 };
 
 /// <summary>
@@ -3517,10 +3587,10 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T absV = Hypot<T>(helper.In.y, helper.In.z);
-		T s = sin(absV);
+		T s = std::sin(absV);
-		T c = cos(absV);
+		T c = std::cos(absV);
-		T sh = sinh(helper.In.x);
+		T sh = std::sinh(helper.In.x);
-		T ch = cosh(helper.In.x);
+		T ch = std::cosh(helper.In.x);
 		T d = -m_Weight * sh * s / absV;
 
 		helper.Out.x = m_Weight * c * ch;
@@ -3548,6 +3618,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Hypot" };
+	}
 };
 
 /// <summary>
@@ -3566,10 +3641,10 @@ public:
 		T sysz = SQR(helper.In.y) + SQR(helper.In.z);
 		T absV = std::sqrt(sysz);
 		T ni = m_Weight / (SQR(helper.In.x) + sysz);
-		T s = sin(helper.In.x);
+		T s = std::sin(helper.In.x);
-		T c = cos(helper.In.x);
+		T c = std::cos(helper.In.x);
-		T sh = sinh(absV);
+		T sh = std::sinh(absV);
-		T ch = cosh(absV);
+		T ch = std::cosh(absV);
 		T d = c * sh / absV;
 		T b = -s * sh / absV;
 		T stcv = s * ch;
@@ -3625,10 +3700,10 @@ public:
 		T sysz = SQR(helper.In.y) + SQR(helper.In.z);
 		T absV = std::sqrt(sysz);
 		T ni = m_Weight / (SQR(helper.In.x) + sysz);
-		T s = sin(absV);
+		T s = std::sin(absV);
-		T c = cos(absV);
+		T c = std::cos(absV);
-		T sh = sinh(helper.In.x);
+		T sh = std::sinh(helper.In.x);
-		T ch = cosh(helper.In.x);
+		T ch = std::cosh(helper.In.x);
 		T d = ch * s / absV;
 		T b = sh * s / absV;
 		T stcv = sh * c;
@@ -3683,10 +3758,10 @@ public:
 	{
 		T absV = Hypot<T>(helper.In.y, helper.In.z);
 		T ni = m_Weight / (helper.m_PrecalcSumSquares + SQR(helper.In.z));
-		T s = sin(helper.In.x);
+		T s = std::sin(helper.In.x);
-		T c = cos(helper.In.x);
+		T c = std::cos(helper.In.x);
-		T sh = sinh(absV);
+		T sh = std::sinh(absV);
-		T ch = cosh(absV);
+		T ch = std::cosh(absV);
 		T d = ni * c * sh / absV;
 
 		helper.Out.x =   s * ch * ni;
@@ -3715,6 +3790,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Hypot" };
+	}
 };
 
 /// <summary>
@@ -3732,10 +3812,10 @@ public:
 	{
 		T absV = Hypot<T>(helper.In.y, helper.In.z);
 		T ni = m_Weight / (helper.m_PrecalcSumSquares + SQR(helper.In.z));
-		T s = sin(absV);
+		T s = std::sin(absV);
-		T c = cos(absV);
+		T c = std::cos(absV);
-		T sh = sinh(helper.In.x);
+		T sh = std::sinh(helper.In.x);
-		T ch = cosh(helper.In.x);
+		T ch = std::cosh(helper.In.x);
 		T d = ni * ch * s / absV;
 
 		helper.Out.x =  sh * c * ni;
@@ -3764,6 +3844,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Hypot" };
+	}
 };
 
 /// <summary>
@@ -3780,9 +3865,9 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T absV = Hypot<T>(helper.In.y, helper.In.z);
-		T e = exp(helper.In.x);
+		T e = std::exp(helper.In.x);
-		T s = sin(absV);
+		T s = std::sin(absV);
-		T c = cos(absV);
+		T c = std::cos(absV);
 		T a = e * s / absV;
 
 		helper.Out.x = m_Weight * e * c;
@@ -3809,6 +3894,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Hypot" };
+	}
 };
 
 /// <summary>
@@ -3828,9 +3918,9 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T absV = Hypot<T>(helper.In.y, helper.In.z);
-		T c = m_Weight * atan2(absV, helper.In.x) / absV;
+		T c = m_Weight * std::atan2(absV, helper.In.x) / absV;
 
-		helper.Out.x = log(SQR(helper.In.x) + SQR(absV)) * m_Denom;
+		helper.Out.x = std::log(SQR(helper.In.x) + SQR(absV)) * m_Denom;
 		helper.Out.y = c * helper.In.y;
 		helper.Out.z = c * helper.In.z;
 	}
@@ -3856,9 +3946,14 @@ public:
 		return ss.str();
 	}
 	
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Hypot" };
+	}
+
 	virtual void Precalc() override
 	{
-		m_Denom = T(0.5) / log(m_Base);
+		m_Denom = T(0.5) / std::log(m_Base);
 	}
 
 protected:
@@ -3907,6 +4002,11 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
 };
 
 /// <summary>
@@ -4024,7 +4124,7 @@ public:
 		T xx = (rand.Frand01<T>() - T(0.5)) * 2;
 		T yy = (rand.Frand01<T>() - T(0.5)) * 2;
 		T k = SignNz(yy);
-		T yymax = ((m_A * pow(fabs(xx), m_P) + k * m_B * std::sqrt(fabs(1 - SQR(xx)))) - m_A);
+		T yymax = ((m_A * std::pow(fabs(xx), m_P) + k * m_B * std::sqrt(fabs(1 - SQR(xx)))) - m_A);
 
 		//The function must be in a range 0-1 to work properly.
 		yymax /= Zeps(fabs(m_A) + fabs(m_B));
@@ -4083,6 +4183,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "SignNz", "Zeps" };
+	}
+
 protected:
 	void Init()
 	{
@@ -4190,26 +4295,26 @@ public:
 		{
 			if (tileType < 1)
 			{
-				r0 = pow((pow(fabs(x    ), m_Exponent) + pow(fabs(y    ), m_Exponent)), m_OneOverEx);
+				r0 = std::pow((pow(fabs(x    ), m_Exponent) + std::pow(fabs(y    ), m_Exponent)), m_OneOverEx);
-				r1 = pow((pow(fabs(x - 1), m_Exponent) + pow(fabs(y - 1), m_Exponent)), m_OneOverEx);
+				r1 = std::pow((pow(fabs(x - 1), m_Exponent) + std::pow(fabs(y - 1), m_Exponent)), m_OneOverEx);
 			}
 			else
 			{
-				r0 = pow((pow(fabs(x - 1), m_Exponent) + pow(fabs(y    ), m_Exponent)), m_OneOverEx);
+				r0 = std::pow((pow(fabs(x - 1), m_Exponent) + std::pow(fabs(y    ), m_Exponent)), m_OneOverEx);
-				r1 = pow((pow(fabs(x    ), m_Exponent) + pow(fabs(y - 1), m_Exponent)), m_OneOverEx);
+				r1 = std::pow((pow(fabs(x    ), m_Exponent) + std::pow(fabs(y - 1), m_Exponent)), m_OneOverEx);
 			}
 		}
 		else//Slow drawmode
 		{
 			if (tileType == 1)
 			{
-				r0 = pow((pow(fabs(x    ), m_Exponent) + pow(fabs(y    ), m_Exponent)), m_OneOverEx);
+				r0 = std::pow((std::pow(fabs(x    ), m_Exponent) + std::pow(fabs(y    ), m_Exponent)), m_OneOverEx);
-				r1 = pow((pow(fabs(x - 1), m_Exponent) + pow(fabs(y - 1), m_Exponent)), m_OneOverEx);
+				r1 = std::pow((std::pow(fabs(x - 1), m_Exponent) + std::pow(fabs(y - 1), m_Exponent)), m_OneOverEx);
 			}
 			else
 			{
-				r0 = pow((pow(fabs(x - 1), m_Exponent) + pow(fabs(y    ), m_Exponent)), m_OneOverEx);
+				r0 = std::pow((std::pow(fabs(x - 1), m_Exponent) + std::pow(fabs(y    ), m_Exponent)), m_OneOverEx);
-				r1 = pow((pow(fabs(x    ), m_Exponent) + pow(fabs(y - 1), m_Exponent)), m_OneOverEx);
+				r1 = std::pow((std::pow(fabs(x    ), m_Exponent) + std::pow(fabs(y - 1), m_Exponent)), m_OneOverEx);
 			}
 		}
 
@@ -4378,13 +4483,18 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Round" };
+	}
+
 	virtual void Precalc() override
 	{
 		m_OneOverEx = 1 / m_Exponent;
 		m_AbsSeed = fabs(m_Seed);
 		m_Seed2 = std::sqrt(Zeps(m_AbsSeed + (m_AbsSeed / 2))) / Zeps((m_AbsSeed * T(0.5))) * T(0.25);
-		m_OneOverRmax = 1 / (T(0.5) * (pow(T(2), 1 / m_Exponent) - 1) * m_ArcWidth);
+		m_OneOverRmax = 1 / (T(0.5) * (std::pow(T(2), 1 / m_Exponent) - 1) * m_ArcWidth);
-		m_Scale = (cos(-m_Rotation) - sin(-m_Rotation)) / m_Weight;
+		m_Scale = (std::cos(-m_Rotation) - std::sin(-m_Rotation)) / m_Weight;
 	}
 
 protected:
@@ -4508,7 +4618,7 @@ public:
 		return ss.str();
 	}
 
-	virtual string OpenCLFuncsString() const
+	virtual string OpenCLFuncsString() const override
 	{
 		return
 			"inline real_t GdoffsFcip(real_t a) { return (real_t)((a < 0) ? -((int)(fabs(a)) + 1) : 0) + ((a > 1) ? ((int)(a)) : 0); }\n"
@@ -4561,7 +4671,7 @@ private:
 	static inline T GdoffsFcip(T a) { return T((a < 0) ? -(int(fabs(a)) + 1) : 0) + ((a > 1) ? (int(a)) : 0); }
 	static inline T GdoffsFclp(T a) { return ((a < 0) ? -(fmod(fabs(a), T(1))) : fmod(fabs(a), T(1))); }
 	static inline T GdoffsFscl(T a) { return GdoffsFclp((a + 1) / 2); }
-	static inline T GdoffsFosc(T p, T a) { return GdoffsFscl(-1 * cos(p * a * M_2PI)); }
+	static inline T GdoffsFosc(T p, T a) { return GdoffsFscl(-1 * std::cos(p * a * M_2PI)); }
 	static inline T GdoffsFlip(T a, T b, T c) { return (c * (b - a) + a); }
 
 	T m_DeltaX;//Params.
@@ -4703,6 +4813,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Zeps" };
+	}
+
 protected:
 	void Init()
 	{
@@ -4748,8 +4863,8 @@ public:
 				r *= m_R2 / m_R1;
 				temp = atan2(helper.In.y, c1mx);
 
-				helper.Out.x = m_Weight * (r * cos(temp) - m_C2);
+				helper.Out.x = m_Weight * (r * std::cos(temp) - m_C2);
-				helper.Out.y = m_Weight *  r * sin(temp);
+				helper.Out.y = m_Weight *  r * std::sin(temp);
 			}
 			else
 			{
@@ -4767,8 +4882,8 @@ public:
 				r *= m_R1 / m_R2;
 				temp = atan2(helper.In.y, c1mx);
 
-				helper.Out.x = m_Weight * (r * cos(temp) + m_C1);
+				helper.Out.x = m_Weight * (r * std::cos(temp) + m_C1);
-				helper.Out.y = m_Weight *  r * sin(temp);
+				helper.Out.y = m_Weight *  r * std::sin(temp);
 			}
 			else
 			{
@@ -4887,12 +5002,12 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T arg = helper.m_PrecalcAtanyx + fmod(T(rand.Rand()), T(1 / m_ReInv)) * M_2PI;
-		T lnmod = m_Dist * T(0.5) * log(helper.m_PrecalcSumSquares);
+		T lnmod = m_Dist * T(0.5) * std::log(helper.m_PrecalcSumSquares);
 		T temp = arg * m_ReInv + lnmod * m_Im100;
-		T mod2 = exp(lnmod * m_ReInv - arg * m_Im100);
+		T mod2 = std::exp(lnmod * m_ReInv - arg * m_Im100);
 
-		helper.Out.x = m_Weight * mod2 * cos(temp);
+		helper.Out.x = m_Weight * mod2 * std::cos(temp);
-		helper.Out.y = m_Weight * mod2 * sin(temp);
+		helper.Out.y = m_Weight * mod2 * std::sin(temp);
 		helper.Out.z = (m_VarType == VARTYPE_REG) ? 0 : helper.In.z;
 	}
 
@@ -5006,11 +5121,11 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		T r = helper.m_PrecalcSqrtSumSquares * (m_BlobLow + m_BlobDiff * (T(0.5) + T(0.5) * sin(m_BlobWaves * helper.m_PrecalcAtanxy)));
+		T r = helper.m_PrecalcSqrtSumSquares * (m_BlobLow + m_BlobDiff * (T(0.5) + T(0.5) * std::sin(m_BlobWaves * helper.m_PrecalcAtanxy)));
 
 		helper.Out.x = m_Weight * helper.m_PrecalcSina * r;
 		helper.Out.y = m_Weight * helper.m_PrecalcCosa * r;
-		helper.Out.z = m_Weight * sin(m_BlobWaves * helper.m_PrecalcAtanxy) * r;
+		helper.Out.z = m_Weight * std::sin(m_BlobWaves * helper.m_PrecalcAtanxy) * r;
 	}
 
 	virtual string OpenCLString() const override
@@ -5082,7 +5197,7 @@ public:
 
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
-		T t = Zeps((cos(helper.In.x) + cos(helper.In.y)) / m_Mp + 1);
+		T t = Zeps((std::cos(helper.In.x) + std::cos(helper.In.y)) / m_Mp + 1);
 		T r = m_Weight / t;
 		T tmp = helper.m_PrecalcSumSquares + 1;
 		T x2 = 2 * helper.In.x;
@@ -5137,6 +5252,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "SafeSqrt", "Zeps" };
+	}
+
 	virtual void Precalc() override
 	{
 		m_V = m_Weight / T(M_PI_2);

Source/Ember/VariationsDC.h

@@ -6,8 +6,7 @@ namespace EmberNs
 {
 /// <summary>
 /// DC Bubble.
-/// This accesses the summed output point in a rare and different way
+/// This accesses the summed output point in a rare and different way.
-/// and therefore cannot be made into pre and post variations.
 /// </summary>
 template <typename T>
 class EMBER_API DCBubbleVariation : public ParametricVariation<T>
@@ -30,8 +29,21 @@ public:
 		helper.Out.y = r4_1 * helper.In.y;
 		helper.Out.z = m_Weight * (2 / r4_1 - 1);
 
-		T tempX = helper.Out.x + outPoint.m_X;
+		T sumX, sumY;
-		T tempY = helper.Out.y + outPoint.m_Y;
+
+		if (m_VarType == VARTYPE_PRE)
+		{
+			sumX = helper.In.x;
+			sumY = helper.In.y;
+		}
+		else
+		{
+			sumX = outPoint.m_X;
+			sumY = outPoint.m_Y;
+		}
+
+		T tempX = helper.Out.x + sumX;
+		T tempY = helper.Out.y + sumY;
 
 		outPoint.m_ColorX = fmod(fabs(m_Bdcs * (Sqr<T>(tempX + m_CenterX) + Sqr<T>(tempY + m_CenterY))), T(1.0));
 	}
@@ -56,8 +68,24 @@ public:
 		   << "\t\tvOut.y = r4_1 * vIn.y;\n"
 		   << "\t\tvOut.z = xform->m_VariationWeights[" << varIndex << "] * (2 / r4_1 - 1);\n"
 		   << "\n"
-		   << "\t\treal_t tempX = vOut.x + outPoint->m_X;\n"
+		   << "\t\treal_t sumX, sumY;\n\n";
-		   << "\t\treal_t tempY = vOut.y + outPoint->m_Y;\n"
+
+		if (m_VarType == VARTYPE_PRE)
+		{
+			ss
+				<< "\t\tsumX = vIn.x;\n"
+				<< "\t\tsumY = vIn.y;\n";
+		}
+		else
+		{
+			ss
+				<< "\t\tsumX = outPoint->m_X;\n"
+				<< "\t\tsumY = outPoint->m_Y;\n";
+		}
+
+		ss
+		   << "\t\treal_t tempX = vOut.x + sumX;\n"
+		   << "\t\treal_t tempY = vOut.y + sumY;\n"
 		   << "\n"
 		   << "\t\toutPoint->m_ColorX = fmod(fabs(" << bdcs << " * (Sqr(tempX + " << centerX << ") + Sqr(tempY + " << centerY << "))), (real_t)(1.0));\n"
 		   << "\t}\n";
@@ -65,6 +93,11 @@ public:
 		return ss.str();
 	}
 
