No I havent tried using valgrind.
here is the latest code -
#include "mpi.h"
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#define ARRAYSIZE 200000
#define MASTER 0
int *data;
int main(int argc, char* argv[])
{
int numtasks, taskid, rc, dest, offset, i, j, tag1, tag2, source,
chunksize, namelen;
int mysum, sum;
int update(int myoffset, int chunk, int myid);
char myname[MPI_MAX_PROCESSOR_NAME];
MPI_Status status;
double start, stop, time;
double totaltime;
FILE *fp;
char line[128];
char element;
int n;
int k=0;
/***** Initializations *****/
MPI_Init(&argc, &argv);
MPI_Comm_size(MPI_COMM_WORLD, &numtasks);
if (numtasks % 4 != 0) {
printf("Quitting. Number of MPI tasks must be divisible by 4.\n");
MPI_Abort(MPI_COMM_WORLD, rc);
MPI_Finalize();
}
MPI_Comm_rank(MPI_COMM_WORLD,&taskid);
MPI_Get_processor_name(myname, &namelen);
printf ("MPI task %d has started on host %s...\n", taskid, myname);
chunksize = (ARRAYSIZE / numtasks);
tag2 = 1;
tag1 = 2;
/***** Master task only ******/
if (taskid == MASTER){
/* Initialize the array */
data = malloc(ARRAYSIZE * sizeof(int));
if(NULL == data){
printf("Data null");
}
sum = 0;
for(i=0; i<ARRAYSIZE; i++) {
data[i] = i * 1 ;
sum = sum + data[i];
}
printf("Initialized array sum = %d\n",sum);
/* Send each task its portion of the array - master keeps 1st part */
offset = chunksize;
for (dest=1; dest<numtasks; dest++) {
MPI_Send(&offset, 1, MPI_INT, dest, tag1, MPI_COMM_WORLD);
MPI_Send(&data[offset], chunksize, MPI_INT, dest, tag2, MPI_COMM_WORLD);
printf("Sent %d elements to task %d offset=
%d\n",chunksize,dest,offset);
offset = offset + chunksize;
}
/* Master does its part of the work */
offset = 0;
mysum = update(offset, chunksize, taskid);
/* Wait to receive results from each task */
for (i=1; i<numtasks; i++) {
source = i;
MPI_Recv(&offset, 1, MPI_INT, source, tag1, MPI_COMM_WORLD, &status);
MPI_Recv(&data[offset], chunksize, MPI_INT, source, tag2,
MPI_COMM_WORLD, &status);
}
/* Get final sum and print sample results */
MPI_Reduce(&mysum, &sum, 1, MPI_INT, MPI_SUM, MASTER, MPI_COMM_WORLD);
/* printf("Sample results: \n");
offset = 0;
for (i=0; i<numtasks; i++) {
for (j=0; j<5; j++)
printf(" %d",data[offset+j]);ARRAYSIZE
printf("\n");
offset = offset + chunksize;
}*/
printf("\n*** Final sum= %d ***\n",sum);
} /* end of master section */
/***** Non-master tasks only *****/
if (taskid > MASTER) {
/* Receive my portion of array from the master task */
start= MPI_Wtime();
source = MASTER;
MPI_Recv(&offset, 1, MPI_INT, source, tag1, MPI_COMM_WORLD, &status);
MPI_Recv(&data[offset], chunksize, MPI_INT, source, tag2,MPI_COMM_WORLD,
&status);
mysum = update(offset, chunksize, taskid);
stop = MPI_Wtime();
time = stop -start;
printf("time taken by process %d to recieve elements and caluclate own
sum is = %lf seconds \n", taskid, time);
totaltime = totaltime + time;
/* Send my results back to the master task */
dest = MASTER;
MPI_Send(&offset, 1, MPI_INT, dest, tag1, MPI_COMM_WORLD);
MPI_Send(&data[offset], chunksize, MPI_INT, MASTER, tag2, MPI_COMM_WORLD);
MPI_Reduce(&mysum, &sum, 1, MPI_INT, MPI_SUM, MASTER, MPI_COMM_WORLD);
} /* end of non-master */
// printf("Total time taken for distribution is - %lf seconds",
totaltime);
MPI_Finalize();
} /* end of main */
int update(int myoffset, int chunk, int myid) {
int i,j;
int mysum;
int mydata[myoffset+chunk];
/* Perform addition to each of my array elements and keep my sum */
mysum = 0;
/* printf("task %d has elements:",myid);
for(j = myoffset; j<myoffset+chunk; j++){
printf("\t%d", data[j]);
}
printf("\n");*/
for(i=myoffset; i < myoffset + chunk; i++) {
//data[i] = data[i] + i;
mysum = mysum + data[i];
}
printf("Task %d has sum = %d\n",myid,mysum);
return(mysum);
}
On Thu, Apr 19, 2012 at 5:59 PM, Jeffrey Squyres <jsquy...@cisco.com> wrote:
> Send the most recent version of your code.
