Skip to content

GitLab

Commit 4d47c78d authored by berenger-bramas's avatar berenger-bramas
Browse files

Refactor the merge leaf after sort algorithm. 

And perfromed some basic tests.

git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/scalfmm/scalfmm/trunk@223 2616d619-271b-44dc-8df4-d4a8f33a7222
parent 68b61e3b
Hide whitespace changes
Inline Side-by-side
Showing with 78 additions and 103 deletions
Src/Files/FMpiTreeBuilder.hpp
View file @ 4d47c78d
......@@ -165,12 +165,15 @@ public:
}
// sort particles
FQuickSort<IndexedParticle,MortonIndex, FSize>::QsMpi(realParticlesIndexed, loader.getNumberOfParticles(),outputArray,outputSize);
delete [] (realParticlesIndexed);
//FBitonicSort::sort<IndexedParticle,MortonIndex>( realParticlesIndexed, loader.getNumberOfParticles() );
//outputArray = realParticlesIndexed;
//outputSize = loader.getNumberOfParticles();
if(false){
FQuickSort<IndexedParticle,MortonIndex, FSize>::QsMpi(realParticlesIndexed, loader.getNumberOfParticles(),outputArray,outputSize);
delete [] (realParticlesIndexed);
}
else {
FBitonicSort::sort<IndexedParticle,MortonIndex>( realParticlesIndexed, loader.getNumberOfParticles() );
outputArray = realParticlesIndexed;
outputSize = loader.getNumberOfParticles();
}
}
// be sure there is no splited leaves
// to do that we exchange the first index with the left proc
......@@ -178,37 +181,46 @@ public:
FTRACE( FTrace::FRegion regionTrace("Remove Splited leaves", __FUNCTION__ , __FILE__ , __LINE__) );
MortonIndex otherFirstIndex = -1;
{
FMpi::Request req[2];
int reqiter = 0;
// can I send my first index? == I am not rank 0 & I have data
if( 0 < rank && outputSize){
MPI_Isend( &outputArray[0].index, 1, MPI_LONG_LONG, rank - 1, FMpi::TagExchangeIndexs, MPI_COMM_WORLD, &req[reqiter++]);
}
if( rank != nbProcs - 1){
MPI_Irecv(&otherFirstIndex, 1, MPI_LONG_LONG, rank + 1, FMpi::TagExchangeIndexs, MPI_COMM_WORLD, &req[reqiter++]);
}
MPI_Waitall(reqiter,req,MPI_STATUSES_IGNORE);
// I could not send because I do not have data, so I transmit the data coming
// from my right neigbors
if( 0 < rank && !outputSize){
MPI_Send( &otherFirstIndex, 1, MPI_LONG_LONG, rank - 1, FMpi::TagExchangeIndexs, MPI_COMM_WORLD);
}
if(outputSize != 0 && rank != 0 && rank != nbProcs - 1){
MPI_Sendrecv(&outputArray[0].index, 1, MPI_LONG_LONG, rank - 1, FMpi::TagExchangeIndexs,
&otherFirstIndex, 1, MPI_LONG_LONG, rank + 1, FMpi::TagExchangeIndexs,
MPI_COMM_WORLD, MPI_STATUS_IGNORE);
}
else if( rank == 0){
MPI_Recv(&otherFirstIndex, 1, MPI_LONG_LONG, rank + 1, FMpi::TagExchangeIndexs, MPI_COMM_WORLD, MPI_STATUS_IGNORE);
}
else if( rank == nbProcs - 1){
MPI_Send( &outputArray[0].index, 1, MPI_LONG_LONG, rank - 1, FMpi::TagExchangeIndexs, MPI_COMM_WORLD);
}
else {
MPI_Recv(&otherFirstIndex, 1, MPI_LONG_LONG, rank + 1, FMpi::TagExchangeIndexs, MPI_COMM_WORLD, MPI_STATUS_IGNORE);
MPI_Send(&otherFirstIndex, 1, MPI_LONG_LONG, rank - 1, FMpi::TagExchangeIndexs, MPI_COMM_WORLD);
}
MPI_Request req[2];
int reqiter = 0;
// at this point every one know the first index of his right neighbors
const bool needToRecvBeforeSend = (rank != 0 && ((outputSize && otherFirstIndex == outputArray[0].index ) || !outputSize));
