Hybrid job example

From MediaWiki

(Difference between revisions)
Jump to: navigation, search
(Created page with 'In this exercise user should obtain prepared mpich job, extract archive, list content of files, submit job on PARADOX cluster, monitor his progress with information from queue an…')
Line 1: Line 1:
-
In this exercise user should obtain prepared mpich job, extract archive, list content of files, submit job on PARADOX cluster, monitor his progress with information from queue and when job is done list resaults file.
+
In this exercise user should obtain prepared hybrid job, extract archive, list content of files, submit job on PARADOX cluster, monitor his progress with information from queue and when job is done list resaults file.
1. Login on ui.ipb.ac.rs:
1. Login on ui.ipb.ac.rs:
Line 11: Line 11:
3. Download tgz archive with example files.
3. Download tgz archive with example files.
-
  wget http://wiki.ipb.ac.rs/images/c/cc/Mpich.tgz
+
  wget http://wiki.ipb.ac.rs/images/b/bb/Hybrid.tgz
4. Extract archive :
4. Extract archive :
-
  $ tar xvzf Mpich.tgz
+
  $ tar xvzf Hybrid.tgz
5. Enter Mpich folder
5. Enter Mpich folder
-
  $ cd Mpich
+
  $ cd Hybrid
6. List content of folder:
6. List content of folder:
Line 25: Line 25:
  $ ll
  $ ll
-
7. List content of job.pbs and job.c files:
+
7. List content of job.pbs and hybrid_example.c files:
  $ cat job.pbs
  $ cat job.pbs
Line 31: Line 31:
  #!/bin/bash
  #!/bin/bash
  #PBS -q hpsee
  #PBS -q hpsee
-
  #PBS -l nodes=3:ppn=2
+
  #PBS -l nodes=4:ppn=8
  #PBS -l walltime=10:00:00
  #PBS -l walltime=10:00:00
  #PBS -e ${PBS_JOBID}.err
  #PBS -e ${PBS_JOBID}.err
  #PBS -o ${PBS_JOBID}.out
  #PBS -o ${PBS_JOBID}.out
 +
export OMP_NUM_THREADS=8
   
   
  cd $PBS_O_WORKDIR
  cd $PBS_O_WORKDIR
  chmod +x job
  chmod +x job
-
  cat $PBS_NODEFILE
+
   
-
  ${MPI_MPICH_MPIEXEC} ./job
+
  ${MPI_OPENMPI_MPIEXEC} -np 4 -npernode 1 ./job
-
  $ cat job.c
+
  $ cat hybrid_example.c
-
# include <stdio.h>
+
  #include <mpi.h>       /* MPI Library            */
-
# include <math.h>
+
#include <omp.h>       /* OpenMP Library */
-
  # include "mpi.h"
+
#include <stdio.h>      /* printf()       */
-
+
#include <stdlib.h>     /* EXIT_SUCCESS    */
-
/* Evaluation of the inregration function
+
 
-
    f(x)=1/(1+x*x) (0<=x<=1) */
+
int main (int argc, char *argv[]) {
-
double f(double x)
+
 
-
{
+
    /* Parameters of MPI.                                                       
-
     double value;
+
*/     int M_N;                             /* number of MPI ranks                 
-
    value=4.0/(1.0+x*x);
+
*/     int M_ID;                           /* MPI rank ID                         
-
    return value;
+
*/     int rtn_val;                         /* return value                         
-
}
+
*/     char name[128];                     /*
-
+
MPI_MAX_PROCESSOR_NAME == 128        */
-
int main (int argc, char *argv[])
+
    int namelen;
-
{
+
 
-
    int counter;
+
    /* Parameters of OpenMP.                                                      */
-
    double walltime_start,walltime_end,walltime_diff;
+
    int O_P;                             /* number of OpenMP processors          */
-
+
    int O_T;                             /* number of OpenMP threads              */
-
     int master_node = 0;
+
    int O_ID;                           /* OpenMP thread ID                      */
-
     int number_of_nodes;
+
 
-
     int node_id;
+
    /* Initialize MPI.                                                            */
-
+
    /* Construct the default communicator MPI_COMM_WORLD.                        */
-
     int number_of_intervals=300000000;
+
    rtn_val = MPI_Init(&argc,&argv);
-
    double node_integral;
+
 
