email@example.com (silvestri) writes:
> As a C/C++ programmer (Fortran newbie) I need to port a Fortran-77
> (with extensions for namelist/IO, long variable names and include
> files) source to MS Windows.
All of those extensions are standardized in f90 and are also
very common in f77 compilers. They shouldn't cause much problems.
The original source runs under VMS, with
> differential compile support (cpp) for a number of Unix systems
> (Solaris, Digital Unix, IRIX, UNICOS). Which is the fastest way Compaq
> Visual Fortran or Intel Fortran Compiler for Windows or ...?
I would guess that if someone has gone to the trouble of documenting
that the code uses those extensions, you are probably in pretty good
shape. My general observation is that where the author just
says "well of course this is standard Fortran", you are most likely
to find that the author actually has no idea what standard Fortran
actually is and that the code is full of nonstandard junk. When
the author lists the ways in which the code is nonstandard, then
odds are at least decent that those are the only ways in which it is
From your description, including the list of systems that the code
has run on before, I'd expect very little difficulty from any of
the current compiler offerings. I hesitate to recommend which
might be the fastest one. Since you mention VMS, I might note
the historical comonality between the VMS compiler and CVF, thus
possibly suggesting it. But it isn't a very strong suggestion;
I doubt you will go majorly wrong with any current offering.
(Just don't try to scrounge up something like an old, unsupported
copy of Microsoft Powerstation or some such - that doesn't count
as a current offering).
Richard Maine | Good judgment comes from experience;
email: my first.last at org.domain | experience comes from bad judgment.
org: nasa, domain: gov | -- Mark Twain