if( needToRecvBeforeSend || (rank == nbProcs - 1) ){
// Here we want to send data we do not have
// so we first receive other data and put then into the array
// this case happens only if I have one leaf with index MX
// and if proc[rank - 1].last.index == MX && proc[rank + 1].first.index == MX
MPI_Request requestSendLeaf;
IndexedParticle* sendBuffer = 0;
if(rank != nbProcs - 1 && needToRecvBeforeSend == false){
FSize idxPart = outputSize - 1 ;
while(idxPart >= 0 && outputArray[idxPart].index == otherFirstIndex){
--idxPart;
}
const int particlesToSend = int(outputSize - 1 - idxPart);
if(particlesToSend){
outputSize -= particlesToSend;
sendBuffer = new IndexedParticle[particlesToSend];
memcpy(sendBuffer, &outputArray[idxPart + 1], particlesToSend * sizeof(IndexedParticle));
MPI_Isend( sendBuffer, particlesToSend * sizeof(IndexedParticle), MPI_BYTE, rank + 1, FMpi::TagSplittedLeaf, MPI_COMM_WORLD, &requestSendLeaf);
}
else{
MPI_Isend( 0, 0, MPI_BYTE, rank + 1, FMpi::TagSplittedLeaf, MPI_COMM_WORLD, &requestSendLeaf);
}
}
if( rank != 0 ){
int sendByOther = 0;
MPI_Status probStatus;
......@@ -217,6 +229,7 @@ public:
if(sendByOther){
sendByOther /= sizeof(IndexedParticle);
const IndexedParticle* const reallocOutputArray = outputArray;
const FSize reallocOutputSize = outputSize;
......@@ -225,88 +238,57 @@ public:
FMemUtils::memcpy(&outputArray[sendByOther], reallocOutputArray, reallocOutputSize * sizeof(IndexedParticle));
delete[] reallocOutputArray;
MPI_Recv(outputArray, sizeof(IndexedParticle) * sendByOther, MPI_BYTE, rank - 1, FMpi::TagSplittedLeaf, MPI_COMM_WORLD, &probStatus);
MPI_Recv(outputArray, sizeof(IndexedParticle) * sendByOther, MPI_BYTE, rank - 1, FMpi::TagSplittedLeaf, MPI_COMM_WORLD, MPI_STATUS_IGNORE);
}
else{
MPI_Irecv(0, 0, MPI_BYTE, rank - 1, FMpi::TagSplittedLeaf, MPI_COMM_WORLD, &req[reqiter++]);
MPI_Recv( 0, 0, MPI_BYTE, rank - 1, FMpi::TagSplittedLeaf, MPI_COMM_WORLD, MPI_STATUS_IGNORE);
}
}
if(rank != nbProcs - 1){
FSize idxPart = outputSize - 1 ;
while(idxPart >= 0 && outputArray[idxPart].index == otherFirstIndex){
--idxPart;
}
const int toSend = int(outputSize - 1 - idxPart);
MPI_Isend( &outputArray[idxPart + 1], toSend * sizeof(IndexedParticle), MPI_BYTE, rank + 1, FMpi::TagSplittedLeaf, MPI_COMM_WORLD, &req[reqiter++]);
if( rank != 0 && !needToRecvBeforeSend && (rank != nbProcs - 1)){
int sendByOther = 0;
MPI_Status probStatus;
MPI_Probe(rank - 1, FMpi::TagSplittedLeaf, MPI_COMM_WORLD, &probStatus);
MPI_Get_count( &probStatus, MPI_BYTE, &sendByOther);
if(sendByOther){
sendByOther /= sizeof(IndexedParticle);
char* const tempBuffer = new char[sizeof(IndexedParticle) * sendByOther];
MPI_Irecv(tempBuffer, sizeof(IndexedParticle) * sendByOther, MPI_BYTE, rank - 1, FMpi::TagSplittedLeaf, MPI_COMM_WORLD, &req[reqiter++]);
MPI_Waitall(reqiter,req, MPI_STATUSES_IGNORE);
reqiter = 0;
const IndexedParticle* const reallocOutputArray = outputArray;
const FSize reallocOutputSize = outputSize;
outputSize += sendByOther;
outputArray = new IndexedParticle[outputSize];
