// foreach & tutos

Apps/Tuto/Traversals/CMakeLists.txt

@@ -42,6 +42,9 @@ target_link_libraries( traverse_cells ${CGoGN_LIBS_D} ${CGoGN_EXT_LIBS} )
 add_executable( traverse_neighbours traverse_neighbours.cpp)
 target_link_libraries( traverse_neighbours ${CGoGN_LIBS_D} ${CGoGN_EXT_LIBS} )
 
+add_executable( traverse_attributes traverse_attributes.cpp)
+target_link_libraries( traverse_attributes ${CGoGN_LIBS_D} ${CGoGN_EXT_LIBS} )
+
 QT4_WRAP_UI( show_traversors_ui2 show_traversors2.ui )
 QT4_WRAP_CPP(show_traversors_moc2 show_traversors2.h)
 add_executable( show_traversors2 show_traversors2.cpp ${show_traversors_ui2} ${show_traversors_moc2})

Apps/Tuto/Traversals/traverse_attributes.cpp

+/*******************************************************************************
+* CGoGN: Combinatorial and Geometric modeling with Generic N-dimensional Maps  *
+* version 0.1                                                                  *
+* Copyright (C) 2009-2012, IGG Team, LSIIT, University of Strasbourg           *
+*                                                                              *
+* This library is free software; you can redistribute it and/or modify it      *
+* under the terms of the GNU Lesser General Public License as published by the *
+* Free Software Foundation; either version 2.1 of the License, or (at your     *
+* option) any later version.                                                   *
+*                                                                              *
+* This library is distributed in the hope that it will be useful, but WITHOUT  *
+* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or        *
+* FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License  *
+* for more details.                                                            *
+*                                                                              *
+* You should have received a copy of the GNU Lesser General Public License     *
+* along with this library; if not, write to the Free Software Foundation,      *
+* Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301 USA.           *
+*                                                                              *
+* Web site: http://cgogn.unistra.fr/                                           *
+* Contact information: cgogn@unistra.fr                                        *
+*                                                                              *
+*******************************************************************************/
+
+
+#include "Topology/generic/parameters.h"
+#include "Topology/map/embeddedMap2.h"
+#include "Algo/Tiling/Surface/square.h"
+
+using namespace CGoGN ;
+
+/**
+ * Struct that contains some informations about the types of the manipulated objects
+ * Mainly here to be used by the algorithms that are parameterized by it
+ */
+struct PFP: public PFP_STANDARD
+{
+	// definition of the type of the map
+	typedef EmbeddedMap2 MAP;
+};
+
+// some typedef shortcuts
+typedef PFP::MAP MAP ;				// map type
+typedef PFP::MAP::IMPL MAP_IMPL ;	// map implementation
+typedef PFP::VEC3 VEC3 ;			// type of R³ vector 
+
+
+int main()
+{
+	// declare a map to handle the mesh
+	MAP myMap;
+
+	// add position attribute on vertices and get handler on it
+	VertexAttribute<VEC3, MAP_IMPL> position = myMap.addAttribute<VEC3, VERTEX>("position");
+
+	// create a topo grid of 2x2 squares
+	Algo::Surface::Tilings::Square::Grid<PFP> grid(myMap, 4, 4, true);
+
+	// and embed it using position attribute
+    grid.embedIntoGrid(position, 1.,1.,0.);
+
+	
+	// traversal with begin/end/next on attribute handler
+	for (unsigned int id=position.begin(); id !=position.end(); position.next(id))
+	{
+		std::cout << id << " : " << position[id]<< " / ";
+	};
+	std::cout << std::endl;
+
+	//using foreach function (C++11 lambda expression)
+	foreach_attribute(position,[&](unsigned int id) // for each element of position
+	{
+		std::cout << id << " : " << position[id]<< " / ";
+	});
+
+	// using parallel foreach
+	// parameter position must be captured explicitly even if it used as first parameter of foreach !
+	Parallel::foreach_attribute(position,[&position](unsigned int id, unsigned int thread) // for each elt of the position attribute
+	{
+		position[id] *= 2.0f;
+	},4); // 4:4 thread, false for no need for markers in threaded code.
+
+
+
+	return 0;
+}

Apps/Tuto/Traversals/traverse_cells.cpp

@@ -26,7 +26,7 @@
 #include "Topology/generic/parameters.h"
 #include "Topology/map/embeddedMap2.h"
 #include "Algo/Tiling/Surface/square.h"
-
+#include "Algo/Geometry/area.h"
 
 using namespace CGoGN ;
 
