remove needMarker parameter

Apps/Examples/viewer.cpp

@@ -459,45 +459,41 @@ void Viewer::cb_keyPress(int keycode)
 
 			Utils::Chrono ch;
 			ch.start();
-
-			for (unsigned int i=0; i<4; ++i)
-				Algo::Surface::Geometry::computeNormalVertices<PFP>(myMap, position, normal) ;
-			std::cout << "Algo::Surface::Geometry::Parallel::computeNormalVertices "<< ch.elapsed()<< " ms "<< std::endl;
-
-			ch.start();
-
 			CGoGN::Parallel::NumberOfThreads = 1;
 			for (unsigned int i=0; i<4; ++i)
 				Algo::Surface::Geometry::Parallel::computeNormalVertices<PFP>(myMap, position, normal) ;
 			std::cout << "Algo::Surface::Geometry::Parallel::computeNormalVertices1 "<< ch.elapsed()<< " ms "<< std::endl;
 
-			ch.start();
+//			ch.start();
+//			CGoGN::Parallel::NumberOfThreads = 2;
+//			for (unsigned int i=0; i<4; ++i)
+//				Algo::Surface::Geometry::Parallel::computeNormalVertices<PFP>(myMap, position, normal) ;
+//			std::cout << "Algo::Surface::Geometry::Parallel::computeNormalVertices2 "<< ch.elapsed()<< " ms "<< std::endl;
 
-			CGoGN::Parallel::NumberOfThreads = 2;
-			for (unsigned int i=0; i<4; ++i)
-				Algo::Surface::Geometry::Parallel::computeNormalVertices<PFP>(myMap, position, normal) ;
-			std::cout << "Algo::Surface::Geometry::Parallel::computeNormalVertices2 "<< ch.elapsed()<< " ms "<< std::endl;
+//			ch.start();
+//			CGoGN::Parallel::NumberOfThreads = 3;
+//			for (unsigned int i=0; i<4; ++i)
+//				Algo::Surface::Geometry::Parallel::computeNormalVertices<PFP>(myMap, position, normal) ;
+//			std::cout << "Algo::Surface::Geometry::Parallel::computeNormalVertices3 "<< ch.elapsed()<< " ms "<< std::endl;
 
 			ch.start();
-
-			CGoGN::Parallel::NumberOfThreads = 3;
-			for (unsigned int i=0; i<4; ++i)
-				Algo::Surface::Geometry::Parallel::computeNormalVertices<PFP>(myMap, position, normal) ;
-			std::cout << "Algo::Surface::Geometry::Parallel::computeNormalVertices3 "<< ch.elapsed()<< " ms "<< std::endl;
-
-			ch.start();
-
 			CGoGN::Parallel::NumberOfThreads = 4;
 			for (unsigned int i=0; i<4; ++i)
 				Algo::Surface::Geometry::Parallel::computeNormalVertices<PFP>(myMap, position, normal) ;
 			std::cout << "Algo::Surface::Geometry::Parallel::computeNormalVertices4 "<< ch.elapsed()<< " ms "<< std::endl;
 
-			ch.start();
+//			ch.start();
+//			CGoGN::Parallel::NumberOfThreads = 8;
+//			for (unsigned int i=0; i<4; ++i)
+//				Algo::Surface::Geometry::Parallel::computeNormalVertices<PFP>(myMap, position, normal) ;
+//			std::cout << "Algo::Surface::Geometry::Parallel::computeNormalVertices8 "<< ch.elapsed()<< " ms "<< std::endl;
 
-			CGoGN::Parallel::NumberOfThreads = 8;
+			ch.start();
+			CGoGN::Parallel::NumberOfThreads = 1;
 			for (unsigned int i=0; i<4; ++i)
 				Algo::Surface::Geometry::Parallel::computeNormalVertices<PFP>(myMap, position, normal) ;
-			std::cout << "Algo::Surface::Geometry::Parallel::computeNormalVertices8 "<< ch.elapsed()<< " ms "<< std::endl;
+			std::cout << "Algo::Surface::Geometry::Parallel::computeNormalVertices1 "<< ch.elapsed()<< " ms "<< std::endl;
+
 
 			ch.start();
 
