[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: gfortran and a compilation

Francesco Pietra wrote:
I was just trying to compile dalton-2.0, not gamess, and I absolutely need parallelization for dft calculations. Serial installation would not allow me to deal with the large molecules of my interest.

On Tuesday 27 June 2006 10:18, Stefano Simonucci wrote:
On Tue, 2006-06-27 at 07:16 +0200, Francesco Pietra wrote:
Stefano, accepting your experience I started preparing Makefile.config
indicaring g77 for Fortran and gcc for C (I am at amd64 debian etch,
parallel machine dual opteron) yesterday I went on obtaining successfully
makefile.config and an invite to run make, although none of the mathem
libraries had been found. The program specifically said "should I proceed
without"? I said yes, but I did not run make.
I don't have experience of GAMESS's compilation. I only have experience
of Dalton and, obviously, of my programs. I have etch AMD64 and I know
that g77 and gfortran are two distinct compilers. The installation of
gfortran or g77 is not automatic with the installation of gcc.

I have installed g77 separately, after that gcc was already installed and dalton ./configure was apparently pleased with that.

Moreover maybe you need other mathematical libraries ....

Yes, I have been informed that the highly specialized libraries libatlas have to found (where?)

atlas3-headers - Automatically Tuned Linear Algebra Software,C header files

I have often added (manually) symbolic link to satisfy the dependencies
of configure (for example
"ln -s /usr/lib/libblas.so.3 /usr/lib/libblas.so").
I don't know the reason for which the symbolic links as libblas.so ->
libblas.so.3 are not always created. In these cases, nevertheless, the
configure is not able to find the libraries.

I have not examined carefully these suggestions from you because I am pressed at the moment and _- most importantly- because of my doubts below wheteher I can go on anyway.

What I also understand is that probably my system lacks the parallelization software mpich (that is probably available on internet: for debian?)

mpich-bin - MPI parallel computing system implementation
mpich-mpd-bin - MPI parallel computing system implementation, MPD version
mpich-shmem-bin - MPI parallel computing system implementation, SHMEM version
libmpich-mpd1.0-dev - mpich static libraries and development files
libmpich-mpd1.0c2 - mpich-mpd runtime shared library
libmpich-shmem1.0-dev - mpich static libraries and development files
libmpich-shmem1.0c2 - mpich-shmem runtime shared library
libmpich1.0-dev - mpich static libraries and development files
libmpich1.0c2 - mpich runtime shared library

However, I do not understand whether dft, like nevpt is only distributed in cvs version. In that case I can do nothing.

All these efforts because mpqc - which is parallelized - hangs with my molecules (in spite of running correctly with simple molecules).


Reply to: