This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: Fwd: gfortran - pointers to components of derived type arrays
- From: "Paul Richard Thomas" <paul dot richard dot thomas at gmail dot com>
- To: "Tobias Schlüter" <tobias dot schlueter at physik dot uni-muenchen dot de>
- Cc: "fortran at gcc dot gnu dot org List" <fortran at gcc dot gnu dot org>, paul at codesourcery dot com
- Date: Thu, 12 Apr 2007 20:32:35 +0200
- Subject: Re: Fwd: gfortran - pointers to components of derived type arrays
- Dkim-signature: a=rsa-sha1; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=WwrPp6WtBUCxd/5Gp1t2JxBBnQGIyETtt61ayyp3Lspd2OArvruFm+R9QVzSTrQ8mub/EHw4zDag1VhYp7ffmV/HkKanTDswpv+agDHV1BnWeFqNM2QIFkYGe9yp/D1YKZ8qZzmwU+7MUImCvMnuXP4Cpz7deS6Gy8AGC/X0NRM=
- Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=Hw4MECT69+jXqzdKqn9J6SfH0Md07yk1zaviNk8xyt924VpmRzDet9Ks8BIGzAPSmOw+lDV8vYzzzIfO/rkcEyjeWvDL//tRHNEqL6e5yWlRAYVDXzCP5J20uvOisPVm9yJ/ta493l5gqJSW41zQMEXI5q00RfSYfHo3ReBZ59k=
- References: <339c37f20704100233h3f3f32cfh8101cd00d887c6d0@mail.gmail.com> <200704102145.15107.paul@codesourcery.com> <339c37f20704110815gb29bad9r909f22262a34c1ce@mail.gmail.com> <461E638D.20809@physik.uni-muenchen.de>
Tobi,
A minor remark: it should be possible to keep the old
gfc_conv_descriptor_stride interface alive during the development (just
have it include the necessary multiplication. If that allows you to
avoid a wholesale conversion of the code during development, it should
make the whole thing much easier.
If I have understood what you are suggesting, that would introduce an
extra multiplication for every dimension*array_ref. I am not sure
that I see how that would help. In fact, I am worrying that this is a
general problem with scheme (i); expecially for vectors, where stride
= 1 is ignored. In anycase, were stride in units of bytes, it would
be extracted by the gfc_conv_descriptor_stride interface completely
unmodified.
Anyway, what little time I have had in the last 24 hours has been
devoted to assessing the scale of the job and the various wrinkles
involved. I hope somebody actually uses such pointer/target
combinations!
Thanks
Paul
--
Saint Augustine - "O Lord, help me to be pure, but not yet"