@@ -505,7 +501,7 @@ void Viewer::cb_keyPress(int keycode)
 				Parallel::foreach_cell<VERTEX>(myMap, [&] (Vertex v, unsigned int th)
 				{
 					normal[v] = Algo::Surface::Geometry::vertexNormal<PFP>(myMap,v,position);
-				}, false, FORCE_CELL_MARKING);
+				}, FORCE_CELL_MARKING);
 
 			std::cout << "Parallel::foreach_cell "<< ch.elapsed()<< " ms "<< std::endl;
 

Apps/Tuto/Traversals/traverse_cells.cpp

@@ -108,7 +108,7 @@ int main()
 	{
 		position[v] += VEC3(0.0, 0.0, PFP::REAL(thread) * 0.1f);
 		// WARNING thread vary here from 1 to 4 (and not from 0 to 3) !!!!
-	}, false); // 4:4 thread, false for no need for markers in threaded code.
+	}); // 4:4 thread, false for no need for markers in threaded code.
 
 	std::cout << "After // processing"<< std::endl;
 	foreach_cell<VERTEX>(myMap, [&] (Vertex v) // for each Vertex v of the MAP myMap
@@ -132,7 +132,7 @@ int main()
 	{
 		// for each face add surface to accumulator (-1 because counter between 1-3 not 0-3)
 		surf[thr-1] += Algo::Surface::Geometry::convexFaceArea<PFP>(myMap, f, position);
-	}, false);
+	});
 
 	std::cout << surf[0]<< "/"<< surf[1]<< "/"<< surf[2]<< "/"<< std::endl;
 	std::cout << "Total="<<surf[0]+surf[1]+surf[2]<< std::endl;

include/Algo/Geometry/area.hpp

@@ -236,7 +236,7 @@ void computeAreaFaces(typename PFP::MAP& map, const VertexAttribute<typename PFP
 	CGoGN::Parallel::foreach_cell<FACE>(map,[&](Face f, unsigned int thr)
 	{
 		area[f] = convexFaceArea<PFP>(map, f, position) ;
-	},false,AUTO);
+	});
 }
 
 
@@ -246,7 +246,7 @@ void computeOneRingAreaVertices(typename PFP::MAP& map, const VertexAttribute<ty
 	CGoGN::Parallel::foreach_cell<VERTEX>(map,[&](Vertex v, unsigned int thr)
 	{
 		area[v] = vertexOneRingArea<PFP>(map, v, position) ;
-	},false,FORCE_CELL_MARKING);
+	},FORCE_CELL_MARKING);
 }
 
 template <typename PFP>
@@ -256,7 +256,7 @@ void computeBarycentricAreaVertices(typename PFP::MAP& map, const VertexAttribut
 	{
 		vertex_area[v] = vertexBarycentricArea<PFP>(map, v, position) ;
 	}
-	,false,FORCE_CELL_MARKING);
+	,FORCE_CELL_MARKING);
 }
 
 template <typename PFP>
@@ -265,7 +265,7 @@ void computeVoronoiAreaVertices(typename PFP::MAP& map, const VertexAttribute<ty
 	CGoGN::Parallel::foreach_cell<VERTEX>(map,[&](Vertex v, unsigned int thr)
 	{
 		area[v] = vertexVoronoiArea<PFP>(map, v, position) ;
-	},false,FORCE_CELL_MARKING);
+	},FORCE_CELL_MARKING);
 }
 
