From xemacs-m  Wed Mar 26 18:18:36 1997
Received: from corona.pixar.com (corona.pixar.com [138.72.20.84])
	by xemacs.org (8.8.5/8.8.5) with SMTP id SAA07006
	for <xemacs-beta@xemacs.org>; Wed, 26 Mar 1997 18:18:35 -0600 (CST)
Received: by corona.pixar.com (Smail3.1.29.1 #2)
	id m0wA2sd-0001ZYC; Wed, 26 Mar 97 16:17 PST
Sender: retnuh@pixar.com (Hunter Kelly)
Sender: retnuh@corona
To: wmperry@aventail.com
Cc: Per Abrahamsen <abraham@dina.kvl.dk>, xemacs-beta@xemacs.org
Subject: Re: Speed (was Re: NEWS submission)
References: <199703252230.RAA19898@blight.IntraNet.com> 	<m2d8snfrtr.fsf@altair.xemacs.org> 	<199703252349.PAA17211@newman> 	<kig2093bjx2.fsf@jagor.srce.hr> 	<rxspvwmaovv.fsf@midnight.ecf.teradyne.com> 	<199703261447.GAA18348@newman> 	<rjpvwm7j04.fsf@zuse.dina.kvl.dk> <199703261551.HAA18603@newman>
Mime-Version: 1.0 (generated by tm-edit 7.106)
Content-Type: text/plain; charset=US-ASCII
From: Hunter Kelly <retnuh@corona.pixar.com>
Date: 26 Mar 1997 16:17:20 -0800
In-Reply-To: "William M. Perry"'s message of Wed, 26 Mar 1997 07:51:46 -0800
Message-ID: <yvtrah289yn.fsf@corona.pixar.com>
Lines: 59
X-Mailer: Gnus v5.4.37/XEmacs 19.15(beta104)

On an SGI Indigo 100Mhz R4000 96Megs of ram, the parsing took about a
minute, and drawing took somewhere between 5 and 10 minutes.  This is
with all debugging turned off, compiled with -O2, I believe.
(beta104).

Hunter

"William M. Perry" <wmperry@aventail.com> writes:

> Per Abrahamsen writes:
> >
> >"William M. Perry" <wmperry@aventail.com> writes:
> >
> >> Adrian Aichner writes:
> >> >>>>>> "Hrv" == Hrvoje Niksic <hniksic@srce.hr> writes:
> >> >
> >> >    Hrv> "William M. Perry" <wmperry@aventail.com> writes:
> >> >    >> - Much much much faster
> >> >
> >> >    Hrv> How much is this?  For example, how long does it take to
> >> >    Hrv> render <URL:http://fly.cc.fer.hr/HTML3/WD-css1.html> on a
> >> >    Hrv> computer humans can afford?  That URL used to be my
> >> >
> >> >This is date for a 110MHz SPARCstation 5 running Solaris2.5 with
> >> >64Mbyte memory. This is a computer the company I work for can afford,
> >> >not me :-(
> >> 
> >>   This is a PPRO/200 w/32M of memory, a machine people can pretty easily
> >> afford ($3k).  I also use a 586/100 laptop, which damn near anyone can
> >> afford ($1500 5 months ago)
> >
> >Perhaps.  I'm using an machine identical to Adrian's (110MHz SPARCstation
> >5 running Solaris2.5 with 64Mbyte memory).  The html file is copied to a
> >local (not NFS) disk first.
> >
> >Here parsing takes about 15 seconds, and drawing .. well, it is 19 minutes
> >since it started, and it is not finished yet, so I'll believe Adrian's
> >number.
> 
>   I don't see how these gigantic discrepancies could be Emacs-W3's fault.
> Unless your X server sucks and querying fonts is abominably slow.
> 
>   Are either of you running a XEmacs with debugging or error checking
> turned on?  Emacs-W3 makes extensive use of extents, and this is where you
> get hammered a lot with error checking turned on.
> 
> >If I use Emacs 19.34 instead of XEmacs on the same machine.  parsing
> >takes 6 seconds, and drawing one minute.
> >
> >Something is very wrong with XEmacs performance.
> >
> >If I use a 200 MHz UltraSPARC instead (comparable to a PPRO 200 MHz) the
> >parsing and drawing time are cut to a third.  This makes w3 on Emacs 19.34
> >usable.
> 
>   Could there be something architecture-specific that would account for
> this incredible slowdown?  Anyone else see ridiculous timings like this?
> 
> -Bill P.

