From owner-freebsd-hardware  Tue Oct  1 22:01:17 1996
Return-Path: owner-hardware
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id WAA02808
          for hardware-outgoing; Tue, 1 Oct 1996 22:01:17 -0700 (PDT)
Received: from lserver.infoworld.com (lserver.infoworld.com [192.216.48.4])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id WAA02800
          for <freebsd-hardware@FreeBSD.org>; Tue, 1 Oct 1996 22:01:15 -0700 (PDT)
From: BRETT_GLASS@infoworld.com
Received: from ccgate.infoworld.com (ccgate.infoworld.com [192.216.49.101]) 
	by lserver.infoworld.com (8.7.5/8.7.3/GNAC-GW-1.2) with SMTP id WAA16388; Tue, 1 Oct 1996 22:01:03 -0700 (PDT)
Received: from ccMail by ccgate.infoworld.com (SMTPLINK V2.11)
	id AA844232194; Tue, 01 Oct 96 22:51:08 PST
Date: Tue, 01 Oct 96 22:51:08 PST
Message-Id: <9609018442.AA844232194@ccgate.infoworld.com>
To: richard@pegasus.com (Richard Foulk), freebsd-hardware@FreeBSD.org
Subject: Re: H/W recommendation
Sender: owner-hardware@FreeBSD.org
X-Loop: FreeBSD.org
Precedence: bulk

> This isn't quite the way cache works.  A board with 512K of cache
> won't hold 512K of code, even if that were necessary -- most of the
> code in the kernel or any other large program seldom gets run.

I never stated that the cache would (or should!) hold the whole kernel.
It will, obviously, hold some data and user code. 

But it's a time-tested rule of thumb that, to a first approximation,
your cache should equal the size of the kernel. (In the case of UNIX, the
kernel should be trimmed to include only the devices that are actually
present).