Linux mol biol/biochem/etc software - compiling fortran apps

Martin Kroeker martin at tutor.oc.chemie.th-darmstadt.de
Tue Oct 22 05:23:56 EST 1996


Tim Cutts (tjrc1 at mole.bio.cam.ac.uk) wrote:
: Getting back to the subject at hand, I haven't tried the g77 package
: yet, but I probably wouldn't recommend it for serious work anyway; GNU
: Fortran is not supposed to be very good at all yet, so if you want
: reliable numbers at the end, don't risk it.
There are binary distributions of g77 available for Linux that should
work with any distribution. And yes, g77 appears to be stable and it is
usually faster than f2c/gcc. Where did you get the "not supposed to be very
good at all yet... don't risk it" stuff from ?
If you want reliable numbers in the end, check _any compiler_ , whether
free or commercial, with data where you know the exact results, or at 
least compare results from different compilers.    

Martin
PS: to get back to the original topic, g77 does not yet handle the 'Q'
format descriptor that appears in two statements in raster3d - but i do
not see why this non-standard format is used in this code anyway.
--
Dr.-Ing. Martin Kroeker                 
Inst. f. Organ. Chemie                   martin at oc2.oc.chemie.th-darmstadt.de
Univ. (TH) Darmstadt                           db7p at hrzpub.th-darmstadt.de
Germany                                     




More information about the Bio-soft mailing list