-
    double step;
+
    /* Get a few MPI parameters.                                                  */
-
+
    rtn_val = MPI_Comm_size(MPI_COMM_WORLD,&M_N);   /* get number of MPI ranks  */
-
    double x;
+
    rtn_val = MPI_Comm_rank(MPI_COMM_WORLD,&M_ID);   /* get MPI rank ID          */
-
+
    MPI_Get_processor_name(name,&namelen);
-
    double pi_estimate;
+
    printf("name:%s  M_ID:%d M_N:%d\n", name,M_ID,M_N);
-
    double pi_diff;
+
 
-
    double pi_exact = 3.141592653589793238462643;
+
    /* Get a few OpenMP parameters.                                              */
-
+
    O_P = omp_get_num_procs();         /* get number of OpenMP processors      */
-
/* Establish the MPI environment */
+
    O_T = omp_get_num_threads();        /* get number of OpenMP threads          */
-
    MPI_Init (&argc, &argv);
+
    O_ID = omp_get_thread_num();         /* get OpenMP thread ID                  */
-
/* Get the number of processes */
+
 
-
    MPI_Comm_size (MPI_COMM_WORLD,&number_of_nodes);
+
    printf("name:%s  M_ID:%d O_ID:%d  O_P:%d  O_T:%d\n", name,M_ID,O_ID,O_P,O_T);
-
/* Determine this processes's rank */
+
-
    MPI_Comm_rank (MPI_COMM_WORLD,&node_id);
+
-
+
-
    if (node_id==master_node)
+
-
    {
+
-
      printf("Compiled on %s at %s\n",__DATE__, __TIME__);
+
-
      printf("Number of nodes: %d\n",number_of_nodes);
+
-
      printf ( "Number of intervals: %d\n",number_of_intervals);
+
-
      walltime_start=MPI_Wtime();
+
-
    }
+
-
   
+
-
    MPI_Bcast (&number_of_nodes,1,MPI_INT,master_node,MPI_COMM_WORLD );
+
-
   
+
-
    step=1.0/(double)number_of_intervals;
+
-
    node_integral=0.0;
+
-
    for (counter=node_id+1;counter<=number_of_intervals;counter=counter+number_of_nodes)
+
-
    {
+
-
       x=step*((double)counter-0.5);
+
-
      node_integral=node_integral+f(x);
+
-
    }
+
-
+
-
    printf("\nNode: %d\n",node_id);
+
-
    printf("Node integral: %f\n",node_integral*step);
+
-
+
-
    MPI_Reduce (&node_integral,&pi_estimate,1,MPI_DOUBLE,MPI_SUM,master_node,MPI_COMM_WORLD);
+
   
   
-
/* The master node prints the answer */
+
    /* PARALLEL REGION                                                            */
-
    if (node_id==master_node)
+
    /* Thread IDs range from 0 through omp_get_num_threads()-1.                  */
-
    {
+
    /* We execute identical code in all threads (data parallelization).          */
-
      pi_estimate=pi_estimate*step;
+
    #pragma omp parallel private(O_ID)
-
      pi_diff=fabs(pi_estimate-pi_exact);
+
    {
-
      printf("PI exact  : %24.16f\n", pi_exact);
+
    O_ID = omp_get_thread_num();         /* get OpenMP thread ID                */
-
      printf("PI estimate: %24.16f\n", pi_estimate);
+
    MPI_Get_processor_name(name,&namelen);
-
      printf("PI diff    : %24.16f\n", pi_diff);
+
    printf("parallel region:      name:%s M_ID=%d O_ID=%d\n", name,M_ID,O_ID);
 +
    }
   
   
-
      walltime_end=MPI_Wtime();
+
    /* Terminate MPI.                                                            */
-
      walltime_diff=walltime_end-walltime_start;
+
    rtn_val = MPI_Finalize();
-
      printf ("Wall clock: %f\n", walltime_diff );
+
-
    }
+
   
   
-
/* Shut down MPI */
+
    /* Exit master thread.                                                        */
-
    MPI_Finalize();
+
    printf("name:%s M_ID:%d O_ID:%d  Exits\n", name,M_ID,O_ID);  
-
}
+
    return EXIT_SUCCESS;
 +
}
8. Submit job :
8. Submit job :

Revision as of 21:54, 12 October 2011

In this exercise user should obtain prepared hybrid job, extract archive, list content of files, submit job on PARADOX cluster, monitor his progress with information from queue and when job is done list resaults file.

