From xemacs-m  Thu Mar 13 00:30:57 1997
Received: from altair.xemacs.org (steve@xemacs.miranova.com [206.190.83.19])
	by xemacs.org (8.8.5/8.8.5) with ESMTP id AAA17789
	for <xemacs-beta@xemacs.org>; Thu, 13 Mar 1997 00:30:55 -0600 (CST)
Received: (from steve@localhost)
	by altair.xemacs.org (8.8.5/8.8.5) id WAA21679;
	Wed, 12 Mar 1997 22:42:57 -0800
Mail-Copies-To: never
To: xemacs-beta@xemacs.org
Subject: Re: off-topic - Nt vs Linux
References: <199703130607.BAA00427@beavis.bayserve.net>
X-Url: http://www.miranova.com/%7Esteve/
X-Face: #!T9!#9s-3o8)*uHlX{Ug[xW7E7Wr!*L46-OxqMu\xz23v|R9q}lH?cRS{rCNe^'[`^sr5"
 f8*@r4ipO6Jl!:Ccq<xoV[Qz2u8<8-+Vwf2gzJ44lf_/y9OaQ`@#Q65{U4/TC)i2`~/M&QI$X>p:9I
 OSS'2{-)-4wBnVeg0S\O4Al@)uC[pD|+
X-Attribution: sb
From: Steven L Baur <steve@miranova.com>
In-Reply-To: Jeff Miller's message of Thu, 13 Mar 1997 01:07:48 -0500
Mime-Version: 1.0 (generated by tm-edit 7.105)
Content-Type: text/plain; charset=US-ASCII
Date: 12 Mar 1997 22:42:55 -0800
Message-ID: <m2hgig9tsg.fsf@altair.xemacs.org>
Lines: 41
X-Mailer: Gnus v5.4.25/XEmacs 20.1

Please, no more List Abuse in the subject messages.  This mailing list
is for developing XEmacs.  If you simply must post here, please use
xemacs-beta-discuss@xemacs.org instead.

Jeff Miller writes:

> well, sorta off topic. 

 ...

> I'm hoping and praying that there is just some non-obvious compile
> options or something we've been missing.  I thik the last options I
> used were like "-O4" I think this gave me the best.

I really, *really*, hate to bring this up.  After recent discussion
on the Linux mailing lists about performance on Linux -vs- MS I've
been trying XEmacs compiled with `-fomit-frame-pointer'.  Without a
parachute in every sense of the word because compiling with that
option eliminates the possiblity of using stack backtraces to track
down problems. (If you choose to run beta XEmacsen in this fashion,
please keep a copy of a normal -g compiled one around for reporting
problems, as I will not look at any problem with an XEmacs compiled
in this fashion).

Subjectively it seems to fly.  I haven't run any bench marks yet as
I've only done it with MULE enabled.  I only have a Pentium not a
Pentium Pro, your mileage may vary, don't even dream of reporting a
problem with an XEmacs compiled this way and getting a response to
your bug report.  The exact CFLAGS I've been using are

-m486 -O4 -fno-strength-reduce -malign-loops=2 -malign-jumps=2
	   -malign-functions=2 -fomit-frame-pointer

I would also consider looking at the stack alignment.  The same recent
discussion in the Linux Mailing lists indicates that Gcc can be very
stupid about it, and stupidity is costly in terms of performance.  You
don't say what libc version you're using, upgrade it to 5.4.23.  Only
the very latest 5.4 libcs address the problem of stack misalignment.
-- 
steve@miranova.com baur
Unsolicited commercial e-mail will be billed at $250/message.