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "Sqr", "Zeps" };
+	}
+
 	virtual void Precalc() override
 	{
 		m_Bdcs = 1 / (m_Scale == 0 ? T(10E-6) : m_Scale);
@@ -255,7 +288,7 @@ public:
 		   << "\t\treal_t x, y, z;\n"
 		   << "\t\treal_t p = 2 * MwcNext01(mwc) - 1;\n"
 		   << "\t\treal_t q = 2 * MwcNext01(mwc) - 1;\n"
-		   << "\t\tuint i = MwcNext(mwc) % 3;\n"
+		   << "\t\tuint i = MwcNextRange(mwc, 3);\n"
 		   << "\t\tuint j = MwcNext(mwc) & 1;\n"
 		   << "\n"
 		   << "\t\tswitch (i)\n"
@@ -356,8 +389,7 @@ private:
 
 /// <summary>
 /// DC Cylinder.
-/// This accesses the summed output point in a rare and different way
+/// This accesses the summed output point in a rare and different way.
-/// and therefore cannot be made into pre and post variations.
 /// </summary>
 template <typename T>
 class EMBER_API DCCylinderVariation : public ParametricVariation<T>
@@ -373,16 +405,29 @@ public:
 	virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
 	{
 		T temp = rand.Frand01<T>() * M_2PI;
-		T sr = sin(temp);
+		T sr = std::sin(temp);
-		T cr = cos(temp);
+		T cr = std::cos(temp);
 		T r = m_Blur * (rand.Frand01<T>() + rand.Frand01<T>() + rand.Frand01<T>() + rand.Frand01<T>() - 2);
 
-		helper.Out.x = m_Weight * sin(helper.In.x + r * sr) * m_X;
+		helper.Out.x = m_Weight * std::sin(helper.In.x + r * sr) * m_X;
 		helper.Out.y = r + helper.In.y * m_Y;
-		helper.Out.z = m_Weight * cos(helper.In.x + r * cr);
+		helper.Out.z = m_Weight * std::cos(helper.In.x + r * cr);
 
-		T tempX = helper.Out.x + outPoint.m_X;
+		T sumX, sumY;
-		T tempY = helper.Out.y + outPoint.m_Y;
+
+		if (m_VarType == VARTYPE_PRE)
+		{
+			sumX = helper.In.x;
+			sumY = helper.In.y;
+		}
+		else
+		{
+			sumX = outPoint.m_X;
+			sumY = outPoint.m_Y;
+		}
+
+		T tempX = helper.Out.x + sumX;
+		T tempY = helper.Out.y + sumY;
 
 		outPoint.m_ColorX = fmod(fabs(T(0.5) * (m_Ldcs * ((m_Cosa * tempX + m_Sina * tempY + m_Offset)) + 1)), T(1.0));
 	}
@@ -413,8 +458,24 @@ public:
 		   << "\t\tvOut.y = r + vIn.y * " << y << ";\n"
 		   << "\t\tvOut.z = xform->m_VariationWeights[" << varIndex << "] * cos(vIn.x + r * cr);\n"
 		   << "\n"
-		   << "\t\treal_t tempX = vOut.x + outPoint->m_X;\n"
+		   << "\t\treal_t sumX, sumY;\n\n";
-		   << "\t\treal_t tempY = vOut.y + outPoint->m_Y;\n"
+
+		if (m_VarType == VARTYPE_PRE)
+		{
+			ss
+				<< "\t\tsumX = vIn.x;\n"
+				<< "\t\tsumY = vIn.y;\n";
+		}
+		else
+		{
+			ss
+				<< "\t\tsumX = outPoint->m_X;\n"
+				<< "\t\tsumY = outPoint->m_Y;\n";
+		}
+
+		ss
+		   << "\t\treal_t tempX = vOut.x + sumX;\n"
+		   << "\t\treal_t tempY = vOut.y + sumY;\n"
 		   << "\n"
 		   << "\t\toutPoint->m_ColorX = fmod(fabs((real_t)(0.5) * (" << ldcs << " * ((" << cosa << " * tempX + " << sina << " * tempY + " << offset << ")) + (real_t)(1.0))), (real_t)(1.0));\n"
 		   << "\t}\n";
@@ -631,12 +692,16 @@ public:
 
 		return ss.str();
 	}
+
+	virtual vector<string> OpenCLGlobalFuncNames() const override
+	{
+		return vector<string> { "LRint" };
+	}
 };
 
 /// <summary>
 /// DC Linear.
-/// This accesses the summed output point in a rare and different way
+/// This accesses the summed output point in a rare and different way.
-/// and therefore cannot be made into pre and post variations.
 /// </summary>
 template <typename T>
 class EMBER_API DCLinearVariation : public ParametricVariation<T>
@@ -655,8 +720,21 @@ public:
 		helper.Out.y = m_Weight * helper.In.y;
 		helper.Out.z = m_Weight * helper.In.z;
 
-		T tempX = helper.Out.x + outPoint.m_X;
+		T sumX, sumY;
-		T tempY = helper.Out.y + outPoint.m_Y;
+
+		if (m_VarType == VARTYPE_PRE)
+		{
+			sumX = helper.In.x;
+			sumY = helper.In.y;
+		}
+		else
+		{
+			sumX = outPoint.m_X;
+			sumY = outPoint.m_Y;
+		}
+
+		T tempX = helper.Out.x + sumX;
+		T tempY = helper.Out.y + sumY;
 
 		outPoint.m_ColorX = fmod(fabs(T(0.5) * (m_Ldcs * ((m_Cosa * tempX + m_Sina * tempY + m_Offset)) + T(1.0))), T(1.0));
 	}
@@ -680,8 +758,24 @@ public:
 		   << "\t\tvOut.y = xform->m_VariationWeights[" << varIndex << "] * vIn.y;\n"
 		   << "\t\tvOut.z = xform->m_VariationWeights[" << varIndex << "] * vIn.z;\n"
 		   << "\n"
-		   << "\t\treal_t tempX = vOut.x + outPoint->m_X;\n"
+		   << "\t\treal_t sumX, sumY;\n\n";
-		   << "\t\treal_t tempY = vOut.y + outPoint->m_Y;\n"
+
+		if (m_VarType == VARTYPE_PRE)
+		{
+			ss
+				<< "\t\tsumX = vIn.x;\n"
+				<< "\t\tsumY = vIn.y;\n";
+		}
+		else
+		{
+			ss
+				<< "\t\tsumX = outPoint->m_X;\n"
+				<< "\t\tsumY = outPoint->m_Y;\n";
+		}
+
+		ss
+		   << "\t\treal_t tempX = vOut.x + sumX;\n"
+		   << "\t\treal_t tempY = vOut.y + sumY;\n"
 		   << "\n"
 		   << "\t\toutPoint->m_ColorX = fmod(fabs((real_t)(0.5) * (" << ldcs << " * ((" << cosa << " * tempX + " << sina << " * tempY + " << offset << ")) + (real_t)(1.0))), (real_t)(1.0));\n"
 		   << "\t}\n";
@@ -1034,9 +1128,12 @@ private:
 	T m_X1_m_x0;
 };
 
+MAKEPREPOSTPARVAR(DCBubble, dc_bubble, DC_BUBBLE)
 MAKEPREPOSTPARVAR(DCCarpet, dc_carpet, DC_CARPET)
 MAKEPREPOSTPARVARASSIGN(DCCube, dc_cube, DC_CUBE, ASSIGNTYPE_SUM)
+MAKEPREPOSTPARVAR(DCCylinder, dc_cylinder, DC_CYLINDER)
 MAKEPREPOSTVAR(DCGridOut, dc_gridout, DC_GRIDOUT)
+MAKEPREPOSTPARVAR(DCLinear, dc_linear, DC_LINEAR)
 MAKEPREPOSTPARVAR(DCTriangle, dc_triangle, DC_TRIANGLE)
 MAKEPREPOSTPARVAR(DCZTransl, dc_ztransl, DC_ZTRANSL)
 }

Source/Ember/Xform.h

@@ -915,7 +915,6 @@ public:
 			AllVarsFunc([&] (vector<Variation<T>*>& variations, bool& keepGoing)
 			{
 				for (auto var : variations)
-				//for (size_t i = 0; i < variations.size(); i++)
 				{
 					if (var->m_Weight != 0)//This should never happen, but just to be safe.
 					{
@@ -1273,7 +1272,7 @@ private:
 		if (in == 0)
 			return 0;
 		else
-			return pow(T(10.0), -log(T(1.0) / T(in)) / log(T(2)));
+			return std::pow(T(10.0), -std::log(T(1.0) / T(in)) / std::log(T(2)));
 	}
 
 	vector<T> m_Xaos;//Xaos vector which affects the probability that this xform is chosen. Usually empty.

Source/Ember/XmlToEmber.h

@@ -80,81 +80,100 @@ public:
 		if (!m_Init)
 		{
 			m_BadParamNames.reserve(100);
-			m_BadParamNames.push_back(pair<string, string>("swtin_distort",   "stwin_distort"));//stwin.
+			m_BadParamNames["swtin_distort"] =   "stwin_distort";//stwin.
-			m_BadParamNames.push_back(pair<string, string>("pow_numerator",   "pow_block_numerator"));//pow_block.
+			m_BadParamNames["pow_numerator"] =   "pow_block_numerator";//pow_block.
-			m_BadParamNames.push_back(pair<string, string>("pow_denominator", "pow_block_denominator"));
+			m_BadParamNames["pow_denominator"] = "pow_block_denominator";
-			m_BadParamNames.push_back(pair<string, string>("pow_root",        "pow_block_root"));
+			m_BadParamNames["pow_root"] =        "pow_block_root";
-			m_BadParamNames.push_back(pair<string, string>("pow_correctn",    "pow_block_correctn"));
+			m_BadParamNames["pow_correctn"] =    "pow_block_correctn";
-			m_BadParamNames.push_back(pair<string, string>("pow_correctd",    "pow_block_correctd"));
+			m_BadParamNames["pow_correctd"] =    "pow_block_correctd";
-			m_BadParamNames.push_back(pair<string, string>("pow_power",       "pow_block_power"));
+			m_BadParamNames["pow_power"] =       "pow_block_power";
-			m_BadParamNames.push_back(pair<string, string>("lT", "linearT_powX"));//linearT.
+			m_BadParamNames["lt"] = "linearT_powX";//linearT.
-			m_BadParamNames.push_back(pair<string, string>("lT", "linearT_powY"));
+			m_BadParamNames["lt"] = "linearT_powY";
-			m_BadParamNames.push_back(pair<string, string>("Re_A", "Mobius_Re_A"));//Mobius.
+			m_BadParamNames["re_a"] = "Mobius_Re_A";//Mobius.
-			m_BadParamNames.push_back(pair<string, string>("Im_A", "Mobius_Im_A"));
+			m_BadParamNames["im_a"] = "Mobius_Im_A";
-			m_BadParamNames.push_back(pair<string, string>("Re_B", "Mobius_Re_B"));
+			m_BadParamNames["re_b"] = "Mobius_Re_B";
-			m_BadParamNames.push_back(pair<string, string>("Im_B", "Mobius_Im_B"));
+			m_BadParamNames["im_b"] = "Mobius_Im_B";
-			m_BadParamNames.push_back(pair<string, string>("Re_C", "Mobius_Re_C"));
+			m_BadParamNames["re_c"] = "Mobius_Re_C";
-			m_BadParamNames.push_back(pair<string, string>("Im_C", "Mobius_Im_C"));
+			m_BadParamNames["im_c"] = "Mobius_Im_C";
-			m_BadParamNames.push_back(pair<string, string>("Re_D", "Mobius_Re_D"));
+			m_BadParamNames["re_d"] = "Mobius_Re_D";
-			m_BadParamNames.push_back(pair<string, string>("Im_D", "Mobius_Im_D"));
+			m_BadParamNames["im_d"] = "Mobius_Im_D";
-			m_BadParamNames.push_back(pair<string, string>("rx_sin", "rotate_x_sin"));//rotate_x.
+			m_BadParamNames["rx_sin"] = "rotate_x_sin";//rotate_x.
-			m_BadParamNames.push_back(pair<string, string>("rx_cos", "rotate_x_cos"));
+			m_BadParamNames["rx_cos"] = "rotate_x_cos";
-			m_BadParamNames.push_back(pair<string, string>("ry_sin", "rotate_y_sin"));//rotate_y.
+			m_BadParamNames["ry_sin"] = "rotate_y_sin";//rotate_y.
-			m_BadParamNames.push_back(pair<string, string>("ry_cos", "rotate_y_cos"));
+			m_BadParamNames["ry_cos"] = "rotate_y_cos";
-			m_BadParamNames.push_back(pair<string, string>("intrfr2_a1", "interference2_a1"));//interference2.
+			m_BadParamNames["intrfr2_a1"] = "interference2_a1";//interference2.
-			m_BadParamNames.push_back(pair<string, string>("intrfr2_b1", "interference2_b1"));
+			m_BadParamNames["intrfr2_b1"] = "interference2_b1";
-			m_BadParamNames.push_back(pair<string, string>("intrfr2_c1", "interference2_c1"));
+			m_BadParamNames["intrfr2_c1"] = "interference2_c1";
-			m_BadParamNames.push_back(pair<string, string>("intrfr2_p1", "interference2_p1"));
+			m_BadParamNames["intrfr2_p1"] = "interference2_p1";
-			m_BadParamNames.push_back(pair<string, string>("intrfr2_t1", "interference2_t1"));
+			m_BadParamNames["intrfr2_t1"] = "interference2_t1";
-			m_BadParamNames.push_back(pair<string, string>("intrfr2_a2", "interference2_a2"));
+			m_BadParamNames["intrfr2_a2"] = "interference2_a2";
-			m_BadParamNames.push_back(pair<string, string>("intrfr2_b2", "interference2_b2"));
+			m_BadParamNames["intrfr2_b2"] = "interference2_b2";
-			m_BadParamNames.push_back(pair<string, string>("intrfr2_c2", "interference2_c2"));
+			m_BadParamNames["intrfr2_c2"] = "interference2_c2";
-			m_BadParamNames.push_back(pair<string, string>("intrfr2_p2", "interference2_p2"));
+			m_BadParamNames["intrfr2_p2"] = "interference2_p2";
-			m_BadParamNames.push_back(pair<string, string>("intrfr2_t2", "interference2_t2"));
+			m_BadParamNames["intrfr2_t2"] = "interference2_t2";
-			m_BadParamNames.push_back(pair<string, string>("octa_x", "octagon_x"));//octagon.
+			m_BadParamNames["octa_x"] = "octagon_x";//octagon.
-			m_BadParamNames.push_back(pair<string, string>("octa_y", "octagon_y"));
+			m_BadParamNames["octa_y"] = "octagon_y";
-			m_BadParamNames.push_back(pair<string, string>("octa_z", "octagon_z"));
+			m_BadParamNames["octa_z"] = "octagon_z";
-			m_BadParamNames.push_back(pair<string, string>("bubble_x", "bubble2_x"));//bubble2.
+			m_BadParamNames["bubble_x"] = "bubble2_x";//bubble2.
-			m_BadParamNames.push_back(pair<string, string>("bubble_y", "bubble2_y"));
+			m_BadParamNames["bubble_y"] = "bubble2_y";
-			m_BadParamNames.push_back(pair<string, string>("bubble_z", "bubble2_z"));
+			m_BadParamNames["bubble_z"] = "bubble2_z";
-			m_BadParamNames.push_back(pair<string, string>("cubic3D_xpand", "cubicLattice_3D_xpand"));//cubicLattice_3D.
+			m_BadParamNames["cubic3d_xpand"] = "cubicLattice_3D_xpand";//cubicLattice_3D.
-			m_BadParamNames.push_back(pair<string, string>("cubic3D_style", "cubicLattice_3D_style"));
+			m_BadParamNames["cubic3d_style"] = "cubicLattice_3D_style";
-			m_BadParamNames.push_back(pair<string, string>("splitb_x",  "SplitBrdr_x"));//SplitBrdr.
+			m_BadParamNames["splitb_x"] =  "SplitBrdr_x";//SplitBrdr.
-			m_BadParamNames.push_back(pair<string, string>("splitb_y",  "SplitBrdr_y"));
+			m_BadParamNames["splitb_y"] =  "SplitBrdr_y";
-			m_BadParamNames.push_back(pair<string, string>("splitb_px", "SplitBrdr_px"));
+			m_BadParamNames["splitb_px"] = "SplitBrdr_px";
-			m_BadParamNames.push_back(pair<string, string>("splitb_py", "SplitBrdr_py"));
+			m_BadParamNames["splitb_py"] = "SplitBrdr_py";
-			m_BadParamNames.push_back(pair<string, string>("dc_cyl_offset", "dc_cylinder_offset"));//dc_cylinder.
+			m_BadParamNames["dc_cyl_offset"] = "dc_cylinder_offset";//dc_cylinder.
-			m_BadParamNames.push_back(pair<string, string>("dc_cyl_angle",  "dc_cylinder_angle"));
+			m_BadParamNames["dc_cyl_angle"] =  "dc_cylinder_angle";
-			m_BadParamNames.push_back(pair<string, string>("dc_cyl_scale",  "dc_cylinder_scale"));
+			m_BadParamNames["dc_cyl_scale"] =  "dc_cylinder_scale";
-			m_BadParamNames.push_back(pair<string, string>("cyl_x",         "dc_cylinder_x"));
+			m_BadParamNames["cyl_x"] =         "dc_cylinder_x";
-			m_BadParamNames.push_back(pair<string, string>("cyl_y",         "dc_cylinder_y"));
+			m_BadParamNames["cyl_y"] =         "dc_cylinder_y";
-			m_BadParamNames.push_back(pair<string, string>("cyl_blur",      "dc_cylinder_blur"));
+			m_BadParamNames["cyl_blur"] =      "dc_cylinder_blur";
-			m_BadParamNames.push_back(pair<string, string>("mobius_radius",   "mobius_strip_radius"));//mobius_strip.
+			m_BadParamNames["mobius_radius"] =   "mobius_strip_radius";//mobius_strip.
-			m_BadParamNames.push_back(pair<string, string>("mobius_width",    "mobius_strip_width"));
+			m_BadParamNames["mobius_width"] =    "mobius_strip_width";
-			m_BadParamNames.push_back(pair<string, string>("mobius_rect_x",   "mobius_strip_rect_x"));
+			m_BadParamNames["mobius_rect_x"] =   "mobius_strip_rect_x";
-			m_BadParamNames.push_back(pair<string, string>("mobius_rect_y",   "mobius_strip_rect_y"));
+			m_BadParamNames["mobius_rect_y"] =   "mobius_strip_rect_y";
-			m_BadParamNames.push_back(pair<string, string>("mobius_rotate_x", "mobius_strip_rotate_x"));
+			m_BadParamNames["mobius_rotate_x"] = "mobius_strip_rotate_x";
-			m_BadParamNames.push_back(pair<string, string>("mobius_rotate_y", "mobius_strip_rotate_y"));
+			m_BadParamNames["mobius_rotate_y"] = "mobius_strip_rotate_y";
-			m_BadParamNames.push_back(pair<string, string>("bwraps2_cellsize",    "bwraps_cellsize"));//bwraps2.
+			m_BadParamNames["bwraps2_cellsize"] =    "bwraps_cellsize";//bwraps2.
-			m_BadParamNames.push_back(pair<string, string>("bwraps2_space",       "bwraps_space"));
+			m_BadParamNames["bwraps2_space"] =       "bwraps_space";
-			m_BadParamNames.push_back(pair<string, string>("bwraps2_gain",        "bwraps_gain"));
+			m_BadParamNames["bwraps2_gain"] =        "bwraps_gain";
-			m_BadParamNames.push_back(pair<string, string>("bwraps2_inner_twist", "bwraps_inner_twist"));
+			m_BadParamNames["bwraps2_inner_twist"] = "bwraps_inner_twist";
-			m_BadParamNames.push_back(pair<string, string>("bwraps2_outer_twist", "bwraps_outer_twist"));
+			m_BadParamNames["bwraps2_outer_twist"] = "bwraps_outer_twist";
-			m_BadParamNames.push_back(pair<string, string>("bwraps7_cellsize",    "bwraps_cellsize"));//bwraps7.
+			m_BadParamNames["bwraps7_cellsize"] =    "bwraps_cellsize";//bwraps7.
-			m_BadParamNames.push_back(pair<string, string>("bwraps7_space",       "bwraps_space"));
+			m_BadParamNames["bwraps7_space"] =       "bwraps_space";
-			m_BadParamNames.push_back(pair<string, string>("bwraps7_gain",        "bwraps_gain"));
+			m_BadParamNames["bwraps7_gain"] =        "bwraps_gain";
-			m_BadParamNames.push_back(pair<string, string>("bwraps7_inner_twist", "bwraps_inner_twist"));
+			m_BadParamNames["bwraps7_inner_twist"] = "bwraps_inner_twist";
-			m_BadParamNames.push_back(pair<string, string>("bwraps7_outer_twist", "bwraps_outer_twist"));
+			m_BadParamNames["bwraps7_outer_twist"] = "bwraps_outer_twist";
-			m_BadParamNames.push_back(pair<string, string>("pre_bwraps2_cellsize",    "pre_bwraps_cellsize"));
+			m_BadParamNames["pre_bwraps2_cellsize"] =    "pre_bwraps_cellsize";
-			m_BadParamNames.push_back(pair<string, string>("pre_bwraps2_space",       "pre_bwraps_space"));
+			m_BadParamNames["pre_bwraps2_space"] =       "pre_bwraps_space";
-			m_BadParamNames.push_back(pair<string, string>("pre_bwraps2_gain",        "pre_bwraps_gain"));
+			m_BadParamNames["pre_bwraps2_gain"] =        "pre_bwraps_gain";
-			m_BadParamNames.push_back(pair<string, string>("pre_bwraps2_inner_twist", "pre_bwraps_inner_twist"));
+			m_BadParamNames["pre_bwraps2_inner_twist"] = "pre_bwraps_inner_twist";
-			m_BadParamNames.push_back(pair<string, string>("pre_bwraps2_outer_twist", "pre_bwraps_outer_twist"));
+			m_BadParamNames["pre_bwraps2_outer_twist"] = "pre_bwraps_outer_twist";
-			m_BadParamNames.push_back(pair<string, string>("post_bwraps2_cellsize",    "post_bwraps_cellsize"));
+			m_BadParamNames["post_bwraps2_cellsize"] =    "post_bwraps_cellsize";
-			m_BadParamNames.push_back(pair<string, string>("post_bwraps2_space",       "post_bwraps_space"));
+			m_BadParamNames["post_bwraps2_space"] =       "post_bwraps_space";
-			m_BadParamNames.push_back(pair<string, string>("post_bwraps2_gain",        "post_bwraps_gain"));
+			m_BadParamNames["post_bwraps2_gain"] =        "post_bwraps_gain";
-			m_BadParamNames.push_back(pair<string, string>("post_bwraps2_inner_twist", "post_bwraps_inner_twist"));
+			m_BadParamNames["post_bwraps2_inner_twist"] = "post_bwraps_inner_twist";
-			m_BadParamNames.push_back(pair<string, string>("post_bwraps2_outer_twist", "post_bwraps_outer_twist"));
+			m_BadParamNames["post_bwraps2_outer_twist"] = "post_bwraps_outer_twist";
+			m_BadParamNames["hexa3d_majp"] =  "hexaplay3D_majp";
+			m_BadParamNames["hexa3d_scale"] = "hexaplay3D_scale";
+			m_BadParamNames["hexa3d_zlift"] = "hexaplay3D_zlift";
+			m_BadParamNames["nb_numedges"] =		"nBlur_numEdges";
+			m_BadParamNames["nb_numstripes"] =		"nBlur_numStripes";
+			m_BadParamNames["nb_ratiostripes"] =	"nBlur_ratioStripes";
+			m_BadParamNames["nb_ratiohole"] =		"nBlur_ratioHole";
+			m_BadParamNames["nb_circumcircle"] =	"nBlur_circumCircle";
+			m_BadParamNames["nb_adjusttolinear"] = "nBlur_adjustToLinear";
+			m_BadParamNames["nb_equalblur"] =		"nBlur_equalBlur";
+			m_BadParamNames["nb_exactcalc"] =		"nBlur_exactCalc";
+			m_BadParamNames["nb_highlightedges"] = "nBlur_highlightEdges";
+			m_BadParamNames["octapol_r"] = "octapol_radius";
+			m_BadParamNames["number_of_stripes"] = "bubbleT3D_number_of_stripes";
+			m_BadParamNames["ratio_of_stripes"]  = "bubbleT3D_ratio_of_stripes";
+			m_BadParamNames["angle_of_hole"]	 = "bubbleT3D_angle_of_hole";
+			m_BadParamNames["exponentZ"]		 = "bubbleT3D_exponentZ";
+			m_BadParamNames["_symmetryZ"]		 = "bubbleT3D_symmetryZ";
+			m_BadParamNames["_modusBlur"]		 = "bubbleT3D_modusBlur";
 
