ABINIT Discussion Forums

Hello,

This is related to a previous post on the forum (viewtopic.php?f=9&t=356). I have similar experience and would appreciate if anyone can suggest a solution.

I have a large supercell with 128 atoms and the problems arises when the code starts to write the WF file. I am using paral_kgb option to speed up the calculation which is painfully slow otherwise. The code is able to write the _DEN file (with MPI I/O) however it fails to write the WF file. Details of compilation are mentioned below.

I was thinking if its possible to run the calculation in two steps : 1st step is the usual SCF calculation using band/k-point/FFT parallelization (paral_kgb =1, ...) and just write the _DEN file and then run a second non-SCF calculation (using the density from the previous step) with just k-point parallelization to write the WF file.

The code works well with k-point parallelization and can write decently big WF files without any problem. So my intention is to circumvent the MPI I/O problem by resorting to use regular FORTRAN routines. However since simple k-point parallelization would take a long time to converge the electronic steps I thought of this idea. This may seem strange so please let me know if there is a caveat to this idea.

I tried to run this two step calculation (setting ndtset 2) but unfortunately when code encounters the following in the input file

#parallelization variables
paral_kgb1 1
wfoptalg1=4
nloalg1=4
fftalg1=401
intxc1=0
fft_opt_lob1=2

npkpt1 4 <------------- problem is reported for this variable
npband1 9
npfft1 2
bandpp1 2
#accesswff 1 1 1 1
istwfk1 1 1 1 1

I get an error which says ...

intagm : ERROR -
In the input file, there is an occurence of the
keyword "npkpt", appended with the digit "1".
This is forbidden when ndtset==0 .
Action : remove this occurence, or change ndtset.

I don't know how to solve the source of this error. Any help is very much appreciated.

I can post the input file if that can help.

best regards,
Anurag

I am running ABINIT 6.0.4 compiled on a linux cluster with Intel 10.0 fortran compiler and open mpi 1.4.1. ABINIT was configured as follows:

$ ./configure --prefix=/global/home/users/achaudhry --enable-mpi-fft=yes --enable-mpi-io=yes --enable-fttw=yes --enable-64bit-flags=yes FC=mpif90 F77=mpif90 --enable-scalapack CC=mpicc CXX=mpiCC --with-mpi-runner=/global/software/centos-5.x86_64/modules/openmpi/1.4.1-intel/bin/mpirun --with-mpi-level=2 CC_LIBS=-lmpi CXX_LIBS=-lmpi++ -lmpi --with-fc-vendor=intel --with-linalg-includes=-I/global/software/centos-5.x86_64/modules/mkl/10.0.4.023/include --with-linalg-libs=-L/global/software/centos-5.x86_64/modules/mkl/10.0.4.023/lib/em64t -lmkl_intel_lp64 -lmkl_intel_thread -lmkl_intel_solver_lp64 -lmkl_lapack -lmkl_core -lguide -lpthread --with-scalapack-libs=-L/global/software/centos-5.x86_64/modules/mkl/10.0.4.023/lib/em64t -lmkl_scalapack_lp64 -lmkl_blacs_openmpi_lp64 -lmkl_lapack -lmkl_intel_lp64 -lmkl_intel_thread -lmkl_lapack -lmkl_core -liomp5 -lpthread --disable-wannier90