 } // namespace Parallel

include/Algo/Geometry/centroid.hpp

@@ -187,7 +187,7 @@ void computeCentroidFaces(typename PFP::MAP& map, const V_ATT& position, F_ATT&
 	CGoGN::Parallel::foreach_cell<FACE>(map,[&](Face f, unsigned int thr)
 	{
 		face_centroid[f] = faceCentroid<PFP>(map, f, position) ;
-	},false,AUTO);
+	});
 }
 
 template <typename PFP, typename V_ATT, typename F_ATT>
@@ -196,7 +196,7 @@ void computeCentroidELWFaces(typename PFP::MAP& map, const V_ATT& position, F_AT
 	CGoGN::Parallel::foreach_cell<FACE>(map,[&](Face f, unsigned int thr)
 	{
 		face_centroid[f] = faceCentroidELW<PFP>(map, f, position) ;
-	},false,AUTO);
+	});
 
 }
 
@@ -207,7 +207,7 @@ void computeNeighborhoodCentroidVertices(typename PFP::MAP& map,
 	CGoGN::Parallel::foreach_cell<VERTEX>(map,[&](Vertex v, unsigned int thr)
 	{
 		vertex_centroid[v] = vertexNeighborhoodCentroid<PFP>(map, v, position) ;
-	},false,FORCE_CELL_MARKING);
+	},FORCE_CELL_MARKING);
 }
 
 } // namespace Parallel
@@ -297,7 +297,7 @@ void computeCentroidVolumes(typename PFP::MAP& map, const V_ATT& position, W_ATT
 	CGoGN::Parallel::foreach_cell<VOLUME>(map,[&](Vol v, unsigned int thr)
 	{
 		vol_centroid[v] = Surface::Geometry::volumeCentroid<PFP,V_ATT>(map, v, position, thr) ;
-	},true,AUTO);
+	});
 }
 
 template <typename PFP, typename V_ATT, typename W_ATT>
@@ -306,7 +306,7 @@ void computeCentroidELWVolumes(typename PFP::MAP& map, const V_ATT& position, W_
 	CGoGN::Parallel::foreach_cell<VOLUME>(map,[&](Vol v, unsigned int thr)
 	{
 		vol_centroid[v] = Surface::Geometry::volumeCentroidELW<PFP,V_ATT>(map, v, position, thr) ;
-	},true,AUTO);
+	});
 }
 
 

include/Algo/Geometry/curvature.hpp

@@ -667,7 +667,7 @@ void computeCurvatureVertices_NormalCycles(
 	CGoGN::Parallel::foreach_cell<VERTEX>(map,[&](Vertex v, unsigned int threadID)
 	{
 		computeCurvatureVertex_NormalCycles<PFP>(map, v, radius, position, normal, edgeangle, kmax, kmin, Kmax, Kmin, Knormal, threadID) ;
-	},true,FORCE_CELL_MARKING);
+	},FORCE_CELL_MARKING);
 }
 
 template <typename PFP>
@@ -696,7 +696,7 @@ void computeCurvatureVertices_NormalCycles_Projected(
 	CGoGN::Parallel::foreach_cell<VERTEX>(map,[&](Vertex v, unsigned int threadID)
 	{
 		computeCurvatureVertex_NormalCycles_Projected<PFP>(map, v, radius, position, normal, edgeangle, kmax, kmin, Kmax, Kmin, Knormal, threadID) ;
-	},true,FORCE_CELL_MARKING);
+	},FORCE_CELL_MARKING);
 }
 
 
@@ -714,7 +714,7 @@ void computeCurvatureVertices_QuadraticFitting(
 	CGoGN::Parallel::foreach_cell<VERTEX>(map,[&](Vertex v, unsigned int threadID)
 	{
 		computeCurvatureVertex_QuadraticFitting<PFP>(map, v, position, normal, kmax, kmin, Kmax, Kmin, threadID) ;
-	},true,FORCE_CELL_MARKING);
+	},FORCE_CELL_MARKING);
 
