[RegCNET] error in parallel run

[bixq]

Hi,

On Thu, 29 Oct 2009, rajesh j wrote:

>
> While trying to run RegCM3 on a Linux Cluster with 8 processors, the
> following error message came. I could not resolve it. Please give a
> mention if anybody is familiar with this. Earlier I have installed the
> evaluation version of MPI from intel site successfully. The error
> message is given below. Thank you in advance
>
>
>
> ...............................................................................
>
> Common2/split.cb(53): warning #6375: Because of COMMON, the alignment of object is inconsistent with its type   [G1]
>
>       real*8  g1

I guess these messages related to Common2/split.cb are just warning, to 
avoid these warning, you can move the iw2 line and let it become the last
line of the common /cdiag/


> /opt/intel/impi/3.2.2.006//bin64/mpiifort -o regcm -convert big_endian
> aermix.o aeroppt.o aerout.o albedov.o balanc.o balbak.o bconst.o
> bdyin.o bdyuv.o bdyval.o blhnew.o BLOCKDATA001.o bmpara.o bndry.o
> carbon.o cdiv.o chdrydep.o chemtap.o chrsetc.o chsrfem.o cldefr.o
> cldems.o cldfrac.o co2.o colmod3.o conadv.o condch.o condcq.o condtq.o
>
> ifort: command line warning #10006: ignoring unknown option '-Msignextend'
>
> ifort: command line warning #10006: ignoring unknown option '-Mcray=pointer'
>
> ifort: command line warning #10159: invalid argument for option '-tp'
>
> ifort: command line warning #10006: ignoring unknown option '-Mnoframe'
>
> ifort: command line warning #10006: ignoring unknown option '-fastsse'

Just delete those compiling option in the Makefile you used.

>
> /opt/intel/impi/3.2.2.006/lib64/libmpi.a(simple_fpmi.o)(.text+0x429): In function `PMI_Init':
>
> : warning: Using 'dlopen' in statically linked applications requires at
> runtime the shared libraries from the glibc version used for linking
>
> /opt/intel/impi/3.2.2.006/lib64/libmpi.a(rdma_iba_util.o)(.text+0x22b): In function `MPIDI_CH3I_RDMA_util_get_ia_addr':
>
> : warning: Using 'getaddrinfo' in statically linked applications
> requires at runtime the shared libraries from the glibc version used
> for linking
>
> /opt/intel/impi/3.2.2.006/lib64/libmpi.a(sock.o)(.text+0x400c): In function `MPIDU_Sock_get_host_description':
>
> : warning: Using 'gethostbyaddr' in statically linked applications
> requires at runtime the shared libraries from the glibc version used
> for linking
>
> /opt/intel/impi/3.2.2.006/lib64/libmpi.a(simple_pmi.o)(.text+0x30e3): In function `iPMI_Init':
>
> : warning: Using 'gethostbyname' in statically linked applications
> requires at runtime the shared libraries from the glibc version used
> for linking
>
>  The number of CPU is not well set
>
>  NCPU =           1    nproc =          8

I think this is the key point, you are using NPROC=8 in regcm.param2, but 
just use 1 CPU when you submit the job.

Note: you'd better run the commands in regcm.x step by step, and use

mpirun -np 8 ./regcm          as the last command
                              (instead of
                                           ./regcm < regcm.in ) 
>
> [unset]: aborting job:
>
> application called MPI_Abort(MPI_COMM_WORLD, 0) - process 0
>
> [user at cluster Run-parallel]$
>
>
>
> Mr. Rajesh J.
>
> Junior Research Fellow,
>
> Department of Atmospheric Sciences,
>
> School of Marine Sciences,
>
> Cochin University of Science and Technology, Lakeside Campus,
>
> Finearts Avenue, Cochin 682016
>
> Phone: 9846812663
>
> e-mail: rajeshj at cusat.ac.in
>
>
>

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   Dr. Xunqiang Bi         email:bixq at ictp.it
   Earth System Physics Group
   The Abdus Salam ICTP
   Strada Costiera, 11
   P.O. BOX 586, 34100 Trieste, ITALY
   Tel: +39-040-2240302  Fax: +39-040-2240449
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~