From owner-freebsd-current  Sat Nov 11 22:38:32 1995
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.6.12/8.6.6) id WAA25677
          for current-outgoing; Sat, 11 Nov 1995 22:38:32 -0800
Received: from asstdc.scgt.oz.au (root@asstdc.scgt.oz.au [202.14.234.65])
          by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id WAA25672
          for <freebsd-current@freefall.FreeBSD.org>; Sat, 11 Nov 1995 22:38:26 -0800
Received: (from imb@localhost)
	by asstdc.scgt.oz.au (8.6.12/BSD-4.4)
	id RAA26875; Sun, 12 Nov 1995 17:38:12 +1100
From: michael butler <imb@scgt.oz.au>
Message-Id: <199511120638.RAA26875@asstdc.scgt.oz.au>
Subject: Re: cvs commit: src/usr.sbin/pppd RELNOTES
To: hsu@cs.hut.fi (Heikki Suonsivu)
Date: Sun, 12 Nov 1995 17:38:11 +1100 (EST)
Cc: freebsd-current@freefall.FreeBSD.org
In-Reply-To: <199511120341.FAA08651@shadows.cs.hut.fi> from "Heikki Suonsivu" at Nov 12, 95 05:41:34 am
X-Mailer: ELM [version 2.4 PL24beta]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 2460      
Sender: owner-current@FreeBSD.org
Precedence: bulk

Heikki Suonsivu writes:

>    ZIP is computationally expensive .. you'll speed most of your time in the
 
> Huh?
 
> hsu#hauki.clinet.fi Sun 37: time gzip -1 < csh > /dev/null

 .. ah .. this is not the default compression method. You should've said :-)

> This was on a P90, -current from two weeks back.

For comparison and using the default method on a 386DX/40 also running
-current ..

asstdc:~ % time gzip </bin/csh >/dev/null
9.417u 0.109s 0:10.63 89.4% 100+624k 9+0io 10pf+0w
asstdc:~ % time gzip < /bin/csh > /dev/null
9.841u 0.062s 0:09.98 99.1% 99+614k 0+0io 0pf+0w
asstdc:~ % time gzip < /bin/csh > /dev/null
9.100u 0.078s 0:09.30 98.6% 99+616k 0+0io 0pf+0w
asstdc:~ % time gzip < /bin/csh > /dev/null
10.018u 0.108s 0:10.67 94.7% 100+619k 0+0io 0pf+0w
asstdc:~ % time gzip < /bin/csh > /dev/null
9.319u 0.093s 0:10.48 89.6% 101+630k 0+0io 0pf+0w
asstdc:~ % time gzip < /bin/csh > /dev/null
9.230u 0.046s 0:09.37 98.9% 99+617k 0+0io 0pf+0w

 => mean of ~9.49 user secs.

asstdc:~ % time compress </bin/csh >/dev/null
3.315u 0.108s 0:03.42 99.7% 19+727k 2+0io 1pf+0w
asstdc:~ % time compress < /bin/csh > /dev/null
2.822u 0.093s 0:02.90 100.3% 20+728k 0+0io 0pf+0w
asstdc:~ % time compress < /bin/csh > /dev/null
2.900u 0.094s 0:03.03 98.6% 20+733k 0+0io 0pf+0w
asstdc:~ % time compress < /bin/csh > /dev/null
2.816u 0.101s 0:02.91 100.0% 19+729k 0+0io 0pf+0w
asstdc:~ % time compress < /bin/csh > /dev/null
2.810u 0.117s 0:02.91 100.3% 20+729k 0+0io 0pf+0w
asstdc:~ % time compress < /bin/csh > /dev/null
3.349u 0.109s 0:03.44 100.0% 18+730k 0+0io 0pf+0w
asstdc:~ % time compress < /bin/csh > /dev/null
2.862u 0.109s 0:02.99 98.9% 19+729k 0+0io 0pf+0w

 => mean of ~2.982 user secs.

Without the qualification of not using the default method, ZIP is ~300% more
computationally expensive. We are both right .. just talking about quite
different things.

The difference is size is also quite marked .. the default gives ..

-r-xr-xr-x  1 imb  staff  237568 Nov 12 17:27 csh
-r-xr-xr-x  1 imb  staff  124868 Nov 12 17:26 csh.gz

At an estimated 3000 bytes per second down a 28k8 modem link and with my
386, this would translate into 9 seconds of CPU for 41 seconds of transfer
time or a load average of ~0.22 not counting any other system or interrupt
overhead. This is getting rather steep as I have 3 such modems directly
attached to this box and the reason why I was concerned. There is a very
fixed limit where it runs out of "puff" :-(

	michael