Latest modules

CMakeLists.txt

@@ -153,6 +153,10 @@ else()
 	message( FATAL_ERROR "LibHilbert is required in order to install OpenFPM")
 endif()
 
+if(SUITESPARSE_FOUND)
+	set(DEFINE_HAVE_SUITESPARSE "#define HAVE_SUITESPARSE")
+endif()
+
 if(APPLE)
 	set(DEFINE_HAVE_OSX "#define HAVE_OSX")
 endif()

example/Debugging/0_classes/main.cpp

@@ -43,7 +43,7 @@ int main(int argc, char* argv[])
 	// randomly in the domain, we create a Box that define our domain, boundary conditions and ghost
 	//
 	openfpm_init(&argc,&argv);
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 	
 	typedef Point<2,float> s;
 
@@ -87,7 +87,7 @@ int main(int argc, char* argv[])
 	// decomposition
 	//
 	{
-		vector_dist<2,float, Point_test<float>, CartDecomposition<2,float> > vd(4096,box,bc,g);
+		vector_dist<2,float, Point_test<float> > vd(4096,box,bc,g);
 
 		//
 		// ### WIKI 6 ###
@@ -112,8 +112,8 @@ int main(int argc, char* argv[])
 	// we create, now two of them using new
 	//
 
-	vector_dist<2,float, Point_test<float>, CartDecomposition<2,float> > * vd1 = new vector_dist<2,float, Point_test<float>, CartDecomposition<2,float> >(4096,box,bc,g);
-	vector_dist<2,float, Point_test<float>, CartDecomposition<2,float> > * vd2 = new vector_dist<2,float, Point_test<float>, CartDecomposition<2,float> >(4096,box,bc,g);
+	vector_dist<2,float, Point_test<float> > * vd1 = new vector_dist<2,float, Point_test<float>, CartDecomposition<2,float> >(4096,box,bc,g);
+	vector_dist<2,float, Point_test<float> > * vd2 = new vector_dist<2,float, Point_test<float>, CartDecomposition<2,float> >(4096,box,bc,g);
 
 	//
 	// ### WIKI 8 ###

example/Debugging/1_classes/main.cpp

@@ -36,7 +36,7 @@ int main(int argc, char* argv[])
 	// * throw_on_allocation throw when one allocation is reached, producing the termination of the program and a core dump (if no try catch is set-up)
 	//
 	openfpm_init(&argc,&argv);
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 
 	throw_on_alloc(10);
 	// message_on_alloc(10);

example/Grid/0_simple/main.cpp

@@ -270,7 +270,7 @@ PX,Y Processor X, sub-domain Y</pre><p>The point # has</p>
 	//! \cond [reduce] \endcond
 
 	// Get the VCluster object
-	Vcluster & vcl = create_vcluster();
+	Vcluster<> & vcl = create_vcluster();
 
 	// queue an operation of sum for the counter count
 	vcl.sum(count);

example/Grid/2_solve_eq/main.cpp

@@ -393,7 +393,7 @@ int main(int argc, char* argv[])
 
 	// Get the maximum across processor to calculate the norm infinity of the error
 	// Norm infinity of the error is the maximum error across all the grid points
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 	v_cl.max(error);
 	v_cl.execute();
 

example/Numerics/Vortex_in_cell/main_vic_petsc_opt.cpp

@@ -406,7 +406,7 @@ void comp_vel(grid_type_s & gr_ps,
 		solError serr;
 		serr = solver.get_residual_error(phi_s[i],b);
 
-		Vcluster & v_cl = create_vcluster();
+		Vcluster<> & v_cl = create_vcluster();
 		if (v_cl.getProcessUnitID() == 0)
 		{std::cout << "Solved component " << i << "  Error: " << serr.err_inf << std::endl;}
 
