Hi,
I have built openmpi-v2.x-dev-1468-g6011906 on my machines (Solaris 10
Sparc, Solaris 10 x86_64, and openSUSE Linux 12.1 x86_64) with gcc-5.1.0
and Sun C 5.13. Unfortunately I have a problem with "--host" for a MPMD
program. The behaviour is different on different machines. Why do I need
two host names on my Sparc machine, although I only start one slave process?
tyr hello_2 126 ompi_info | grep -e "OPAL repo revision:" -e "C compiler
absolute:"
OPAL repo revision: v2.x-dev-1468-g6011906
C compiler absolute: /usr/local/gcc-5.1.0/bin/gcc
tyr hello_2 127 mpiexec --oversubscribe -np 1 --host tyr hello_2_mpi : -np 1
--host tyr hello_2_slave_mpi
--------------------------------------------------------------------------
While computing bindings, we found no available cpus on
the following node:
Node: tyr
Please check your allocation.
--------------------------------------------------------------------------
tyr hello_2 128 mpiexec --oversubscribe -np 1 --host tyr hello_2_mpi : -np 1
--host tyr,tyr hello_2_slave_mpi
Process 0 of 2 running on tyr.informatik.hs-fulda.de
Now 1 slave tasks are sending greetings.
Process 1 of 2 running on tyr.informatik.hs-fulda.de
Greetings from task 1:
message type: 3
msg length: 136 characters
message:
hostname: tyr.informatik.hs-fulda.de
operating system: SunOS
release: 5.10
processor: sun4u
tyr hello_2 129
Why does the program break with an error on my Linux and Solaris x86_64
machines, if I try the same thing? I get the same error message on my
Solaris x86_64 machine, which I've added for Linux.
linpc1 fd1026 107 ompi_info | grep -e "OPAL repo revision:" -e "C compiler
absolute:"
OPAL repo revision: v2.x-dev-1468-g6011906
C compiler absolute: /usr/local/gcc-5.1.0/bin/gcc
linpc1 fd1026 108 mpiexec --oversubscribe -np 1 --host linpc1 hello_2_mpi : -np
1 --host linpc1 hello_2_slave_mpi
--------------------------------------------------------------------------
While computing bindings, we found no available cpus on
the following node:
Node: linpc1
Please check your allocation.
--------------------------------------------------------------------------
linpc1 fd1026 109 mpiexec --oversubscribe -np 1 --host linpc1 hello_2_mpi : -np
1 --host linpc1,linpc1 hello_2_slave_mpi
--------------------------------------------------------------------------
A requested component was not found, or was unable to be opened. This
means that this component is either not installed or is unable to be
used on your system (e.g., sometimes this means that shared libraries
that the component requires are unable to be found/loaded). Note that
Open MPI stopped checking at the first component that it did not find.
Host: linpc1
Framework: ess
Component: pmi
--------------------------------------------------------------------------
[linpc1:24949] [[INVALID],INVALID] ORTE_ERROR_LOG: Not found in file
../../openmpi-1.10.3rc4/orte/runtime/orte_init.c at line 129
--------------------------------------------------------------------------
A requested component was not found, or was unable to be opened. This
means that this component is either not installed or is unable to be
used on your system (e.g., sometimes this means that shared libraries
that the component requires are unable to be found/loaded). Note that
Open MPI stopped checking at the first component that it did not find.
