|
List Info
Thread: LAM: problem with non-blocking
|
|
| LAM: problem with non-blocking |
|
2007-03-28 00:08:45 |
|
Hi. now im doing parallel GA for my final project using C++. But now im having a problem with my program. when im running my program this kind of error happened :
One of the processes started by mpirun has exited with a nonzero exit
code. This typically indicates that the
process
finished in error.
If your
process
did not finish in error, be sure to include a "return
0
" or "exit(0)" in your C code before exiting the application.
PID 26349 failed on node n0 (10.1.1.1) due to signal 13
.
I hope there is someone can help me
explain whats wrong with my program and whats the maeaning by signal 13. Thanks
Here is algorithm of my program (most of my parallel program at function
Objective1) :
.............................................................................................................................................................................................................
int main (argc, argv) { //all in main is master part
MPI::Init(argc, argv);
master = id = 0;
if (id == master) {
ga(); //calling the function GA
MPI::Finalize();
} //end master
} // end main
.
.
.
void ga(); {
.
.
.
float Objective1();
} //function ga
void
svm_train(); //function svm_train
float Objective1(); float Objective1(GAGenome& g)
{
;
int p, id, slave;
int master = 0;
int tag = 15;
MPI::Status status;
id = MPI::COMM_WORLD.Get_rank();
p = MPI::COMM_WORLD.Get_size();
MPI::Request send[p], recv[p];
GA1DBinaryStringGenome & genome = (GA1DBinaryStringGenome &)g;
fstream file_index(indexf,ios::trunc|ios::in|ios::out);//create file for index output
double score=0.0;
if (id == master) {
if(!file_index)
; {
cerr << "Cannot open index1 file " << file_index << " for input.n";
; exit(1);
}
cout<<"chromosome value "<<++bilang<<": ";
for (int i=0; i<genome.length(); i++)
; {
cout<<(genome.gene(i));
file_index<<(
genome.gene(i));
; }
cout<<" ";
file_index<<endl;
file_index.close();
; for (slave = 1; slave < p; slave++) {
; //cout<<"sending to slave "<<slave<<"; from "<<id<<endl;
; send[slave] = MPI::COMM_WORLD.Isend(&indexf, 1, MPI::CHAR, bilang, tag);
send[slave] = MPI::COMM_WORLD.Isend(&file_index, 1, MPI::CHAR, bilang, tag);
}//end for
; }//end master
else {
recv[slave] = MPI::COMM_WORLD.Irecv(&indexf, 1, MPI::CHAR, master, tag);
recv[slave] = MPI::COMM_WORLD.Irecv(&file_index, 1, MPI::CHAR, master, tag);
featureselection_train(); //select features (genes) from s2AMLALL_train.fs.data
featureselection_test(); //select features (genes) from s2AMLALL_test.fs.data
svm_train();
send[slave] = MPI::COMM_WORLD.Isend(&cvscore, 1, MPI: OUBLE, master, tag);
}//end slaves
; if (id == master) {
for (slave = 1; slave < bilang; slave++) {
recv[slave] = MPI::COMM_WORLD.Irecv(&cvscore, 1, MPI::CHAR, slave, tag);
}//end for
; //MPI::Request::Waitall(p, recv);
//svm__test();
//score=testscore;
cout<<" cvs in fitness: "<<cvscore<<" "<<endl;
; score=cvscore;
; }//end master
return score;
}
|
| Re: LAM: problem with non-blocking |
|
2007-03-28 04:06:02 |
|
Hello.
I9;m not very familiar with the C++ calls, but some points to watch out for over a quick scan of the code:
1) Do the slaves ever call MPI::Finalize(); ?
2) Do the slaves ever enter
ga() or Objective1(); ? main() only has the master doing anything..
On 3/28/07,
Muhammad Faiz Misman < faizmisman gmail.com">faizmismangmail.com> wrote:
Hi. now im doing parallel GA for my final project using C++. But now im having a problem with my program. when im running my program this kind of error happened :
One of the processes started by mpirun has exited with a nonzero exit
code. This typically indicates that the
process
finished in error.