@@ -665,7 +665,7 @@ template<typename vector, typename grid> void check_point_and_save(vector & part
 																   grid & g_dvort,
 																   size_t i)
 {
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 
 	if (v_cl.getProcessingUnits() < 24)
 	{
@@ -775,7 +775,7 @@ int main(int argc, char* argv[])
 	Vector<double,PETSC_BASE> x_;
 
 	// Parallel object
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 
 	// print out the spacing of the grid
 	if (v_cl.getProcessUnitID() == 0)

example/VCluster/0_simple/main.cpp

@@ -61,7 +61,7 @@ int main(int argc, char* argv[])
 
 	//! \cond [create] \endcond
 
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 	long int N_prc = v_cl.getProcessingUnits();
 
 	//! \cond [create] \endcond

example/VCluster/1_semantic/main.cpp

@@ -39,7 +39,7 @@ int main(int argc, char* argv[])
 	//! \cond [initialization] \endcond
 
 	openfpm_init(&argc,&argv);
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 	
 	//! \cond [initialization] \endcond
 

example/VCluster/2_serial_and_parallel/main.cpp

@@ -124,7 +124,7 @@ int main(int argc, char* argv[])
 	//! \cond [initialization] \endcond
 
 	openfpm_init(&argc,&argv);
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 	
 	//! \cond [initialization] \endcond
 

example/Vector/1_celllist/main.cpp

@@ -34,7 +34,7 @@ int main(int argc, char* argv[])
 	//! \cond [Initialization and parameters] \endcond
 
 	openfpm_init(&argc,&argv);
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 
 	// we will place the particles on a grid like way with 128 particles on each direction
 	size_t sz[3] = {128,128,128};

example/Vector/3_molecular_dynamic_gpu_opt/Makefile

@@ -5,6 +5,7 @@ CC=nvcc #${CC_SCOREP}
 CC_MPI=mpic++
 
 LDIR =
+OPT=
 
 OBJ_GPU = main_gpu.o
 OBJ_CPU = main_cpu.o
@@ -12,14 +13,17 @@ OBJ_CPU_BEST = main_cpu_best.o
 
 all: md_dyn_gpu md_dyn_cpu md_dyn_cpu_best
 
+md_dyn_gpu_test: OPT += -DTEST_RUN
+md_dyn_gpu_test: all
+
 %.o: %.cu
-	$(CC) -O3 -g -c -isystem=/home/i-bird/MPI/include --std=c++11 -o $@ $< $(INCLUDE_PATH_NVCC)
+	$(CC) $(OPT)  -O3 -g -c -isystem=/home/i-bird/MPI/include --std=c++11 -o $@ $< $(INCLUDE_PATH_NVCC)
 
 %.o: %.cpp
-	$(CC_MPI) -O3 -g -c --std=c++11 -o $@ $< $(INCLUDE_PATH)
+	$(CC_MPI) $(OPT) -O3 -g -c --std=c++11 -o $@ $< $(INCLUDE_PATH)
 
 md_dyn_gpu: $(OBJ_GPU)
-	$(CC) -o $@ $^ $(CFLAGS) $(LIBS_PATH_NVCC) $(LIBS) -L/home/i-bird/MPI/lib  -L/usr/local/cuda/lib64 -lcudart -lmpi -L/usr/local/cuda/extras/CUPTI/lib64 -lhdf5
+	$(CC) -o $@ $^ $(CFLAGS) $(LIBS_PATH) $(LIBS) -L/home/i-bird/MPI/lib  -L/usr/local/cuda/lib64 -lcudart -lmpi -L/usr/local/cuda/extras/CUPTI/lib64 -lhdf5
 
 md_dyn_cpu: $(OBJ_CPU)
 	$(CC_MPI) -o $@ $^ $(CFLAGS) $(LIBS_PATH) $(LIBS)
@@ -27,7 +31,7 @@ md_dyn_cpu: $(OBJ_CPU)
 md_dyn_cpu_best: $(OBJ_CPU_BEST)
 	$(CC_MPI) -o $@ $^ $(CFLAGS) $(LIBS_PATH) $(LIBS)
 
