This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: Limiting compile time memory usage?
- From: Harald Anlauf <anlauf at gmx dot de>
- To: Bud Davis <bdavis9659 at sbcglobal dot net>
- Cc: fortran at gcc dot gnu dot org
- Date: Sat, 23 Dec 2006 22:48:13 +0100 (MET)
- Subject: Re: Limiting compile time memory usage?
- References: <489342.60233.qm@web81206.mail.mud.yahoo.com>
Bud Davis wrote:
i am curious. are you encountering problems compiling
code on a machine with limited memory ?
Yes. My "biggest" machine at home has only 256MB.
When gfortran (Linux/x86) reaches 500 MB virtual memory,
it brings the machine almost down to its knees.
I sometimes have to set "ulimit -v 450000" to prevent this.
or does gfortran use excessive amounts of memory when
compiling specific routines ?
Also yes. It is particularly bad when compiling even
a rather small ("top-level") file of a larger but strongly
modularized project that heavily uses OO techniques in
Fortran 90. Compared to the compilers I use regularly
(xlf, Intel ifort 8 and 9, g95) it appears to require the most
virtual memory even at -O0, no debugging, no bounds-checking.
(One program of the project that I am engaged with has
less than 100 lines, including ~20 "use, only" statements
(single!). I guess that it is the heavy use of overloading
that makes memory usage almost explode.
in my experience, i have never seen a file of fortran
source that causes compile problems (due to memory
usage) in the last say 5 - 10 years.
Are you sure we're talking about the same compiler?
gfortran is younger than that according to the wiki.
Cheers,
-ha