From mboxrd@z Thu Jan  1 00:00:00 1970
Received: (from majordomo@localhost) by pauillac.inria.fr (8.7.6/8.7.3) id XAA17592; Thu, 29 Mar 2001 23:34:41 +0200 (MET DST)
X-Authentication-Warning: pauillac.inria.fr: majordomo set sender to owner-caml-list@pauillac.inria.fr using -f
Received: from concorde.inria.fr (concorde.inria.fr [192.93.2.39]) by pauillac.inria.fr (8.7.6/8.7.3) with ESMTP id XAA17551 for <caml-list@pauillac.inria.fr>; Thu, 29 Mar 2001 23:34:40 +0200 (MET DST)
Received: from barry.mail.mindspring.net (barry.mail.mindspring.net [207.69.200.25])
	by concorde.inria.fr (8.11.1/8.10.0) with ESMTP id f2TLYc906786
	for <caml-list@inria.fr>; Thu, 29 Mar 2001 23:34:38 +0200 (MET DST)
Received: from dylan (1Cust94.tnt3.tucson.az.da.uu.net [63.11.144.94])
	by barry.mail.mindspring.net (8.9.3/8.8.5) with SMTP id QAA00998
	for <caml-list@inria.fr>; Thu, 29 Mar 2001 16:34:37 -0500 (EST)
Message-ID: <001301c0b898$1d482120$210148bf@dylan>
From: "David McClain" <dmcclain1@mindspring.com>
To: <caml-list@inria.fr>
Subject: Fw: [Caml-list] Complex numbers in OCaml
Date: Thu, 29 Mar 2001 14:35:01 -0700
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 5.00.2919.6600
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6600
Sender: owner-caml-list@pauillac.inria.fr
Precedence: bulk


----- Original Message -----
From: "David McClain" <dmcclain1@mindspring.com>
To: "Xavier Leroy" <Xavier.Leroy@inria.fr>
Sent: Thursday, March 29, 2001 2:34 PM
Subject: Re: [Caml-list] Complex numbers in OCaml


> > As Markus mentioned, I'm considering extending the Bigarray module to
> > handle complex numbers as well.  This would help interfacing with
> > Fortran, and provide essentially the same data representation than in
> > C and Fortran.
>
> If you do this you need to be aware that:
>
> 1. Most C representations use arrays in which real and imaginary parts are
> in adjacent memory addresses, interleaved throughout the entire array, and
>
> 2. Most Fortran representations use separate arrays for the real parts and
> the imaginary parts.
>
> I say "most" based purely on my own personal experiences with all the glue
> code I have had to write to interface to the IMSL, NAG, MatLab, BLAS, and
> MKL code.
>
> My experiments here have shown me that there is a real speed penalty on
> Pentium II and III architectures for the separate array versions of code,
> contrasted with the interleaved structure versions. I can only assume that
> this has to do with data locality and cache performance.
>
> I ultimately gave up trying to come up with a "standard" array
> representation, except within my own little universe. I was quite
surprised
> to find that a data structure as fundamental as an array of numbers can
have
> so many different implementations.
>
> - DM
>

-------------------
To unsubscribe, mail caml-list-request@inria.fr.  Archives: http://caml.inria.fr