 			m_FlattenNames.reserve(24);
 			m_FlattenNames.push_back("pre_crop");
@@ -318,7 +337,7 @@ public:
 			for (size_t i = 1; i < emberSize; i++)
 			{
 				//Only do this adjustment if not in compat mode..
-				if (embers[i - 1].m_AffineInterp != INTERP_COMPAT && embers[i - 1].m_AffineInterp != INTERP_OLDER)
+				if (embers[i - 1].m_AffineInterp != AFFINE_INTERP_COMPAT && embers[i - 1].m_AffineInterp != AFFINE_INTERP_OLDER)
 				{
 					while (embers[i].m_Rotate < embers[i - 1].m_Rotate - 180)
 						embers[i].m_Rotate += 360;
@@ -524,7 +543,7 @@ private:
 			if		(ParseAndAssign(curAtt->name, attStr, "time",                  currentEmber.m_Time,                ret)) { }
 			else if (ParseAndAssign(curAtt->name, attStr, "scale",				   currentEmber.m_PixelsPerUnit,	   ret)) { currentEmber.m_OrigPixPerUnit = currentEmber.m_PixelsPerUnit; }
 			else if (ParseAndAssign(curAtt->name, attStr, "rotate",                currentEmber.m_Rotate,              ret)) { }
-			else if (ParseAndAssign(curAtt->name, attStr, "zoom",                  currentEmber.m_Zoom,                ret)) { }
+			else if (ParseAndAssign(curAtt->name, attStr, "zoom",				   currentEmber.m_Zoom,				   ret)) { ClampGteRef<T>(currentEmber.m_Zoom, 0); }
 			else if (ParseAndAssign(curAtt->name, attStr, "filter",                currentEmber.m_SpatialFilterRadius, ret)) { }
 			else if (ParseAndAssign(curAtt->name, attStr, "temporal_filter_width", currentEmber.m_TemporalFilterWidth, ret)) { }
 			else if (ParseAndAssign(curAtt->name, attStr, "temporal_filter_exp",   currentEmber.m_TemporalFilterExp,   ret)) { }
@@ -576,13 +595,13 @@ private:
 			else if (!Compare(curAtt->name, "interpolation_space") || !Compare(curAtt->name, "interpolation_type"))
 			{
 				if (!_stricmp("linear", attStr))
-					currentEmber.m_AffineInterp = INTERP_LINEAR;
+					currentEmber.m_AffineInterp = AFFINE_INTERP_LINEAR;
 				else if (!_stricmp("log", attStr))
-					currentEmber.m_AffineInterp = INTERP_LOG;
+					currentEmber.m_AffineInterp = AFFINE_INTERP_LOG;
 				else if (!_stricmp("old", attStr))
-					currentEmber.m_AffineInterp = INTERP_COMPAT;
+					currentEmber.m_AffineInterp = AFFINE_INTERP_COMPAT;
 				else if (!_stricmp("older", attStr))
-					currentEmber.m_AffineInterp = INTERP_OLDER;
+					currentEmber.m_AffineInterp = AFFINE_INTERP_OLDER;
 				else
 					m_ErrorReport.push_back(string(loc) + " : Unrecognized interpolation type " + string(attStr));
 			}
@@ -593,7 +612,7 @@ private:
 			}
 			else if (!Compare(curAtt->name, "version"))
 			{
-				if (ToLower(string(attStr)).find_first_of("ember") != string::npos)
+				if (ToLower(string(attStr)).find("ember") != string::npos)
 					fromEmber = true;
 			}
 			else if (!Compare(curAtt->name, "size"))
@@ -1050,8 +1069,8 @@ private:
 			}
 		}
 
-		//if (!newLinear)
+		if (!fromEmber && !newLinear)
-		//	currentEmber.Flatten(m_FlattenNames);
+			currentEmber.Flatten(m_FlattenNames);
 
 		for (i = 0; i < currentEmber.XformCount(); i++)
 			if (soloXform >= 0 && i != soloXform)
@@ -1342,16 +1361,17 @@ private:
 	/// Some Apophysis plugins use an inconsistent naming scheme for the parametric variation variables.
 	/// This function identifies and converts them to Ember's consistent naming convention.
 	/// </summary>
-	/// <param name="vec">The vector of corrected names to search</param>
+	/// <param name="names">The map of corrected names to search</param>
-	/// <param name="att">The current Xml node to check</param>
+	/// <param name="name">The current Xml node to check</param>
 	/// <returns>The corrected name if one was found, else the passed in name.</returns>
-	static string GetCorrectedParamName(vector<pair<string, string>>& vec, const char* name)
+	static string GetCorrectedParamName(const unordered_map<string, string>& names, const char* name)
 	{
-		for (auto& v : vec)
+		auto& newName = names.find(ToLower(name));
-			if (!_stricmp(v.first.c_str(), name))
-				return v.second;
 
-		return name;
+		if (newName != names.end())
+			return newName->second;
+		else
+			return name;
 	}
 
 	/// <summary>
@@ -1503,7 +1523,7 @@ private:
 	}
 
 	static bool m_Init;
-	static vector<pair<string, string>> m_BadParamNames;
+	static unordered_map<string, string> m_BadParamNames;
 	static vector<pair<pair<string, string>, vector<string>>> m_BadVariationNames;
 	VariationList<T> m_VariationList;//The variation list used to make copies of variations to populate the embers with.
 	PaletteList<T> m_PaletteList;

Source/EmberAnimate/EmberAnimate.cpp

@@ -29,7 +29,7 @@ bool EmberAnimate(EmberOptions& opt)
 	Timing t;
 	bool unsorted = false;
 	uint channels, padding;
-	size_t i;
+	size_t i, firstUnsortedIndex = 0;
 	string inputPath = GetPath(opt.Input());
 	vector<Ember<T>> embers;
 	XmlToEmber<T> parser;
@@ -128,6 +128,12 @@ bool EmberAnimate(EmberOptions& opt)
 	if (!ParseEmberFile(parser, opt.Input(), embers))
 		return false;
 
+	if (embers.size() <= 1)
+	{
+		cout << "Read " << embers.size() << " embers from file. At least 2 required to animate, exiting." << endl;
+		return false;
+	}
+
 	if (opt.Format() != "jpg" &&
 		opt.Format() != "png" &&
 		opt.Format() != "ppm" &&
@@ -206,7 +212,18 @@ bool EmberAnimate(EmberOptions& opt)
 	for (i = 0; i < embers.size(); i++)
 	{
 		if (i > 0 && embers[i].m_Time <= embers[i - 1].m_Time)
+		{
+			if (!unsorted)
+				firstUnsortedIndex = i;
+
 			unsorted = true;
+		}
+
+		if (i > 0 && embers[i].m_Time == embers[i - 1].m_Time)
+		{
+			cout << "Image " << i << " time of " << embers[i].m_Time << " equaled previous image time of " << embers[i - 1].m_Time << ". Adjusting up by 1." << endl;
+			embers[i].m_Time++;
+		}
 
 		if (opt.Supersample() > 0)
 			embers[i].m_Supersample = opt.Supersample();
@@ -215,8 +232,8 @@ bool EmberAnimate(EmberOptions& opt)
 			embers[i].m_SubBatchSize = opt.SubBatchSize();
 
 		embers[i].m_Quality *= T(opt.QualityScale());
-		embers[i].m_FinalRasW = uint(T(embers[i].m_FinalRasW) * opt.SizeScale());
+		embers[i].m_FinalRasW = size_t(T(embers[i].m_FinalRasW) * opt.SizeScale());
-		embers[i].m_FinalRasH = uint(T(embers[i].m_FinalRasH) * opt.SizeScale());
+		embers[i].m_FinalRasH = size_t(T(embers[i].m_FinalRasH) * opt.SizeScale());
 		embers[i].m_PixelsPerUnit *= T(opt.SizeScale());
 
 		//Cast to double in case the value exceeds 2^32.
@@ -251,17 +268,18 @@ bool EmberAnimate(EmberOptions& opt)
 
 	if (unsorted)
 	{
-		cout << "Embers were unsorted by time. First out of order index was " << i << ". Sorting." << endl;
+		cout << "Embers were unsorted by time. First out of order index was " << firstUnsortedIndex << ". Sorting." << endl;
 		std::sort(embers.begin(), embers.end(), &CompareEmbers<T>);
 	}
 
 	if (!opt.Time() && !opt.Frame())
 	{
 		if (opt.FirstFrame() == UINT_MAX)
-			opt.FirstFrame(int(embers[0].m_Time));
+			opt.FirstFrame(size_t(embers[0].m_Time));
 
 		if (opt.LastFrame() == UINT_MAX)
-			opt.LastFrame(ClampGte<size_t>(size_t(embers.back().m_Time - 1), opt.FirstFrame()));
+			opt.LastFrame(ClampGte<size_t>(size_t(embers.back().m_Time),//Make sure time - 1 is positive before converting to size_t.
+				opt.FirstFrame() + opt.Dtime()));//Make sure the final value is at least first frame + dtime.
 	}
 
 	if (!opt.Out().empty())
@@ -271,7 +289,7 @@ bool EmberAnimate(EmberOptions& opt)
 	}
 
