From owner-freebsd-hackers  Wed May 23 23:19:44 2001
Delivered-To: freebsd-hackers@freebsd.org
Received: from m44.spnet.com (m44.spnet.com [207.181.251.44])
	by hub.freebsd.org (Postfix) with ESMTP id 5737D37B43E
	for <hackers@FreeBSD.ORG>; Wed, 23 May 2001 23:19:41 -0700 (PDT)
	(envelope-from elh_fbsd@spnet.com)
Received: from spnet.com (localhost.spnet [127.0.0.1])
	by m44.spnet.com (8.11.3/8.11.3) with ESMTP id f4O6JJ900472;
	Wed, 23 May 2001 23:19:19 -0700 (PDT)
	(envelope-from elh_fbsd@spnet.com)
Message-Id: <200105240619.f4O6JJ900472@m44.spnet.com>
X-Mailer: exmh version 2.3.1 01/18/2001 with nmh-1.0.4
To: Mike Silbersack <silby@silby.com>
Cc: Ed Hudson <elh@spnet.com>, hackers@FreeBSD.ORG
Subject: Re: general speed differences between 4.1.1-RELEASE and 4.3-RELEASE 
In-Reply-To: Message from Mike Silbersack <silby@silby.com> 
   of "Thu, 24 May 2001 00:48:02 CDT." <20010524004057.V55532-100000@achilles.silby.com> 
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Wed, 23 May 2001 23:19:19 -0700
From: Ed Hudson <elh_fbsd@spnet.com>
Sender: owner-freebsd-hackers@FreeBSD.ORG
Precedence: bulk
List-ID: <freebsd-hackers.FreeBSD.ORG>
List-Archive: <http://docs.freebsd.org/mail/> (Web Archive)
List-Help: <mailto:majordomo?subject=help> (List Instructions)
List-Subscribe: <mailto:majordomo?subject=subscribe%20freebsd-hackers>
List-Unsubscribe: <mailto:majordomo?subject=unsubscribe%20freebsd-hackers>
X-Loop: FreeBSD.ORG

Mike Silbersack <silby@silby.com> writes:
> Write caching is now off by default.  man ata to see how to turn it back
> on.

	Mr. Silbersack, thank you very much.  you've restored my
	systems to their pre-4.3 stunningly fast behavior.

	to the hackers group, i apologize for polluting the air waves with
	my original plea for help.

> <begin divergance>
> 
> Seeing messages like this is becoming quite common on the lists since the
> change was made.  On the other hand, I don't recall seeing any message

	for my money, hw.ata.wc=1 soft updates OFF is a better performing
	choice than hw.ata.wc=0 and soft updates ON.
	(soft updates are great, but i really dislike the performance
	stalls that it (or async mode) engenders with big copies/etc
	for other processes).

	if the freebsd group still chooses to keep hw.ata.wc=0 as
	the default, i would urge perhaps a different mode for base
	system installs.  the reason i used '8192m' partition in
	my original mail was because the wait for a 60gig newfs was
	pretty painful.


	in my original message i made a reference to linux.
	i'm a die-hard freebsd fan, but for work reasons etc i take
	a look at linux from time to time.  in the same system that
	i mentioned originally, i've tried the experiment of powering
	off a machine (with no halt) after the first reboot after
	a virgin suse-7.1 install.  linux couldn't put itself back together
	without manual intervention - the fsck seemed to just freak.

	even with the historic hw.ata.wc=1 effective mode of freebsd-*,
	i've never seen such bad behavior as i had with the linux-7.1
	system.


	THANKS again to all you folks for the outstanding performance
	and reliability that freebsd attains.
	
		-elh



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message