Message-Id: <199411171125.DAA06673@faerie.CS.Berkeley.EDU>
From: aoki@CS.Berkeley.EDU (Paul M. Aoki)
To: bmcuyper@etro1.vub.ac.be (Bernard De Cuyper)
Cc: postgres-arch@postgres.Berkeley.EDU
Subject: Re: Speeding up Postgres 4.2 
Reply-To: aoki@CS.Berkeley.EDU (Paul M. Aoki)
In-reply-to: Your message of Thu, 17 Nov 94 12:17:51 +0100 
	     <9411171117.AA04308@etro6.vub.ac.be> 
Date: Thu, 17 Nov 94 03:25:14 -0800
Sender: aoki@postgres.Berkeley.EDU

bmcuyper@etro1.vub.ac.be (Bernard De Cuyper) writes:
> Thanks a lot, you should ask the guy to come back form illustra, and
> put this feature as standart, it is really working. It give the same
> impression that the first step on the moon. I use "main memory"
> mapping on Sun/Solaris2.3 and I plan to make some extra cache
> relations but it seems also interesting for more complex tasks to
> support labelling in images.

just as long as you know that this is like putting files on a memory
fs or sunos tmpfs -- if you reboot the machine, the main memory data
goes away..

(i've been of the opinion that nvram boards are a better answer for the
disk-write latency problem.. but of course those cost thousands of
dollars..)
--
  Paul M. Aoki          |  University of California at Berkeley
  aoki@CS.Berkeley.EDU  |  Dept. of EECS, Computer Science Division (#1776) 
                        |  Berkeley, CA 94720-1776