On 16 September 2010 20:09, Barry Smith <bsmith at mcs.anl.gov> wrote: >> >> How could I know? If any of my codes works in Apple, I would say it is >> just for accident :-). Anyway, I expect to workaround any issue. > > ?How shall I test it? Remember I need line by line instructions :-) >
Give me some time to push my new setup.py and manage the merge in -dev, and I'll let you know how to test 3.1. For petsc-dev, you can try this: cd $PETSC_DIR python setup.py sdist pip-python install --install-option=--home=$HOME dist/petsc-3.2.dev0.tar.gz Next, go $HOME/lib/python/petsc and inspect the stuff in include and lib. Perhaps you could setenv PETSC_DIR=$HOME/lib/python/petsc and unset PETSC_ARCH, and try to run 'make test' in the full source tree. And remember that Fortran stuff is disabled. PS: mpi4py now contains info about mpicxx and mpif90. Perhaps I should I use them, in the near future we could have f2py wrappers to Fortran routines involving PETSc objects... Python is so slow for number crunching... -- Lisandro Dalcin --------------- CIMEC (INTEC/CONICET-UNL) Predio CONICET-Santa Fe Colectora RN 168 Km 472, Paraje El Pozo Tel: +54-342-4511594 (ext 1011) Tel/Fax: +54-342-4511169
