From: Neelanjana Sengupta (senguptan_at_gmail.com)
Date: Wed Oct 12 2011 - 02:29:16 CDT
Hi Matteo and others,
I have actually been trying out things from the notes.
Since the ++mpiexec option had worked in a previous cluster (but it isn't
working here), here's one thing I tried:
Create a mympiexec in home directory:
shift; shift; exec ibrun $**
*charmrun +p12 namd2 ++verbose ++mpiexec ++remote-shell
/home1/nsengupta/mympiexec min01.inp > min01.out*
I now get:
*Charmrun> charmrun started...
Charmrun> mpiexec started
Charmrun> node programs all started
Charmrun> Couldn't find mpiexec program '/home1/nsengupta/mympiexec'!
Charmrun> error 0 attaching to node:
Timeout waiting for node-program to connect*
The error message is similar if I do not give an mpiexec file; I get *Charmrun>
Couldn't find mpiexec program 'mpiexec'!*
Can someone please help?
On Wed, Oct 12, 2011 at 12:07 AM, Matteo Rotter <matteo.rotter_at_uniud.it>wrote:
> read the notes
> Quoting Neelanjana Sengupta <senguptan_at_gmail.com>:
> Dear NAMD experts,
>> We have installed NAMD2.8 in our cluster (Intel Xeon X5670 processors;
>> running RHEL5.0), which presently has rsh enabled on each node. I have a
>> .nodelist file of this type in my home directory:
>> *group main
>> host cn001
>> host cn002
>> My submit script contains the foll:
>> *charmrun +p12 namd2 min01.inp > min01.out*
>> However, rsh is refusing the connection, and I get error messages of this
>> kind for each core:
>> *trying normal rsh (/usr/bin/rsh)
>> connect to address 10.1.128.12 port 544: Connection refused
>> Permission denied.
>> Charmrun> Error 1 returned from rsh (cn001:0)*
>> Can someone throw some light on this?
>> Thanks and regards,
> SEMEL (SErvizio di Messaging ELettronico) - AINF, Universita' di Udine
This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:20:53 CST