The ABINIT Users Mailing List ( CLOSED )

["Chun Li " <chun.li@mf.mpg.de>]

Dear abinit users,

I'm trying to compile abinip on a PC with linux OS. Because there are 2
CPUs, I just want to test the parallel version of abinit.

I set up my makefile_macro file with the following text(modified the
file in the directory of
~ABINIT/Machine_dept_files/Intel_P6/makefile_macros.PGI_Lam):

# This file contains the definitions needed for every possible script #
to be used with the ABINIT makefiles.  It is very much
machine-dependent. # It is read by makemake and makearch.

# This version for a cluster of 16 Pentium II processors at Georgetown.
# It is based on the PGI compiler, and the LAM library of MPI routines.
# It was communicated by A. Quandt. See explanations at the end of this
# files, including modifications of .cshrc

# Explanations about this file can be found in the  makefile_macros_help
file, # in the Infos subdirectory

####################################################################
# For the sequential version

# Machine type
MACHINE=P6
# Fortran optimized compilation FC=/usr/local/software/lam/bin/hf77
FFLAGS=-Mextend -Mfree -O2
FFLAGS_LIBS= -O2
# C preprocessor, used to preprocess the fortran source.
CPP=/usr/bin/cpp CPP_FLAGS=-P # C optimized compilation.
CC=/usr/local/software/bin/pgcc CFLAGS=-O # Location of perl . Used to
generate the script fldiff, in ~ABINIT/Utilities .
PERL=/usr/local/bin/perl # List of machine-dependent routines
MACHINE_DEP_C_SEQ_SUBS_LIST=etime.o

####################################################################
# For the parallel version

# Compiler flags and definitions
 FFLAGS_PAR=$(FFLAGS) -I /usr/local/software/lam/include
# List of machine-dependent routines
MACHINE_DEP_C_PAR_SUBS_LIST=etime.par
# Location of the MPI library
MPI_A=/usr/local/software/lam/lib/libmpi.a
# MPI_A=/usr/local/pgi/linux86/lib/libpghpf_mpi.a
# Include blas, lapack, and any other libraries here
 LIBS_PAR=$(LIBS) $(MPI_A)

####################################################################
# Additional explanations (provided by A. Quandt)

# The fortran compilation uses the Portland Group (PGI) compiler. #As
you can see, we first tried to locate all the libraries necessary to 
#operate with mpirun, but we found out that the standard MPI library of 
#the pgf90 alone was somewhat incomplete. Therefore, we used a wrapper 
#called hf77, which provided the original pgf90 compiler with all the 
#LAM and MPI features that it probably ever needed. The way to do this 
#was to specify FC=hf77 instead of FC=pgf90 in makefile_macros (see
above), 
#and to create a .cshrc along the following lines :

#setenv PGI /usr/global/pgi
#set path = ( . $PGI/linux86/bin /usr/global/bin $path )
#set path = ( $path /usr/global/lam-6.3-b2/bin )
#setenv LM_LICENSE_FILE $PGI/license.dat
#setenv LAMHOME /usr/global/lam-6.3-b2
#setenv LAMHF77 $PGI/linux86/bin/pgf90
#setenv MANPATH $PGI/man:/usr/local/man:/usr/bin/man:/usr/man
#setenv MANPATH
${MANPATH}:/usr/X11R6/man:/usr/lib/perl5/man:/usr/global/man
#setenv MANPATH ${MANPATH}:/usr/global/lam-6.3-b2/man

#After generating the code via : 

#make abinip 

#the program was made running by editing a cluster file, 
#booting the cluster ( lamboot cluster_file ), and issuing : 

#mpirun -w -c #number_of_processors (../)abinip < X.files (>& log &) 

#(and of course wiping the cluster lateron ...).

But when I ran the command "make abinip", after some compile
information, the program received no response for so long time (many
hours)! Then nothing happened. (below are the last lines)

...... 
         if [ ! -f makefile_macros ] ; then \
          cp /home/shasta/cli/Documents/test433/makefile_macros . ; \
         fi
cd ./Lib_blas ; \
         if [ ! -f object_list ] ; then \
          cp /home/shasta/cli/Documents/test433/Lib_blas/object_list . ;
\
         fi ; \
         /home/shasta/cli/Documents/test433/makearch Lib_blas blas
no_cpp

 makearch : included makefile_macros in Lib_blas/tmp_makefile  makearch
: reached end, the Lib_blas/tmp_makefile file has been generated

cd ./Lib_blas; \
         make -f tmp_makefile blas.a
VPATH=/home/shasta/cli/Documents/test433/Lib_blas
make[2]: Entering directory
`/home/shasta/cli/Documents/test433/Lib_blas'
make[2]: `blas.a' is up to date.
make[2]: Leaving directory `/home/shasta/cli/Documents/test433/Lib_blas'
cd . ; \
         /usr/local/software/lam/bin/hf77  -o abinip Src_main/abinip.o
Src_9drive/ar_9drive.a Src_8seqpar/ar_8abinip.a Src_7suscep/ar_7suscep.a
Src_6response/ar_6response.a Src_5common/ar_5common.a
Src_4iowfdenpot/ar_4iowfdenpot.a Src_4wfs/ar_4wfs.a Src_3gw/ar_3gw.a
Src_3iovars/ar_3iovars.a Src_3paw/ar_3paw.a
Src_3recipspace/ar_3recipspace.a Src_3xc/ar_3xc.a Src_3xml/ar_3xml.a
Src_2nonlocal/ar_2nonlocal.a Src_2ffts/ar_2ffts.a Src_2psp/ar_2psp.a
Src_2geometry/ar_2geometry.a Src_2parser/ar_2parser.a
Src_2spacepar/ar_2spaceparp.a Src_1util/ar_1util.a
Src_1contract/ar_1contract.a Src_1managempi/ar_1managempip.a
Src_0basis/ar_0basis.a Src_defs/ar_defs.a Lib_fftnew/fftnewp.a
Lib_numrecip/numrecip.a Lib_lapack/lapack.a Lib_blas/blas.a
/usr/local/software/lam/lib/libmpi.a ; \
         chmod ugo+rx abinip
(The procedure idled here!!!)

I really don't know what the reason is? Could you help me, or just give
me suggestions? By the way, I found no abinip file genarated, while the
hf77 is still running.

Best regards.

Chun Li