1. Login on ui.ipb.ac.rs: 

$ ssh ngrkic@ui.ipb.ac.rs

2. Navigate to your folder in nfs filesystem. 

$ cd /nfs/ngrkic

3. Download tgz archive with example files. 

wget http://wiki.ipb.ac.rs/images/b/bb/Hybrid.tgz

4. Extract archive : 

$ tar xvzf Hybrid.tgz

5. Enter Mpich folder 

$ cd Hybrid

6. List content of folder: 

$ ll

7. List content of job.pbs and hybrid_example.c files: 

$ cat job.pbs

#!/bin/bash
#PBS -q hpsee
#PBS -l nodes=4:ppn=8
#PBS -l walltime=10:00:00
#PBS -e ${PBS_JOBID}.err
#PBS -o ${PBS_JOBID}.out
export OMP_NUM_THREADS=8

cd $PBS_O_WORKDIR
chmod +x job

${MPI_OPENMPI_MPIEXEC} -np 4 -npernode 1 ./job

$ cat hybrid_example.c

  #include <mpi.h>        /* MPI Library             */

#include <omp.h> /* OpenMP Library */ #include <stdio.h> /* printf() */ #include <stdlib.h> /* EXIT_SUCCESS */ 

	int main (int argc, char *argv[]) {
  
	    /* Parameters of MPI.
  • / int M_N; /* number of MPI ranks
  • / int M_ID; /* MPI rank ID
  • / int rtn_val; /* return value
  • / char name[128]; /*

MPI_MAX_PROCESSOR_NAME == 128 */ int namelen;

/* Parameters of OpenMP. */ int O_P; /* number of OpenMP processors */ int O_T; /* number of OpenMP threads */ int O_ID; /* OpenMP thread ID */

/* Initialize MPI. */ /* Construct the default communicator MPI_COMM_WORLD. */ rtn_val = MPI_Init(&argc,&argv);

/* Get a few MPI parameters. */ rtn_val = MPI_Comm_size(MPI_COMM_WORLD,&M_N); /* get number of MPI ranks */ rtn_val = MPI_Comm_rank(MPI_COMM_WORLD,&M_ID); /* get MPI rank ID */ MPI_Get_processor_name(name,&namelen); printf("name:%s M_ID:%d M_N:%d\n", name,M_ID,M_N);

/* Get a few OpenMP parameters. */ O_P = omp_get_num_procs(); /* get number of OpenMP processors */ O_T = omp_get_num_threads(); /* get number of OpenMP threads */ O_ID = omp_get_thread_num(); /* get OpenMP thread ID */

printf("name:%s M_ID:%d O_ID:%d O_P:%d O_T:%d\n", name,M_ID,O_ID,O_P,O_T);

/* PARALLEL REGION */ /* Thread IDs range from 0 through omp_get_num_threads()-1. */ 

	    /* We execute identical code in all threads (data parallelization).           */
	    #pragma omp parallel private(O_ID)
	    {
	    O_ID = omp_get_thread_num();          /* get OpenMP thread ID                 */
	    MPI_Get_processor_name(name,&namelen);
	    printf("parallel region:       name:%s M_ID=%d O_ID=%d\n", name,M_ID,O_ID);
	    }

	    /* Terminate MPI.                                                             */
	    rtn_val = MPI_Finalize();

	    /* Exit master thread.                                                        */
	    printf("name:%s M_ID:%d O_ID:%d   Exits\n", name,M_ID,O_ID);

return EXIT_SUCCESS; 

	}

8. Submit job : 

qsub job.pbs

qsub will print output : 

<jobID>.ce64.ipb.ac.rs

9. Monitor your job : 

qstat <jobID>.ce64.ipb.ac.rs

10.When job is done list content of <jobID>.ce64.ipb.ac.rs.out file : 

cat <jobID>.ce64.ipb.ac.rs.out

n06.ipb.ac.rs
n06.ipb.ac.rs
n08.ipb.ac.rs
n08.ipb.ac.rs
n13.ipb.ac.rs
n13.ipb.ac.rs

Node: 2
Node integral: 0.523599

Node: 1
Node integral: 0.523599

Node: 3
Node integral: 0.523599

Node: 4
Node integral: 0.523599

Node: 5
Node integral: 0.523599
Compiled on Oct 12 2011 at 11:34:11
Number of nodes: 6
Number of intervals: 300000000

Node: 0
Node integral: 0.523599
PI exact   :       3.1415926535897931
PI estimate:       3.1415926535899650
PI diff    :       0.0000000000001719
Wall clock: 1.894531
Retrieved from "http://wiki.ipb.ac.rs/index.php/Hybrid_job_example"
Personal tools