[CIG-MC] MPI_TYPE_DEPTH environmet variable

Alina Hale a.hale at usyd.edu.au
Thu Jan 8 21:05:46 PST 2009


Hello,

 

I'm trying to run CitcomS on an Altix Ice supercomputer. I've been
working through the user manual but have a problem. When I try to run
example1.cfg, I get an error message.

 

i.e. using the command:

citcoms example1.cfg --solver.output.output_format=hdf5

 

I get the error:

MPI Error.  The program attempted to construct a derived datatype with
depth 10, but the maximum allowed depth is 8.  You can increase the
allowed depth via the MPI_TYPE_DEPTH environmet variable.

MPI Error.  The program attempted to construct a derived datatype with
depth 10, but the maximum allowed depth is 8.  You can increase the
allowed depth via the MPI_TYPE_DEPTH environmet variable.

MPI Error, rank:2, function:MPI_TYPE_STRUCT, Invalid datatype argument
MPI Error.  The program attempted to construct a derived datatype with
depth 10, but the maximum allowed depth is 8.  You can increase the
allowed depth via the MPI_TYPE_DEPTH environmet variable.

MPI Error, rank:3, function:MPI_TYPE_STRUCT, Invalid datatype argument
MPI Error, rank:1, function:MPI_TYPE_STRUCT, Invalid datatype argument
MPI Error.  The program attempted to construct a derived datatype with
depth 10, but the maximum allowed depth is 8.  You can increase the
allowed depth via the MPI_TYPE_DEPTH environmet variable.

MPI Error, rank:0, function:MPI_TYPE_STRUCT, Invalid datatype argument

--pyre-start: mpirun: exit -9

/sw/apps/CitcomS/bin/citcoms: /sw/apps/CitcomS/bin/pycitcoms: exit 1

 

Please can you tell me what I need to do? I understand that I need to
set the env virable MPI_TYPE_DEPTH=10 and rerun the program, but where
do I set this?

 

Many  thanks,

Alina

 

-----------------------------------------

A. J. Hale

 

Australian Postdoctoral Research Fellow (APD)

School of Geosciences 

Madsen Building 

University of Sydney  

NSW  2006, AUSTRALIA

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://geodynamics.org/pipermail/cig-mc/attachments/20090109/c16fb4ef/attachment.htm 


More information about the CIG-MC mailing list