>
> Have you tried running it through a memory-checking debugger, like
> valgrind?
>
>
> On Apr 19, 2012, at 4:24 AM, Rohan Deshpande wrote:
>
> > Hi Pascal,
> >
> > The offset is received from the master task. so no need to initialize
> for non master tasks?
> >
> > not sure what u meant exactly.
> >
> > Thanks
> >
> >
> >
> > On Thu, Apr 19, 2012 at 3:36 PM, <pascal.dev...@bull.net> wrote:
> > I do not see where you initialize the offset on the "Non-master tasks".
> This could be the problem.
> >
> > Pascal
> >
> > users-boun...@open-mpi.org a écrit sur 19/04/2012 09:18:31 :
> >
> > > De : Rohan Deshpande <rohan...@gmail.com>
> > > A : Open MPI Users <us...@open-mpi.org>
> > > Date : 19/04/2012 09:18
> > > Objet : Re: [OMPI users] machine exited on signal 11 (Segmentation
> fault).
> > > Envoyé par : users-boun...@open-mpi.org
> > >
> > > Hi Jeffy,
> > >
> > > I checked the SEND RECV buffers and it looks ok to me. The code I
> > > have sent works only when I statically initialize the array.
> > >
> > > The code fails everytime I use malloc to initialize the array.
> > >
> > > Can you please look at code and let me know what is wrong.
> >
> > > On Wed, Apr 18, 2012 at 8:11 PM, Jeffrey Squyres <jsquy...@cisco.com>
> wrote:
> > > As a guess, you're passing in a bad address.
> > >
> > > Double check the buffers that you're sending to MPI_SEND/MPI_RECV/etc.
> > >
> > >
> > > On Apr 17, 2012, at 10:43 PM, Rohan Deshpande wrote:
> > >
> > > > After using malloc i am getting following error
> > > >
> > > > *** Process received signal ***
> > > > Signal: Segmentation fault (11)
> > > > Signal code: Address not mapped (1)
> > > > Failing at address: 0x1312d08
> > > > [ 0] [0x5e840c]
> > > > [ 1] /usr/local/lib/openmpi/mca_btl_tcp.so(+0x5bdb) [0x119bdb]
> > > > /usr/local/lib/libopen-pal.so.0(+0x19ce0) [0xb2cce0]
> > > > /usr/local/lib/libopen-pal.so.0(opal_event_loop+0x27) [0xb2cf47]
> > > > /usr/local/lib/libopen-pal.so.0(opal_progress+0xda) [0xb200ba]
> > > > /usr/local/lib/openmpi/mca_pml_ob1.so(+0x3f75) [0xa9ef75]
> > > > [ 6] /usr/local/lib/libmpi.so.0(MPI_Recv+0x136) [0xea7c46]
> > > > [ 7] mpi_array(main+0x501) [0x8048e25]
> > > > [ 8] /lib/libc.so.6(__libc_start_main+0xe6) [0x2fece6]
> > > > [ 9] mpi_array() [0x8048891]
> > > > *** End of error message ***
> > > > [machine4][[3968,1],0][btl_tcp_frag.c:216:mca_btl_tcp_frag_recv]
> > > mca_btl_tcp_frag_recv: readv failed: Connection reset by peer (104)
> > > >
> --------------------------------------------------------------------------
> > > > mpirun noticed that process rank 1 with PID 2936 on node machine4
> > > exited on signal 11 (Segmentation fault).
