From xemacs-m  Sun Dec 15 18:46:28 1996
Received: from altair.xemacs.org (steve@xemacs.miranova.com [206.190.83.19]) by xemacs.cs.uiuc.edu (8.8.3/8.8.3) with ESMTP id SAA19874 for <xemacs-beta@xemacs.org>; Sun, 15 Dec 1996 18:46:26 -0600 (CST)
Received: (from steve@localhost)
          by altair.xemacs.org (8.8.4/8.8.4)
	  id QAA24773; Sun, 15 Dec 1996 16:56:28 -0800
Sender: steve@xemacs.org
To: xemacs-beta@xemacs.org
Subject: Re: 19.15-b4 bench results
References: <199612160002.RAA28973@branagh.lanl.gov>
X-Url: http://www.miranova.com/%7Esteve/
Mail-Copies-To: never
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: John Turner's message of Sun, 15 Dec 1996 17:02:05 -0700
Mime-Version: 1.0 (generated by tm-edit 7.96)
Content-Type: text/plain; charset=US-ASCII
Date: 15 Dec 1996 16:56:28 -0800
Message-ID: <m2engrfgsj.fsf@altair.xemacs.org>
Lines: 75
X-Mailer: Red Gnus v0.72/XEmacs 19.15

>>>>> "John" == John Turner <turner@xdiv.lanl.gov> writes:

Thanks for the report John.

John> Some very interesting results.  I don't have access to 19.34 on
John> the machine I use XEmacs on, so I only compare to 19.14.  My
John> goal here was mainly to isolate my own perf. problems, so that
John> didn't matter.

John> Turns out it may shed light on the "hang until C-l or other input"
John> thing as well.

John> Platform is UltraSPARC 170E, Solaris 2.5.1, CDE.  In all cases I've
John> deleted the Call Count and Elapsed Time columns to save width, since I
John> only did (bench 1).

John> First, 19.14.  Here are the results with -q using the precompiled
John> binaries for Solaris 2.4:

It looks like you compiled under greater optimization.  You left one
thing out.  Did the display lock up?  Try stripping all but the Tower
of Hanoi and Large File scrolling, and running (bench 5) or (bench 10).

John> Now's where it gets interesting.  Here's 19.15-b4 with my .emacs
John> loaded:

 ...

John> First of all, it's obvious that at least some aspects of my
John> "loaded" XEmacs are considerably slower than the stripped-down
John> one.  Hanoi is slower by a factor of 2.5x, font-lock by 2x,
John> frame creation by 2x, etc.

Fascinating.  But be sure to run the tests multiple times.  elp uses
wallclock times, and if your system was under different levels of
load, that alone could count for the variation.

 ...

John> So I looked at the commonly-blamed packages, lazy-lock,
John> func-menu, and truncate-lines t.  I turned each of them off
John> individually, and guess what?  Test 3 still always locked up
John> until C-l, with other tests doing so occasionally, and times for
John> the tests that succeeded were not appreciably different from
John> when each pkg was turned on.

Keep turning things off then.  If it isn't lazy-lock, or func-menu it
must be something else then.  I hope what you meant by `turned off',
was exiting the program, commenting out the lines in .emacs that turn
on the specific package and restarting.  All bets are off otherwise.

John> Wondering what's going on,

If you didn't lock up when running in a naked environment, but you did
when your .emacs is loaded, then you need to isolate which package is
doing the damage.

I think if we only get as far as isolating that, it would be an
enormous improvement.

I've made one more personal observation in the matter of performance.
We seem to be doing certain things several times.  I've already seen
situations with multiple frames where every frame gets completely
redrawn twice on each keystroke.

The example I just found is that for each keystroke, the translation
against the X keysym code is done twice (keymap.c:define_key_parser()).
(Especially) In an editor, the only thing as precious as the text
being edited is the user's keystrokes in data entry.  So this bears
some looking into.
-- 
steve@miranova.com baur
Unsolicited commercial e-mail will be billed at $250/message.
"That Bill Clinton.  He probably doesn't know how to log on to the
Internet."  -- Rush Limbaugh, noted Computer Expert