 	//Final setup steps before running.
-	padding = uint(log10(double(embers.size()))) + 1;
+	padding = uint(std::log10(double(embers.size()))) + 1;
 
 	for (auto& r : renderers)
 	{
@@ -284,7 +302,7 @@ bool EmberAnimate(EmberOptions& opt)
 		r->Transparency(opt.Transparency());
 		r->NumChannels(channels);
 		r->BytesPerChannel(opt.BitsPerChannel() / 8);
-		r->Priority(eThreadPriority(Clamp<int>(int(opt.Priority()), int(eThreadPriority::LOWEST), int(eThreadPriority::HIGHEST))));
+		r->Priority(eThreadPriority(Clamp<intmax_t>(intmax_t(opt.Priority()), intmax_t(eThreadPriority::LOWEST), intmax_t(eThreadPriority::HIGHEST))));
 	}
 
 	std::function<void (vector<byte>&, string, EmberImageComments, size_t, size_t, size_t)> saveFunc = [&](vector<byte>& finalImage,

Source/EmberCL/DEOpenCLKernelCreator.cpp

@@ -787,7 +787,8 @@ string DEOpenCLKernelCreator::CreateGaussianDEKernelNoLocalCache(size_t ss)
 		"\n";
 
 		if (doScf)
-		os << "		filterSelect *= scfact;\n";
+		os <<
+		"			filterSelect *= scfact;\n";
 	}
 	else
 	{
@@ -830,7 +831,7 @@ string DEOpenCLKernelCreator::CreateGaussianDEKernelNoLocalCache(size_t ss)
 		"			}\n"
 		"		}\n"//bucket.w != 0.
 		"	}\n"//In bounds.
-		"\n"
+		//"\n"
 		//"	barrier(CLK_GLOBAL_MEM_FENCE);\n"//Just to be safe.
 		"}\n";
 

Source/EmberCL/EmberCLFunctions.h

@@ -149,8 +149,8 @@ static const char* CalcAlphaFunctionString =
 static const char* CurveAdjustFunctionString =
 "static inline void CurveAdjust(__constant real4reals_bucket* csa, float* a, uint index)\n"
 "{\n"
-"	uint tempIndex = (uint)Clamp(*a, 0.0, (float)COLORMAP_LENGTH_MINUS_1);\n"
+"	uint tempIndex = (uint)clamp(*a, (float)0.0, (float)COLORMAP_LENGTH_MINUS_1);\n"
-"	uint tempIndex2 = (uint)Clamp(csa[tempIndex].m_Real4.x, 0.0, (real_t)COLORMAP_LENGTH_MINUS_1);\n"
+"	uint tempIndex2 = (uint)clamp((float)csa[tempIndex].m_Real4.x, (float)0.0, (float)COLORMAP_LENGTH_MINUS_1);\n"
 "\n"
 "	*a = (float)round(csa[tempIndex2].m_Reals[index]);\n"
 "}\n";
@@ -194,135 +194,6 @@ static const char* RandFunctionString =
 	"}\n"
 	"\n";
 
-/// <summary>
-/// OpenCL equivalent of the global ClampRef().
-/// </summary>
-static const char* ClampRealFunctionString =
-	"inline real_t Clamp(real_t val, real_t min, real_t max)\n"
-	"{\n"
-	"	if (val < min)\n"
-	"		return min;\n"
-	"	else if (val > max)\n"
-	"		return max;\n"
-	"	else\n"
-	"		return val;\n"
-	"}\n"
-	"\n"
-	"inline void ClampRef(real_t* val, real_t min, real_t max)\n"
-	"{\n"
-	"	if (*val < min)\n"
-	"		*val = min;\n"
-	"	else if (*val > max)\n"
-	"		*val = max;\n"
-	"}\n"
-	"\n"
-	"inline real_t ClampGte(real_t val, real_t gte)\n"
-	"{\n"
-	"	return (val < gte) ? gte : val;\n"
-	"}\n"
-	"\n";
-
-/// <summary>
-/// OpenCL equivalent of the global LRint().
-/// </summary>
-static const char* InlineMathFunctionsString = 
-	"inline real_t LRint(real_t x)\n"
-	"{\n"
-	"    intPrec temp = (x >= 0.0 ? (intPrec)(x + 0.5) : (intPrec)(x - 0.5));\n"
-	"    return (real_t)temp;\n"
-	"}\n"
-	"\n"
-	"inline real_t Round(real_t r)\n"
-	"{\n"
-	"	return (r > 0.0) ? floor(r + 0.5) : ceil(r - 0.5);\n"
-	"}\n"
-	"\n"
-	"inline real_t Sign(real_t v)\n"
-	"{\n"
-	"	return (v < 0.0) ? -1 : (v > 0.0) ? 1 : 0.0;\n"
-	"}\n"
-	"\n"
-	"inline real_t SignNz(real_t v)\n"
-	"{\n"
-	"	return (v < 0.0) ? -1.0 : 1.0;\n"
-	"}\n"
-	"\n"
-	"inline real_t Sqr(real_t v)\n"
-	"{\n"
-	"	return v * v;\n"
-	"}\n"
-	"\n"
-	"inline real_t SafeSqrt(real_t x)\n"
-	"{\n"
-	"	if (x <= 0.0)\n"
-	"		return 0.0;\n"
-	"\n"
-	"	return sqrt(x);\n"
-	"}\n"
-	"\n"
-	"inline real_t Cube(real_t v)\n"
-	"{\n"
-	"	return v * v * v;\n"
-	"}\n"
-	"\n"
-	"inline real_t Hypot(real_t x, real_t y)\n"
-	"{\n"
-	"	return sqrt(SQR(x) + SQR(y));\n"
-	"}\n"
-	"\n"
-	"inline real_t Spread(real_t x, real_t y)\n"
-	"{\n"
-	"	return Hypot(x, y) * ((x) > 0.0 ? 1.0 : -1.0);\n"
-	"}\n"
-	"\n"
-	"inline real_t Powq4(real_t x, real_t y)\n"
-	"{\n"
-	"	return pow(fabs(x), y) * SignNz(x);\n"
-	"}\n"
-	"\n"
-	"inline real_t Powq4c(real_t x, real_t y)\n"
-	"{\n"
-	"	return y == 1.0 ? x : Powq4(x, y);\n"
-	"}\n"
-	"\n"
-	"inline real_t Zeps(real_t x)\n"
-	"{\n"
-	"	return x == 0.0 ? EPS : x;\n"
-	"}\n"
-	"\n"
-	"inline real_t Lerp(real_t a, real_t b, real_t p)\n"
-	"{\n"
-	"	return a + (b - a) * p;\n"
-	"}\n"
-	"\n"
-	"inline real_t Fabsmod(real_t v)\n"
-	"{\n"
-	"	real_t dummy;\n"
-	"\n"
-	"	return modf(v, &dummy);\n"
-	"}\n"
-	"\n"
-	"inline real_t Fosc(real_t p, real_t amp, real_t ph)\n"
-	"{\n"
-	"	return 0.5 - cos(p * amp + ph) * 0.5;\n"
-	"}\n"
-	"\n"
-	"inline real_t Foscn(real_t p, real_t ph)\n"
-	"{\n"
-	"	return 0.5 - cos(p + ph) * 0.5;\n"
-	"}\n"
-	"\n"
-	"inline real_t LogScale(real_t x)\n"
-	"{\n"
-	"	return x == 0.0 ? 0.0 : log((fabs(x) + 1) * M_E) * SignNz(x) / M_E;\n"
-	"}\n"
-	"\n"
-	"inline real_t LogMap(real_t x)\n"
-	"{\n"
-	"	return x == 0.0 ? 0.0 : (M_E + log(x * M_E)) * 0.25 * SignNz(x);\n"
-	"}\n"
-	"\n";
-
 /// <summary>
 /// OpenCL equivalent Renderer::AddToAccum().
 /// </summary>

Source/EmberCL/EmberCLPch.h

@@ -44,9 +44,11 @@
 #include <cstdlib>
 #include <fstream>
 #include <iostream>
+#include <set>
 #include <string>
 #include <iterator>
 #include <time.h>
+#include <unordered_map>
 
 #ifdef _WIN32
 	#if defined(BUILDING_EMBERCL)

Source/EmberCL/EmberCLStructs.h

@@ -13,15 +13,6 @@
 
 namespace EmberCLns
 {
-//These two must always match.
-#ifdef WIN32
-	#define ALIGN __declspec(align(16))
-#else
-	#define ALIGN __attribute__ ((aligned (16)))
-#endif
-
-#define ALIGN_CL "((aligned (16)))"//The extra parens are necessary.
-
 /// <summary>
 /// Various constants needed for rendering.
 /// </summary>
@@ -32,20 +23,23 @@ static string ConstantDefinesString(bool doublePrecision)
 	if (doublePrecision)
 	{
 		os << "#if defined(cl_amd_fp64)\n"//AMD extension available?
-		   << "	#pragma OPENCL EXTENSION cl_amd_fp64 : enable\n"
+			  "	#pragma OPENCL EXTENSION cl_amd_fp64 : enable\n"
-		   << "#endif\n"
+			  "#endif\n"
-		   << "#if defined(cl_khr_fp64)\n"//Khronos extension available?
+			  "#if defined(cl_khr_fp64)\n"//Khronos extension available?
-		   << "	#pragma OPENCL EXTENSION cl_khr_fp64 : enable\n"
+			  "	#pragma OPENCL EXTENSION cl_khr_fp64 : enable\n"
-		   << "#endif\n"
+			  "#endif\n"
-		   << "#pragma OPENCL EXTENSION cl_khr_int64_base_atomics : enable\n"//Only supported on nVidia.
+			  "#pragma OPENCL EXTENSION cl_khr_int64_base_atomics : enable\n"//Only supported on nVidia.
-		   << "typedef long intPrec;\n"
+			  "typedef long intPrec;\n"
-		   << "typedef uint atomi;\n"//Same size as real_bucket_t, always 4 bytes.
+			  "typedef uint atomi;\n"//Same size as real_bucket_t, always 4 bytes.
-		   << "typedef double real_t;\n"
+			  "typedef double real_t;\n"
-		   << "typedef float real_bucket_t;\n"//Assume buckets are always float, even though iter calcs are in double.
+			  "typedef float real_bucket_t;\n"//Assume buckets are always float, even though iter calcs are in double.
-		   << "typedef double4 real4;\n"
+			  "typedef double2 real2;\n"
-		   << "typedef float4 real4_bucket;\n"//And here too.
+			  "typedef double4 real4;\n"
-		   << "#define EPS (DBL_EPSILON)\n"
+			  "typedef float4 real4_bucket;\n"//And here too.
-		   ;
+			  "#define EPS (DBL_EPSILON)\n"
+			  "#define TLOW (DBL_MIN)\n"
+			  "#define TMAX (DBL_MAX)\n"
+			  ;
 	}
 	else
 	{
@@ -53,9 +47,12 @@ static string ConstantDefinesString(bool doublePrecision)
 			  "typedef uint atomi;\n"
 			  "typedef float real_t;\n"
 			  "typedef float real_bucket_t;\n"
+			  "typedef float2 real2;\n"
 			  "typedef float4 real4;\n"
 			  "typedef float4 real4_bucket;\n"
 			  "#define EPS (FLT_EPSILON)\n"
+			  "#define TLOW (FLT_MIN)\n"
+			  "#define TMAX (FLT_MAX)\n"
 			  ;
 	}
 

Source/EmberCL/FinalAccumOpenCLKernelCreator.cpp

@@ -186,7 +186,6 @@ string FinalAccumOpenCLKernelCreator::CreateFinalAccumKernelString(bool earlyCli
 
 	os <<
 		ConstantDefinesString(m_DoublePrecision) <<
-		ClampRealFunctionString <<
 		UnionCLStructString <<
 		RgbToHsvFunctionString <<
 		HsvToRgbFunctionString <<
@@ -484,7 +483,6 @@ string FinalAccumOpenCLKernelCreator::CreateGammaCorrectionKernelString(bool alp
 
 	os <<
 		ConstantDefinesString(m_DoublePrecision) <<
-		ClampRealFunctionString <<
 		UnionCLStructString <<
 		RgbToHsvFunctionString <<
 		HsvToRgbFunctionString <<

Source/EmberCL/FunctionMapper.cpp

@@ -0,0 +1,151 @@
+#include "EmberCLPch.h"
+#include "FunctionMapper.h"
+
+namespace EmberCLns
+{
+std::unordered_map<string, string> FunctionMapper::m_GlobalMap;
+
+FunctionMapper::FunctionMapper()
+{
+	if (m_GlobalMap.empty())
+	{
+		m_GlobalMap["LRint"] =
+			"inline real_t LRint(real_t x)\n"
+			"{\n"
+			"    intPrec temp = (x >= 0.0 ? (intPrec)(x + 0.5) : (intPrec)(x - 0.5));\n"
+			"    return (real_t)temp;\n"
+			"}\n";
+
+		m_GlobalMap["Round"] =
+			"inline real_t Round(real_t r)\n"
+			"{\n"
+			"	return (r > 0.0) ? floor(r + 0.5) : ceil(r - 0.5);\n"
+			"}\n";
+			
+		m_GlobalMap["Sign"] =
+			"inline real_t Sign(real_t v)\n"
+			"{\n"
+			"	return (v < 0.0) ? -1 : (v > 0.0) ? 1 : 0.0;\n"
+			"}\n";
+
+		m_GlobalMap["SignNz"] =
+			"inline real_t SignNz(real_t v)\n"
+			"{\n"
+			"	return (v < 0.0) ? -1.0 : 1.0;\n"
+			"}\n";
+
+		m_GlobalMap["Sqr"] =
+			"inline real_t Sqr(real_t v)\n"
+			"{\n"
+			"	return v * v;\n"
+			"}\n";
+
+		m_GlobalMap["SafeSqrt"] =
+			"inline real_t SafeSqrt(real_t x)\n"
+			"{\n"
+			"	if (x <= 0.0)\n"
+			"		return 0.0;\n"
+			"\n"
+			"	return sqrt(x);\n"
+			"}\n";
+
+		m_GlobalMap["Cube"] =
+			"inline real_t Cube(real_t v)\n"
+			"{\n"
+			"	return v * v * v;\n"
+			"}\n";
+
+		m_GlobalMap["Hypot"] =
+			"inline real_t Hypot(real_t x, real_t y)\n"
+			"{\n"
+			"	return sqrt(SQR(x) + SQR(y));\n"
+			"}\n";
+
+		m_GlobalMap["Spread"] =
+			"inline real_t Spread(real_t x, real_t y)\n"
+			"{\n"
+			"	return Hypot(x, y) * ((x) > 0.0 ? 1.0 : -1.0);\n"
+			"}\n";
+
+		m_GlobalMap["Powq4"] =
+			"inline real_t Powq4(real_t x, real_t y)\n"
+			"{\n"
+			"	return pow(fabs(x), y) * SignNz(x);\n"
+			"}\n";
+
+		m_GlobalMap["Powq4c"] =
+			"inline real_t Powq4c(real_t x, real_t y)\n"
+			"{\n"
+			"	return y == 1.0 ? x : Powq4(x, y);\n"
+			"}\n";
+
+		m_GlobalMap["Zeps"] =
+			"inline real_t Zeps(real_t x)\n"
+			"{\n"
+			"	return x == 0.0 ? EPS : x;\n"
+			"}\n";
+
+		m_GlobalMap["Lerp"] =
+			"inline real_t Lerp(real_t a, real_t b, real_t p)\n"
+			"{\n"
+			"	return a + (b - a) * p;\n"
+			"}\n";
+
+		m_GlobalMap["Fabsmod"] =
+			"inline real_t Fabsmod(real_t v)\n"
+			"{\n"
+			"	real_t dummy;\n"
+			"\n"
+			"	return modf(v, &dummy);\n"
+			"}\n";
+
+		m_GlobalMap["Fosc"] =
+			"inline real_t Fosc(real_t p, real_t amp, real_t ph)\n"
+			"{\n"
+			"	return 0.5 - cos(p * amp + ph) * 0.5;\n"
+			"}\n";
+
+		m_GlobalMap["Foscn"] =
+			"inline real_t Foscn(real_t p, real_t ph)\n"
+			"{\n"
+			"	return 0.5 - cos(p + ph) * 0.5;\n"
+			"}\n";
+
+		m_GlobalMap["LogScale"] =
+			"inline real_t LogScale(real_t x)\n"
+			"{\n"
+			"	return x == 0.0 ? 0.0 : log((fabs(x) + 1) * M_E) * SignNz(x) / M_E;\n"
+			"}\n";
+
+		m_GlobalMap["LogMap"] =
+			"inline real_t LogMap(real_t x)\n"
+			"{\n"
+			"	return x == 0.0 ? 0.0 : (M_E + log(x * M_E)) * 0.25 * SignNz(x);\n"
+			"}\n";
+
+		m_GlobalMap["ClampGte"] =
+			"inline real_t ClampGte(real_t val, real_t gte)\n"
+			"{\n"
+			"	return (val < gte) ? gte : val;\n"
+			"}\n";
+
+		m_GlobalMap["Swap"] =
+			"inline void Swap(real_t* val1, real_t* val2)\n"
+			"{\n"
+			"	real_t tmp = *val1;\n"
+			"	*val1 = *val2;\n"
+			"	*val2 = tmp;\n"
+			"}\n";
+	}
+}
+
+const string* FunctionMapper::GetGlobalFunc(const string& func)
+{
+	auto& text = m_GlobalMap.find(func);
+	
+	if (text != m_GlobalMap.end())
+		return &text->second;
+	else
+		return nullptr;
+}
+}

