Change jobs to print more informations and change timers.

Doc/noDist/implicit/implicit.org

@@ -496,6 +496,9 @@ Mais c'est données n'impliquent pas de forcément des transitions de données m
 - Reflexion à propos du graphe de flux de données pour Treematch
 - Ajout de tests avec le noyau Chebyshev et la versin mpi implicite
 - Symétriser l'algorithme implicite au niveau des P2P
+- Modifier les jobs pour qu'il utilisent la même graine et générent le même ensemble de particules
+- Post traiter les traces d'une exécution pour créer des graphiques.
+      - Exploiter les scripts de Samuel
 
 
 ** Et après ?
@@ -516,8 +519,10 @@ Mais c'est données n'impliquent pas de forcément des transitions de données m
 	- Ne pas allouer les cellules symboliques si ce n'est pas necessaire
 	- Distribuer l'arbre
 - Post traiter les traces d'une exécution pour créer des graphiques.
+      - Modifier la génération du trace.rec
+            - Pour récupérer le temps d'attente active des comm mpi
+            - Pour qu'il réagisse bien en distribué
 - Valider les résultats mpi explicite avec l'algorithme sequentiel plutôt que l'algorithme mpi non bloqué
 	- Reproduire l'arbre globale
 	- Ne comparer que les cellules du nœud
-- Modifier les jobs pour qu'il utilisent la même graine et générent le même ensemble de particules
 - État de l'art load balancing sur non-uniforme

Tests/GroupTree/testBlockedImplicitAlgorithm.cpp

@@ -161,6 +161,7 @@ int main(int argc, char* argv[]){
     GroupAlgorithm groupalgo(&groupedTree,&groupkernel, distributedMortonIndex);
 	FTic timerExecute;
 	groupalgo.execute();
+	mpiComm.global().barrier();
 	double elapsedTime = timerExecute.tacAndElapsed();
 	timeAverage(mpi_rank, nproc, elapsedTime);
 	
@@ -226,16 +227,14 @@ void timeAverage(int mpi_rank, int nproc, double elapsedTime)
 	if(mpi_rank == 0)
 	{
 		double sumElapsedTime = elapsedTime;
-		std::cout << "Executing time node 0 (implicit) : " << sumElapsedTime << "s" << std::endl;
 		for(int i = 1; i < nproc; ++i)
 		{
 			double tmp;
 			MPI_Recv(&tmp, 1, MPI_DOUBLE, i, 0, MPI_COMM_WORLD, 0);
-			sumElapsedTime += tmp;
-			std::cout << "Executing time node " << i << " (implicit) : " << tmp << "s" << std::endl;
+			if(tmp > sumElapsedTime)
+				sumElapsedTime = tmp;
 		}
-		sumElapsedTime = sumElapsedTime / (double)nproc;
-		std::cout << "Average time per node (implicit) : " << sumElapsedTime << "s" << std::endl;
+		std::cout << "Average time per node (implicit Cheby) : " << sumElapsedTime << "s" << std::endl;
 	}
 	else
 	{

Tests/GroupTree/testBlockedImplicitChebyshev.cpp

@@ -140,6 +140,7 @@ int main(int argc, char* argv[]){
     GroupAlgorithm groupalgo(&groupedTree,&groupkernel, distributedMortonIndex);
 	FTic timerExecute;
 	groupalgo.execute(operationsToProceed);
+	mpiComm.global().barrier();
 	double elapsedTime = timerExecute.tacAndElapsed();
 	timeAverage(mpi_rank, nproc, elapsedTime);
 	
@@ -258,15 +259,13 @@ void timeAverage(int mpi_rank, int nproc, double elapsedTime)
 	if(mpi_rank == 0)
 	{
 		double sumElapsedTime = elapsedTime;
-		std::cout << "Executing time node 0 (implicit Cheby) : " << sumElapsedTime << "s" << std::endl;
 		for(int i = 1; i < nproc; ++i)
 		{
 			double tmp;
 			MPI_Recv(&tmp, 1, MPI_DOUBLE, i, 0, MPI_COMM_WORLD, 0);
-			sumElapsedTime += tmp;
-			std::cout << "Executing time node " << i << " (implicit Cheby) : " << tmp << "s" << std::endl;
+			if(tmp > sumElapsedTime)
+				sumElapsedTime = tmp;
 		}
-		sumElapsedTime = sumElapsedTime / (double)nproc;
 		std::cout << "Average time per node (implicit Cheby) : " << sumElapsedTime << "s" << std::endl;
 	}
 	else

