From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <garrigue@math.nagoya-u.ac.jp>
X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on yquem.inria.fr
X-Spam-Level: *
X-Spam-Status: No, score=1.1 required=5.0 tests=AWL,DNS_FROM_RFC_ABUSE,
	SPF_SOFTFAIL autolearn=disabled version=3.1.3
Received: from mail3-relais-sop.national.inria.fr (mail3-relais-sop.national.inria.fr [192.134.164.104])
	by yquem.inria.fr (Postfix) with ESMTP id 7B2B7BC6C
	for <caml-list@yquem.inria.fr>; Fri, 18 Jan 2008 03:15:17 +0100 (CET)
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: Ao8CADeaj0eCNhAB/2dsb2JhbACtZQ
X-IronPort-AV: E=Sophos;i="4.25,213,1199660400"; 
   d="scan'208";a="8022522"
Received: from kurims.kurims.kyoto-u.ac.jp ([130.54.16.1])
  by mail3-smtp-sop.national.inria.fr with ESMTP; 18 Jan 2008 03:15:15 +0100
Received: from localhost (orion [130.54.16.5])
	by kurims.kurims.kyoto-u.ac.jp (8.13.8/8.13.8) with ESMTP id m0I2FBEE017240;
	Fri, 18 Jan 2008 11:15:11 +0900 (JST)
Date: Fri, 18 Jan 2008 11:15:03 +0900 (JST)
Message-Id: <20080118.111503.185813743.garrigue@math.nagoya-u.ac.jp>
To: benjamin.canou@gmail.com
Cc: caml-list@yquem.inria.fr
Subject: Re: [Caml-list] Strange performances
From: Jacques Garrigue <garrigue@math.nagoya-u.ac.jp>
In-Reply-To: <1200619933.6383.47.camel@benjamin-laptop>
References: <1200619933.6383.47.camel@benjamin-laptop>
X-Mailer: Mew version 4.2 on Emacs 22.1 / Mule 5.0 (SAKAKI)
Mime-Version: 1.0
Content-Type: Text/Plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Spam: no; 0.00; bytecode:01 segfaults:01 bytecode:01 bug:01 failwith:01 endline:01 pred:01 ocamlopt:01 ocamlc:01 ocaml:01 -version:01 toplevel:01 123.:98 123.:98 rec:01 

The problem seems to be in list_of_string.
If I write:

let list_of_string s =
  let rec list_of_string s i =
    if i < String.length s then s.[i] :: list_of_string s (succ i) else []
  in list_of_string s 0

I get a 10 time increase in speed for bytecode, and 5 times better for
native code, as one would expect. Out-of-bounds exceptions are
intended as fatal errors, do not try to catch them...

By the way, on my machine your version doesn't even work in native
code, I only get segfaults. This is allowed behaviour for
out-of-bounds access.

Jacques Garrigue

From: Benjamin Canou <benjamin.canou@gmail.com>
>   Hi,
> 
> The code following my message is way faster in bytecode than in native
> code. Is there a good reason for that or is it a bug ?
> Note : It is a (way too, I know) naive implementation of the well known
> string suite 1, 11, 21, 1211, 111221, ...
> 
>   Benjamin Canou.
> 
> === code ===
> 
> let list_of_string s =
>   let rec list_of_string s i =
>     try s.[i] :: list_of_string s (succ i)
>     with _ -> []
>   in list_of_string s 0
> 
> let rec trans = function
>   | '1' :: '1' :: '1' :: tl -> "31" ^ trans tl
>   | '1' :: '1' :: tl -> "21" ^ trans tl
>   | '1' :: tl -> "11" ^ trans tl
>   | '2' :: '2' :: '2' :: tl -> "32" ^ trans tl
>   | '2' :: '2' :: tl -> "22" ^ trans tl
>   | '2' :: tl -> "12" ^ trans tl
>   | '3' :: '3' :: '3' :: tl -> "33" ^ trans tl
>   | '3' :: '3' :: tl -> "23" ^ trans tl
>   | '3' :: tl -> "13" ^ trans tl
>   | [] -> ""
>   | _ -> failwith "bad input"
> 
> let rec print n s =
>   print_endline s ;
>   if n > 0 then print (pred n) (trans (list_of_string s))
>     
> let _ = print 30 "1"
> 
> === perfs ===
> 
> benjamin@benjamin-laptop:~/Work/Stuff$ ocamlopt 123.ml -o 123
> benjamin@benjamin-laptop:~/Work/Stuff$ time ./123
> [...]
> real    0m5.245s
> user    0m4.944s
> sys     0m0.016s
> benjamin@benjamin-laptop:~/Work/Stuff$ ocamlc 123.ml -o 123
> benjamin@benjamin-laptop:~/Work/Stuff$ time ./123
> [...]
> real    0m1.097s
> user    0m0.840s
> sys     0m0.008s
> benjamin@benjamin-laptop:~/Work/Stuff$ ocaml -version
> The Objective Caml toplevel, version 3.09.2