Source/EmberCL/FunctionMapper.h

@@ -0,0 +1,21 @@
+#pragma once
+
+#include "EmberCLPch.h"
+
+namespace EmberCLns
+{
+/// <summary>
+/// Functionality to map OpenCL function names to their full function body program strings.
+/// This is used to ensure only the functions that are needed by a program are included once
+/// in the program string.
+/// </summary>
+class EMBERCL_API FunctionMapper
+{
+public:
+	FunctionMapper();
+	static const string* GetGlobalFunc(const string& func);
+
+private:
+	static std::unordered_map<string, string> m_GlobalMap;
+};
+}

Source/EmberCL/IterOpenCLKernelCreator.cpp

@@ -2,7 +2,6 @@
 #include "IterOpenCLKernelCreator.h"
 
 //#define STRAIGHT_RAND 1
-#define USE_CASE 1
 
 namespace EmberCLns
 {
@@ -39,16 +38,20 @@ template <typename T> const string& IterOpenCLKernelCreator<T>::IterEntryPoint()
 /// <param name="doAccum">Debugging parameter to include or omit accumulating to the histogram. Default: true.</param>
 /// <returns>The kernel string</returns>
 template <typename T>
-string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, string& parVarDefines, bool lockAccum, bool doAccum)
+string IterOpenCLKernelCreator<T>::CreateIterKernelString(const Ember<T>& ember, string& parVarDefines, bool lockAccum, bool doAccum)
 {
 	bool doublePrecision = typeid(T) == typeid(double);
 	size_t i, v, varIndex, varCount, totalXformCount = ember.TotalXformCount();
 	ostringstream kernelIterBody, xformFuncs, os;
 	vector<Variation<T>*> variations;
 
-	xformFuncs << "\n" << parVarDefines << endl;
+	xformFuncs << VariationStateString(ember);
+	xformFuncs << parVarDefines << endl;
 	ember.GetPresentVariations(variations);
-	for (auto var : variations) if (var) xformFuncs << var->OpenCLFuncsString();
+
+	for (auto var : variations)
+		if (var)
+			xformFuncs << var->OpenCLFuncsString();
 
 	for (i = 0; i < totalXformCount; i++)
 	{
@@ -62,7 +65,7 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
 
 		v = varIndex = varCount = 0;
 		xformFuncs <<
-			"void Xform" << i << "(__constant XformCL* xform, __constant real_t* parVars, Point* inPoint, Point* outPoint, uint2* mwc)\n" <<
+			"void Xform" << i << "(__constant XformCL* xform, __constant real_t* parVars, Point* inPoint, Point* outPoint, uint2* mwc, VariationState* varState)\n" <<
 			"{\n"
 			"	real_t transX, transY, transZ;\n"
 			"	real4 vIn, vOut = 0.0;\n";
@@ -215,8 +218,7 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
 
 	os <<
 		ConstantDefinesString(doublePrecision) <<
-		InlineMathFunctionsString <<
+		GlobalFunctionsString(ember) <<
-		ClampRealFunctionString <<
 		RandFunctionString <<
 		PointCLStructString <<
 		XformCLStructString <<
@@ -237,12 +239,13 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
 		"	__constant EmberCL* ember,\n"
 		"	__constant XformCL* xforms,\n"
 		"	__constant real_t* parVars,\n"
-		"	__global uchar* xformDistributions,\n"//Using uchar is quicker than uint. Can't be constant because the size can be too large to fit when using xaos.//FINALOPT
+		"	__global uchar* xformDistributions,\n"//Using uchar is quicker than uint. Can't be constant because the size can be too large to fit when using xaos.
 		"	__constant CarToRasCL* carToRas,\n"
 		"	__global real4reals_bucket* histogram,\n"
 		"	uint histSize,\n"
 		"	__read_only image2d_t palette,\n"
 		"	__global Point* points\n"
+		//"	uint startRender\n"
 		"\t)\n"
 		"{\n"
 		"	bool fuse, ok;\n"
@@ -261,8 +264,9 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
 		"		CLK_ADDRESS_CLAMP_TO_EDGE |\n"//Clamp to edge
 		"		CLK_FILTER_NEAREST;\n"//Don't interpolate
 		"	uint threadXY = (THREAD_ID_X + THREAD_ID_Y);\n"
-		"	uint threadXDivRows = (THREAD_ID_X / (NTHREADS / THREADS_PER_WARP));\n"
+		"	uint threadXDivRows = (THREAD_ID_X / NWARPS);\n"
 		"	uint threadsMinus1 = NTHREADS - 1;\n"
+		"	VariationState varState;\n"
 		;
 
 	os <<
@@ -278,6 +282,9 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
 		"	{\n"
 		"		fuse = true;\n"
 		"		itersToDo = fuseCount;\n"
+		//Calling MwcNextNeg1Pos1() twice is deliberate. The first call to mwc is not very random since it just does
+		//an xor. So it must be called twice to get it in a good random state.
+		"		firstPoint.m_X = MwcNextNeg1Pos1(&mwc);\n"
 		"		firstPoint.m_X = MwcNextNeg1Pos1(&mwc);\n"
 		"		firstPoint.m_Y = MwcNextNeg1Pos1(&mwc);\n"
 		"		firstPoint.m_Z = 0.0;\n"
@@ -290,7 +297,13 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
 		"		itersToDo = iterCount;\n"
 		"		firstPoint = points[pointsIndex];\n"
 		"	}\n"
-		"\n";
+		"\n"
+		;
+
+	auto varStateString = VariationStateInitString(ember);
+
+	if (!varStateString.empty())
+		os << varStateString << "\n\n";
 
 		//This is done once initially here and then again after each swap-sync in the main loop.
 		//This along with the randomness that the point shuffle provides gives sufficient randomness
@@ -336,7 +349,6 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
 
 		for (i = 0; i < ember.XformCount(); i++)
 		{
-#ifdef USE_CASE
 			if (i == 0)
 			{
 			os <<
@@ -347,7 +359,7 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
 			os <<
 		"				case " << i << ":\n"
 		"				{\n" <<
-		"					Xform" << i << "(&(xforms[" << i << "]), parVars, &firstPoint, &secondPoint, &mwc);\n" <<
+		"					Xform" << i << "(&(xforms[" << i << "]), parVars, &firstPoint, &secondPoint, &mwc, &varState);\n" <<
 		"					break;\n"
 		"				}\n";
 
@@ -356,19 +368,6 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
 			os <<
 		"			}\n";
 			}
-#else
-			if (i == 0)
-				os <<
-		"			if (secondPoint.m_LastXfUsed == " << i << ")\n";
-			else
-				os <<
-		"			else if (secondPoint.m_LastXfUsed == " << i << ")\n";
-
-		os <<
-		"			{\n" <<
-		"				Xform" << i << "(&(xforms[" << i << "]), parVars, &firstPoint, &secondPoint, &mwc);\n" <<
-		"			}\n";
-#endif
 		}
 
 		os <<
@@ -438,7 +437,7 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
 			"		if ((xforms[" << finalIndex << "].m_Opacity == 1) || (MwcNext01(&mwc) < xforms[" << finalIndex << "].m_Opacity))\n"
 			"		{\n"
 			"			tempPoint.m_LastXfUsed = secondPoint.m_LastXfUsed;\n"
-			"			Xform" << finalIndex << "(&(xforms[" << finalIndex << "]), parVars, &secondPoint, &tempPoint, &mwc);\n"
+			"			Xform" << finalIndex << "(&(xforms[" << finalIndex << "]), parVars, &secondPoint, &tempPoint, &mwc, &varState);\n"
 			"			secondPoint = tempPoint;\n"
 			"		}\n"
 			"\n";
@@ -543,6 +542,50 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
 	return os.str();
 }
 
+/// <summary>
+/// Return a string containing all of the global functions needed by the passed in ember.
+/// </summary>
+/// <param name="ember">The ember to create the global function strings from</param>
+/// <returns>String of all global function names and bodies</returns>
+template <typename T>
+string IterOpenCLKernelCreator<T>::GlobalFunctionsString(const Ember<T>& ember)
+{
+	size_t i, j, xformCount = ember.TotalXformCount();
+	vector<string> funcNames;//Can't use a set here because they sort and we must preserve the insertion order due to nested function calls.
+	ostringstream os;
+	static string zeps = "Zeps";
+
+	for (i = 0; i < xformCount; i++)
+	{
+		if (auto xform = ember.GetTotalXform(i))
+		{
+			size_t varCount = xform->TotalVariationCount();
+
+			if (xform->NeedPrecalcAngles())
+				if (!Contains(funcNames, zeps))
+					funcNames.push_back(zeps);
+
+			for (j = 0; j < varCount; j++)
+			{
+				if (auto var = xform->GetVariation(j))
+				{
+					auto names = var->OpenCLGlobalFuncNames();
+
+					for (auto& name : names)
+						if (!Contains(funcNames, name))
+							funcNames.push_back(name);
+				}
+			}
+		}
+	}
+
+	for (auto& funcName : funcNames)
+		if(auto text = m_FunctionMapper.GetGlobalFunc(funcName))
+			os << *text << endl;
+
+	return os.str();
+}
+
 /// <summary>
 /// Create an OpenCL string of #defines and a corresponding host side vector for parametric variation values.
 /// Parametric variations present a special problem in the iteration code.
@@ -571,9 +614,9 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
 /// #define CURL_C2_2 3
 /// #define BLOB_LOW_3 4
 /// #define BLOB_HIGH_3 5
-/// #define BLOB_WAVES_ 6
+/// #define BLOB_WAVES_3 6
 ///
-/// The variations the use these #defines by first looking up the index of the
+/// The variations use these #defines by first looking up the index of the
 /// xform they belong to in the parent ember and generating the OpenCL string based on that
 /// in their overridden OpenCLString() functions.
 /// Template argument expected to be float or double.
@@ -583,10 +626,9 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
 /// <param name="doVals">True if the vector should be populated, else false. Default: true.</param>
 /// <param name="doString">True if the string should be populated, else false. Default: true.</param>
 template <typename T>
-void IterOpenCLKernelCreator<T>::ParVarIndexDefines(Ember<T>& ember, pair<string, vector<T>>& params, bool doVals, bool doString)
+void IterOpenCLKernelCreator<T>::ParVarIndexDefines(const Ember<T>& ember, pair<string, vector<T>>& params, bool doVals, bool doString)
 {
 	size_t i, j, k, size = 0, xformCount = ember.TotalXformCount();
-	Xform<T>* xform;
 	ostringstream os;
 
 	if (doVals)
@@ -594,23 +636,26 @@ void IterOpenCLKernelCreator<T>::ParVarIndexDefines(Ember<T>& ember, pair<string
 
 	for (i = 0; i < xformCount; i++)
 	{
-		if ((xform = ember.GetTotalXform(i)))
+		if (auto xform = ember.GetTotalXform(i))
 		{
 			size_t varCount = xform->TotalVariationCount();
 
 			for (j = 0; j < varCount; j++)
 			{
-				if (ParametricVariation<T>* parVar = dynamic_cast<ParametricVariation<T>*>(xform->GetVariation(j)))
+				if (auto parVar = dynamic_cast<ParametricVariation<T>*>(xform->GetVariation(j)))
 				{
 					for (k = 0; k < parVar->ParamCount(); k++)
 					{
-						if (doString)
+						if (!parVar->Params()[k].IsState())
-							os << "#define " << ToUpper(parVar->Params()[k].Name()) << "_" << i << " " << size << endl;//Uniquely identify this param in this variation in this xform.
+						{
+							if (doString)
+								os << "#define " << ToUpper(parVar->Params()[k].Name()) << "_" << i << " " << size << endl;//Uniquely identify this param in this variation in this xform.
 
-						if (doVals)
+							if (doVals)
-							params.second.push_back(parVar->Params()[k].ParamVal());
+								params.second.push_back(parVar->Params()[k].ParamVal());
 
-						size++;
+							size++;
+						}
 					}
 				}
 			}
@@ -624,6 +669,69 @@ void IterOpenCLKernelCreator<T>::ParVarIndexDefines(Ember<T>& ember, pair<string
 	}
 }
 
+/// <summary>
+/// Create the string needed for the struct whose values will change between each iteration.
+/// This is only needed for variations whose state changes.
+/// If none are present, the struct will be empty.
+/// </summary>
+/// <param name="ember">The ember to generate the variation state struct string for</param>
+/// <returns>The variation state struct string</returns>
+template <typename T>
+string IterOpenCLKernelCreator<T>::VariationStateString(const Ember<T>& ember)
+{
+	ostringstream os;
+
+	os << "typedef struct __attribute__ " ALIGN_CL " _VariationState\n{";
+
+	for (size_t i = 0; i < ember.TotalXformCount(); i++)
+	{
+		if (auto xform = ember.GetTotalXform(i))
+		{
+			for (size_t j = 0; j < xform->TotalVariationCount(); j++)
+			{
+				if (auto var = xform->GetVariation(j))
+				{
+					os << var->StateOpenCLString();
+				}
+			}
+		}
+	}
+
+	os << "\n} VariationState;\n\n";
+
+	return os.str();
+}
+
+/// <summary>
+/// Create the string needed for the initial state of the struct whose values will change between each iteration.
+/// This is only needed for variations whose state changes.
+/// If none are present, the returned init string will be empty.
+/// This will be called at the beginning of each kernel.
+/// </summary>
+/// <param name="ember">The ember to generate the variation state struct init string for</param>
+/// <returns>The variation state struct init string</returns>
+template <typename T>
+string IterOpenCLKernelCreator<T>::VariationStateInitString(const Ember<T>& ember)
+{
+	ostringstream os;
+
+	for (size_t i = 0; i < ember.TotalXformCount(); i++)
+	{
+		if (auto xform = ember.GetTotalXform(i))
+		{
+			for (size_t j = 0; j < xform->TotalVariationCount(); j++)
+			{
+				if (auto var = xform->GetVariation(j))
+				{
+					os << var->StateInitOpenCLString();
+				}
+			}
+		}
+	}
+
+	return os.str();
+}
+
 /// <summary>
 /// Determine whether the two embers passed in differ enough
 /// to require a rebuild of the iteration code.
@@ -641,7 +749,7 @@ void IterOpenCLKernelCreator<T>::ParVarIndexDefines(Ember<T>& ember, pair<string
 /// <param name="ember2">The second ember to compare</param>
 /// <returns>True if a rebuild is required, else false</returns>
 template <typename T>
-bool IterOpenCLKernelCreator<T>::IsBuildRequired(Ember<T>& ember1, Ember<T>& ember2)
+bool IterOpenCLKernelCreator<T>::IsBuildRequired(const Ember<T>& ember1, const Ember<T>& ember2)
 {
 	size_t i, j, xformCount = ember1.TotalXformCount();
 
@@ -688,7 +796,7 @@ bool IterOpenCLKernelCreator<T>::IsBuildRequired(Ember<T>& ember1, Ember<T>& emb
 /// </summary>
 /// <returns>The kernel string</returns>
 template <typename T>
-string IterOpenCLKernelCreator<T>::CreateZeroizeKernelString()
+string IterOpenCLKernelCreator<T>::CreateZeroizeKernelString() const
 {
 	ostringstream os;
 
@@ -707,8 +815,16 @@ string IterOpenCLKernelCreator<T>::CreateZeroizeKernelString()
 	return os.str();
 }
 
+/// <summary>
+/// Create the histogram summing kernel string.
+/// This is used when running with multiple GPUs. It takes
+/// two histograms present on a single device, source and dest,
+/// and adds the values of source to dest.
+/// It optionally sets all values of source to zero.
+/// </summary>
+/// <returns>The kernel string</returns>
 template <typename T>
-string IterOpenCLKernelCreator<T>::CreateSumHistKernelString()
+string IterOpenCLKernelCreator<T>::CreateSumHistKernelString() const
 {
 	ostringstream os;
 