Tests/GroupTree/testBlockedMpiAlgorithm.cpp

@@ -149,8 +149,8 @@ int main(int argc, char* argv[]){
     GroupAlgorithm groupalgo(mpiComm.global(), &groupedTree,&groupkernel);
 	FTic timerExecute;
     groupalgo.execute();
-	double elapsedTime = timerExecute.tacAndElapsed();
     mpiComm.global().barrier();
+	double elapsedTime = timerExecute.tacAndElapsed();
 	timeAverage(mpiComm.global().processId(), mpiComm.global().processCount(), elapsedTime);
 
     groupedTree.forEachCellLeaf<GroupContainerClass>([&](GroupCellClass cell, GroupContainerClass* leaf){
@@ -214,19 +214,18 @@ void timeAverage(int mpi_rank, int nproc, double elapsedTime)
 	if(mpi_rank == 0)
 	{
 		double sumElapsedTime = elapsedTime;
-		std::cout << "Executing time node 0 (explicit) : " << sumElapsedTime << "s" << std::endl;
 		for(int i = 1; i < nproc; ++i)
 		{
 			double tmp;
 			MPI_Recv(&tmp, 1, MPI_DOUBLE, i, 0, MPI_COMM_WORLD, 0);
-			sumElapsedTime += tmp;
-			std::cout << "Executing time node " << i << " (explicit) : " << tmp << "s" << std::endl;
+			if(tmp > sumElapsedTime)
+				sumElapsedTime = tmp;
 		}
-		sumElapsedTime = sumElapsedTime / (double)nproc;
-		std::cout << "Average time per node (explicit) : " << sumElapsedTime << "s" << std::endl;
+		std::cout << "Average time per node (implicit Cheby) : " << sumElapsedTime << "s" << std::endl;
 	}
 	else
 	{
 		MPI_Send(&elapsedTime, 1, MPI_DOUBLE, 0, 0, MPI_COMM_WORLD);
 	}
+	MPI_Barrier(MPI_COMM_WORLD);
 }

Tests/GroupTree/testBlockedMpiChebyshev.cpp

@@ -170,7 +170,7 @@ int main(int argc, char* argv[]){
         GroupAlgorithm groupalgo(mpiComm.global(), &groupedTree,&groupkernel);
         timer.tic();
         groupalgo.execute();
-
+		mpiComm.global().barrier();
         timer.tac();
 		timeAverage(mpiComm.global().processId(), mpiComm.global().processCount(), timer.elapsed());
         //std::cout << "Done  " << "(@Algorithm = " << timer.elapsed() << "s)." << std::endl;
@@ -293,16 +293,14 @@ void timeAverage(int mpi_rank, int nproc, double elapsedTime)
 	if(mpi_rank == 0)
 	{
 		double sumElapsedTime = elapsedTime;
-		std::cout << "Executing time node 0 (explicit Cheby) : " << sumElapsedTime << "s" << std::endl;
 		for(int i = 1; i < nproc; ++i)
 		{
 			double tmp;
 			MPI_Recv(&tmp, 1, MPI_DOUBLE, i, 0, MPI_COMM_WORLD, 0);
-			sumElapsedTime += tmp;
-			std::cout << "Executing time node " << i << " (explicit Cheby) : " << tmp << "s" << std::endl;
+			if(tmp > sumElapsedTime)
+				sumElapsedTime = tmp;
 		}
-		sumElapsedTime = sumElapsedTime / (double)nproc;
-		std::cout << "Average time per node (explicit Cheby) : " << sumElapsedTime << "s" << std::endl;
+		std::cout << "Average time per node (implicit Cheby) : " << sumElapsedTime << "s" << std::endl;
 	}
 	else
 	{

compile_dag_result.sh

@@ -4,7 +4,7 @@ export GROUP_SIZE=50
 export TREE_HEIGHT=5
 export NB_NODE=4
 #export NB_PARTICLE_PER_NODE=$(( (`awk "BEGIN{print 8 ** ($TREE_HEIGHT-1)}"` / $NB_NODE) ))
-export NB_PARTICLE_PER_NODE=5000
+export NB_PARTICLE_PER_NODE=50000
 export STARPU_NCPU=1
 export STARPU_FXT_PREFIX=`pwd`/
 