If your
process
did not finish in error, be sure to include a "return
0
" or "exit(0)" in your C code before exiting the application.
PID 26349 failed on node n0 (10.1.1.1) due to signal 13
.
I hope there is someone can help me
explain whats wrong with my program and whats the maeaning by signal 13. Thanks
Here is algorithm of my program (most of my parallel program at function
Objective1) :
.............................................................................................................................................................................................................
int main (argc, argv) { //all in main is master part
MPI::Init(argc, argv);
master = id = 0;
if (id == master) {
ga(); //calling the function GA
MPI::Finalize();
} //end master
} // end main
.
.
.
void ga(); {
.
.
.
float Objective1();
} //function ga
void
svm_train(); //function svm_train
float Objective1(); float Objective1(GAGenome& g)
{
;
int p, id, slave;
int master = 0;
int tag = 15;
MPI::Status status;
id = MPI::COMM_WORLD.Get_rank();
p = MPI::COMM_WORLD.Get_size();
MPI::Request send[p], recv[p];
GA1DBinaryStringGenome & genome = (GA1DBinaryStringGenome &)g;
fstream file_index(indexf,ios::trunc|ios::in|ios::out);//create file for index output
double score=0.0;
if (id == master) {
if(!file_index)
; {
cerr << "Cannot open index1 file " << file_index << " for input.n";
; exit(1);
}
cout<<"chromosome value "<<++bilang<<": ";
for (int i=0; i<genome.length(); i++)
; {
cout<<(genome.gene(i));
file_index<<(
genome.gene(i));
; }
cout<<" ";
file_index<<endl;
file_index.close();
; for (slave = 1; slave < p; slave++) {
; //cout<<"sending to slave "<<slave<<"; from "<<id<<endl;
; send[slave] = MPI::COMM_WORLD.Isend(&indexf, 1, MPI::CHAR, bilang, tag);
send[slave] = MPI::COMM_WORLD.Isend(&file_index, 1, MPI::CHAR, bilang, tag);
}//end for
; }//end master
else {
recv[slave] = MPI::COMM_WORLD.Irecv(&indexf, 1, MPI::CHAR, master, tag);
recv[slave] = MPI::COMM_WORLD.Irecv(&file_index, 1, MPI::CHAR, master, tag);
featureselection_train(); //select features (genes) from s2AMLALL_train.fs.data
featureselection_test(); //select features (genes) from s2AMLALL_test.fs.data
svm_train();
send[slave] = MPI::COMM_WORLD.Isend(&cvscore, 1, MPI:OUBLE, master, tag);
}//end slaves
; if (id == master) {
for (slave = 1; slave < bilang; slave++) {
recv[slave] = MPI::COMM_WORLD.Irecv(&cvscore, 1, MPI::CHAR, slave, tag);
}//end for
; //MPI::Request::Waitall(p, recv);
//svm__test();
//score=testscore;
cout<<" cvs in fitness: "<<cvscore<<" "<<endl;
; score=cvscore;
; }//end master
return score;
}
_______________________________________________
This list is archived at http://www.lam-mpi.org/MailArchives/lam/
--
(N)E
|
| Re: LAM: problem with non-blocking |
|
2007-03-30 04:11:33 |
|
Yes you're right. the slave never enter the function GA() and Objective1(), thanks for your help. Is anybody can help me on giving some advice or idea on how im gonna call the slave in function
Objective1() ? because in my program, main() only for master.
On 3/28/07, Elie Choueiri
< elie.choueirigmail.com">elie.choueirigmail.com> wrote:
Hello.
I'm not very familiar with the C++ calls, but some points to watch out for over a quick scan of the code:
1) Do the slaves ever call MPI::Finalize(); ?
2) Do the slaves ever enter
ga() or Objective1(); ? main() only has the master doing anything..
On 3/28/07,
Muhammad Faiz Misman < faizmismangmail.com" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">faizmismangmail.com> wrote:
Hi. now im doing parallel GA for my final project using C++. But now im having a problem with my program. when im running my program this kind of error happened :
One of the processes started by mpirun has exited with a nonzero exit
code. This typically indicates that the
process
finished in error.
