From mboxrd@z Thu Jan 1 00:00:00 1970 Received: (from weis@localhost) by pauillac.inria.fr (8.7.6/8.7.3) id JAA10333 for caml-redistribution@pauillac.inria.fr; Mon, 8 Mar 1999 09:00:36 +0100 (MET) Resent-Message-Id: <199903080800.JAA10333@pauillac.inria.fr> Received: from nez-perce.inria.fr (nez-perce.inria.fr [192.93.2.78]) by pauillac.inria.fr (8.7.6/8.7.3) with ESMTP id CAA28996 for ; Sat, 6 Mar 1999 02:17:12 +0100 (MET) Received: from pegasus.azstarnet.com (pegasus.azstarnet.com [169.197.56.194]) by nez-perce.inria.fr (8.8.7/8.8.7) with ESMTP id CAA04703 for ; Sat, 6 Mar 1999 02:17:09 +0100 (MET) Received: from dylan (dialup17ip055.tus.azstarnet.com [169.197.38.55]) by pegasus.azstarnet.com (8.9.1a/8.9.1a) with SMTP id SAA12917 for ; Fri, 5 Mar 1999 18:16:52 -0700 (MST) X-Sent-via: StarNet http://www.azstarnet.com/ Message-ID: <000b01be676f$375fd250$210148bf@dylan> From: "David McClain" To: "Liste CAML" Subject: Speedup from 2.01 to 2.02 Date: Fri, 5 Mar 1999 18:18:15 -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 4.72.3110.5 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3 Resent-From: weis@pauillac.inria.fr Resent-Date: Mon, 8 Mar 1999 09:00:36 +0100 Resent-To: caml-redistribution@pauillac.inria.fr I have just recompiled my BlurFit routine under OcamlOpt 2.02. This version is a completely new and far more sophisticated version than the one I developed in December in only 3 days. However, its timing is still dominated by large numeric array operations, and I would expect no increase in performce for those inner tight loops. The non-numeric overhead has been measured at 0.28 sec/iteration under OCamlOpt 2.01 and 0.6 sec/iteration for the numeric operations. Under OcamlOpt 2.02 the same program executes in 0.85 sec/iteration, for an apparent speedup of roughly 10% in the non-numeric OCAML overhead code. I think this is very good indeed. But I am posting this for the benefit of the developers who may be interested in how their changes affect real-life programs. I would be interested to hear from others regarding the speed improvements. David McClain Sr. Scientist Raytheon Missile Systems Co. Tucson, AZ http://www.azstarnet.com/~dmcclain/homepage.htm