[Pw_forum] PAW parallel error

[Jelle van Sijl]

Dear All, 
first of, happy 2009!

I have compiled QE-4.0.4 (parallel) with intel fortran (v.11) and the
appropriate blas and lapack libraries (tried it with and without MKL).
Jobs that use pseudopotentials from the online database run smoothly. 
I am now testing a PAW pseudo potential made with ld1. The calculation
proceeds perfectly on QE-4.0.4 compiled in serial on my desktop, but
dies in parallel right at the start:

Program PWSCF     v.4.0.4  starts ...
     Today is  2Jan2009 at 8:31:38 

     Parallel version (MPI)

     Number of processors in use:       4
     R & G space division:  proc/pool =    4

     For Norm-Conserving or Ultrasoft (Vanderbilt) Pseudopotentials or
PAW

     Current dimensions of program pwscf are:
     Max number of different atomic species (ntypx) = 10
     Max number of k-points (npk) =  40000
     Max angular momentum in pseudopotentials (lmaxx) =  3
rank 0 in job 1 caused collective abort of all ranks

And the following error message appears:

########################################################################################################################
# FROM IOTK LIBRARY, VERSION 1.1.0development
# UNRECOVERABLE ERROR (ierr=1)
# ERROR IN: iotk_getline (iotk_scan.spp:812)
# CVS Revision: 1.7 
# 
iostat=-1
# ERROR IN: iotk_scan_tag (iotk_scan.spp:581)
# CVS Revision: 1.7 
# ERROR IN: iotk_scan (iotk_scan.spp:742)
# CVS Revision: 1.7 
# ERROR IN: iotk_scan_begin (iotk_scan.spp:94)
# CVS Revision: 1.7 
########################################################################################################################
application called MPI_Abort(MPI_COMM_WORLD, 1) - process 0[cli_0]:
aborting job:
application called MPI_Abort(MPI_COMM_WORLD, 1) - process 0

One of the issues may be that environmental variables are not passed to
all slave jobes on all nodes (from QE-wiki, comment by Axel I think).
However, even using a single node and 1 processor results in this error.
Another possible solution involving modifying a string in
xml_io_base.f90 (in ./Modules) did not work because that particular
change has already been included in this release.
I am at a bit at a loss how to resolve it, hope someone here can help
me.

cheers,

Jelle van Sijl


-- 
drs. Jelle van Sijl
PhD student, Department of Petrology
Faculty of Earth and Life Sciences (FALW)
VU University Amsterdam, De Boelelaan 1085
1081 HV Amsterdam, The Netherlands
Phone +31 (0)20 5987403
http://www.falw.vu.nl/petrology