Host: linpc1
Framework: ess
Component: pmi
--------------------------------------------------------------------------
[linpc1:24950] [[INVALID],INVALID] ORTE_ERROR_LOG: Not found in file
../../openmpi-1.10.3rc4/orte/runtime/orte_init.c at line 129
--------------------------------------------------------------------------
It looks like orte_init failed for some reason; your parallel process is
likely to abort. There are many reasons that a parallel process can
fail during orte_init; some of which are due to configuration or
environment problems. This failure appears to be an internal failure;
here's some additional information (which may only be relevant to an
Open MPI developer):
orte_ess_base_open failed
--> Returned value Not found (-13) instead of ORTE_SUCCESS
--------------------------------------------------------------------------
--------------------------------------------------------------------------
It looks like orte_init failed for some reason; your parallel process is
likely to abort. There are many reasons that a parallel process can
fail during orte_init; some of which are due to configuration or
environment problems. This failure appears to be an internal failure;
here's some additional information (which may only be relevant to an
Open MPI developer):
orte_ess_base_open failed
--> Returned value Not found (-13) instead of ORTE_SUCCESS
--------------------------------------------------------------------------
--------------------------------------------------------------------------
It looks like MPI_INIT failed for some reason; your parallel process is
likely to abort. There are many reasons that a parallel process can
fail during MPI_INIT; some of which are due to configuration or environment
problems. This failure appears to be an internal failure; here's some
additional information (which may only be relevant to an Open MPI
developer):
ompi_mpi_init: ompi_rte_init failed
--> Returned "Not found" (-13) instead of "Success" (0)
--------------------------------------------------------------------------
--------------------------------------------------------------------------
It looks like MPI_INIT failed for some reason; your parallel process is
likely to abort. There are many reasons that a parallel process can
fail during MPI_INIT; some of which are due to configuration or environment
problems. This failure appears to be an internal failure; here's some
additional information (which may only be relevant to an Open MPI
developer):
ompi_mpi_init: ompi_rte_init failed
--> Returned "Not found" (-13) instead of "Success" (0)
--------------------------------------------------------------------------
*** An error occurred in MPI_Init
*** on a NULL communicator
*** MPI_ERRORS_ARE_FATAL (processes in this communicator will now abort,
*** and potentially your MPI job)
[linpc1:24949] Local abort before MPI_INIT completed successfully; not able to
aggregate error messages, and not able to guarantee that all other processes
were killed!
*** An error occurred in MPI_Init
*** on a NULL communicator
*** MPI_ERRORS_ARE_FATAL (processes in this communicator will now abort,
*** and potentially your MPI job)
[linpc1:24950] Local abort before MPI_INIT completed successfully; not able to
aggregate error messages, and not able to guarantee that all other processes
were killed!
-------------------------------------------------------
Primary job terminated normally, but 1 process returned
a non-zero exit code.. Per user-direction, the job has been aborted.
-------------------------------------------------------
--------------------------------------------------------------------------
mpiexec detected that one or more processes exited with non-zero status, thus
causing
the job to be terminated. The first process to do so was:
Process name: [[2401,1],0]
Exit code: 1
--------------------------------------------------------------------------
linpc1 fd1026 110
I would be grateful if somebody can fix the problem. Please let me
know, if you need more information. Thank you very much for any help
in advance.
Kind regards
Siegmar
/* Another MPI-version of the "hello world" program, which delivers
* some information about its machine and operating system. In this
* version the functions "master" and "slave" from "hello_1_mpi.c"
* are implemented as independant processes. This is the file for the
* "master".
*
*
* Compiling:
* Store executable(s) into local directory.
* mpicc -o <program name> <source code file name>
*
* Store executable(s) into predefined directories.
* make
*
* Make program(s) automatically on all specified hosts. You must
* edit the file "make_compile" and specify your host names before
* you execute it.
* make_compile
*
* Running:
* LAM-MPI:
* mpiexec -boot -np <number of processes> <program name>
* or
* mpiexec -boot \
* -host <hostname> -np <number of processes> <program name> : \
* -host <hostname> -np <number of processes> <program name>
* or
* mpiexec -boot [-v] -configfile <application file>
* or
* lamboot [-v] [<host file>]
* mpiexec -np <number of processes> <program name>
* or
* mpiexec [-v] -configfile <application file>
* lamhalt
*
* OpenMPI:
* "host1", "host2", and so on can all have the same name,
* if you want to start a virtual computer with some virtual
* cpu's on the local host. The name "localhost" is allowed
* as well.
*
* mpiexec -np <number of processes> <program name>
* or
* mpiexec --host <host1,host2,...> \
* -np <number of processes> <program name>
* or
* mpiexec -hostfile <hostfile name> \
* -np <number of processes> <program name>
* or
* mpiexec -app <application file>
*
* Cleaning:
* local computer:
* rm <program name>
* or
* make clean_all
* on all specified computers (you must edit the file "make_clean_all"
* and specify your host names before you execute it.
