From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-0.9 required=5.0 tests=BAYES_00,FORGED_GMAIL_RCVD,
	FREEMAIL_FROM autolearn=no autolearn_force=no version=3.4.4
X-Received: by 2002:ac8:4253:: with SMTP id r19mr284807qtm.116.1588194486462;
        Wed, 29 Apr 2020 14:08:06 -0700 (PDT)
X-Received: by 2002:a9d:6315:: with SMTP id q21mr28752715otk.16.1588194486135;
 Wed, 29 Apr 2020 14:08:06 -0700 (PDT)
Path: 
 eternal-september.org!reader01.eternal-september.org!feeder.eternal-september.org!news.gegeweb.eu!gegeweb.org!fdn.fr!proxad.net!feeder1-2.proxad.net!209.85.160.216.MISMATCH!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.ada
Date: Wed, 29 Apr 2020 14:08:05 -0700 (PDT)
In-Reply-To: <ly8sjg49h0.fsf@pushface.org>
Complaints-To: groups-abuse@google.com
Injection-Info: google-groups.googlegroups.com; posting-host=176.151.7.118;
 posting-account=hya6vwoAAADTA0O27Aq3u6Su3lQKpSMz
NNTP-Posting-Host: 176.151.7.118
References: <6c2c0e35-af07-4169-8be5-464ec7fd0fd5@googlegroups.com>
 <cc95457e-b41c-422f-8972-a108af2204fa@googlegroups.com>
 <ly8sjg49h0.fsf@pushface.org>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <5abd8196-b4cf-4394-9e5a-af034797f785@googlegroups.com>
Subject: Re: GNAT vs Matlab - operation on   multidimensional complex matrices
From: vincent.diemunsch@gmail.com
Injection-Date: Wed, 29 Apr 2020 21:08:06 +0000
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Xref: reader01.eternal-september.org comp.lang.ada:58528
Date: 2020-04-29T14:08:05-07:00
List-Id: <comp.lang.ada>

Le mardi 31 mars 2020 21:20:45 UTC+2, Simon Wright a =C3=A9crit=C2=A0:
> Shark8 <onewingedshark@gmail.com> writes:
>=20
> > Ok then; the following should have convention Fortran IIRC:
> >
> > type tCubeReal is array (1..NumChannels, 1..NumAngles, 1..NumRanges) of=
 mReal
> >   with Convention =3D> Fortran;
>=20
> I wondered about this; the code ran considerably slower with Convention
> =3D> Fortran.

In Row Major (Convention Fortran) one needs to iterate on the first dimensi=
on, then the second and finally the third, to stay on contiguous regions in=
 memory (and hence in the cache). So the loop must put the first dimension =
in the center :

for p in M=E2=80=99Range(3) loop
  for n in M=E2=80=99Range (2) loop
     for m in M=E2=80=99Range (1) loop
         ... M(m, n, p)

This is somehow counter-intuitive. But Row Major (like Little-endian with w=
hich it is closely related) has more properties (like extracting a submatri=
x or conforming to classical linear algebra) and is therefore preferred by =
scientists.

Regards,

Vincent