@@ -739,7 +855,7 @@ string IterOpenCLKernelCreator<T>::CreateSumHistKernelString()
 /// <param name="ember">The ember to create the projection string for</param>
 /// <returns>The kernel string</returns>
 template <typename T>
-string IterOpenCLKernelCreator<T>::CreateProjectionString(Ember<T>& ember)
+string IterOpenCLKernelCreator<T>::CreateProjectionString(const Ember<T>& ember) const
 {
 	size_t projBits = ember.ProjBits();
 	ostringstream os;

Source/EmberCL/IterOpenCLKernelCreator.h

@@ -3,6 +3,7 @@
 #include "EmberCLPch.h"
 #include "EmberCLStructs.h"
 #include "EmberCLFunctions.h"
+#include "FunctionMapper.h"
 
 /// <summary>
 /// IterOpenCLKernelCreator class.
@@ -29,20 +30,24 @@ public:
 	const string& SumHistKernel() const;
 	const string& SumHistEntryPoint() const;
 	const string& IterEntryPoint() const;
-	string CreateIterKernelString(Ember<T>& ember, string& parVarDefines, bool lockAccum = false, bool doAccum = true);
+	string CreateIterKernelString(const Ember<T>& ember, string& parVarDefines, bool lockAccum = false, bool doAccum = true);
-	static void ParVarIndexDefines(Ember<T>& ember, pair<string, vector<T>>& params, bool doVals = true, bool doString = true);
+	string GlobalFunctionsString(const Ember<T>& ember);
-	static bool IsBuildRequired(Ember<T>& ember1, Ember<T>& ember2);
+	static void ParVarIndexDefines(const Ember<T>& ember, pair<string, vector<T>>& params, bool doVals = true, bool doString = true);
+	static string VariationStateString(const Ember<T>& ember);
+	static string VariationStateInitString(const Ember<T>& ember);
+	static bool IsBuildRequired(const Ember<T>& ember1, const Ember<T>& ember2);
 
 private:
-	string CreateZeroizeKernelString();
+	string CreateZeroizeKernelString() const;
-	string CreateSumHistKernelString();
+	string CreateSumHistKernelString() const;
-	string CreateProjectionString(Ember<T>& ember);
+	string CreateProjectionString(const Ember<T>& ember) const;
 
 	string m_IterEntryPoint;
 	string m_ZeroizeKernel;
 	string m_ZeroizeEntryPoint;
 	string m_SumHistKernel;
 	string m_SumHistEntryPoint;
+	FunctionMapper m_FunctionMapper;
 };
 
 #ifdef OPEN_CL_TEST_AREA

Source/EmberCL/RendererCL.cpp

@@ -354,7 +354,7 @@ bool RendererCL<T, bucketT>::WritePoints(size_t device, vector<PointCL<T>>& vec)
 
 #ifdef TEST_CL
 template <typename T, typename bucketT>
-bool RendererCL<T, bucketT>::WriteRandomPoints()
+bool RendererCL<T, bucketT>::WriteRandomPoints(size_t device)
 {
 	size_t size = IterGridKernelCount();
 	vector<PointCL<T>> vec(size);
@@ -368,7 +368,7 @@ bool RendererCL<T, bucketT>::WriteRandomPoints()
 		vec[i].m_LastXfUsed = 0;
 	}
 
-	return WritePoints(vec);
+	return WritePoints(device, vec);
 }
 #endif
 
@@ -613,6 +613,8 @@ vector<string> RendererCL<T, bucketT>::ErrorReport()
 /// <summary>
 /// Set the vector of random contexts on every device.
 /// Call the base, and reset the seeds vector.
+/// Used on the command line when the user wants a specific set of seeds to start with to
+/// produce an exact result. Mostly for debugging.
 /// </summary>
 /// <param name="randVec">The vector of random contexts to assign</param>
 /// <returns>True if the size of the vector matched the number of threads used for rendering and writing seeds to OpenCL succeeded, else false.</returns>
@@ -707,9 +709,10 @@ bool RendererCL<T, bucketT>::ResetBuckets(bool resetHist, bool resetAccum)
 /// <summary>
 /// Perform log scale density filtering on the primary device.
 /// </summary>
+/// <param name="forceOutput">Whether this output was forced due to an interactive render</param>
 /// <returns>True if success and not aborted, else false.</returns>
 template <typename T, typename bucketT>
-eRenderStatus RendererCL<T, bucketT>::LogScaleDensityFilter()
+eRenderStatus RendererCL<T, bucketT>::LogScaleDensityFilter(bool forceOutput)
 {
 	return RunLogScaleFilter();
 }
@@ -804,10 +807,10 @@ EmberStats RendererCL<T, bucketT>::Iterate(size_t iterCount, size_t temporalSamp
 		{
 			auto& wrapper = device->m_Wrapper;
 
-			if (b && !(b = wrapper.WriteBuffer		(m_EmberBufferName, reinterpret_cast<void*>(&m_EmberCL),		 sizeof(m_EmberCL))))						     { this->m_ErrorReport.push_back(loc); }
+			if (b && !(b = wrapper.WriteBuffer		(m_EmberBufferName,	   reinterpret_cast<void*>(&m_EmberCL),		   sizeof(m_EmberCL))))								 { this->m_ErrorReport.push_back(loc); }
-			if (b && !(b = wrapper.WriteBuffer		(m_XformsBufferName, reinterpret_cast<void*>(m_XformsCL.data()), sizeof(m_XformsCL[0]) * m_XformsCL.size())))	 { this->m_ErrorReport.push_back(loc); }
+			if (b && !(b = wrapper.WriteBuffer		(m_XformsBufferName,   reinterpret_cast<void*>(m_XformsCL.data()), sizeof(m_XformsCL[0]) * m_XformsCL.size())))		 { this->m_ErrorReport.push_back(loc); }
-			if (b && !(b = wrapper.AddAndWriteBuffer(m_DistBufferName, reinterpret_cast<void*>(const_cast<byte*>(XformDistributions())), XformDistributionsSize()))) { this->m_ErrorReport.push_back(loc); }//Will be resized for xaos.
+			if (b && !(b = wrapper.AddAndWriteBuffer(m_DistBufferName,	   reinterpret_cast<void*>(const_cast<byte*>(XformDistributions())), XformDistributionsSize()))) { this->m_ErrorReport.push_back(loc); }//Will be resized for xaos.
-			if (b && !(b = wrapper.WriteBuffer		(m_CarToRasBufferName, reinterpret_cast<void*>(&m_CarToRasCL),	 sizeof(m_CarToRasCL))))					     { this->m_ErrorReport.push_back(loc); }
+			if (b && !(b = wrapper.WriteBuffer		(m_CarToRasBufferName, reinterpret_cast<void*>(&m_CarToRasCL),	   sizeof(m_CarToRasCL))))							 { this->m_ErrorReport.push_back(loc); }
 
 			if (b && !(b = wrapper.AddAndWriteImage("Palette", CL_MEM_READ_ONLY, m_PaletteFormat, m_Dmap.m_Entries.size(), 1, 0, m_Dmap.m_Entries.data()))) { this->m_ErrorReport.push_back(loc); }
 
@@ -966,7 +969,7 @@ bool RendererCL<T, bucketT>::RunIter(size_t iterCount, size_t temporalSample, si
 		{
 			cl_uint argIndex = 0;
 #ifdef TEST_CL
-			fuse = 0;
+			uint fuse = 0;
 #else
 			uint fuse = uint((m_Devices[dev]->m_Calls % fuseFreq) == 0u ? FuseCount() : 0u);
 #endif

Source/EmberCL/RendererCL.h

@@ -129,7 +129,7 @@ public:
 	bool ClearAccum();
 	bool WritePoints(size_t device, vector<PointCL<T>>& vec);
 #ifdef TEST_CL
-	bool WriteRandomPoints();
+	bool WriteRandomPoints(size_t device);
 #endif
 	const string& IterKernel() const;
 	const string& DEKernel() const;
@@ -159,7 +159,7 @@ protected:
 	//Protected virtual functions overridden from Renderer.
 	virtual bool Alloc(bool histOnly = false) override;
 	virtual bool ResetBuckets(bool resetHist = true, bool resetAccum = true) override;
-	virtual eRenderStatus LogScaleDensityFilter() override;
+	virtual eRenderStatus LogScaleDensityFilter(bool forceOutput = false) override;
 	virtual eRenderStatus GaussianDensityFilter() override;
 	virtual eRenderStatus AccumulatorToFinalImage(byte* pixels, size_t finalOffset) override;
 	virtual EmberStats Iterate(size_t iterCount, size_t temporalSample) override;

Source/EmberGenome/EmberGenome.cpp

@@ -321,8 +321,8 @@ bool EmberGenome(EmberOptions& opt)
 			embers[i].DeleteMotionElements();
 		}
 
-		firstFrame = uint(opt.FirstFrame() == UINT_MAX ? embers[0].m_Time : opt.FirstFrame());
+		firstFrame = size_t(opt.FirstFrame() == UINT_MAX ? embers[0].m_Time : opt.FirstFrame());
-		lastFrame  = uint(opt.LastFrame()  == UINT_MAX ? embers.back().m_Time : opt.LastFrame());
+		lastFrame  = size_t(opt.LastFrame()  == UINT_MAX ? embers.back().m_Time : opt.LastFrame());
 
 		if (lastFrame < firstFrame)
 			lastFrame = firstFrame;
@@ -335,7 +335,7 @@ bool EmberGenome(EmberOptions& opt)
 
 			for (i = 0; i < embers.size(); i++)
 			{
-				if (ftime == uint(embers[i].m_Time))
+				if (ftime == size_t(embers[i].m_Time))
 				{
 					interpolated = embers[i];
 					exactTimeMatch = true;
@@ -349,7 +349,7 @@ bool EmberGenome(EmberOptions& opt)
 
 				for (i = 0; i < embers.size(); i++)
 				{
-					if (ftime == uint(embers[i].m_Time - 1))
+					if (ftime == size_t(embers[i].m_Time - 1))
 					{
 						exactTimeMatch = true;
 						break;
@@ -357,7 +357,7 @@ bool EmberGenome(EmberOptions& opt)
 				}
 
 				if (!exactTimeMatch)
-					interpolated.m_AffineInterp = INTERP_LINEAR;
+					interpolated.m_AffineInterp = AFFINE_INTERP_LINEAR;
 			}
 
 			if (pTemplate)
@@ -400,23 +400,12 @@ bool EmberGenome(EmberOptions& opt)
 
 			if (i < embers.size() - 1)
 			{
-				vector<Ember<T>> interpEmbers;
-
-				interpEmbers.push_back(embers[i]);
-				interpEmbers.push_back(embers[i + 1]);
-
-				if (opt.Loops() > 0)
-				{
-					//We might have looped a non-integral number of times, so store the last result as our flame to interpolate from.
-					interpEmbers[i] = result;
-				}
-
 				for (frame = 0; frame < opt.Frames(); frame++)
 				{
 					seqFlag = (frame == 0 || frame == opt.Frames() - 1);
 					blend = frame / T(opt.Frames());
 					result.Clear();
-					tools.SpinInter(&interpEmbers[i], pTemplate, result, frameCount++, seqFlag, blend);
+					tools.SpinInter(&embers[i], pTemplate, result, frameCount++, seqFlag, blend);
 					cout << emberToXml.ToString(result, opt.Extras(), opt.PrintEditDepth(), !opt.NoEdits(), false, opt.HexPalette());
 				}
 			}
@@ -500,7 +489,7 @@ bool EmberGenome(EmberOptions& opt)
 
 			oldX = embers[i].m_CenterX;
 			oldY = embers[i].m_CenterY;
-			embers[i].m_FinalRasH = uint(T(embers[i].m_FinalRasH) / T(opt.Frames()));
+			embers[i].m_FinalRasH = size_t(T(embers[i].m_FinalRasH) / T(opt.Frames()));
 
 			embers[i].m_CenterY = embers[i].m_CenterY - ((opt.Frames() - 1) * embers[i].m_FinalRasH) /
 				(2 * embers[i].m_PixelsPerUnit * pow(T(2.0), embers[i].m_Zoom));
@@ -596,7 +585,7 @@ bool EmberGenome(EmberOptions& opt)
 						mutMeth = MUTATE_NOT_SPECIFIED;
 					}
 
-					os << tools.Mutate(orig, mutMeth, vars, opt.Symmetry(), T(opt.Speed()));
+					os << tools.Mutate(orig, mutMeth, vars, opt.Symmetry(), T(opt.Speed()), MAX_CL_VARS);
 
 					//Scan string returned for 'mutate color'.
 					if (strstr(os.str().c_str(), "mutate color"))
@@ -647,7 +636,7 @@ bool EmberGenome(EmberOptions& opt)
 				{
 					os << "random";
 					randomMode = true;
-					tools.Random(orig, vars, opt.Symmetry(), 0);
+					tools.Random(orig, vars, opt.Symmetry(), 0, MAX_CL_VARS);
 					aselp0 = nullptr;
 					aselp1 = nullptr;
 				}

Source/EmberRender/EmberRender.cpp

@@ -149,7 +149,7 @@ bool EmberRender(EmberOptions& opt)
 
 	//Final setup steps before running.
 	os.imbue(std::locale(""));
-	padding = uint(log10(double(embers.size()))) + 1;
+	padding = uint(std::log10(double(embers.size()))) + 1;
 	renderer->EarlyClip(opt.EarlyClip());
 	renderer->YAxisUp(opt.YAxisUp());
 	renderer->LockAccum(opt.LockAccum());
@@ -158,7 +158,7 @@ bool EmberRender(EmberOptions& opt)
 	renderer->Transparency(opt.Transparency());
 	renderer->NumChannels(channels);
 	renderer->BytesPerChannel(opt.BitsPerChannel() / 8);
-	renderer->Priority(eThreadPriority(Clamp<int>(int(opt.Priority()), int(eThreadPriority::LOWEST), int(eThreadPriority::HIGHEST))));
+	renderer->Priority(eThreadPriority(Clamp<intmax_t>(intmax_t(opt.Priority()), intmax_t(eThreadPriority::LOWEST), intmax_t(eThreadPriority::HIGHEST))));
 	renderer->Callback(opt.DoProgress() ? progress.get() : nullptr);
 
 	for (i = 0; i < embers.size(); i++)
@@ -176,8 +176,8 @@ bool EmberRender(EmberOptions& opt)
 
 		embers[i].m_TemporalSamples = 1;//Force temporal samples to 1 for render.
 		embers[i].m_Quality *= T(opt.QualityScale());
-		embers[i].m_FinalRasW = uint(T(embers[i].m_FinalRasW) * opt.SizeScale());
+		embers[i].m_FinalRasW = size_t(T(embers[i].m_FinalRasW) * opt.SizeScale());
-		embers[i].m_FinalRasH = uint(T(embers[i].m_FinalRasH) * opt.SizeScale());
+		embers[i].m_FinalRasH = size_t(T(embers[i].m_FinalRasH) * opt.SizeScale());
 		embers[i].m_PixelsPerUnit *= T(opt.SizeScale());
 
 		if (embers[i].m_FinalRasW == 0 || embers[i].m_FinalRasH == 0)
@@ -309,9 +309,11 @@ bool EmberRender(EmberOptions& opt)
 		{
 			if (auto rendererCL = dynamic_cast<RendererCL<T, float>*>(renderer.get()))
 			{
-				cout << "Iteration kernel: \n" <<
+				cout << "Iteration kernel:\n" <<
 				rendererCL->IterKernel() << "\n\n" <<
+				"Density filter kernel:\n" <<
 				rendererCL->DEKernel() << "\n\n" <<
+				"Final accumulation kernel:\n" <<
 				rendererCL->FinalAccumKernel() << endl;
 			}
 		}

Source/EmberTester/EmberTester.cpp

@@ -136,6 +136,12 @@ void MakeTestAllVarsRegPrePost(vector<Ember<T>>& embers)
 
 	while (index < varList.RegSize())
 	{
+		/*if (index != VAR_SYNTH)
+		{
+			index++;
+			continue;
+		}
+*/
 		Ember<T> ember1;
 		unique_ptr<Variation<T>> regVar(varList.GetVariationCopy(index, VARTYPE_REG));
 		unique_ptr<Variation<T>> preVar(varList.GetVariationCopy("pre_" + regVar->Name()));
@@ -422,7 +428,7 @@ bool TestVarUnique()
 
 	for (size_t i = 0; i < vl.Size(); i++)
 	{
-		Variation<T>* var = vl.GetVariation(i);
+		auto var = vl.GetVariation(i);
 
 		if (std::find(ids.begin(), ids.end(), var->VariationId()) != ids.end())
 		{
@@ -490,7 +496,7 @@ bool TestVarPrecalcEqual(const Variation<sT>* var1, const Variation<dT>* var2)
 }
 
 template <typename sT, typename dT>
-bool TestVarEqual(Variation<sT>* var1, Variation<dT>* var2)
+bool TestVarEqual(const Variation<sT>* var1, const Variation<dT>* var2)
 {
 	bool success = true;
 
@@ -530,8 +536,8 @@ bool TestVarEqual(Variation<sT>* var1, Variation<dT>* var2)
 		success = false;
 	}
 
-	ParametricVariation<sT>* parVar1 = dynamic_cast<ParametricVariation<sT>*>(var1);
+	auto parVar1 = dynamic_cast<const ParametricVariation<sT>*>(var1);
-	ParametricVariation<dT>* parVar2 = dynamic_cast<ParametricVariation<dT>*>(var2);
+	auto parVar2 = dynamic_cast<const ParametricVariation<dT>*>(var2);
 
 	if (parVar1 && parVar2)
 	{
@@ -585,10 +591,13 @@ bool TestVarEqual(Variation<sT>* var1, Variation<dT>* var2)
 				}
 			}
 
-			if (!IsClose<sT>(params1[i].ParamVal(), (sT)params2[i].ParamVal(), sT(1e-4)))
+			if (!params1[i].IsState() && !params2[i].IsState())//Don't compare state params, they can be different if set to random vals.
 			{
-				cout << "Param " << params1[i].Name() << " Val were not equal: " << params1[i].ParamVal() << " != " << params2[i].ParamVal() << endl;
+				if (!IsClose<sT>(params1[i].ParamVal(), (sT)params2[i].ParamVal(), sT(1e-4)))
-				success = false;
+				{
+					cout << "Param " << params1[i].Name() << " Val were not equal: " << params1[i].ParamVal() << " != " << params2[i].ParamVal() << endl;
+					success = false;
+				}
 			}
 		}
 	}
@@ -669,7 +678,7 @@ bool TestVarCopy()
 
 	for (size_t i = 0; i < vlf.Size(); i++)
 	{
-		Variation<sT>* var = vlf.GetVariation(i);
+		auto var = vlf.GetVariation(i);
 		Variation<dT>* destVar = NULL;
 		unique_ptr<Variation<sT>> copyVar(var->Copy());//Test Copy().
 		