 }
 

include/Algo/Geometry/normal.hpp

@@ -235,7 +235,7 @@ void computeNormalVertices(typename PFP::MAP& map, const V_ATT& position, V_ATT&
 	CGoGN::Parallel::foreach_cell<VERTEX>(map,[&](Vertex v, unsigned int /*thr*/)
 	{
 		normal[v] = vertexNormal<PFP>(map, v, position) ;
-	},true,FORCE_CELL_MARKING);
+	},FORCE_CELL_MARKING);
 }
 
 template <typename PFP, typename V_ATT, typename F_ATT>
@@ -244,7 +244,7 @@ void computeNormalFaces(typename PFP::MAP& map, const V_ATT& position, F_ATT& no
 	CGoGN::Parallel::foreach_cell<FACE>(map,[&](Face f, unsigned int thr)
 	{
 		normal[f] = faceNormal<PFP>(map, f, position) ;
-	},true,AUTO);
+	});
 }
 
 
@@ -254,7 +254,7 @@ void computeAnglesBetweenNormalsOnEdges(typename PFP::MAP& map, const V_ATT& pos
 	CGoGN::Parallel::foreach_cell<EDGE>(map,[&](Edge e, unsigned int thr)
 	{
 		angles[e] = computeAngleBetweenNormalsOnEdge<PFP>(map, e, position) ;
-	},true,AUTO);
+	});
 }
 
 } // namespace Parallel

include/Topology/generic/attributeHandler.hpp

@@ -332,7 +332,7 @@ void foreach_attribute(ATTR& attribute, FUNC func, unsigned int nbthread)
 
 	for (unsigned int i = 0; i < nbth; ++i)
 	{
-		tfs[i] = new ThreadFunctionAttrib<FUNC>(func, vd[i], sync1,sync2,finished,1+i);
+		tfs[i] = new ThreadFunctionAttrib<FUNC>(func, vd[i], sync1,sync2,finished,i);
 		threads[i] = new boost::thread( boost::ref( *(tfs[i]) ) );
 	}
 

include/Topology/generic/cellmarker.h

@@ -45,10 +45,11 @@ class CellMarkerGen
 protected:
 	AttributeMultiVector<MarkerBool>* m_markVector ;
 	unsigned int m_cell ;
+	unsigned int m_thread;
 
 public:
 	CellMarkerGen(unsigned int cell, unsigned int thread = 0) :
-		m_cell(cell)
+		m_cell(cell),m_thread(thread)
 	{}
 
 	virtual ~CellMarkerGen()
@@ -86,7 +87,7 @@ public:
 	{
 		if(!m_map.template isOrbitEmbedded<CELL>())
 			m_map.template addEmbedding<CELL>() ;
-		m_markVector = m_map.template askMarkVector<CELL>();
+		m_markVector = m_map.template askMarkVector<CELL>(m_thread);
 	}
 
 	CellMarkerBase(const MAP& map, unsigned int thread = 0) :
@@ -95,20 +96,20 @@ public:
 	{
 		if(!m_map.template isOrbitEmbedded<CELL>())
 			m_map.template addEmbedding<CELL>() ;
-		m_markVector = m_map.template askMarkVector<CELL>();
+		m_markVector = m_map.template askMarkVector<CELL>(m_thread);
 	}
 
 	virtual ~CellMarkerBase()
 	{
 		if (GenericMap::alive(&m_map))
-			m_map.template releaseMarkVector<CELL>(m_markVector);
+			m_map.template releaseMarkVector<CELL>(m_markVector,m_thread);
 	}
 
 	void update()
 	{
 		if(!m_map.template isOrbitEmbedded<CELL>())
 			m_map.template addEmbedding<CELL>() ;
-		m_markVector = m_map.template askMarkVector<CELL>();
+		m_markVector = m_map.template askMarkVector<CELL>(m_thread);
 	}
 
 

include/Topology/generic/dartmarker.h

@@ -44,12 +44,14 @@ class DartMarkerGen
 
 protected:
 	AttributeMultiVector<MarkerBool>* m_markVector;
+	unsigned int m_thread;
 
 public:
 	/**
 	 * constructor
 	 */
-	DartMarkerGen(unsigned int thread=0)
+	DartMarkerGen(unsigned int thread=0):
+		m_thread(thread)
 	{}
 
 	virtual ~DartMarkerGen()
@@ -77,26 +79,26 @@ public:
 		DartMarkerGen(thread),
 		m_map(map)
 	{
-		m_markVector = m_map.template askMarkVector<DART>();
+		m_markVector = m_map.template askMarkVector<DART>(m_thread);
 	}
 
