[mpich-discuss] Problem with running mpiexec on more than oneWindows XP machines

[Jayesh Krishna]

Hi,
 Can you try providing the complete path for cpi.exe (mpiexec -n 2
-machinefile mf.txt c:\progra~1\mpich2\examples\cpi.exe)?
 
Regards,
Jayesh

  _____  

From: Sole Eko [mailto:sole.eko at gmail.com] 
Sent: Thursday, February 19, 2009 3:39 AM
To: Jayesh Krishna
Subject: Re: [mpich-discuss] Problem with running mpiexec on more than
oneWindows XP machines


Hi Jayesh,

Thanks for advices.

I tried all that you recommended and still the same result..
I also tried to run:
    On machine1, mpiexec -n 1 -host machine2_ipaddress hostname 
    On machine2, mpiexec -n 1 -host machine1_ipaddress hostname
and it worked well.. but when I try to run on both machines, like

mpiexec -n 2 -machinefile mf.txt cpi.exe

and what is happening is this: (I turned on debugging mode of smpd: smpd
-d , to see what is going on) and everything seams to be all right till I
enter number of intervals on, lets say, machine1, when I hit ENTER button
there is somethin happening on machine1, but on machine2 nothing is listed
in console that runs smpd -d in mid time, last few lines in that console
are those:
......
[02:2504]...../smpd_post_read_command
[02:2504]..../smpd_state_reading_cmd
[02:2504].../smpd_handle_op_read
[02:2504]...sock_waiting for the next event.
[02:2504]...SOCK_OP_WRITE
[02:2504]...\smpd_handle_op_write
[02:2504]....\smpd_state_writing_cmd
[02:2504].....wrote command
[02:2504].....command written to pmi: "cmd=result src=1 dest=2 tag=5
cmd_tag=2 c
md_orig=barrier ctx_key=0 result=DBS_SUCCESS "
[02:2504].....\smpd_free_command
[02:2504]......\smpd_init_command
[02:2504]....../smpd_init_command
[02:2504]...../smpd_free_command
[02:2504]..../smpd_state_writing_cmd
[02:2504].../smpd_handle_op_write
[02:2504]...sock_waiting for the next event.

and, like I said, nothing else is listed on machine2 after hitting enter
on machine1.
if I cancel this job on machine1 with cotrl+c -> smpd seams to work just
fine on both machines..

what could be wrong?

Regards, 
Sole


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/mpich-discuss/attachments/20090219/75db1b05/attachment.htm>