@@ -1126,6 +1135,36 @@ bool TestConstants()
 	return success;
 }
 
+bool TestGlobalFuncs()
+{
+	bool success = true;
+	VariationList<float> vlf;
+	vector<string> funcs;
+	FunctionMapper mapper;
+
+	for (size_t i = 0; i < vlf.Size(); i++)
+	{
+		auto var = vlf.GetVariation(i);
+		
+		funcs = var->OpenCLGlobalFuncNames();
+
+		for (auto& func : funcs)
+		{
+			if (!mapper.GetGlobalFunc(func))
+			{
+				cout << "Variation " << var->Name() << " used unknown global funcion " << func << endl;
+				success = false;
+			}
+			else
+			{
+				//cout << "Variation " << var->Name() << " used valid global funcion " << func << endl;
+			}
+		}
+	}
+
+	return success;
+}
+
 void PrintAllVars()
 {
 	uint i = 0;
@@ -1506,17 +1545,20 @@ void TestVarsSimilar()
 #ifdef TEST_CL
 
 template <typename T>
-void TestAllVarsCLBuild(bool printSuccess = true)
+bool TestAllVarsCLBuild(size_t platform, size_t device, bool printSuccess = true)
 {
+	bool success = true;
 	vector<Ember<T>> embers;
 	QTIsaac<ISAAC_SIZE, ISAAC_INT> rand;
-	RendererCL<T> renderer;
+	vector<pair<size_t, size_t>> devices{ std::make_pair(platform, device) };
+	RendererCL<T, float> renderer(devices);
+
 	const char* loc = __FUNCTION__;
 
-	if (!renderer.Init(1, 0, false, 0))
+	if (!renderer.Ok())
 	{
 		cout << loc << "Creating RendererCL failed, tests will not be run." << endl;
-		return;
+		return false;
 	}
 
 	MakeTestAllVarsRegPrePost(embers);
@@ -1531,12 +1573,18 @@ void TestAllVarsCLBuild(bool printSuccess = true)
 				cout << loc << ": Build succeeded for ember " << it.m_Name << endl;
 		}
 		else
-			cout << loc << ": OpenCL program build failed:\n" << renderer.ErrorReport() << endl;
+		{
+			success = false;
+			cout << loc << ": OpenCL program build failed for ember " << it.m_Name << ":\n" << renderer.ErrorReportString() << endl;
+			break;
+		}
 	}
+
+	return success;
 }
 
 template <typename T>
-void TestCpuGpuResults()
+void TestCpuGpuResults(size_t platform, size_t device)
 {
 	bool breakOnBad = true;
 	int i = 0;//(int)VAR_TARGET;//Start at the one you want to test.
@@ -1547,10 +1595,8 @@ void TestCpuGpuResults()
 	VariationList<T> vlf;
 	QTIsaac<ISAAC_SIZE, ISAAC_INT> rand;
 	vector<PointCL<T>> points;
-	RendererCL<T> renderer;
+	vector<pair<size_t, size_t>> devices{ std::make_pair(platform, device) };
-
+	RendererCL<T, float> renderer(devices);
-	if (!renderer.Init(1, 0, false, 0))
-		return;
 
 	points.resize(renderer.TotalIterKernelCount());
 
@@ -1637,7 +1683,7 @@ void TestCpuGpuResults()
 }
 
 template <typename T>
-void TestGpuVectorRead()
+void TestGpuVectorRead(size_t platform, size_t device)
 {
 	T minx = TMAX, miny = TMAX, minz = TMAX, mincolorx = TMAX;
 	T maxx = TLOW, maxy = TLOW, maxz = TLOW, maxcolorx = TLOW;
@@ -1649,10 +1695,8 @@ void TestGpuVectorRead()
 	VariationList<T> vlf;
 	QTIsaac<ISAAC_SIZE, ISAAC_INT> rand;
 	vector<PointCL<T>> points;
-	RendererCL<T> renderer;
+	vector<pair<size_t, size_t>> devices{ std::make_pair(platform, device) };
-
+	RendererCL<T, float> renderer(devices);
-	if (!renderer.Init(1, 0, false, 0))
-		return;
 
 	points.resize(renderer.TotalIterKernelCount());
 
@@ -1918,17 +1962,110 @@ void TestThreadedKernel()
 	}
 }
 
+template <typename T>
+void DistribTester()
+{
+	size_t i;
+	size_t distribCount = 1;
+	size_t xformCount = 3;
+	vector<byte> m_XformDistributions;
+	//const Xform<T>* xforms = 3;
+	size_t j = 0;
+	vector<T> weights { T(0.333333), T(1.0), T(0.25) };
+	double tempDensity = 0, currentDensityLimit = 0, densityPerElement;
+
+	if (m_XformDistributions.size() < CHOOSE_XFORM_GRAIN * distribCount)
+		m_XformDistributions.resize(CHOOSE_XFORM_GRAIN * distribCount);
+
+	if (m_XformDistributions.size() < CHOOSE_XFORM_GRAIN * distribCount)
+		return;
+
+	for (size_t distrib = 0; distrib < distribCount; distrib++)
+	{
+		double totalDensity = 0;
+
+		//First find the total densities of all xforms.
+		for (i = 0; i < xformCount; i++)
+		{
+			T d = weights[i];
+
+			totalDensity += d;
+		}
+
+		//Original returned false if all were 0, but it's allowed here
+		//which will just end up setting all elements to 0 which means
+		//only the first xform will get used.
+
+		//Calculate how much of a fraction of a the total density each element represents.
+		j = 0;
+		tempDensity = 0;
+		currentDensityLimit = 0;
+		densityPerElement = totalDensity / CHOOSE_XFORM_GRAIN;
+
+		//Assign xform indices in order to each element of m_XformDistributions.
+		//The number of elements assigned a given index is proportional to that xform's
+		//density relative to the sum of all densities.
+		for (i = 0; i < xformCount; i++)
+		{
+			T temp = weights[i];
+			currentDensityLimit += temp;
+
+			//Populate points corresponding to this xform's weight/density.
+			//Also check that j is within the bounds of the distribution array just to be safe in the case of a rounding error.
+			while (tempDensity < currentDensityLimit && j < CHOOSE_XFORM_GRAIN)
+			{
+#ifdef _DEBUG
+				//Ensure distribution contains no out of bounds indices.
+				if (byte(i) >= xformCount)
+					throw "Out of bounds xform index in selection distribution.";
+#endif
+				//printf("offset = %d, xform = %d, running sum = %f\n", j, i, tempDensity);
+				m_XformDistributions[(distrib * CHOOSE_XFORM_GRAIN) + j] = byte(i);
+				tempDensity += densityPerElement;
+				j++;
+			}
+		}
+
+
+		//Flam3 did this, which gives the same result.
+		//T t = xforms[0].m_Weight;
+		//
+		//if (distrib > 0)
+		//	t *= xforms[distrib - 1].Xaos(0);
+		//
+		//T r = 0;
+		//
+		//for (i = 0; i < CHOOSE_XFORM_GRAIN; i++)
+		//{
+		//	while (r >= t)
+		//	{
+		//		j++;
+		//
+		//		if (distrib > 0)
+		//			t += xforms[j].m_Weight * xforms[distrib - 1].Xaos(j);
+		//		else
+		//			t += xforms[j].m_Weight;
+		//	}
+		//
+		//	m_XformDistributions[(distrib * CHOOSE_XFORM_GRAIN) + i] = j;
+		//	r += densityPerElement;
+		//}
+	}
+}
+
 int _tmain(int argc, _TCHAR* argv[])
 {
 	//int i;
+	bool b = true;
 	Timing t(4);
 	QTIsaac<ISAAC_SIZE, ISAAC_INT> rand(1, 2, 3);
 	mt19937 meow(1729);
-	MakeTestAllVarsRegPrePostComboFile("testallvarsout.flame");
+	
+	/*MakeTestAllVarsRegPrePostComboFile("testallvarsout.flame");
 	return 0;
 
 
-	/*TestThreadedKernel();
+	TestThreadedKernel();
 
 	PaletteList<float> palf;
 	Palette<float>* pal = palf.GetRandomPalette();
@@ -1939,7 +2076,7 @@ int _tmain(int argc, _TCHAR* argv[])
 
 	for (int i = 0; i < 10; i++)
 	{
-		cout << "log10(" << d << ") = " << std::max<uint>(1u, uint(log10(d)) + 1u) << endl;
+		cout << "log10(" << d << ") = " << std::max<uint>(1u, uint(std::log10(d)) + 1u) << endl;
 		d *= 10;
 	}
 
@@ -2041,7 +2178,7 @@ int _tmain(int argc, _TCHAR* argv[])
 
 	//cd2 = sin(cd);
 	
-	/*t.Tic();
+	t.Tic();
 	TestCasting();
 	t.Toc("TestCasting()");
 	
@@ -2088,6 +2225,7 @@ int _tmain(int argc, _TCHAR* argv[])
 	TestVarCopy<double, float>();
 	t.Toc("TestVarCopy<double, float>()");
 #endif
+	
 	t.Tic();
 	TestVarRegPrePost();
 	t.Toc("TestVarRegPrePost()");
@@ -2121,17 +2259,22 @@ int _tmain(int argc, _TCHAR* argv[])
 	t.Toc("TestConstants()");
 	
 	t.Tic();
+	TestGlobalFuncs();
+	t.Toc("TestGlobalFuncs()");
+	
+	/*t.Tic();
 	TestXformsInOutPoints();
 	t.Toc("TestXformsInOutPoints()");
 	
 	t.Tic();
 	TestVarTime<float>();
 	t.Toc("TestVarTime()");
+	*/
 
 	t.Tic();
 	TestOperations<float>();
 	t.Toc("TestOperations()");
-	*/
+	
 	//t.Tic();
 	//TestVarsSimilar<float>();
 	//t.Toc("TestVarsSimilar()");
@@ -2141,18 +2284,34 @@ int _tmain(int argc, _TCHAR* argv[])
 	//TestCpuGpuResults<float>();
 	//t.Toc("TestCpuGpuResults<float>()");
 	
-	t.Tic();
+	//t.Tic();
-	TestAllVarsCLBuild<float>();
+	//b = TestAllVarsCLBuild<float>(0, 0, true);
-	t.Toc("TestAllVarsCLBuild<float>()");
+	//t.Toc("TestAllVarsCLBuild<float>()");
+
+	if (b)
+	{
+		t.Tic();
+		b = TestAllVarsCLBuild<float>(1, 0, true);
+		t.Toc("TestAllVarsCLBuild<float>()");
+	}
 
 #ifdef DO_DOUBLE
 	//t.Tic();
 	//TestCpuGpuResults<double>();
 	//t.Toc("TestCpuGpuResults<double>()");
+	if (b)
+	{
+		t.Tic();
+		TestAllVarsCLBuild<double>(0, 0, true);
+		t.Toc("TestAllVarsCLBuild<double>()");
 
-	t.Tic();
+		if (b)
-	TestAllVarsCLBuild<double>();
+		{
-	t.Toc("TestAllVarsCLBuild<double>()");
+			t.Tic();
+			TestAllVarsCLBuild<double>(1, 0, true);
+			t.Toc("TestAllVarsCLBuild<double>()");
+		}
+	}
 #endif
 #endif
 

Source/Fractorium/EmberFile.h

@@ -114,7 +114,7 @@ public:
 			{
 				if (i != j && m_Embers[i].m_Name == m_Embers[j].m_Name)
 				{
-					m_Embers[j].m_Name = m_Embers[j].m_Name + "_" + ToString(++x).toStdString();
+					m_Embers[j].m_Name = IncrementTrailingUnderscoreInt(QString::fromStdString(m_Embers[j].m_Name)).toStdString();
 					j = 0;
 				}
 			}
@@ -130,6 +130,33 @@ public:
 		return "Flame_" + QDateTime(QDateTime::currentDateTime()).toString("yyyy-MM-dd-hhmmss");
 	}
 
+	/// <summary>
+	/// Return a copy of the string which ends with _# where # is the
+	/// previous number at that position incremented by one.
+	/// If the original string did not end with _#, the returned
+	/// string will just have _1 appended to it.
+	/// </summary>
+	/// <param name="str">The string to process</param>
+	/// <returns>The original string with the number after the final _ character incremented by one</returns>
+	static QString IncrementTrailingUnderscoreInt(const QString& str)
+	{
+		bool ok = false;
+		size_t num = 0;
+		QString endSection;
+		QString ret = str;
+		int lastUnderscore = str.lastIndexOf('_');
+
+		if (lastUnderscore != -1)
+		{
+			endSection = str.section('_', -1);
+			num = endSection.toULongLong(&ok);
+			ret.chop(str.size() - lastUnderscore);
+		}
+
+		ret += "_" + QString::number(num + 1);
+		return ret;
+	}
+
 	/// <summary>
 	/// Ensures a given input filename is unique by appending a count to the end.
 	/// </summary>
@@ -149,7 +176,8 @@ public:
 
 		do
 		{
-			newPath = path + base + "_" + ToString(counter++) + "." + extension;
+			base = IncrementTrailingUnderscoreInt(base);
+			newPath = path + base + "." + extension;
 		}
 		while (QFile::exists(newPath));
 		
@@ -164,7 +192,7 @@ public:
 	/// <returns>The default ember name</returns>
 	static QString DefaultEmberName(uint i)
 	{
-		return DefaultFilename() + "-" + ToString(i);
+		return DefaultFilename() + "_" + ToString(i);
 	}
 
 	QString m_Filename;

Source/Fractorium/Fractorium.cpp

@@ -331,6 +331,8 @@ bool Fractorium::eventFilter(QObject* o, QEvent* e)
 	}
 	else if (QKeyEvent* ke = dynamic_cast<QKeyEvent*>(e))
 	{
+		bool shift = QGuiApplication::keyboardModifiers().testFlag(Qt::ShiftModifier);
+
 		if (ke->key() >= Qt::Key_F1 && ke->key() <= Qt::Key_F32)
 		{
 			int val = ke->key() - (int)Qt::Key_F1;
@@ -340,7 +342,8 @@ bool Fractorium::eventFilter(QObject* o, QEvent* e)
 		}
 		else if (o == ui.LibraryTree)
 		{
-			if (ke->key() == Qt::Key_Delete && e->type() == QEvent::KeyRelease)
+			//Require shift for deleting to prevent it from triggering when the user enters delete in the edit box.
+			if (ke->key() == Qt::Key_Delete && e->type() == QEvent::KeyRelease && shift)
 			{
 				auto p = GetCurrentEmberIndex();
 

