This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: New loop optimizations in gfortran 4.4
- From: "Richard Guenther" <richard dot guenther at gmail dot com>
- To: "Paul Richard Thomas" <paul dot richard dot thomas at gmail dot com>
- Cc: "Harald Anlauf" <anlauf at gmx dot de>, fortran at gcc dot gnu dot org
- Date: Tue, 23 Sep 2008 08:55:44 +0200
- Subject: Re: New loop optimizations in gfortran 4.4
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; 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; bh=gIfu9AdgiUKGtTNih1XQcnaAj5L8c8OB6xMtoPuZoj0=; b=XzPoGN7tuTHpCFvp3PqH6T/fpaFJxBjy1zbZQvnQH1vxq8hIjo824R3lfeu+6mc+O+ Rnve3OlF3xzo9VtO6sXgIJ2HLZo2fAWjNGX/I4q2v4cUttKoS1VrTm93jr2sC+1zp16Y scUFsvXyvKm5G3/czInhZC/prrmKBaVr26cBQ=
- Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:date:from:to:subject:cc:in-reply-to:mime-version :content-type:content-transfer-encoding:content-disposition :references; b=Ufo44uf8KG+6y+PNuO3/siDRVVSBOIrlcdcM3rklLTL8RKBMqk94rHuD4ZhQsXP0/U 5acJ6csm4mPTw3wBOQ55tVd9ko9/f5befsssNLlzKblRBuELkN78J8i4/TItZMLEmhT9 AniE2RMOA9RbeH4zH9FUfRDPaWAp/nz/sA9BE=
- References: <48D8152C.6090208@gmx.de> <339c37f20809222206y118205d6v757531dc0e44344d@mail.gmail.com>
On Tue, Sep 23, 2008 at 7:06 AM, Paul Richard Thomas
<paul.richard.thomas@gmail.com> wrote:
> Dear Harald,
>
>> Assuming that the Graphite framework supports such
>> constraints (in principle), are there any (long term)
>> plans to implement similar compiler directives in the
>> gcc collection and in particular in gfortran?
>
> We discussed it very briefly on the list a few months ago but came to
> no conclusion as to who, when and how.
>
> Perhaps, Richard, you can give us a timescale on which it would be
> sensible to start gearing up for Graphite? I have to confess to
> thinking about reworking trans-array.c with a bit of a shudder:-)
I guess you refer to the middle-end array expression work, which is as far
as I understand something different from what Harald wants. Harald
seems to want to annotate loops with instructions for the high level
loop optimizers much like OpenMP annotations. As for the middle-end
array expression stuff I hope to move the existing code to tuples soon ...
Richard.