@@ -55,13 +55,11 @@ int main()
 	VertexAttribute<VEC3, MAP_IMPL> position = myMap.addAttribute<VEC3, VERTEX>("position");
 
 	// create a topo grid of 2x2 squares
-	Algo::Surface::Tilings::Square::Grid<PFP> grid(myMap, 2, 2, true);
+	Algo::Surface::Tilings::Square::Grid<PFP> grid(myMap, 4, 4, true);
 
 	// and embed it using position attribute
     grid.embedIntoGrid(position, 1.,1.,0.);
 
-
-	
 	// ALGO:
 	// foreach vertex of the map
 	//    print position of the vertex
@@ -106,9 +104,41 @@ int main()
 	{
 		std::cout << v << " : " << position[v]<< " / ";
 	});
+	// warning here v is a Vertex and not a Dart (but can cast automatically into)
 
 	std::cout <<  std::endl;
-	// warning here v is a Vertex and not a Dart (but can cast automatically into)
+
+
+	// example of parallel traversal of cells
+	Parallel::foreach_cell<VERTEX>(myMap,[&](Vertex v, unsigned int thread) // for each Vertex v of the MAP myMap
+	{
+		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) !!!!
+	},4,false); // 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
+	{
+		std::cout << position[v]<< " / ";
+	});
+	std::cout <<  std::endl;
+
+
+	// Example with // accumulation
+	// computing the sum of area faces
+	// init nbthread values with 0
+	float surf[4]={0.0f,0.0f,0.0f,0.0f};
+	// traverse face in //
+	Parallel::foreach_cell<FACE>(myMap,[&](Face f, unsigned int thr)
+	{
+		// for each face add surface to accumulator (-1 because counter between 1-4 not 0-3)
+		surf[thr-1] += Algo::Surface::Geometry::convexFaceArea<PFP>(myMap,f,position);
+	},4,false);
+
+	std::cout << surf[0]<< "/"<< surf[1]<< "/"<< surf[2]<< "/"<< surf[3]<< "/"<< std::endl;
+	std::cout << "Total="<<surf[0]+surf[1]+surf[2]+surf[3]<< std::endl;
+
 
 	return 0;
 }

include/Algo/Parallel/parallel_foreach.hpp

@@ -163,66 +163,15 @@ void foreach_cell(MAP& map, std::vector<FunctorMapThreaded<MAP>*>& funcs, bool n
 	for (unsigned int i = 0; i < nbth; ++i)
 		vd[i].reserve(SIZE_BUFFER_THREAD);
 
-	AttributeContainer* cont = NULL;
-	DartMarker<MAP>* dmark = NULL;
-	CellMarker<MAP, ORBIT>* cmark = NULL;
-	const AttributeMultiVector<Dart>* quickTraversal = map.template getQuickTraversal<ORBIT>() ;
-
-	// fill each vd buffers with SIZE_BUFFER_THREAD darts
-	Dart d;
-	unsigned int di=0;
-
-	if(quickTraversal != NULL)
+	unsigned int nb=0;
+	TraversorCell<MAP, ORBIT> trav(map);
+	Cell<ORBIT> cell = trav.begin();
+	Cell<ORBIT> c_end = trav.end();
+	while ((cell.dart != c_end.dart) && (nb < nbth*SIZE_BUFFER_THREAD) )
 	{
-		cont = &(map.template getAttributeContainer<ORBIT>()) ;
-
-		di = cont->begin();
-		unsigned int nb = 0;
-		while ((di != cont->end()) && (nb < nbth*SIZE_BUFFER_THREAD) )
-		{
-			d = quickTraversal->operator[](di);
-
-			vd[nb%nbth].push_back(d);
-			nb++;
-
-			cont->next(di);
-		}
-	}
-	else
-	{
-		if(map.template isOrbitEmbedded<ORBIT>())
-		{
-			cmark = new CellMarker<MAP, ORBIT>(map) ;
-
-			d = map.begin();
-			unsigned int nb = 0;
-			while ((d != map.end()) && (nb < nbth*SIZE_BUFFER_THREAD) )
-			{
-				if ((!map.template isBoundaryMarked<MAP::DIMENSION>(d)) && (!cmark->isMarked(d)))
-				{
-					cmark->mark(d);
-					vd[nb%nbth].push_back(d);
-					nb++;
-				}
-				map.next(d);
-			}
-		}
-		else
-		{
-			dmark = new DartMarker<MAP>(map) ;
-			d = map.begin();
-			unsigned int nb = 0;
-			while ((d != map.end()) && (nb < nbth*SIZE_BUFFER_THREAD) )
-			{
-				if ((!map.template isBoundaryMarked<MAP::DIMENSION>(d)) && (!dmark->isMarked(d)))
-				{
-					dmark->template markOrbit<ORBIT>(d);
-					vd[nb%nbth].push_back(d);
-					nb++;
-				}
-				map.next(d);
-			}
-		}
+		vd[nb%nbth].push_back(cell.dart);
+		nb++;
+		cell = trav.next();
 	}
 
