From owner-freebsd-questions@FreeBSD.ORG Mon Jun 18 10:34:38 2007 Return-Path: X-Original-To: questions@freebsd.org Delivered-To: freebsd-questions@FreeBSD.ORG Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 8D6E216A41F for ; Mon, 18 Jun 2007 10:34:38 +0000 (UTC) (envelope-from perrin@apotheon.com) Received: from host222.ipowerweb.com (host222.ipowerweb.com [66.235.210.10]) by mx1.freebsd.org (Postfix) with SMTP id 6F55F13C455 for ; Mon, 18 Jun 2007 10:34:38 +0000 (UTC) (envelope-from perrin@apotheon.com) Received: (qmail 40839 invoked from network); 18 Jun 2007 10:30:42 -0000 Received: from unknown (HELO demeter.hydra) (24.9.123.251) by host222.ipowerweb.com with SMTP; 18 Jun 2007 10:30:42 -0000 Received: from demeter.hydra (localhost [127.0.0.1]) by demeter.hydra (8.13.6/8.13.6) with ESMTP id l5IAYaM3037958 for ; Mon, 18 Jun 2007 04:34:36 -0600 (MDT) (envelope-from perrin@apotheon.com) Received: (from ren@localhost) by demeter.hydra (8.13.6/8.13.6/Submit) id l5IAYawo037957 for questions@freebsd.org; Mon, 18 Jun 2007 04:34:36 -0600 (MDT) (envelope-from perrin@apotheon.com) X-Authentication-Warning: demeter.hydra: ren set sender to perrin@apotheon.com using -f Date: Mon, 18 Jun 2007 04:34:36 -0600 From: Chad Perrin To: questions@freebsd.org Message-ID: <20070618103436.GC37851@demeter.hydra> Mail-Followup-To: questions@freebsd.org References: <04E232FDCD9FBE43857F7066CAD3C0F137B3FD@svmailmel.bytecraft.internal> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <04E232FDCD9FBE43857F7066CAD3C0F137B3FD@svmailmel.bytecraft.internal> User-Agent: Mutt/1.4.2.2i Cc: Subject: Re: FreeBSD and Robotics X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 18 Jun 2007 10:34:38 -0000 On Mon, Jun 18, 2007 at 01:32:21PM +1000, Murray Taylor wrote: > I can only think of one other point for this... > Interrupt latency. Depending on what you are attempting to do, > the variable nature of interrupt responses could be an issue. > I.e. if the system becomes io bound during a data capture cycle, > and something occurs that requires a response within a very narrow > window, it is possible to miss the window due to other interrupt > processes running. > > For this reason, robotics systems often run on highly optimised > single process systems where there is a 'guaranteed' poll cycle > and / or a very minimal defined interrupt system with minimal > overheads. I suspect that increasing concurrency capability in the near future will change that a fair bit. It'll probably start with highly concurrent embedded and realtime OS development (he said, wildly guessing) and seep out into other areas of computing from there. -- CCD CopyWrite Chad Perrin [ http://ccd.apotheon.org ] awj @reddit: "The terms never and always are never always true."