* make_clean_all
*
*
* File: hello_2_mpi.c Author: S. Gross
* Date: 01.10.2012
*
*/
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "mpi.h"
#define BUF_SIZE 255 /* message buffer size */
#define MAX_TASKS 12 /* max. number of tasks */
#define SENDTAG 1 /* send message command */
#define EXITTAG 2 /* termination command */
#define MSGTAG 3 /* normal message token */
#define ENTASKS -1 /* error: too many tasks */
int main (int argc, char *argv[])
{
int mytid, /* my task id */
ntasks, /* number of parallel tasks */
namelen, /* length of processor name */
num, /* number of chars in buffer */
i; /* loop variable */
char processor_name[MPI_MAX_PROCESSOR_NAME],
buf[BUF_SIZE + 1]; /* message buffer (+1 for '\0') */
MPI_Status stat; /* message details */
MPI_Init (&argc, &argv);
MPI_Comm_rank (MPI_COMM_WORLD, &mytid);
MPI_Comm_size (MPI_COMM_WORLD, &ntasks);
MPI_Get_processor_name (processor_name, &namelen);
/* With the next statement every process executing this code will
* print one line on the display. It may happen that the lines will
* get mixed up because the display is a critical section. In general
* only one process (mostly the process with rank 0) will print on
* the display and all other processes will send their messages to
* this process. Nevertheless for debugging purposes (or to
* demonstrate that it is possible) it may be useful if every
* process prints itself.
*/
fprintf (stdout, "Process %d of %d running on %s\n",
mytid, ntasks, processor_name);
fflush (stdout);
MPI_Barrier (MPI_COMM_WORLD); /* wait for all other processes */
if (ntasks > MAX_TASKS)
{
fprintf (stderr, "Error: Too many tasks. Try again with at most "
"%d tasks.\n", MAX_TASKS);
/* terminate all slave tasks */
for (i = 1; i < ntasks; ++i)
{
MPI_Send ((char *) NULL, 0, MPI_CHAR, i, EXITTAG, MPI_COMM_WORLD);
}
MPI_Finalize ();
exit (ENTASKS);
}
printf ("\n\nNow %d slave tasks are sending greetings.\n\n",
ntasks - 1);
/* request messages from slave tasks */
for (i = 1; i < ntasks; ++i)
{
MPI_Send ((char *) NULL, 0, MPI_CHAR, i, SENDTAG, MPI_COMM_WORLD);
}
/* wait for messages and print greetings */
for (i = 1; i < ntasks; ++i)
{
MPI_Recv (buf, BUF_SIZE, MPI_CHAR, MPI_ANY_SOURCE,
MPI_ANY_TAG, MPI_COMM_WORLD, &stat);
MPI_Get_count (&stat, MPI_CHAR, &num);
buf[num] = '\0'; /* add missing end-of-string */
printf ("Greetings from task %d:\n"
" message type: %d\n"
" msg length: %d characters\n"
" message: %s\n\n",
stat.MPI_SOURCE, stat.MPI_TAG, num, buf);
}
/* terminate all slave tasks */
for (i = 1; i < ntasks; ++i)
{
MPI_Send ((char *) NULL, 0, MPI_CHAR, i, EXITTAG, MPI_COMM_WORLD);
}
MPI_Finalize ();
return EXIT_SUCCESS;
}
/* Another MPI-version of the "hello world" program, which delivers
* some information about its machine and operating system. In this
* version the functions "master" and "slave" from "hello_1_mpi.c"
* are implemented as independant processes. This is the file for the
* "slave".
*
* To simulate "real work" the "slave"-process waits up to MAX_WTIME
* seconds before replying to a message request.
*
*
* Compiling:
* Store executable(s) into local directory.
* mpicc -o <program name> <source code file name>
*
* Store executable(s) into predefined directories.
* make
*
* Make program(s) automatically on all specified hosts. You must
* edit the file "make_compile" and specify your host names before
* you execute it.
