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 RAA19064; Tue, 1 Jul 2003 17:12:38 +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 RAA18906 for ; Tue, 1 Jul 2003 17:12:36 +0200 (MET DST) Received: from gnome03.net.rol.ru (smtp.rol.ru [194.67.21.9]) by concorde.inria.fr (8.11.1/8.11.1) with ESMTP id h61FCYj04604; Tue, 1 Jul 2003 17:12:34 +0200 (MET DST) Received: from 194.186.30.209.adsl-spb.net.rol.ru ([194.186.30.209]:33463 "EHLO snob.metalinkltd.com" whoson: "madlazy@rol.ru" smtp-auth: TLS-CIPHER: TLS-PEER-CN1: ) by gnome03.net.rol.ru with ESMTP id S12590852AbTGAPMd (ORCPT + 1 other); Tue, 1 Jul 2003 19:12:33 +0400 From: Nickolay Kolchin-Semyonov Organization: Metalink LTD To: Xavier Leroy , Christophe Michel Subject: Re: [Caml-list] architecture optimisation at compile time for ocaml 3.06 Date: Tue, 1 Jul 2003 19:08:02 +0400 User-Agent: KMail/1.5.2 Cc: caml-list@inria.fr References: <20030701090701.92742.qmail@web41506.mail.yahoo.com> <20030701152320.B14262@pauillac.inria.fr> In-Reply-To: <20030701152320.B14262@pauillac.inria.fr> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200307011908.02487.snob@metalinkltd.com> X-Spam: no; 0.00; nickolay:01 snob:01 caml-list:01 3.06:01 -specific:01 compiler:01 ocaml:01 command:98 native-code:02 compile:02 optimized:02 wrote:03 optimizing:03 subset:03 efficient:05 Sender: owner-caml-list@pauillac.inria.fr Precedence: bulk On Tuesday 01 July 2003 17:23, Xavier Leroy wrote: > > my system relies on an athlon XP2000+ and I would like to take > > The native-code OCaml compiler has only one code generator for the > whole IA32 family. This code generator is called "i386" for > historical reasons, but it is not particulary optimized for the Intel > 386 processor. Actually, the generated code is targeted for efficient > execution on the PPro-PII-PIII family, and appears to fit the Athlon > quite well too. The next release will include a few P4-specific > tricks that do not decrease performance on other processors. What about "cmov" support? This is not P4 specific, this command present for extremely long time. > > > can anyone help me in optimizing source for athlon ?? > > No :-) Quite frankly, given the subset of the IA32 instruction set > that OCaml generates, I can't see any Athlon-specific optimizations > that would be worth doing. > 3DNow, cache control. This are just first thoughts. Nickolay ------------------- 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