portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-11-03 21:44:33 +00:00
Code Releases Activity

Bug fix in MPI3

Browse Source
This commit is contained in:
paboyle
2017-03-29 04:43:55 -04:00
parent 81ead48850
commit 35695ba57a
1 changed files with 17 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
lib/communicator/Communicator_mpi3.cc
View File

@@ -67,7 +67,7 @@ std::vector<void *> CartesianCommunicator::ShmCommBufs;
int CartesianCommunicator::NodeCount(void) { return GroupSize;}; int CartesianCommunicator::NodeCount(void) { return GroupSize;};
#undef FORCE_COMMS #define FORCE_COMMS
void *CartesianCommunicator::ShmBufferSelf(void) void *CartesianCommunicator::ShmBufferSelf(void)
{ {
return ShmCommBufs[ShmRank]; return ShmCommBufs[ShmRank];
@@ -303,7 +303,7 @@ void CartesianCommunicator::Init(int *argc, char ***argv) {
std::cout<<GridLogMessage<< "Grid MPI-3 configuration: detected "; std::cout<<GridLogMessage<< "Grid MPI-3 configuration: detected ";
std::cout<< WorldSize << " Ranks " ; std::cout<< WorldSize << " Ranks " ;
std::cout<< GroupSize << " Nodes " ; std::cout<< GroupSize << " Nodes " ;
std::cout<< ShmSize << " with ranks-per-node "<<std::endl; std::cout<< " with "<< ShmSize << " ranks-per-node "<<std::endl;
std::cout<<GridLogMessage <<"Grid MPI-3 configuration: allocated shared memory region of size "; std::cout<<GridLogMessage <<"Grid MPI-3 configuration: allocated shared memory region of size ";
std::cout<<std::hex << MAX_MPI_SHM_BYTES <<" ShmCommBuf address = "<<ShmCommBuf << std::dec<<std::endl; std::cout<<std::hex << MAX_MPI_SHM_BYTES <<" ShmCommBuf address = "<<ShmCommBuf << std::dec<<std::endl;
@@ -338,7 +338,7 @@ void CartesianCommunicator::Init(int *argc, char ***argv) {
//////////////////////////////////////////////////////////////////////////////////////////////////////////// ////////////////////////////////////////////////////////////////////////////////////////////////////////////
// Want to implement some magic ... Group sub-cubes into those on same node // Want to implement some magic ... Group sub-cubes into those on same node
//////////////////////////////////////////////////////////////////////////////////////////////////////////// ////////////////////////////////////////////////////////////////////////////////////////////////////////////
void CartesianCommunicator::ShiftedRanks(int dim,int shift,int &source,int &dest) void CartesianCommunicator::ShiftedRanks(int dim,int shift,int &dest,int &source)
{ {
std::vector<int> coor = _processor_coor; // my coord std::vector<int> coor = _processor_coor; // my coord
assert(std::abs(shift) <_processors[dim]); assert(std::abs(shift) <_processors[dim]);
@@ -350,6 +350,7 @@ void CartesianCommunicator::ShiftedRanks(int dim,int shift,int &source,int &dest
coor[dim] = (_processor_coor[dim] - shift + _processors[dim])%_processors[dim]; coor[dim] = (_processor_coor[dim] - shift + _processors[dim])%_processors[dim];
Lexicographic::IndexFromCoor(coor,dest,_processors); Lexicographic::IndexFromCoor(coor,dest,_processors);
dest = LexicographicToWorldRank[dest]; dest = LexicographicToWorldRank[dest];
}// rank is world rank. }// rank is world rank.
int CartesianCommunicator::RankFromProcessorCoor(std::vector<int> &coor) int CartesianCommunicator::RankFromProcessorCoor(std::vector<int> &coor)
@@ -471,12 +472,25 @@ CartesianCommunicator::CartesianCommunicator(const std::vector<int> &processors)
for(int i=0;i<WorldSize;i++){ for(int i=0;i<WorldSize;i++){
int wr = LexicographicToWorldRank[i]; int wr = LexicographicToWorldRank[i];
// int wr = i;
std::vector<int> coor(_ndimension); std::vector<int> coor(_ndimension);
ProcessorCoorFromRank(wr,coor); // from world rank ProcessorCoorFromRank(wr,coor); // from world rank
int ck = RankFromProcessorCoor(coor); int ck = RankFromProcessorCoor(coor);
assert(ck==wr); assert(ck==wr);
if ( wr == WorldRank ) {
for(int j=0;j<coor.size();j++) {
assert(coor[j] == _processor_coor[j]);
}
}
std::cout << GridLogMessage<< " Lexicographic "<<i;
std::cout << " MPI rank "<<wr;
std::cout << " Coor ";
for(int j=0;j<coor.size();j++) std::cout << coor[j];
std::cout<< std::endl;
///////////////////////////////////////////////////// /////////////////////////////////////////////////////
// Check everyone agrees on everyone elses coords // Check everyone agrees on everyone elses coords
///////////////////////////////////////////////////// /////////////////////////////////////////////////////