-run: all
+run: md_dyn_gpu_test
 	mpirun -np 3 ./md_dyn_gpu && mpirun -np 3 ./md_dyn_cpu && mpirun -np 3 ./md_dyn_cpu_best;
 
 .PHONY: clean all run

example/Vector/4_complex_prop/Makefile

@@ -10,7 +10,7 @@ OBJ_SER = main_ser.o
 all: vect_cp vect_ser
 
 %.o: %.cpp
-	$(CC) -O0 -c --std=c++11 -o $@ $< $(INCLUDE_PATH)
+	$(CC) -O3 -c --std=c++11 -o $@ $< $(INCLUDE_PATH)
 
 vect_cp: $(OBJ)
 	$(CC) -o $@ $^ $(CFLAGS) $(LIBS_PATH) $(LIBS)

example/Vector/4_complex_prop/main.cpp

@@ -284,7 +284,7 @@ int main(int argc, char* argv[])
 
 	size_t fg = vd.size_local();
 
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 	if (v_cl.getProcessUnitID() == 0)
 	{
 		for ( ; fg < vd.size_local()+4 ; fg++)

example/Vector/4_complex_prop/main_ser.cpp

@@ -504,7 +504,7 @@ int main(int argc, char* argv[])
 
 	size_t fg = vd.size_local();
 
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 
 	// Only the master processor print
 	if (v_cl.getProcessUnitID() == 0)

example/Vector/4_multiphase_celllist_verlet/main.cpp

@@ -43,7 +43,7 @@ int main(int argc, char* argv[])
 	openfpm_init(&argc,&argv);
 
 	// Vcluster for general usage
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 
 	// The domain
 	Box<3,float> box({0.0,0.0,0.0},{1.0,1.0,1.0});

example/Vector/4_reorder/main_comp_ord.cpp

@@ -235,7 +235,7 @@ int main(int argc, char* argv[])
 	openfpm::vector<openfpm::vector<double>> y;
 
 	openfpm_init(&argc,&argv);
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 
 	// we will use it do place particles on a 40x40x40 Grid like
 	size_t sz[3] = {40,40,40};

example/Vector/4_reorder/main_data_ord.cpp

@@ -50,7 +50,7 @@ int main(int argc, char* argv[])
 	openfpm::vector<openfpm::vector<double>> y;
 
 	openfpm_init(&argc,&argv);
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 
 	// we will use it do place particles on a 10x10x10 Grid like
 	size_t sz[3] = {40,40,40};

example/Vector/5_molecular_dynamic_sym/main.cpp

@@ -284,7 +284,7 @@ int main(int argc, char* argv[])
 	openfpm::vector<openfpm::vector<double>> y;
 
 	openfpm_init(&argc,&argv);
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 
 	// we will use it do place particles on a 10x10x10 Grid like
 	size_t sz[3] = {10,10,10};

example/Vector/5_molecular_dynamic_sym_crs/main.cpp

@@ -357,7 +357,7 @@ int main(int argc, char* argv[])
 	openfpm::vector<openfpm::vector<double>> y;
 
 	openfpm_init(&argc,&argv);
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 
 	// we will use it do place particles on a 10x10x10 Grid like
 	size_t sz[3] = {10,10,10};

example/Vector/7_SPH_dlb_gpu/Makefile

@@ -12,7 +12,7 @@ sph_dlb_test: OPT += -DTEST_RUN
 sph_dlb_test: sph_dlb
 
 %.o: %.cu
-	nvcc -O3 -g -c -isystem=/home/i-bird/MPI/include --std=c++11 -o $@ $< $(INCLUDE_PATH_NVCC)
+	nvcc -O3 $(OPT) -g -c -isystem=/home/i-bird/MPI/include --std=c++11 -o $@ $< $(INCLUDE_PATH_NVCC)
 