 	DartMarkerTmpl(const MAP& map, unsigned int thread = 0) :
 		DartMarkerGen(thread),
 		m_map(const_cast<MAP&>(map))
 	{
-		m_markVector = m_map.template askMarkVector<DART>();
+		m_markVector = m_map.template askMarkVector<DART>(m_thread);
 	}
 
 	virtual ~DartMarkerTmpl()
 	{
 		if (GenericMap::alive(&m_map))
-			m_map.template releaseMarkVector<DART>(m_markVector);
+			m_map.template releaseMarkVector<DART>(m_markVector,m_thread);
 
 	}
 
 	inline void update()
 	{
-		m_markVector = m_map.template askMarkVector<DART>();
+		m_markVector = m_map.template askMarkVector<DART>(m_thread);
 	}
 
 

include/Topology/generic/genericmap.h

@@ -114,7 +114,7 @@ protected:
 
 
 //	std::vector< AttributeMultiVector<MarkerBool>* > m_markVectors[NB_ORBITS] ;
-	std::vector< AttributeMultiVector<MarkerBool>* > m_markVectors_free[NB_ORBITS] ;
+	std::vector< AttributeMultiVector<MarkerBool>* > m_markVectors_free[NB_ORBITS][NB_THREAD] ;
 	boost::mutex m_MarkerStorageMutex[NB_ORBITS];
 
 	/**
@@ -278,13 +278,13 @@ public:
 	 * @brief ask for a marker attribute
 	 */
 	template <unsigned int ORBIT>
-	AttributeMultiVector<MarkerBool>* askMarkVector() ;
+	AttributeMultiVector<MarkerBool>* askMarkVector(unsigned int thread=0) ;
 
 	/**
 	 * @brief release allocated marker attribute
 	 */
 	template <unsigned int ORBIT>
-	void releaseMarkVector(AttributeMultiVector<MarkerBool>* amv);
+	void releaseMarkVector(AttributeMultiVector<MarkerBool>* amv, unsigned int thread=0);
 
 
 	/**

include/Topology/generic/genericmap.hpp

@@ -155,31 +155,31 @@ inline AttributeMultiVectorGen* GenericMap::getAttributeVectorGen(unsigned int o
 
 
 template <unsigned int ORBIT>
-AttributeMultiVector<MarkerBool>* GenericMap::askMarkVector()
+AttributeMultiVector<MarkerBool>* GenericMap::askMarkVector(unsigned int thread)
 {
 	assert(isOrbitEmbedded<ORBIT>() || !"Invalid parameter: orbit not embedded") ;
 
-	boost::mutex::scoped_lock lockMV(m_MarkerStorageMutex[ORBIT]);
-	if (!m_markVectors_free[ORBIT].empty())
+	if (!m_markVectors_free[ORBIT][thread].empty())
 	{
-		AttributeMultiVector<MarkerBool>* amv = m_markVectors_free[ORBIT].back();
-		m_markVectors_free[ORBIT].pop_back();
+		AttributeMultiVector<MarkerBool>* amv = m_markVectors_free[ORBIT][thread].back();
+		m_markVectors_free[ORBIT][thread].pop_back();
+		return amv;
+	}
+	else
+	{
+		boost::mutex::scoped_lock lockMV(m_MarkerStorageMutex[ORBIT]);
+		AttributeMultiVector<MarkerBool>* amv = m_attribs[ORBIT].addAttribute<MarkerBool>("") ;
 		return amv;
 	}
-	//else add attribute
-	AttributeMultiVector<MarkerBool>* amv = m_attribs[ORBIT].addAttribute<MarkerBool>("") ;
-//	std::cout << "ADD ATTRIBUTE"<< std::endl;
-	return amv;
 }
 
 
 template <unsigned int ORBIT>
-inline void GenericMap::releaseMarkVector(AttributeMultiVector<MarkerBool>* amv)
+inline void GenericMap::releaseMarkVector(AttributeMultiVector<MarkerBool>* amv, unsigned int thread)
 {
 	assert(isOrbitEmbedded<ORBIT>() || !"Invalid parameter: orbit not embedded") ;
 
