From owner-freebsd-current  Mon Jun 12 19:50:19 1995
Return-Path: current-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id TAA11844
          for current-outgoing; Mon, 12 Jun 1995 19:50:19 -0700
Received: from haven.uniserve.com (haven.uniserve.com [198.53.215.121])
          by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id TAA11730
          for <current@FreeBSD.ORG>; Mon, 12 Jun 1995 19:50:00 -0700
Received: by haven.uniserve.com id <30732>; Mon, 12 Jun 1995 19:50:57 +0100
Date: 	Mon, 12 Jun 1995 19:50:50 -0700 (PDT)
From: Tom Samplonius <tom@uniserve.com>
To: Atsushi Murai <amurai@spec.co.jp>
cc: current@FreeBSD.ORG
Subject: Re: GENERIC kernel & some basic UNIX pointers
In-Reply-To: <9506130247.AA00039@tama3.spec.co.jp.spec.co.jp>
Message-ID: <Pine.BSF.3.91.950612194519.16318B-100000@haven.uniserve.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: current-owner@FreeBSD.ORG
Precedence: bulk


On Tue, 13 Jun 1995, Atsushi Murai wrote:

> But Predictor-1 will try compression whole date when talk to peer ppp.
> This will give us two advantage as follows:
> 
>      1. Decrese system over head due to reducing a sio H/W interruption.
>         (H/W interruption is required big cpu time especially latest cpu)

  What about the extra CPU time and extra latency required for the actual 
compression?

>      2. For ISDN TA that doen's has a own H/W compression can get double
>         speed with not compression data. ( i.e. 38400bps -> 7KB/sec )

  I've looked at this application myself.  Especially over 56k or 64k 
links.  However, dedicated units still provide better performance (ex. 
Gandalf 5250i with "advertised" compression up to 8:1, and in reality 
isn't as much, but is still very good).

Tom