 %.o: %.cpp
 	$(CC) -O3 $(OPT) -g -c --std=c++11 -o $@ $< $(INCLUDE_PATH)

example/Vector/7_SPH_dlb_gpu/main.cu

@@ -438,7 +438,7 @@ template<typename CellList> inline void calc_forces(particles & vd, CellList & N
 
 	calc_forces_gpu<<<part.wthr,part.thr>>>(vd.toKernel(),NN.toKernel(),W_dap,cbar);
 
-	max_visc = reduce<red,_max_>(vd);
+	max_visc = reduce_local<red,_max_>(vd);
 }
 
 template<typename vector_type>
@@ -460,8 +460,8 @@ void max_acceleration_and_velocity(particles & vd, real_number & max_acc, real_n
 
 	max_acceleration_and_velocity_gpu<<<part.wthr,part.thr>>>(vd.toKernel());
 
-	max_acc = reduce<red,_max_>(vd);
-	max_vel = reduce<red2,_max_>(vd);
+	max_acc = reduce_local<red,_max_>(vd);
+	max_vel = reduce_local<red2,_max_>(vd);
 
 	Vcluster<> & v_cl = create_vcluster();
 	v_cl.max(max_acc);

example/Vector/7_SPH_dlb_gpu_opt/Makefile

@@ -12,7 +12,7 @@ sph_dlb_test: OPT += -DTEST_RUN
 sph_dlb_test: sph_dlb
 
 %.o: %.cu
-	nvcc -O3 -g -c -isystem=/home/i-bird/MPI/include --std=c++11 -o $@ $< $(INCLUDE_PATH_NVCC)
+	nvcc -O3 $(OPT)  -g -c -isystem=/home/i-bird/MPI/include --std=c++11 -o $@ $< $(INCLUDE_PATH_NVCC)
 
 %.o: %.cpp
 	$(CC) -O3 $(OPT) -g -c --std=c++11 -o $@ $< $(INCLUDE_PATH)

example/Vector/7_SPH_dlb_opt/main.cpp

@@ -641,7 +641,7 @@ void max_acceleration_and_velocity(particles & vd, double & max_acc, double & ma
 	max_acc = sqrt(max_acc);
 	max_vel = sqrt(max_vel);
 
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 	v_cl.max(max_acc);
 	v_cl.max(max_vel);
 	v_cl.execute();
@@ -757,7 +757,7 @@ void verlet_int(particles & vd, double dt, double & max_disp)
 
 	/*! \cond [max_across_proc] \endcond */
 
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 	v_cl.max(max_disp);
 	v_cl.execute();
 
@@ -848,7 +848,7 @@ void euler_int(particles & vd, double dt, double & max_disp)
 		++part;
 	}
 
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 	v_cl.max(max_disp);
 	v_cl.execute();
 
@@ -1040,7 +1040,7 @@ int main(int argc, char* argv[])
 	double max_disp;
 	while (t <= t_end)
 	{
-		Vcluster & v_cl = create_vcluster();
+		Vcluster<> & v_cl = create_vcluster();
 		timer it_time;
 
 		/*! \cond [update_verlet] \endcond */

example/Vector/8_DEM/main.cpp

@@ -194,7 +194,7 @@ int main(int argc, char* argv[])
 #endif
 	double mu = 0.5;
 	double max_vel;
-	Vcluster & v_cl = create_vcluster();
+	Vcluster<> & v_cl = create_vcluster();
 
 	Box<3,double> domain({0.0,0.0,0.0},{17.04,6.0,6.6});
 	Ghost<3,double> g(cutoff + skin);

openfpm_numerics @ 1f2e7ab7

-Subproject commit e9a2083b6c4d9ce84519bc8163fbff032db4ff1c
+Subproject commit 1f2e7ab78f971a280cdeba27105beeafd88c7ab1