-	boost::mutex::scoped_lock lockMV(m_MarkerStorageMutex[ORBIT]);
-	m_markVectors_free[ORBIT].push_back(amv);
+	m_markVectors_free[ORBIT][thread].push_back(amv);
 }
 
 

include/Topology/generic/traversor/traversorCell.h

@@ -137,7 +137,7 @@ namespace Parallel
  * @param nbth number of used thread (0:for traversal, [1,nbth-1] for func computing
 */
 template <unsigned int ORBIT, typename MAP, typename FUNC>
-void foreach_cell(MAP& map, FUNC func, bool needMarkers = true, TraversalOptim opt = AUTO, unsigned int nbth = NumberOfThreads);
+void foreach_cell(MAP& map, FUNC func, TraversalOptim opt = AUTO, unsigned int nbth = NumberOfThreads);
 
 } // namespace Parallel
 

include/Topology/generic/traversor/traversorCell.hpp

@@ -686,7 +686,7 @@ public:
 
 
 template <TraversalOptim OPT, unsigned int ORBIT, typename MAP, typename FUNC>
-void foreach_cell_tmpl(MAP& map, FUNC func, bool needMarkers, unsigned int nbth)
+void foreach_cell_tmpl(MAP& map, FUNC func, unsigned int nbth)
 {
 	// buffer for cell traversing
 	std::vector< Cell<ORBIT> >* vd = new std::vector< Cell<ORBIT> >[nbth];
@@ -759,7 +759,7 @@ void foreach_cell_tmpl(MAP& map, FUNC func, bool needMarkers, unsigned int nbth)
 }
 
 template <unsigned int ORBIT, typename MAP, typename FUNC>
-void foreach_cell(MAP& map, FUNC func, bool needMarkers, TraversalOptim opt, unsigned int nbth)
+void foreach_cell(MAP& map, FUNC func, TraversalOptim opt, unsigned int nbth)
 {
 	if (nbth < 2)
 	{
@@ -769,17 +769,17 @@ void foreach_cell(MAP& map, FUNC func, bool needMarkers, TraversalOptim opt, uns
 	switch(opt)
 	{
 		case FORCE_DART_MARKING:
-			foreach_cell_tmpl<FORCE_DART_MARKING,ORBIT,MAP,FUNC>(map,func,needMarkers,nbth-1);
+			foreach_cell_tmpl<FORCE_DART_MARKING,ORBIT,MAP,FUNC>(map,func,nbth-1);
 			break;
 		case FORCE_CELL_MARKING:
-			foreach_cell_tmpl<FORCE_CELL_MARKING,ORBIT,MAP,FUNC>(map,func,needMarkers,nbth-1);
+			foreach_cell_tmpl<FORCE_CELL_MARKING,ORBIT,MAP,FUNC>(map,func,nbth-1);
 			break;
 		case FORCE_QUICK_TRAVERSAL:
-			foreach_cell_tmpl<FORCE_QUICK_TRAVERSAL,ORBIT,MAP,FUNC>(map,func,needMarkers,nbth-1);
+			foreach_cell_tmpl<FORCE_QUICK_TRAVERSAL,ORBIT,MAP,FUNC>(map,func,nbth-1);
 			break;
 		case AUTO:
 		default:
-			foreach_cell_tmpl<AUTO,ORBIT,MAP,FUNC>(map,func,needMarkers,nbth-1);
+			foreach_cell_tmpl<AUTO,ORBIT,MAP,FUNC>(map,func,nbth-1);
 			break;
 	}
 }