 	boost::barrier sync1(nbth+1);
@@ -244,8 +193,6 @@ void foreach_cell(MAP& map, std::vector<FunctorMapThreaded<MAP>*>& funcs, bool n
 		tfs[i] = new ThreadFunction<MAP>(funcs[i], vd[i],sync1,sync2, finished,1+i);
 		threads[i] = new boost::thread( boost::ref( *(tfs[i]) ) );
 	}
-//		threads[i] = new boost::thread(ThreadFunction<MAP>(funcs[i], vd[i],sync1,sync2, finished,1+i));
-
 
 	// and continue to traverse the map
 	std::vector<Dart>* tempo = new std::vector<Dart>[nbth];
@@ -253,79 +200,27 @@ void foreach_cell(MAP& map, std::vector<FunctorMapThreaded<MAP>*>& funcs, bool n
 	for (unsigned int i = 0; i < nbth; ++i)
 		tempo[i].reserve(SIZE_BUFFER_THREAD);
 
-
-	if (cont)
+	while (cell.dart != c_end.dart)
 	{
-		while (di != cont->end())
-		{
-			for (unsigned int i = 0; i < nbth; ++i)
-				tempo[i].clear();
-			unsigned int nb = 0;
-			while ((di != cont->end()) && (nb < nbth*SIZE_BUFFER_THREAD) )
-			{
-				d = quickTraversal->operator[](di);
-
-				tempo[nb%nbth].push_back(d);
-				nb++;
+		for (unsigned int i = 0; i < nbth; ++i)
+			tempo[i].clear();
+		nb = 0;
 
-				cont->next(di);
-			}
-			sync1.wait();
-			for (unsigned int i = 0; i < nbth; ++i)
-				vd[i].swap(tempo[i]);
-			sync2.wait();
-		}
-	}
-	else if (cmark)
-	{
-		while (d != map.end())
+		while ((cell.dart != c_end.dart) && (nb < nbth*SIZE_BUFFER_THREAD) )
 		{
-			for (unsigned int i = 0; i < nbth; ++i)
-				tempo[i].clear();
-			unsigned int nb = 0;
-			while ((d != map.end()) && (nb < nbth*SIZE_BUFFER_THREAD) )
-			{
-				if (!map.template isBoundaryMarked<MAP::DIMENSION>(d) && !cmark->isMarked(d))
-				{
-					cmark->mark(d);
-					tempo[nb%nbth].push_back(d);
-					nb++;
-				}
-				map.next(d);
-			}
-			sync1.wait();
-			for (unsigned int i = 0; i < nbth; ++i)
-				vd[i].swap(tempo[i]);
-			sync2.wait();
-		}
-	}
-	else
-	{
-		while (d != map.end())
-		{
-			for (unsigned int i = 0; i < nbth; ++i)
-				tempo[i].clear();
-			unsigned int nb = 0;
-			while ((d != map.end()) && (nb < nbth*SIZE_BUFFER_THREAD) )
-			{
-				if ((!map.template isBoundaryMarked<MAP::DIMENSION>(d)) && (!dmark->isMarked(d)))
-				{
-					dmark->template markOrbit<ORBIT>(d);
-					tempo[nb%nbth].push_back(d);
-					nb++;
-				}
-				map.next(d);
-			}
-			sync1.wait();
-			for (unsigned int i = 0; i < nbth; ++i)
-				vd[i].swap(tempo[i]);
-			sync2.wait();
+			tempo[nb%nbth].push_back(cell);
+			nb++;
+			cell = trav.next();
 		}
+		sync1.wait(); // wait for all thread to finish its vector
+		for (unsigned int i = 0; i < nbth; ++i)
+			vd[i].swap(tempo[i]);
+		sync2.wait(); // everybody refilled then go
 	}
 
-	sync1.wait();
-	finished = true;
-	sync2.wait();
+	sync1.wait();// wait for all thread to finish its vector
+	finished = true; // say finsih to everyone
+	sync2.wait(); //wait to everybody to say finished !
 