If your
process
did not finish in error, be sure to include a "return
0
" or "exit(0)" in your C code before exiting the application.
PID 26349 failed on node n0 (10.1.1.1) due to signal 13
.
I hope there is someone can help me
explain whats wrong with my program and whats the maeaning by signal 13. Thanks
Here is algorithm of my program (most of my parallel program at function
Objective1) :
.............................................................................................................................................................................................................
int main (argc, argv) { //all in main is master part
MPI::Init(argc, argv);
master = id = 0;
if (id == master) {
ga(); //calling the function GA
MPI::Finalize();
} //end master
} // end main
.
.
.
void ga(); {
.
.
.
float Objective1();
} //function ga
void
svm_train(); //function svm_train
float Objective1(); float Objective1(GAGenome& g)
{
;
int p, id, slave;
int master = 0;
int tag = 15;
MPI::Status status;
id = MPI::COMM_WORLD.Get_rank();
p = MPI::COMM_WORLD.Get_size();
MPI::Request send[p], recv[p];
GA1DBinaryStringGenome & genome = (GA1DBinaryStringGenome &)g;
fstream file_index(indexf,ios::trunc|ios::in|ios::out);//create file for index output
double score=0.0;
if (id == master) {
if(!file_index)
; {
cerr << "Cannot open index1 file " << file_index << " for input.n";
; exit(1);
}
cout<<"chromosome value "<<++bilang<<": ";
for (int i=0; i<genome.length(); i++)
; {
cout<<(genome.gene(i));
file_index<<(
genome.gene(i));
; }
cout<<" ";
file_index<<endl;
file_index.close();
; for (slave = 1; slave < p; slave++) {
; //cout<<"sending to slave "<<slave<<"; from "<<id<<endl;
; send[slave] = MPI::COMM_WORLD.Isend(&indexf, 1, MPI::CHAR, bilang, tag);
send[slave] = MPI::COMM_WORLD.Isend(&file_index, 1, MPI::CHAR, bilang, tag);
}//end for
; }//end master
else {
recv[slave] = MPI::COMM_WORLD.Irecv(&indexf, 1, MPI::CHAR, master, tag);
recv[slave] = MPI::COMM_WORLD.Irecv(&file_index, 1, MPI::CHAR, master, tag);
featureselection_train(); //select features (genes) from s2AMLALL_train.fs.data
featureselection_test(); //select features (genes) from s2AMLALL_test.fs.data
svm_train();
send[slave] = MPI::COMM_WORLD.Isend(&cvscore, 1, MPI:OUBLE, master, tag);
}//end slaves
; if (id == master) {
for (slave = 1; slave < bilang; slave++) {
recv[slave] = MPI::COMM_WORLD.Irecv(&cvscore, 1, MPI::CHAR, slave, tag);
}//end for
; //MPI::Request::Waitall(p, recv);
//svm__test();
//score=testscore;
cout<<" cvs in fitness: "<<cvscore<<" "<<endl;
; score=cvscore;
; }//end master
return score;
}
_______________________________________________
This list is archived at http://www.lam-mpi.org/MailArchives/lam/
--
(N)E
_______________________________________________
This list is archived at http://www.lam-mpi.org/MailArchives/lam/
|
| Re: LAM: problem with non-blocking |
|
2007-03-30 04:27:06 |
|
Organize your code differently.
Instead of saying id = MPI::COMM_WORLD.Get_rank();
in Objective1(), use it in main().
Then you can say:
if (id == 0) { // master
// do master work
}
else
{ // slave work
}
and if master and slave both call the same function anyway, you can ignore the if's and just call the function..
On 3/30/07, Muhammad Faiz Misman < faizmismangmail.com">faizmismangmail.com> wrote:
Yes you're right. the slave never enter the function GA() and Objective1(), thanks for your help. Is anybody can help me on giving some advice or idea on how im gonna call the slave in function
Objective1() ? because in my program, main() only for master.