> > > >
> > > > Can someone help please.
> > > >
> > > > Thanks
> > > >
> > > >
> > > >
> > > > On Tue, Apr 17, 2012 at 6:01 PM, Jeffrey Squyres <jsquy...@cisco.com>
> wrote:
> > > > Try malloc'ing your array instead of creating it statically on the
> > > stack. Something like:
> > > >
> > > > int *data;
> > > >
> > > > int main(..) {
> > > > {
> > > > data = malloc(ARRAYSIZE * sizeof(int));
> > > > if (NULL == data) {
> > > > perror("malloc");
> > > > exit(1);
> > > > }
> > > > // ...
> > > > }
> > > >
> > > >
> > > > On Apr 17, 2012, at 5:05 AM, Rohan Deshpande wrote:
> > > >
> > > > >
> > > > > Hi,
> > > > >
> > > > > I am trying to distribute large amount of data using MPI.
> > > > >
> > > > > When I exceed the certain data size the segmentation fault occurs.
> > > > >
> > > > > Here is my code,
> > > > >
> > > > >
> > > > > #include "mpi.h"
> > > > > #include <stdio.h>
> > > > > #include <stdlib.h>
> > > > > #include <string.h>
> > > > > #define ARRAYSIZE 2000000
> > > > > #define MASTER 0
> > > > >
> > > > > int data[ARRAYSIZE];
> > > > >
> > > > >
> > > > > int main(int argc, char* argv[])
> > > > > {
> > > > > int numtasks, taskid, rc, dest, offset, i, j, tag1, tag2,
> > > source, chunksize, namelen;
> > > > > int mysum, sum;
> > > > > int update(int myoffset, int chunk, int myid);
> > > > > char myname[MPI_MAX_PROCESSOR_NAME];
> > > > > MPI_Status status;
> > > > > double start, stop, time;
> > > > > double totaltime;
> > > > > FILE *fp;
> > > > > char line[128];
> > > > > char element;
> > > > > int n;
> > > > > int k=0;
> > > > >
> > > > >
> > > > >
> > > > > /***** Initializations *****/
> > > > > MPI_Init(&argc, &argv);
> > > > > MPI_Comm_size(MPI_COMM_WORLD, &numtasks);
> > > > > MPI_Comm_rank(MPI_COMM_WORLD,&taskid);
> > > > > MPI_Get_processor_name(myname, &namelen);
> > > > > printf ("MPI task %d has started on host %s...\n", taskid, myname);
> > > > > chunksize = (ARRAYSIZE / numtasks);
> > > > > tag2 = 1;
> > > > > tag1 = 2;
> > > > >
> > > > >
> > > > > /***** Master task only ******/
> > > > > if (taskid == MASTER){
> > > > >
> > > > > /* Initialize the array */
> > > > > sum = 0;
> > > > > for(i=0; i<ARRAYSIZE; i++) {
> > > > > data[i] = i * 1 ;
> > > > > sum = sum + data[i];
> > > > > }
> > > > > printf("Initialized array sum = %d\n",sum);
> > > > >
> > > > > /* Send each task its portion of the array - master keeps 1st
> part */
> > > > > offset = chunksize;
> > > > > for (dest=1; dest<numtasks; dest++) {
> > > > > MPI_Send(&offset, 1, MPI_INT, dest, tag1, MPI_COMM_WORLD);
> > > > > MPI_Send(&data[offset], chunksize, MPI_INT, dest, tag2,
> > > MPI_COMM_WORLD);
> > > > > printf("Sent %d elements to task %d offset= %d
> > > \n",chunksize,dest,offset);
> > > > > offset = offset + chunksize;
> > > > > }
> > > > >
> > > > > /* Master does its part of the work */
> > > > > offset = 0;
> > > > > mysum = update(offset, chunksize, taskid);
> > > > >
> > > > > /* Wait to receive results from each task */
