From owner-freebsd-hackers  Mon Mar  4 04:24:14 1996
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id EAA26676
          for hackers-outgoing; Mon, 4 Mar 1996 04:24:14 -0800 (PST)
Received: from dyson.iquest.net ([198.70.144.127])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id EAA26671
          for <hackers@FreeBSD.ORG>; Mon, 4 Mar 1996 04:24:11 -0800 (PST)
Received: (from root@localhost) by dyson.iquest.net (8.6.11/8.6.9) id HAA01155; Mon, 4 Mar 1996 07:19:31 GMT
From: "John S. Dyson" <toor@dyson.iquest.net>
Message-Id: <199603040719.HAA01155@dyson.iquest.net>
Subject: Re: pwd_mkdb - really slow :(
To: bde@zeta.org.au (Bruce Evans)
Date: Mon, 4 Mar 1996 07:19:30 +0000 ()
Cc: julian@ref.tfs.com, rashid@rk.ios.com, hackers@FreeBSD.ORG
In-Reply-To: <199603041002.VAA05866@godzilla.zeta.org.au> from "Bruce Evans" at Mar 4, 96 09:02:38 pm
X-Mailer: ELM [version 2.4 PL24 ME8]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-hackers@FreeBSD.ORG
Precedence: bulk

>
> same block is written many times.  db often does this.  I think it can
> be programmed to do more buffering in memory, but often isn't.  kvm_mkdb
> was speeded up by a factor of about 5 by rewriting it to use mmap and
> more buffering; now it is only about twice as slow as the old version
> would be with a working disk cache ;-).
> 
> >it might also be quicker to build the entire databse in ram 
> >and bang it out to disk as a single write.. :)
> 
> Not much, if write() is suitably buffered.
> 
Ahh... Come on, you know that the disk cache works, it simply does not
write-behind as much as you would like.  It read-caches very nicely.

John