FMemUtils::memcpy(&outputArray[sendByOther], reallocOutputArray, reallocOutputSize * sizeof(IndexedParticle));
delete[] reallocOutputArray;
memcpy(outputArray, tempBuffer, sendByOther * sizeof(IndexedParticle));
delete[] tempBuffer;
}
else{
MPI_Irecv( 0, 0, MPI_BYTE, rank - 1, FMpi::TagSplittedLeaf, MPI_COMM_WORLD, &req[reqiter++]);
}
}
if(rank != nbProcs - 1 && needToRecvBeforeSend == true){
MPI_Send( outputArray, int(outputSize * sizeof(IndexedParticle)), MPI_BYTE, rank + 1, FMpi::TagSplittedLeaf, MPI_COMM_WORLD);
delete[] outputArray;
outputArray = 0;
outputSize = 0;
}
{
FTRACE( FTrace::FRegion regionWaitAllTrace("Wait all", __FUNCTION__ , __FILE__ , __LINE__) );
MPI_Waitall(reqiter,req,MPI_STATUSES_IGNORE);
else if(rank != nbProcs - 1){
MPI_Wait( &requestSendLeaf, MPI_STATUS_IGNORE);
delete[] sendBuffer;
sendBuffer = 0;
}
}
// We now copy the data from a sorted type into real particles array + counter
{
FTRACE( FTrace::FRegion regionTrace("Remove Splited leaves", __FUNCTION__ , __FILE__ , __LINE__) );
// We now copy the data from a sorted type into real particles array + counter
nbLeavesInIntervals = 0;
if(outputSize){
intervals = new char[outputSize * (sizeof(ParticleClass) + sizeof(int))];
nbLeavesInIntervals = 0;
if(outputSize){
intervals = new char[outputSize * (sizeof(ParticleClass) + sizeof(int))];
MortonIndex previousIndex = -1;
char* writeIndex = intervals;
int* writeCounter = 0;
MortonIndex previousIndex = -1;
char* writeIndex = intervals;
int* writeCounter = 0;
for( FSize idxPart = 0; idxPart < outputSize ; ++idxPart){
if( outputArray[idxPart].index != previousIndex ){
previousIndex = outputArray[idxPart].index;
++nbLeavesInIntervals;
for( FSize idxPart = 0; idxPart < outputSize ; ++idxPart){
if( outputArray[idxPart].index != previousIndex ){
previousIndex = outputArray[idxPart].index;
++nbLeavesInIntervals;
writeCounter = reinterpret_cast<int*>( writeIndex );
writeIndex += sizeof(int);
writeCounter = reinterpret_cast<int*>( writeIndex );
writeIndex += sizeof(int);
(*writeCounter) = 0;
}
(*writeCounter) = 0;
}
memcpy(writeIndex, &outputArray[idxPart].particle, sizeof(ParticleClass));
memcpy(writeIndex, &outputArray[idxPart].particle, sizeof(ParticleClass));
writeIndex += sizeof(ParticleClass);
++(*writeCounter);
writeIndex += sizeof(ParticleClass);
++(*writeCounter);
}
}
delete [] outputArray;
}
delete [] outputArray;
return true;
}
......@@ -408,13 +390,6 @@ public:
const FSize iNeedToSendRightCount = currentLeafsOnMyLeft + currentNbLeafs - correctRightLeavesIndex;
endForMe -= iNeedToSendRightCount;
}
printf("iNeedToSendToLeft %s\n", iNeedToSendToLeft?"True":"False");
printf("iNeedToSendToRight %s\n", iNeedToSendToRight?"True":"False");
printf("currentNbLeafs %lld\n",currentNbLeafs);
printf("iNeedToSendLeftCount %lld\n",iNeedToSendLeftCount);
printf("correctLeftLeavesNumber %lld\n",correctLeftLeavesNumber);
printf("currentLeafsOnMyLeft %lld\n",currentLeafsOnMyLeft);
printf("endForMe %lld\n",endForMe);
// We have to jump the correct number of leaves
for(FSize idxLeaf = Max(iNeedToSendLeftCount,0) ; idxLeaf < endForMe ; ++idxLeaf){
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment