English version
Accueil     À propos     Téléchargement     Ressources     Contactez-nous    

Ce site est rarement mis à jour. Pour les informations les plus récentes, rendez-vous sur le nouveau site OCaml à l'adresse ocaml.org.

Browse thread
[Caml-list] interesting array efficiency observations
[ Home ] [ Index: by date | by threads ]
[ Search: ]

[ Message by date: previous | next ] [ Message in thread: previous | next ] [ Thread: previous | next ]
Date: 2004-05-09 (00:48)
From: Yamagata Yoriyuki <yoriyuki@m...>
Subject: Re: [Caml-list] interesting array efficiency observations
From: Christophe TROESTLER <Christophe.Troestler@umh.ac.be>
Subject: Re: [Caml-list] interesting array efficiency observations
Date: Sun, 09 May 2004 02:25:35 +0200 (CEST)

> On Sun, 09 May 2004, Yamagata Yoriyuki <yoriyuki@mbg.ocn.ne.jp> wrote:
> > 
> > From: briand@aracnet.com
> >
> > > I was investigating the use of 1-D bigarray's vs Array and noticed
> > >  that the 1-D bigarray using c_layout seems to be slower than Array.
> > 
> > Yes, it is expected, since a.{j} expands a call of a C function, while
> > a.(j) expands inlined assembly codes.
> Is it?  Ocamlopt does inline the call when the bigarray type is fully
> known, doesn't it.

Ah...  You are right.  Thanks.

$ cat test.ml
open Bigarray
type t = (int32, int32_elt, c_layout) Array1.t
let get (a:t) i = a.{i}

$ cat test.s
(... snip ...)
        movl    %eax, %edx
.L101:  movl    young_ptr, %eax
        subl    $12, %eax
        movl    %eax, young_ptr
        cmpl    young_limit, %eax
        jb      .L102
        leal    4(%eax), %eax
        movl    $2303, -4(%eax)
        movl    $int32_ops, (%eax)
        sarl    $1, %ebx
        movl    20(%edx), %ecx
        cmpl    %ebx, %ecx
        jbe     .L104
        movl    4(%edx), %ecx
        movl    (%ecx, %ebx, 4), %ebx
        movl    %ebx, 4(%eax)
.L102:  call    caml_call_gc
.L103:  jmp     .L101
.L104:  call    caml_array_bound_error
        .align  16
        .globl  Test__entry
        .type   Test__entry,@function

To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners