Installation of code-saturne V2RC1 in parallel
Posted: Thu Jul 22, 2010 12:50 pm
Hi,
I installed code-saturne on a cluster.
The first test I did was to run a calcul in batch and it was OK.
The same calcul using qsub on one proc is also OK.
My problem is when I submit the calcul on many proc. (eg : nodes=1:ppn=8)
I obtain the following error (:
------------------------------------------------------------------------------
Total number of processors: 8
Parallel Code_Saturne with partitioning in 8 sub-domains
Total number of processors: 8
Code_Saturne is running
***********************
Working directory (to be periodically cleaned) :
/home/sylvain.serra/tmp_Saturne/study.CASE.07191111
Kernel version: /share/apps/code_saturne/2.0/ncs-2.0.0-rc1/bin
Preprocessor: /share/apps/code_saturne/2.0/librairies/ecs-2.0.0-rc1/bin
********************************************
Preparing calculation
********************************************
********************************************
Starting calculation
********************************************
[chapo41.local:03137] pls:tm: failed to poll for a spawned proc, return status = 17002
[chapo41.local:03137] [0,0,0] ORTE_ERROR_LOG: In errno in file rmgr_urm.c at line 462
[chapo41.local:03137] mpiexec: spawn failed with errno=-11
Error running the calculation.
Check Kernel log (listing) and error* files for details
********************************************
Error in calculation stage.
********************************************
---------------------------------------------------------------------------------------
In the lispre.xxx and listpart.xxx files, it seems OK
.-----------------------.
| |
| Preprocessor finish |
| |
`-----------------------'
.------------------------.
| Partitioner finish |
`------------------------'
and there are neither error.xxx file nor listing file...
Anybody could help me?
Best regards,
Sylvain
I installed code-saturne on a cluster.
The first test I did was to run a calcul in batch and it was OK.
The same calcul using qsub on one proc is also OK.
My problem is when I submit the calcul on many proc. (eg : nodes=1:ppn=8)
I obtain the following error (:
------------------------------------------------------------------------------
Total number of processors: 8
Parallel Code_Saturne with partitioning in 8 sub-domains
Total number of processors: 8
Code_Saturne is running
***********************
Working directory (to be periodically cleaned) :
/home/sylvain.serra/tmp_Saturne/study.CASE.07191111
Kernel version: /share/apps/code_saturne/2.0/ncs-2.0.0-rc1/bin
Preprocessor: /share/apps/code_saturne/2.0/librairies/ecs-2.0.0-rc1/bin
********************************************
Preparing calculation
********************************************
********************************************
Starting calculation
********************************************
[chapo41.local:03137] pls:tm: failed to poll for a spawned proc, return status = 17002
[chapo41.local:03137] [0,0,0] ORTE_ERROR_LOG: In errno in file rmgr_urm.c at line 462
[chapo41.local:03137] mpiexec: spawn failed with errno=-11
Error running the calculation.
Check Kernel log (listing) and error* files for details
********************************************
Error in calculation stage.
********************************************
---------------------------------------------------------------------------------------
In the lispre.xxx and listpart.xxx files, it seems OK
.-----------------------.
| |
| Preprocessor finish |
| |
`-----------------------'
.------------------------.
| Partitioner finish |
`------------------------'
and there are neither error.xxx file nor listing file...
Anybody could help me?
Best regards,
Sylvain