 	//wait for all theads to be finished
 	for (unsigned int i = 0; i < nbth; ++i)
@@ -338,12 +233,6 @@ void foreach_cell(MAP& map, std::vector<FunctorMapThreaded<MAP>*>& funcs, bool n
 	delete[] threads;
 	delete[] vd;
 	delete[] tempo;
-
-	if (cmark != NULL)
-		delete cmark;
-
-	if (dmark != NULL)
-		delete dmark;
 }
 
 template <typename MAP>

include/Topology/generic/attributeHandler.h

@@ -324,6 +324,23 @@ assert(sizeof(FROM_T) == sizeof(TO_T));
 ::new(p) TO_T(); // use of placement new
 }
 
+/**
+ * apply function on each element of attribute
+ * Warning attr must also be captured by lambda funct
+ */
+template <typename ATTR, typename FUNC>
+inline void foreach_attribute(ATTR& attr, FUNC func)
+{
+	for (unsigned int id=attr.begin(); id != attr.end(); attr.next(id))
+		func(id);
+}
+
+namespace Parallel
+{
+template <typename ATTR, typename FUNC>
+void foreach_attribute(ATTR& attribute, FUNC func, unsigned int nbth=0);
+}
+
 } // namespace CGoGN
 
 #include "Topology/generic/attributeHandler.hpp"

include/Topology/generic/attributeHandler.hpp

@@ -22,6 +22,10 @@
 *                                                                              *
 *******************************************************************************/
 
+#include <boost/thread.hpp>
+#include <boost/thread/barrier.hpp>
+#include <vector>
+
 namespace CGoGN
 {
 
@@ -263,4 +267,116 @@ inline void AttributeHandler<T, ORBIT, MAP_IMPL>::next(unsigned int& iter) const
 	m_map->template getAttributeContainer<ORBIT>().next(iter) ;
 }
 
+
+namespace Parallel
+{
+
+template <typename FUNC>
+class ThreadFunctionAttrib
+{
+protected:
+	std::vector<unsigned int>& m_ids;
+	boost::barrier& m_sync1;
+	boost::barrier& m_sync2;
+	bool& m_finished;
+	unsigned int m_id;
+	FUNC m_lambda;
+public:
+	ThreadFunctionAttrib(FUNC func, std::vector<unsigned int>& vid, boost::barrier& s1, boost::barrier& s2, bool& finished, unsigned int id):
+		m_ids(vid), m_sync1(s1), m_sync2(s2), m_finished(finished), m_id(id), m_lambda(func)
+	{
+	}
+
+	ThreadFunctionAttrib(const ThreadFunctionAttrib& tf):
+		m_ids(tf.m_ids), m_sync1(tf.m_sync1), m_sync2(tf.m_sync2), m_finished(tf.m_finished), m_id(tf.m_id), m_lambda(tf.m_lambda){}
+
+	void operator()()
+	{
+		while (!m_finished)
+		{
+			for (std::vector<unsigned int>::const_iterator it = m_ids.begin(); it != m_ids.end(); ++it)
+				m_lambda(*it,m_id);
+			m_sync1.wait();
+			m_sync2.wait();
+		}
+	}
+};
+
+
+
+template <typename ATTR, typename FUNC>
+void foreach_attribute(ATTR& attribute, FUNC func, unsigned int nbth)
+{
+	if (nbth==0)
+		nbth = boost::thread::hardware_concurrency();
+
+	std::vector< unsigned int >* vd = new std::vector< unsigned int >[nbth];
+
+	for (unsigned int i = 0; i < nbth; ++i)
+		vd[i].reserve(SIZE_BUFFER_THREAD);
+
+	unsigned int nb = 0;
+	unsigned int attIdx = attribute.begin();
+	while ((attIdx != attribute.end()) && (nb < nbth*SIZE_BUFFER_THREAD) )
+	{
+		vd[nb%nbth].push_back(attIdx);
+		nb++;
+		attribute.next(attIdx);
+	}
+	boost::barrier sync1(nbth+1);
+	boost::barrier sync2(nbth+1);
+	bool finished=false;
+
+
+	boost::thread** threads = new boost::thread*[nbth];
+	ThreadFunctionAttrib<FUNC>** tfs = new ThreadFunctionAttrib<FUNC>*[nbth];
+
+	for (unsigned int i = 0; i < nbth; ++i)
+	{
+		tfs[i] = new ThreadFunctionAttrib<FUNC>(func, vd[i], sync1,sync2,finished,1+i);
+		threads[i] = new boost::thread( boost::ref( *(tfs[i]) ) );
+	}
+
+	// and continue to traverse the map
+	std::vector< unsigned int >* tempo = new std::vector< unsigned int >[nbth];
+	for (unsigned int i = 0; i < nbth; ++i)
+		tempo[i].reserve(SIZE_BUFFER_THREAD);
+
+	while (attIdx != attribute.end())
+	{
+		for (unsigned int i = 0; i < nbth; ++i)
+			tempo[i].clear();
+		unsigned int nb = 0;
+
+		while ((attIdx != attribute.end()) && (nb < nbth*SIZE_BUFFER_THREAD) )
+		{
+			tempo[nb%nbth].push_back(attIdx);
+			nb++;
+			attribute.next(attIdx);
+		}
+		sync1.wait();// wait for all thread to finish its vector
+		for (unsigned int i = 0; i < nbth; ++i)
+			vd[i].swap(tempo[i]);
+		sync2.wait();// everybody refilled then go
+	}
+
+	sync1.wait();// wait for all thread to finish its vector
+	finished = true;// say finsih to everyone
+	sync2.wait(); // just wait for last barrier wait !
+
+	//wait for all theads to be finished
+	for (unsigned int i = 0; i < nbth; ++i)
+	{
+		threads[i]->join();
+		delete threads[i];
+		delete tfs[i];
+	}
+	delete[] tfs;
+	delete[] threads;
+	delete[] vd;
+	delete[] tempo;
+}
+
+}
+
 } //namespace CGoGN