@@ -79,7 +79,7 @@ chebyshev_kernel()
 	#Get task information
 	cp -f $SCALFMM_SIMGRIDOUT scalfmm_explicit.out
 
-	mpiexec -n $NB_NODE ./Tests/Release/testBlockedImplicitChebyshev -nb $NB_PARTICLE_PER_NODE -bs $GROUP_SIZE -h $TREE_HEIGHT
+	mpiexec -n $NB_NODE ./Tests/Release/testBlockedImplicitChebyshev -nb $NB_PARTICLE_PER_NODE -bs $GROUP_SIZE -h $TREE_HEIGHT > pieuvre
 	if [ $? -ne 0 ]; then
 		echo
 		echo " /!\\Error on implicit Chebyshev"
@@ -96,6 +96,6 @@ chebyshev_kernel()
 	./Tests/Release/compareDAGmapping -e scalfmm_explicit.out -i scalfmm_implicit.out -h $TREE_HEIGHT > narval
 }
 
-test_kernel
+#test_kernel
 chebyshev_kernel
 

jobs/explicit_mpi_chebyshev.sh

@@ -28,9 +28,12 @@ export NB_PARTICLE_PER_NODE=5000000
 export STARPU_FXT_PREFIX=`pwd`/
 echo "===== Explicit MPI ===="
 echo "my jobID: " $SLURM_JOB_ID
+echo "Model: cube"
 echo "Nb node: " $NB_NODE
+echo "Nb thread: " $STARPU_NCPU
 echo "Tree height: " $TREE_HEIGHT
 echo "Group size: " $GROUP_SIZE
+echo "Algorithm: explicit"
 echo "Particle per node: " $NB_PARTICLE_PER_NODE
 echo "Total particles: " $(($NB_PARTICLE_PER_NODE*$NB_NODE))
 mpiexec -n $NB_NODE ./Build/Tests/Release/testBlockedMpiChebyshev -nb $NB_PARTICLE_PER_NODE -bs $GROUP_SIZE -h $TREE_HEIGHT -no-validation | grep Average

jobs/implicit_chebyshev.sh

@@ -28,9 +28,12 @@ export NB_PARTICLE_PER_NODE=5000000
 export STARPU_FXT_PREFIX=`pwd`/
 echo "===== Implicit MPI ===="
 echo "my jobID: " $SLURM_JOB_ID
+echo "Model: cube"
 echo "Nb node: " $NB_NODE
+echo "Nb thread: " $STARPU_NCPU
 echo "Tree height: " $TREE_HEIGHT
 echo "Group size: " $GROUP_SIZE
+echo "Algorithm: implicit"
 echo "Particle per node: " $NB_PARTICLE_PER_NODE
 echo "Total particles: " $(($NB_PARTICLE_PER_NODE*$NB_NODE))
 mpiexec -n $NB_NODE ./Build/Tests/Release/testBlockedImplicitChebyshev -nb $NB_PARTICLE_PER_NODE -bs $GROUP_SIZE -h $TREE_HEIGHT -no-validation | grep Average

jobs/starpu_chebyshev.sh

@@ -27,9 +27,12 @@ export NB_PARTICLE_PER_NODE=50000000
 export STARPU_FXT_PREFIX=`pwd`/
 echo "===== StarPU only ====="
 echo "my jobID: " $SLURM_JOB_ID
+echo "Model: cube"
 echo "Nb node: " $NB_NODE
+echo "Nb thread: " $STARPU_NCPU
 echo "Tree height: " $TREE_HEIGHT
 echo "Group size: " $GROUP_SIZE
+echo "Algorithm: starpu"
 echo "Particle per node: " $NB_PARTICLE_PER_NODE
 echo "Total particles: " $(($NB_PARTICLE_PER_NODE*$NB_NODE))
 mpiexec -n $NB_NODE ./Build/Tests/Release/testBlockedChebyshev -nb $NB_PARTICLE_PER_NODE -bs $GROUP_SIZE -h $TREE_HEIGHT -no-validation | grep Kernel