> > > > > for (i=1; i<numtasks; i++) {
> > > > > source = i;
> > > > > MPI_Recv(&offset, 1, MPI_INT, source, tag1, MPI_COMM_WORLD,
> &status);
> > > > > MPI_Recv(&data[offset], chunksize, MPI_INT, source, tag2,
> > > > > MPI_COMM_WORLD, &status);
> > > > > }
> > > > >
> > > > > /* Get final sum and print sample results */
> > > > > MPI_Reduce(&mysum, &sum, 1, MPI_INT, MPI_SUM, MASTER,
> MPI_COMM_WORLD);
> > > > > /* printf("Sample results: \n");
> > > > > offset = 0;
> > > > > for (i=0; i<numtasks; i++) {
> > > > > for (j=0; j<5; j++)
> > > > > printf(" %d",data[offset+j]);ARRAYSIZE
> > > > > printf("\n");
> > > > > offset = offset + chunksize;
> > > > > }*/
> > > > > printf("\n*** Final sum= %d ***\n",sum);
> > > > >
> > > > > } /* end of master section */
> > > > >
> > > > >
> > > > > #include <stdlib.h>
> > > > > /***** Non-master tasks only *****/
> > > > >
> > > > > if (taskid > MASTER) {
> > > > >
> > > > > /* Receive my portion of array from the master task */
> > > > > start= MPI_Wtime();
> > > > > source = MASTER;
> > > > > MPI_Recv(&offset, 1, MPI_INT, source, tag1, MPI_COMM_WORLD,
> &status);
> > > > > MPI_Recv(&data[offset], chunksize, MPI_INT, source,
> > > tag2,MPI_COMM_WORLD, &status);
> > > > >
> > > > > mysum = update(offset, chunksize, taskid);
> > > > > stop = MPI_Wtime();
> > > > > time = stop -start;
> > > > > printf("time taken by process %d to recieve elements and
> > > caluclate own sum is = %lf seconds \n", taskid, time);
> > > > > totaltime = totaltime + time;
> > > > >
> > > > > /* Send my results back to the master task */
> > > > > dest = MASTER;
> > > > > MPI_Send(&offset, 1, MPI_INT, dest, tag1, MPI_COMM_WORLD);
> > > > > MPI_Send(&data[offset], chunksize, MPI_INT, MASTER, tag2,
> > > MPI_COMM_WORLD);
> > > > >
> > > > > MPI_Reduce(&mysum, &sum, 1, MPI_INT, MPI_SUM, MASTER,
> MPI_COMM_WORLD);
> > > > >
> > > > > } /* end of non-master */
> > > > >
> > > > > // printf("Total time taken for distribution is - %lf
> > > seconds", totaltime);
> > > > > MPI_Finalize();
> > > > >
> > > > > } /* end of main */
> > > > >
> > > > >
> > > > > int update(int myoffset, int chunk, int myid) {
> > > > > int i,j;
> > > > > int mysum;
> > > > > int mydata[myoffset+chunk];
> > > > > /* Perform addition to each of my array elements and keep my sum
> */
> > > > > mysum = 0;
> > > > > /* printf("task %d has elements:",myid);
> > > > > for(j = myoffset; j<myoffset+chunk; j++){
> > > > > printf("\t%d", data[j]);
> > > > > }
> > > > > printf("\n");*/
> > > > > for(i=myoffset; i < myoffset + chunk; i++) {
> > > > >
> > > > > //data[i] = data[i] + i;
> > > > > mysum = mysum + data[i];
> > > > > }
> > > > > printf("Task %d has sum = %d\n",myid,mysum);
> > > > > return(mysum);
> > > > > }
> > > > >
> > > > >
> > > > > When I run it with ARRAYSIZE = 2000000 The program works fine.
> > > But when I increase the size ARRAYSIZE = 20000000. The program ends
> > > with segmentation fault.