Source/Fractorium/Fractorium.ui

@@ -5046,6 +5046,9 @@
                <verstretch>0</verstretch>
               </sizepolicy>
              </property>
+             <property name="mouseTracking">
+              <bool>true</bool>
+             </property>
              <property name="focusPolicy">
               <enum>Qt::WheelFocus</enum>
              </property>
@@ -5065,10 +5068,10 @@
               <set>QAbstractItemView::NoEditTriggers</set>
              </property>
              <property name="selectionMode">
-              <enum>QAbstractItemView::NoSelection</enum>
+              <enum>QAbstractItemView::SingleSelection</enum>
              </property>
              <property name="selectionBehavior">
-              <enum>QAbstractItemView::SelectItems</enum>
+              <enum>QAbstractItemView::SelectRows</enum>
              </property>
              <property name="indentation">
               <number>12</number>

Source/Fractorium/FractoriumLibrary.cpp

@@ -213,6 +213,12 @@ void FractoriumEmberController<T>::EmberTreeItemChanged(QTreeWidgetItem* item, i
 	
 		if (emberItem)
 		{
+			if (emberItem->text(0).isEmpty())//Prevent empty string.
+			{
+				emberItem->UpdateEditText();
+				return;
+			}
+
 			string oldName = emberItem->GetEmber()->m_Name;//First preserve the previous name.
 
 			tree->blockSignals(true);

Source/Fractorium/FractoriumMenus.cpp

@@ -65,10 +65,10 @@ void FractoriumEmberController<T>::NewFlock(size_t count)
 
 	for (size_t i = 0; i < count; i++)
 	{
-		m_SheepTools->Random(ember, m_FilteredVariations, static_cast<intmax_t>(QTIsaac<ISAAC_SIZE, ISAAC_INT>::GlobalRand->Frand<T>(-2, 2)), 0);
+		m_SheepTools->Random(ember, m_FilteredVariations, static_cast<intmax_t>(QTIsaac<ISAAC_SIZE, ISAAC_INT>::GlobalRand->Frand<T>(-2, 2)), 0, MAX_CL_VARS);
 		ParamsToEmber(ember);
 		ember.m_Index = i;
-		ember.m_Name = m_EmberFile.m_Filename.toStdString() + "-" + ToString(i + 1ULL).toStdString();
+		ember.m_Name = m_EmberFile.m_Filename.toStdString() + "_" + ToString(i + 1ULL).toStdString();
 		m_EmberFile.m_Embers.push_back(ember);
 	}
 
@@ -126,7 +126,7 @@ void FractoriumEmberController<T>::NewRandomFlameInCurrentFile()
 	Ember<T> ember;
 
 	StopPreviewRender();
-	m_SheepTools->Random(ember, m_FilteredVariations, static_cast<int>(QTIsaac<ISAAC_SIZE, ISAAC_INT>::GlobalRand->Frand<T>(-2, 2)), 0);
+	m_SheepTools->Random(ember, m_FilteredVariations, static_cast<int>(QTIsaac<ISAAC_SIZE, ISAAC_INT>::GlobalRand->Frand<T>(-2, 2)), 0, MAX_CL_VARS);
 	ParamsToEmber(ember);
 	ember.m_Name = EmberFile<T>::DefaultEmberName(m_EmberFile.Size() + 1).toStdString();
 	ember.m_Index = m_EmberFile.Size();

Source/Fractorium/FractoriumParams.cpp

@@ -79,7 +79,7 @@ void Fractorium::InitParamsUI()
 	SetupCombo(table, this, row, 1, m_TemporalFilterTypeCombo, comboVals, SIGNAL(currentIndexChanged(const QString&)), SLOT(OnTemporalFilterTypeComboCurrentIndexChanged(const QString&)));
 
 	SetupSpinner<DoubleSpinBox, double>(table, this, row, 1, m_DEFilterMinRadiusSpin, spinHeight,    0, 25,   1, SIGNAL(valueChanged(double)), SLOT(OnDEFilterMinRadiusWidthChanged(double)), true,   0,   0,   0);
-	SetupSpinner<DoubleSpinBox, double>(table, this, row, 1, m_DEFilterMaxRadiusSpin, spinHeight,    0, 25,   1, SIGNAL(valueChanged(double)), SLOT(OnDEFilterMaxRadiusWidthChanged(double)), true, 9.0, 9.0,   0);
+	SetupSpinner<DoubleSpinBox, double>(table, this, row, 1, m_DEFilterMaxRadiusSpin, spinHeight,    0, 25,   1, SIGNAL(valueChanged(double)), SLOT(OnDEFilterMaxRadiusWidthChanged(double)), true, 0.0, 9.0,   0);
 	SetupSpinner<DoubleSpinBox, double>(table, this, row, 1, m_DECurveSpin,			  spinHeight, 0.15,  5, 0.1, SIGNAL(valueChanged(double)), SLOT(OnDEFilterCurveWidthChanged(double)),     true, 0.4, 0.4, 0.4);
 
 	//Iteration.
@@ -89,7 +89,7 @@ void Fractorium::InitParamsUI()
 	SetupSpinner<SpinBox, int>(			table, this, row, 1, m_FuseSpin,			spinHeight, 1,      1000,   5, SIGNAL(valueChanged(int)),	 SLOT(OnFuseChanged(int)),			  true,	   15,	  15, 15);
 	SetupSpinner<DoubleSpinBox, double>(table, this, row, 1, m_QualitySpin,			spinHeight, 1,      dmax,  50, SIGNAL(valueChanged(double)), SLOT(OnQualityChanged(double)),	  true,    10,	  10, 10);
 	SetupSpinner<SpinBox, int>(         table, this, row, 1, m_SupersampleSpin,		spinHeight, 1,         4,   1, SIGNAL(valueChanged(int)),	 SLOT(OnSupersampleChanged(int)),	  true,     1,	   1,  1);
-	SetupSpinner<SpinBox, int>(         table, this, row, 1, m_TemporalSamplesSpin, spinHeight, 1,      5000,  50, SIGNAL(valueChanged(int)),	 SLOT(OnTemporalSamplesChanged(int)), true,  1000);
+	SetupSpinner<SpinBox, int>(         table, this, row, 1, m_TemporalSamplesSpin, spinHeight, 1,      5000,   1, SIGNAL(valueChanged(int)),	 SLOT(OnTemporalSamplesChanged(int)), true,  1000);
 
 	comboVals.clear();
 	comboVals.push_back("Step");
@@ -412,8 +412,8 @@ void Fractorium::OnFuseChanged(int d) { m_Controller->FuseChanged(d); }
 /// the rendering process is continued, else it's reset.
 /// </summary>
 /// <param name="d">The quality in terms of iterations per pixel</param>
-template <typename T> void FractoriumEmberController<T>::QualityChanged(double d) { Update([&] { m_Ember.m_Quality = d; }, true, d > m_Ember.m_Quality ? KEEP_ITERATING : FULL_RENDER); }
+template <typename T> void FractoriumEmberController<T>::QualityChanged(double d) { /*Update([&] { m_Ember.m_Quality = d; }, true, d > m_Ember.m_Quality ? KEEP_ITERATING : FULL_RENDER);*/ }
-void Fractorium::OnQualityChanged(double d) { m_Controller->QualityChanged(d); }
+void Fractorium::OnQualityChanged(double d) { /*m_Controller->QualityChanged(d);*/ }
 
 /// <summary>
 /// Set the supersample.
@@ -452,11 +452,11 @@ void FractoriumEmberController<T>::AffineInterpTypeChanged(int i)
 	Update([&]
 	{
 		if (i == 0)
-			m_Ember.m_AffineInterp = INTERP_LINEAR;
+			m_Ember.m_AffineInterp = AFFINE_INTERP_LINEAR;
 		else if (i == 1)
-			m_Ember.m_AffineInterp = INTERP_LOG;
+			m_Ember.m_AffineInterp = AFFINE_INTERP_LOG;
 		else
-			m_Ember.m_AffineInterp = INTERP_LINEAR;
+			m_Ember.m_AffineInterp = AFFINE_INTERP_LINEAR;
 	}, true, NOTHING);
 }
 

Source/Fractorium/FractoriumRender.cpp

@@ -301,7 +301,33 @@ bool FractoriumEmberController<T>::Render()
 	bool success = true;
 	GLWidget* gl = m_Fractorium->ui.GLDisplay;
 	RendererCL<T, float>* rendererCL = nullptr;
-	eProcessAction action = CondenseAndClearProcessActions();
+	eProcessAction qualityAction, action;
+
+	//Quality is the only parameter we update inside the timer.
+	//This is to allow the user to rapidly increase the quality spinner
+	//without fully resetting the render. Instead, it will just keep iterating
+	//where it last left off in response to an increase.
+	T d = T(m_Fractorium->m_QualitySpin->value());
+
+	if (d < m_Ember.m_Quality)//Full restart if quality decreased.
+	{
+		m_Ember.m_Quality = d;
+		qualityAction = eProcessAction::FULL_RENDER;
+	}
+	else if (d > m_Ember.m_Quality && ProcessState() == ACCUM_DONE)//If quality increased, keep iterating after current render finishes.
+	{
+		m_Ember.m_Quality = d;
+		qualityAction = eProcessAction::KEEP_ITERATING;
+	}
+	else if (IsClose(d, m_Ember.m_Quality))
+		qualityAction = eProcessAction::NOTHING;
+
+	if (qualityAction == eProcessAction::FULL_RENDER)
+		Update([&] {});//Stop the current render, a full restart is needed.
+	else if (qualityAction == eProcessAction::KEEP_ITERATING)
+		m_ProcessActions.push_back(qualityAction);//Special, direct call to avoid resetting the render inside Update() because only KEEP_ITERATING is needed.
+
+	action = CondenseAndClearProcessActions();//Combine with all other previously requested actions.
 
 	if (m_Renderer->RendererType() == OPENCL_RENDERER)
 		rendererCL = dynamic_cast<RendererCL<T, float>*>(m_Renderer.get());
@@ -391,7 +417,7 @@ bool FractoriumEmberController<T>::Render()
 				//Only certain stats can be reported with OpenCL.
 				if (m_Renderer->RendererType() == OPENCL_RENDERER)
 				{
-					m_Fractorium->m_RenderStatusLabel->setText("Iters: " + iters + ". Scaled quality: " + scaledQuality + ". Total time: " + QString::fromStdString(renderTime));
+					m_Fractorium->m_RenderStatusLabel->setText("Iters: " + iters + ". Scaled quality: " + scaledQuality + ". Total time: " + QString::fromStdString(renderTime) + ".");
 				}
 				else
 				{
@@ -399,7 +425,7 @@ bool FractoriumEmberController<T>::Render()
 					QString badVals = ToString<qulonglong>(stats.m_Badvals);
 					QString badPercent = QLocale::system().toString(percent * 100, 'f', 2);
 
-					m_Fractorium->m_RenderStatusLabel->setText("Iters: " + iters + ". Scaled quality: " + scaledQuality + ". Bad values: " + badVals + " (" + badPercent + "%). Total time: " + QString::fromStdString(renderTime));
+					m_Fractorium->m_RenderStatusLabel->setText("Iters: " + iters + ". Scaled quality: " + scaledQuality + ". Bad values: " + badVals + " (" + badPercent + "%). Total time: " + QString::fromStdString(renderTime) + ".");
 				}
 				
 				if (m_LastEditWasUndoRedo && (m_UndoIndex == m_UndoList.size() - 1))//Traversing through undo list, reached the end, so put back in regular edit mode.

Source/Fractorium/OptionsDialog.cpp

@@ -18,8 +18,8 @@ FractoriumOptionsDialog::FractoriumOptionsDialog(FractoriumSettings* settings, Q
 	m_Settings = settings;
 	QTableWidget* table = ui.OptionsXmlSavingTable;
 	ui.ThreadCountSpin->setRange(1, Timing::ProcessorCount());
-	connect(ui.OpenCLCheckBox, SIGNAL(stateChanged(int)),		 this, SLOT(OnOpenCLCheckBoxStateChanged(int)),	 Qt::QueuedConnection);
+	connect(ui.OpenCLCheckBox, SIGNAL(stateChanged(int)),	  this, SLOT(OnOpenCLCheckBoxStateChanged(int)),  Qt::QueuedConnection);
-	connect(ui.DeviceTable,	   SIGNAL(cellChanged(int, int)),	 this, SLOT(OnDeviceTableCellChanged(int, int)), Qt::QueuedConnection);
+	connect(ui.DeviceTable,	   SIGNAL(cellChanged(int, int)), this, SLOT(OnDeviceTableCellChanged(int, int)), Qt::QueuedConnection);
 
 	SetupSpinner<SpinBox, int>(table, this, row, 1, m_XmlTemporalSamplesSpin, spinHeight,  1,	1000, 100, "", "", true, 1000);
 	SetupSpinner<SpinBox, int>(table, this, row, 1, m_XmlQualitySpin,		  spinHeight,  1, 200000,  50, "", "", true, 1000);

Source/Fractorium/QssDialog.cpp

@@ -70,7 +70,6 @@ QssDialog::QssDialog(Fractorium* parent) :
 
 	connect(ui->QssLoadButton, SIGNAL(clicked()), this, SLOT(LoadButton_clicked()), Qt::QueuedConnection);
 	connect(ui->QssSaveButton, SIGNAL(clicked()), this, SLOT(SaveButton_clicked()), Qt::QueuedConnection);
-	connect(ui->QssSaveDefaultButton, SIGNAL(clicked()), this, SLOT(SaveDefaultButton_clicked()), Qt::QueuedConnection);
 	connect(ui->QssBasicButton, SIGNAL(clicked()), this, SLOT(BasicButton_clicked()), Qt::QueuedConnection);
 	connect(ui->QssMediumButton, SIGNAL(clicked()), this, SLOT(MediumButton_clicked()), Qt::QueuedConnection);
 	connect(ui->QssAdvancedButton, SIGNAL(clicked()), this, SLOT(AdvancedButton_clicked()), Qt::QueuedConnection);
@@ -310,7 +309,7 @@ void QssDialog::accept()
 	if (m_Theme)
 		m_Parent->m_Settings->Theme(m_Theme->objectName());
 
-	SaveDefaultButton_clicked();
+	SaveAsDefault();
 	QDialog::accept();
 }
 
@@ -506,11 +505,20 @@ void QssDialog::LoadButton_clicked()
 /// <summary>
 /// Save the stylesheet to disk.
 /// Called when the user clicks the save button.
+/// The user cannot save to default.qss, as it's a special placeholder.
+/// When they exit the dialog by clicking OK, the currently displayed stylesheet
+/// will be saved to default.qss.
 /// </summary>
 void QssDialog::SaveButton_clicked()
 {
 	auto path = SaveFile();
 
+	if (path.toLower().endsWith("default.qss"))
+	{
+		QMessageBox::critical(this, "File save error", "Stylesheet cannot be saved to default.qss. Save it to a different file name, then exit the dialog by clicking OK which will set it as the default.");
+		return;
+	}
+
 	if (!path.isEmpty())
 	{
 		ofstream of(path.toStdString());
@@ -522,16 +530,16 @@ void QssDialog::SaveButton_clicked()
 			of.close();
 		}
 		else
-			QMessageBox::critical(this, "File open error", "Failed to open " + path + ", style will not be set as default");
+			QMessageBox::critical(this, "File save error", "Failed to save " + path + ", style will not be set as default");
 	}
 }
 
 /// <summary>
 /// Save the stylesheet to the default.qss on disk.
 /// This will be loaded the next time Fractorium runs.
-/// Called when the user clicks the save as default button.
+/// Called when the user clicks ok.
 /// </summary>
-void QssDialog::SaveDefaultButton_clicked()
+void QssDialog::SaveAsDefault()
 {
 	auto path = m_Parent->m_SettingsPath + "/default.qss";
 	ofstream of(path.toStdString());
@@ -543,7 +551,7 @@ void QssDialog::SaveDefaultButton_clicked()
 		of.close();
 	}
 	else
-		QMessageBox::critical(this, "File open error", "Failed to open " + path + ", style will not be set as default");
+		QMessageBox::critical(this, "File save error", "Failed to save " + path + ", style will not be set as default");
 }
 
 /// <summary>

Source/Fractorium/QssDialog.h

@@ -107,12 +107,12 @@ private slots:
 
 	void LoadButton_clicked();
 	void SaveButton_clicked();
-	void SaveDefaultButton_clicked();
 	void BasicButton_clicked();
 	void MediumButton_clicked();
 	void AdvancedButton_clicked();
 
 private:
+	void SaveAsDefault();
 	void InsertCssProperty(const QString &name, const QString &value);
 	void SetupFileDialog();
 	QString OpenFile();

Source/Fractorium/QssDialog.ui

@@ -47,13 +47,6 @@
          </property>
         </widget>
        </item>
-       <item>
-        <widget class="QPushButton" name="QssSaveDefaultButton">
-         <property name="text">
-          <string>Save as default</string>
-         </property>
-        </widget>
-       </item>
        <item>
         <widget class="QPushButton" name="QssBasicButton">
          <property name="text">

Source/Fractorium/qcssparser.cpp

@@ -39,7 +39,6 @@
 **
 ****************************************************************************/
 #include "FractoriumPch.h"
-#include "qcssparser.h"
 #include "qcssscanner.h"
 
 /// <summary>