The ABINIT Users Mailing List ( CLOSED )

[stewart@cnf.cornell.edu]

Hi Nuno, 

  It sounds like there is a memory allocation issue somewhere.  I compiled 
abinit 4.4.4 using g95 and I didn't run into any problems with the 5 
standard tests.  Does your g95 version pass the standard tests?  I have 
included the makefile_macros that I use to compile abinit with g95 below on 
a Red Hat Linux Enterprise 3U4 Xeon node.  However, if you are having 
problems using PGI as well, it could be an issue with your input file.  
Please post this and I can try running it to see if I get the same problem.  
Also please provide the operating system, processor, and any other relevant 
information for your workstation. 

Another issue could be the experimental nature of g95.  The version 
available changes rapidly and some bug fixes could lead to new ones.  The 
version I am using is a little old but fairly stable
G95 (GCC 4.0.0 20050129 (experimental) (g95!) Feb 23 2005). 

The makefile macro entries as listed below for the serial version, abinis. 

                                                     Best regards, 

                                                     Derek 

#####################################
Derek Stewart, Ph. D.
Scientific Computation Associate
250 Duffield Hall
Cornell Nanoscale Facility (CNF)
Ithaca, NY 14853
stewart (at) cnf.cornell.edu
http://www.nnin.org/nnin_compsim.html 

Relevant Makefile macros entries 

# Machine type
MACHINE=P6
# Fortran optimized compilation
FC=/usr/local/bin/g95
FFLAGS=-O2 -ffree-form -fbounds-check
FFLAGS_Src_9drive =-O0 -ffree-form -fbounds-check
FFLAGS_LIBS= -O2 -fbounds-check 

# C preprocessor, used to preprocess the fortran source.
CPP=/usr/bin/cpp
CPP_FLAGS=-P -traditional -D__IFC 

# C optimized compilation.
CC=cc
CFLAGS=-O 

# Location of perl . Used to generate the script fldiff, in 
~ABINIT/Utilities .
PERL=/usr/bin/perl 

# List of machine-dependent routines
MACHINE_DEP_C_SEQ_SUBS_LIST=etime.o 

####################################################################
# For the parallel version : MPICH / MYRINET 

# Compiler flags and definitions
FFLAGS_PAR= $(FFLAGS) -I /usr/include 

# List of machine-dependent routines
MACHINE_DEP_C_PAR_SUBS_LIST=etime.par
# Location of the MPI library
MPI_A=/usr/lib/liblam.a /usr/lib/libmpi.a /usr/lib/liblamf77mpi.a 

LIBS=-L /usr/lib/gcc-lib/i386-redhat-linux/3.2.3/ -L 
/usr/local/lib/ATLAS_SSE2_4 -llapack_atlas_imp2 -lf77blas -lcblas -latlas 
-lg2c 

# Include blas, lapack, and any other libraries here
LIBS_PAR=$(LIBS) $(MPI_A) 

Nuno A. G. Bandeira writes: 

> Hello, Dr Stewart, 
>
> Thank you for your advice. However I'm still not getting there. 
>
> I ran one of my input files and I got this error message: 
>
>  Biggest array : f_fftgr(disk), with    122.0723 MBytes. 
>
>  memana : ERROR -
>   Test failed to allocate an array of    122.072 Mbytes
>   It is not worth to continue
>   Action : modify input variable to fit the available memory,
>   or increase limit on maximal array size. 
>
> I have 2Gb RAM on my pc so I don't know any reason why the program should 
> stop. 
>
> The version of abinit I'm using is 4.5.2 but 4.4.4 produces similar 
> results. The program simply stops because it complains of not having 
> enough memory. 
>
> PGI 6.0.2 compilates also run disastrously 
>
> --
> Nuno A. G. Bandeira, AMRSC
> Graduate researcher and molecular sculptor
> Theoretical and Inorganic Chemistry Group,
> Faculty of Science
> University of Lisbon - C8 building, Campo Grande,
> 1749-016 Lisbon,Portugal
> http://cqb.fc.ul.pt/intheochem/nuno.html
> Doctoral student @ IST,Lisbon
> --