> > > > > I am running it on a cluster (machine 4 is master, machine 5,6
> > > are slaves) and np=20
> > > > >
> > > > > MPI task 0 has started on host machine4
> > > > > MPI task 2 has started on host machine4
> > > > > MPI task 3 has started on host machine4
> > > > > MPI task 14 has started on host machine4
> > > > > MPI task 8 has started on host machine6
> > > > > MPI task 10 has started on host machine6
> > > > > MPI task 13 has started on host machine4
> > > > > MPI task 4 has started on host machine5
> > > > > MPI task 6 has started on host machine5
> > > > > MPI task 7 has started on host machine5
> > > > > MPI task 16 has started on host machine5
> > > > > MPI task 11 has started on host machine6
> > > > > MPI task 12 has started on host machine4
> > > > > MPI task 5 has started on hostmachine5
> > > > > MPI task 17 has started on host machine5
> > > > > MPI task 18 has started on host machine5
> > > > > MPI task 15 has started on host machine4
> > > > > MPI task 19 has started on host machine5
> > > > > MPI task 1 has started on host machine4
> > > > > MPI task 9 has started on host machine6
> > > > > Initialized array sum = 542894464
> > > > > Sent 1000000 elements to task 1 offset= 1000000
> > > > > Task 1 has sum = 1055913696
> > > > > time taken by process 1 to recieve elements and caluclate own
> > > sum is = 0.249345 seconds
> > > > > Sent 1000000 elements to task 2 offset= 2000000
> > > > > Sent 1000000 elements to task 3 offset= 3000000
> > > > > Task 2 has sum = 328533728
> > > > > time taken by process 2 to recieve elements and caluclate own
> > > sum is = 0.274285 seconds
> > > > > Sent 1000000 elements to task 4 offset= 4000000
> > > > >
> --------------------------------------------------------------------------
> > > > > mpirun noticed that process rank 3 with PID 5695 on node
> > > machine4 exited on signal 11 (Segmentation fault).
> > > > >
> > > > > Any idea what could be wrong here?
> > > > >
> > > > >
> > > > > --
> > > > >
> > > > > Best Regards,
> > > > >
> > > > > ROHAN DESHPANDE
> > > > >
> > > > >
> > > > >
> > > > > _______________________________________________
> > > > > users mailing list
> > > > > us...@open-mpi.org
> > > > > http://www.open-mpi.org/mailman/listinfo.cgi/users
> > > >
> > > >
> > > > --
> > > > Jeff Squyres
> > > > jsquy...@cisco.com
> > > > For corporate legal information go to: http://www.cisco.com/web/
> > > about/doing_business/legal/cri/
> > > >
> > > >
> > > > _______________________________________________
> > > > users mailing list
> > > > us...@open-mpi.org
> > > > http://www.open-mpi.org/mailman/listinfo.cgi/users
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > > _______________________________________________
> > > > users mailing list
> > > > us...@open-mpi.org
> > > > http://www.open-mpi.org/mailman/listinfo.cgi/users
> > >
> > >
> > > --
> > > Jeff Squyres
> > > jsquy...@cisco.com
> > > For corporate legal information go to: http://www.cisco.com/web/
> > > about/doing_business/legal/cri/
> > >
> > >
> > > _______________________________________________
> > > users mailing list
> > > us...@open-mpi.org
> > > http://www.open-mpi.org/mailman/listinfo.cgi/users
> > >
> > >
> > >
> > > --
> > >
> > > Best Regards,
> > >
> > > ROHAN DESHPANDE
> > >
> >
> > > _______________________________________________
> > > users mailing list
> > > us...@open-mpi.org
> > > http://www.open-mpi.org/mailman/listinfo.cgi/users
> >
> > _______________________________________________
> > users mailing list
> > us...@open-mpi.org
> > http://www.open-mpi.org/mailman/listinfo.cgi/users
> >
> >
> >
> > --
> >
> > Best Regards,
> >
> > ROHAN DESHPANDE
> >
> >
> >
> > _______________________________________________
> > users mailing list
> > us...@open-mpi.org
> > http://www.open-mpi.org/mailman/listinfo.cgi/users
>
>
> --
> Jeff Squyres
> jsquy...@cisco.com
> For corporate legal information go to:
> http://www.cisco.com/web/about/doing_business/legal/cri/
>
>
> _______________________________________________
> users mailing list
> us...@open-mpi.org
> http://www.open-mpi.org/mailman/listinfo.cgi/users
>
--
Best Regards,
ROHAN DESHPANDE