* make_compile
*
* Running:
* LAM-MPI:
* mpiexec -boot -np <number of processes> <program name>
* or
* mpiexec -boot \
* -host <hostname> -np <number of processes> <program name> : \
* -host <hostname> -np <number of processes> <program name>
* or
* mpiexec -boot [-v] -configfile <application file>
* or
* lamboot [-v] [<host file>]
* mpiexec -np <number of processes> <program name>
* or
* mpiexec [-v] -configfile <application file>
* lamhalt
*
* OpenMPI:
* "host1", "host2", and so on can all have the same name,
* if you want to start a virtual computer with some virtual
* cpu's on the local host. The name "localhost" is allowed
* as well.
*
* mpiexec -np <number of processes> <program name>
* or
* mpiexec --host <host1,host2,...> \
* -np <number of processes> <program name>
* or
* mpiexec -hostfile <hostfile name> \
* -np <number of processes> <program name>
* or
* mpiexec -app <application file>
*
* Cleaning:
* local computer:
* rm <program name>
* or
* make clean_all
* on all specified computers (you must edit the file "make_clean_all"
* and specify your host names before you execute it.
* make_clean_all
*
*
* File: hello_2_slave.c Author: S. Gross
* Date: 01.10.2012
*
*/
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <time.h>
#include <unistd.h>
#include <sys/utsname.h>
#include "mpi.h"
#define BUF_SIZE 255 /* message buffer size */
#define MAX_TASKS 12 /* max. number of tasks */
#define MAX_WTIME 30 /* max. waiting time */
#define SENDTAG 1 /* send message command */
#define EXITTAG 2 /* termination command */
#define MSGTAG 3 /* normal message token */
#define ENTASKS -1 /* error: too many tasks */
int main (int argc, char *argv[])
{
struct utsname sys_info; /* system information */
int mytid, /* my task id */
ntasks, /* number of parallel tasks */
namelen, /* length of processor name */
more_to_do;
char processor_name[MPI_MAX_PROCESSOR_NAME],
buf[BUF_SIZE]; /* message buffer */
MPI_Status stat; /* message details */
MPI_Init (&argc, &argv);
MPI_Comm_rank (MPI_COMM_WORLD, &mytid);
MPI_Comm_size (MPI_COMM_WORLD, &ntasks);
MPI_Get_processor_name (processor_name, &namelen);
/* With the next statement every process executing this code will
* print one line on the display. It may happen that the lines will
* get mixed up because the display is a critical section. In general
* only one process (mostly the process with rank 0) will print on
* the display and all other processes will send their messages to
* this process. Nevertheless for debugging purposes (or to
* demonstrate that it is possible) it may be useful if every
* process prints itself.
*/
fprintf (stdout, "Process %d of %d running on %s\n",
mytid, ntasks, processor_name);
fflush (stdout);
MPI_Barrier (MPI_COMM_WORLD); /* wait for all other processes */
srand ((unsigned int) time ((time_t *) NULL) * mytid * mytid);
more_to_do = 1;
while (more_to_do == 1)
{
/* wait for a message from the master task */
MPI_Recv (buf, BUF_SIZE, MPI_CHAR, 0, MPI_ANY_TAG,
MPI_COMM_WORLD, &stat);
if (stat.MPI_TAG != EXITTAG)
{
uname (&sys_info);
strcpy (buf, "\n hostname: ");
strncpy (buf + strlen (buf), sys_info.nodename,
BUF_SIZE - strlen (buf));
strncpy (buf + strlen (buf), "\n operating system: ",
BUF_SIZE - strlen (buf));
strncpy (buf + strlen (buf), sys_info.sysname,
BUF_SIZE - strlen (buf));
strncpy (buf + strlen (buf), "\n release: ",
BUF_SIZE - strlen (buf));
strncpy (buf + strlen (buf), sys_info.release,
BUF_SIZE - strlen (buf));
strncpy (buf + strlen (buf), "\n processor: ",
BUF_SIZE - strlen (buf));
strncpy (buf + strlen (buf), sys_info.machine,
BUF_SIZE - strlen (buf));
sleep (rand () % MAX_WTIME);
MPI_Send (buf, strlen (buf), MPI_CHAR, stat.MPI_SOURCE,
MSGTAG, MPI_COMM_WORLD);
}
else
{
more_to_do = 0; /* terminate */
}
}
MPI_Finalize ();
return EXIT_SUCCESS;
}
