Re: Zero-size arrays
- From: Tobias Burnus <burnus@xxxxxxxx>
- Date: Fri, 4 Apr 2008 15:27:54 -0700 (PDT)
On 4 Apr., 19:21, Gordon Sande <g.sa...@xxxxxxxxxxxxxxxx> wrote:
On 2008-04-04 12:26:22 -0300, Tobias Burnus <bur...@xxxxxxxx> said:
Can one ask, and hope for a straight answer, whether these are commercial
Fortran compilers or not?
The latest ICE was in sunf95, but ICE, rejects valid etc. problems
were also found in ifort, gfortran, pathscale and some more which I do
not remember.
I have to admit that I do not see much difference between opensource
and commercial compilers in this regard; it depends more on the
individual vendor and the version (e.g. ifort 8 was quite instable and
also gfortran 4.1 had several minor problems while ifort 9.1/10/10.1
or gfortran 4.3 work quite well).
The default compilers are ifort (for serial and MPI builds), gfortran,
sunf95, openf95 and g95; after every check in, the program is
automatically build with those compilers. I think the developer
themselves mostly use gfortran and ifort. The code is essentially
Fortran 90/95, but with configure check also things like Fortran
2003's command-line arguments or type(c_prt) are used. The idea of
using allocatable components was dropped due to compatibility issues.
The observation is that GCC is probably quite heavily used but that Gfortran
and G95 and much less well exercised. There is considerable less development
effort spent on the Fortran components of GCC than on the C related components.
If I understand it many of the GGC C developers are in fact full time
professional compiler writers paid by their organizations to contribute to
GCC while the Fortran components are being done by volunteers (as a hobby
like the signature below).
While this is true, many also work only on the middle end
(optimization etc.) and target specific code, which is used by for
Fortran and C alike. Additionally, there are occasionally patches by
middle-end developers. OpenMP (2.5 and the being-developed 3.0
support) is also from a middle-end person. The cray pointers and
Bind(C) development was paid and gfortran had and presumably will have
a Google Summer of Code student. But still, most of the development
happens in the spare time.
With regards to the usage: I think gfortran and g95 have the advantage
that they are free; that means that they are much more tested than
many commercial compilers (though they have a head start by being
longer on the market. And that the gfortran testsuite is run on many
systems from embedded to big Unix machines also helped/helps to find
bugs.
James Van Buskirk wrote:
I worry that Fortran is taking a road towards being a
language that's expensive to write a compiler for and requires
expensive training to be able to really use its features and yet is
still denigrated by CS types (CS majors really are taught that
Fortran is FORTRAN 77 and to scoff at the mere mention of the
language.)
I have the feeling that for computer science Fortran plays essentially
no role. It is much more used in Physics, Chemistry, Geology,
Meteorology etc. than in CS. In my CS course we only used Java and
Haskell, while in my (obligatory) computational physics course Fortran
or C were used.
Tobias
.
- Follow-Ups:
- Re: Zero-size arrays
- From: Gerry Ford
- Re: Zero-size arrays
- References:
- Zero-size arrays
- From: relaxmike
- Re: Zero-size arrays
- From: Tobias Burnus
- Re: Zero-size arrays
- From: Gordon Sande
- Zero-size arrays
- Prev by Date: Re: Fortran 77 parser
- Next by Date: Re: Fortran 77 parser
- Previous by thread: Re: Zero-size arrays
- Next by thread: Re: Zero-size arrays
- Index(es):
Relevant Pages
|
|
