From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <nicolas.pouillard@gmail.com>
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,SPF_NEUTRAL 
	autolearn=disabled version=3.1.3
Received: from mail2-relais-roc.national.inria.fr (mail2-relais-roc.national.inria.fr [192.134.164.83])
	by yquem.inria.fr (Postfix) with ESMTP id 9A3E8BC6B
	for <caml-list@yquem.inria.fr>; Mon,  5 Nov 2007 11:50:05 +0100 (CET)
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AgAAAHODLkfAXQInh2dsb2JhbACOcAIBCAopgQ8
X-IronPort-AV: E=Sophos;i="4.21,371,1188770400"; 
   d="scan'208";a="3906752"
Received: from concorde.inria.fr ([192.93.2.39])
  by mail2-smtp-roc.national.inria.fr with ESMTP; 05 Nov 2007 11:50:05 +0100
Received: from mail2-relais-roc.national.inria.fr (mail2-relais-roc.national.inria.fr [192.134.164.83])
	by concorde.inria.fr (8.13.6/8.13.6) with ESMTP id lA5Ao5mK011132
	(version=TLSv1/SSLv3 cipher=RC4-SHA bits=128 verify=OK)
	for <caml-list@inria.fr>; Mon, 5 Nov 2007 11:50:05 +0100
X-IronPort-AV: E=Sophos;i="4.21,371,1188770400"; 
   d="scan'208";a="3906751"
Received: from peray.inria.fr (HELO ausone.inria.fr) ([128.93.8.98])
  by mail2-relais-roc.national.inria.fr with SMTP; 05 Nov 2007 11:50:04 +0100
Received: by ausone.inria.fr (sSMTP sendmail emulation); Mon, _d Nov 2007 11:49:26 +0100
From: "Nicolas Pouillard" <nicolas.pouillard@gmail.com>
Cc: caml-list <caml-list@inria.fr>
Subject: Re: [Caml-list] Concurrent ocamlbuild
To: jon <jon@ffconsultancy.com>
References: <200711041309.44590.jon@ffconsultancy.com> <472E0A0D.3010106@dslextreme.com> <200711051011.43263.jon@ffconsultancy.com>
In-Reply-To: <200711051011.43263.jon@ffconsultancy.com>
Date: Mon, 05 Nov 2007 11:49:26 +0100
Message-Id: <1194259044-sup-1150@ausone.inria.fr>
User-Agent: Sup/0.2
Content-Type: text/plain; charset=UTF-8
X-Miltered: at concorde with ID 472EF55D.000 by Joe's j-chkmail (http://j-chkmail . ensmp . fr)!
X-Spam: no; 0.00; makefile:01 makefile:01 compilation:01 byte-code:01 compilation:01 byte-code:01 parallelism:01 dependencies:01 inlining:01 compiler:01 cmx:01 parallelism:01 inlining:01 mylib:01 ocaml:01 

Excerpts from jon's message of Mon Nov 05 11:11:42 +0100 2007:
> On Sunday 04 November 2007 18:06, Erick wrote:
> > Did you see the -j N option?
> >
> >  > ocamlbuild -h
> >
> > ...
> >   -j <N>                      Allow N jobs at once (0 for unlimited)
> > ...
> 
> I hadn't noticed that, thanks. However, it doesn't seem to work (at least for 
> this project). My old Makefile takes 21s and ocamlbuild with -j anything 
> still takes at least 45s. Moreover, the Makefile isn't as parallel as it 
> could be...

Is it native compilation, byte-code compilation?

The  byte-code  one  offer  more opportunities for parallelism than the native
one  if  we  respect  dependencies  to  allow cross module inlining. Since the
compiler  does not complain about not found ".cmx" you can do more parallelism
in  your  Makefile.  However  that's risky and can leads to worst performances
(less inlining).

Sometimes  using  libraries  (.mllib)  can  give  more  parallelism  than  the
traditional ocamlbuild discovery of sources.

So you can try...

$ cat mylib.mllib
# Put here all modules needed by your main program
# You can gather them by looking at the link command produced by ocamlbuild
A B C D E

$ cat myocamlbuild.ml
open Ocamlbuild_plugin;;
open Ocamlbuild_pack;;
dispatch begin function
| After_rules ->
    ocaml_lib "mylib"
| _ -> ()
end;;

$ cat _tags
<myprog.{byte,native}>: use_mylib

$ ocamlbuild mylib.cma mylib.cmxa myprog.byte myprog.native

HTH,
-- 
Nicolas Pouillard aka Ertai