[comp.lang.fortran] Porting from Snun SPARC to IBM RS6000/520 & HP9000/720

playman@cbnewsm.att.com (paul.arthur.layman) (05/24/91)

Over the next couple of weeks I hope to bench mark a couple of rather
large, floating point intensive, fortran codes (some small bits of c)
developed on a SUN SPARC on IBM RS6000/520 and HP9000/720 machines.
Benchmarking will require porting to these platforms.  I would
appretiate any pointers/horror stories you may have on performing ports
to these platforms.

Thanks in advance,
Paul Layman (p.a.layman@att.com)

system@alchemy.chem.utoronto.ca (System Admin (Mike Peterson)) (05/25/91)

In article <1991May23.182102.6851@cbnewsm.att.com> playman@cbnewsm.att.com (paul.arthur.layman) writes:
>
>Over the next couple of weeks I hope to bench mark a couple of rather
>large, floating point intensive, fortran codes (some small bits of c)
>developed on a SUN SPARC on IBM RS6000/520 and HP9000/720 machines.
>Benchmarking will require porting to these platforms.  I would
>appretiate any pointers/horror stories you may have on performing ports
>to these platforms.

I have had to do this too, and if your program use any of the BSD f77
library routines, you are going to have fun - the IBM has a few of them,
(e.g. has 'system', 'getenv', but not 'dtime'/'etime' (use 'mclock'
instead)), and the HP has a grand total of 0 of them (though it will
quietly link in the C routines with the same names, which will core dump
your program). Also, the HP will intermix unit 6 and 7 output (stdout
and stderr). A friend of mine couldn't generate a private library for
the HP 'ld', but didn't have enough time to figure out what was wrong.
The IBM doesn't call the FORTRAN compiler 'f77', which will probably
break your Makefiles (we created a macro for the compiler name in the
Makefile, and also created a 'f77' command on the IBM, but you also have
to modify some file in /etc that describes the default options for
the 'xlf' compiler - copy the 'xlf' options to a new block called 'f77').
-- 
Mike Peterson, System Administrator, U/Toronto Department of Chemistry
E-mail: system@alchemy.chem.utoronto.ca
Tel: (416) 978-7094                  Fax: (416) 978-8775