[mpich-discuss] mpiexec and LSF

[Peer-Joachim Koch]

Thanks a lot !

I'll rewrite the wrapper script and see if it works!

Bye, Peer

Am 18.02.2011 16:58, schrieb Pavan Balaji:
>
> On 02/18/2011 03:11 AM, Dr.Peer-Joachim Koch wrote:
>> We are using LSF as batch system. Unfortunately the syntax of
>> mpiexec has changed. The LSF wrapper script is creating a
>> command line for mpiexec in the following form:
>>
>> mpiexec -n 1 -host foo master : -n 4 -host mysmp slave
>
> You can use:
>
> % mpiexec -n 1 ./master : -n 4 ./slave
>
>> What would be the correct syntax for the new mpiexec ?
>> -host -> -hosts thats clear, but how should the
>> master slave syntax correct using the new mpiexec ?
>>
>> Also I saw the option "-rmk lsf" - what is the advantage
>> and how can I use it in the LSF context ?
>
> Using the LSF RMK will allow mpiexec to automatically query for the
> allocated nodes. So you don't need to provide the hosts.
>
> Btw, even if you don't provide -rmk lsf, mpiexec will try to auto-detect
> LSF anyway.
>
> -- Pavan
>


-- 
Mit freundlichem Gruß
     Peer-Joachim Koch
_________________________________________________________
Max-Planck-Institut fuer Biogeochemie
Dr. Peer-Joachim Koch
Hans-Knöll Str.10             Telefon: ++49 3641 57-6705
D-07745 Jena                  Telefax: ++49 3641 57-7705

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5879 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://lists.mcs.anl.gov/pipermail/mpich-discuss/attachments/20110218/3d66c620/attachment.bin>