On 3/28/07, Elie Choueiri
< elie.choueirigmail.com" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">elie.choueirigmail.com> wrote:
Hello.
I'm not very familiar with the C++ calls, but some points to watch out for over a quick scan of the code:
1) Do the slaves ever call MPI::Finalize(); ?
2) Do the slaves ever enter
ga() or Objective1(); ? main() only has the master doing anything..
On 3/28/07,
Muhammad Faiz Misman < faizmismangmail.com" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">faizmismangmail.com> wrote:
Hi. now im doing parallel GA for my final project using C++. But now im having a problem with my program. when im running my program this kind of error happened :
One of the processes started by mpirun has exited with a nonzero exit
code. This typically indicates that the
process
finished in error.
If your
process
did not finish in error, be sure to include a "return
0
" or "exit(0)" in your C code before exiting the application.
PID 26349 failed on node n0 (10.1.1.1) due to signal 13
.
I hope there is someone can help me
explain whats wrong with my program and whats the maeaning by signal 13. Thanks
Here is algorithm of my program (most of my parallel program at function
Objective1) :
.............................................................................................................................................................................................................
int main (argc, argv) { //all in main is master part
MPI::Init(argc, argv);
master = id = 0;
if (id == master) {
ga(); //calling the function GA
MPI::Finalize();
} //end master
} // end main
.
.
.
void ga(); {
.
.
.
float Objective1();
} //function ga
void
svm_train(); //function svm_train
float Objective1(); float Objective1(GAGenome& g)
{
;
int p, id, slave;
int master = 0;
int tag = 15;
MPI::Status status;
id = MPI::COMM_WORLD.Get_rank();
p = MPI::COMM_WORLD.Get_size();
MPI::Request send[p], recv[p];
GA1DBinaryStringGenome & genome = (GA1DBinaryStringGenome &)g;
fstream file_index(indexf,ios::trunc|ios::in|ios::out);//create file for index output
double score=0.0;
if (id == master) {
if(!file_index)
; {
cerr << "Cannot open index1 file " << file_index << " for input.n";
; exit(1);
}
cout<<"chromosome value "<<++bilang<<": ";
for (int i=0; i<genome.length(); i++)
; {
cout<<(genome.gene(i));
file_index<<(
genome.gene(i));
; }
cout<<" ";
file_index<<endl;
file_index.close();
; for (slave = 1; slave < p; slave++) {
; //cout<<"sending to slave "<<slave<<"; from "<<id<<endl;
; send[slave] = MPI::COMM_WORLD.Isend(&indexf, 1, MPI::CHAR, bilang, tag);
send[slave] = MPI::COMM_WORLD.Isend(&file_index, 1, MPI::CHAR, bilang, tag);
}//end for
; }//end master
else {
recv[slave] = MPI::COMM_WORLD.Irecv(&indexf, 1, MPI::CHAR, master, tag);
recv[slave] = MPI::COMM_WORLD.Irecv(&file_index, 1, MPI::CHAR, master, tag);
featureselection_train(); //select features (genes) from s2AMLALL_train.fs.data
featureselection_test(); //select features (genes) from s2AMLALL_test.fs.data
svm_train();
send[slave] = MPI::COMM_WORLD.Isend(&cvscore, 1, MPI:OUBLE, master, tag);
}//end slaves
; if (id == master) {
for (slave = 1; slave < bilang; slave++) {
recv[slave] = MPI::COMM_WORLD.Irecv(&cvscore, 1, MPI::CHAR, slave, tag);
}//end for
; //MPI::Request::Waitall(p, recv);
//svm__test();
//score=testscore;
cout<<" cvs in fitness: "<<cvscore<<" "<<endl;
; score=cvscore;
; }//end master
return score;
}
_______________________________________________
This list is archived at http://www.lam-mpi.org/MailArchives/lam/
--
(N)E
_______________________________________________
This list is archived at http://www.lam-mpi.org/MailArchives/lam/
_______________________________________________
This list is archived at http://www.lam-mpi.org/MailArchives/lam/
--
(N)E
|
[1-4]
|
|