include/Topology/generic/cells.h

@@ -65,6 +65,12 @@ typedef Cell<EDGE>   Edge;
 typedef Cell<FACE>   Face;
 typedef Cell<VOLUME> Vol;  // not Volume because of the namespace Volume
 
+
+namespace Parallel
+{
+const unsigned int SIZE_BUFFER_THREAD = 8192;
+}
+
 }
 
 #endif /* CELLS_H_ */

include/Topology/generic/traversor/traversorCell.h

@@ -87,7 +87,11 @@ inline void foreach_cell_until(const MAP& map, FUNC f, bool forceDartMarker = fa
 			break;
 }
 
-
+namespace Parallel
+{
+template <unsigned int ORBIT, typename MAP, typename FUNC>
+void foreach_cell(MAP& map, FUNC func, unsigned int nbth=0, bool needMarkers=true);
+}
 
 template <typename MAP>
 class TraversorV : public TraversorCell<MAP, VERTEX>
@@ -121,6 +125,9 @@ public:
 	{}
 };
 
+
+
+
 } // namespace CGoGN
 
 #include "Topology/generic/traversor/traversorCell.hpp"

include/Topology/generic/traversor/traversorCell.hpp

@@ -21,6 +21,9 @@
 * Contact information: cgogn@unistra.fr                                        *
 *                                                                              *
 *******************************************************************************/
+#include <boost/thread.hpp>
+#include <boost/thread/barrier.hpp>
+#include <vector>
 
 namespace CGoGN
 {
@@ -161,4 +164,129 @@ void TraversorCell<MAP, ORBIT>::skip(Cell<ORBIT> c)
 		cmark->mark(c) ;
 }
 
+
+
+namespace Parallel
+{
+
+/// internal functor for boost call
+template <unsigned int ORBIT, typename FUNC>
+class ThreadFunction
+{
+protected:
+	typedef Cell<ORBIT> CELL;
+	std::vector<CELL>& m_cells;
+	boost::barrier& m_sync1;
+	boost::barrier& m_sync2;
+	bool& m_finished;
+	unsigned int m_id;
+	FUNC m_lambda;
+public:
+	ThreadFunction(FUNC func, std::vector<CELL>& vd, boost::barrier& s1, boost::barrier& s2, bool& finished, unsigned int id):
+		m_cells(vd), m_sync1(s1), m_sync2(s2), m_finished(finished), m_id(id), m_lambda(func)
+	{
+	}
+
+	ThreadFunction(const ThreadFunction<ORBIT, FUNC>& tf):
+		m_cells(tf.m_cells), m_sync1(tf.m_sync1), m_sync2(tf.m_sync2), m_finished(tf.m_finished), m_id(tf.m_id), m_lambda(tf.m_lambda){}
+
+	void operator()()
+	{
+		while (!m_finished)
+		{
+			for (typename std::vector<CELL>::const_iterator it = m_cells.begin(); it != m_cells.end(); ++it)
+				m_lambda(*it,m_id);
+			m_sync1.wait(); // wait every body has finished
+			m_sync2.wait(); // wait vectors has been refilled
+		}
+	}
+};
+