This is the mail archive of the
mailing list for the GCC project.
Re: 64 bit fortran unformatted on Opteron -- incompatible with 32unformatted
- From: Kyon <kyon at verizon dot net>
- To: gcc-help at gcc dot gnu dot org
- Date: Tue, 15 Jun 2004 23:15:05 -0700
- Subject: Re: 64 bit fortran unformatted on Opteron -- incompatible with 32unformatted
- Organization: K 9 Bone Admiration Society
- Reply-to: tunaboy2 at verizon dot net
You could stick with the coding which people could read on 32 bit
systems( which is what Portland group has done, and what is done on the SGI
mips machines). This is to use a 4 byte integer for the record length, which
is a mighty large single record. This would only inconvenience the pioneers,
who expect to have arrows in their back anyway. It would also allow all the
existing g77, etc. generated unformatted files to be used as is.
Having arrived on the frontier at this point, obviously I would be making
this arguement, but as part of the appeal of the opterons is that you can go
gently into the 64bit computing world, moving toward a consistent format
with the existing standard makes sense to me (especially since I can't
imagine having a single record of such a large size).
Is this code I could possibly get in and hack myself? I suspect not, since c
is my second language, but if it would be something I could get involved in
and make a positive contribution towards, I would attempt it. ( I might
also see if I could implement a byteswapping option). However, having said
that, and looking realisticly at my aptitude and my amount of free time
available, it's kind of like my 8 year old offering to drive me somewhere.
At any rate, I appreciate your response, and all the work that has gone into
those compilers. My contributions have been limited to purchases of books
from FSF, and I do appreciate them.