From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me X-Spam-Level: X-Spam-Status: No, score=-0.8 required=5.0 tests=BAYES_00,INVALID_DATE autolearn=no autolearn_force=no version=3.4.4 X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: f569d,7d68662d79c59f1c X-Google-Attributes: gidf569d,public X-Google-Thread: 109fba,46882e3fad98420e X-Google-Attributes: gid109fba,public X-Google-Thread: 1108a1,9292211c2d4756a8 X-Google-Attributes: gid1108a1,public X-Google-Thread: 103376,7d68662d79c59f1c X-Google-Attributes: gid103376,public X-Google-Thread: 102b75,7d68662d79c59f1c X-Google-Attributes: gid102b75,public X-Google-ArrivalTime: 1995-01-04 14:05:03 PST Path: nntp.gmd.de!newsserver.jvnc.net!nntpserver.pppl.gov!princeton!gw1.att.com!csn!magnus.acs.ohio-state.edu!math.ohio-state.edu!howland.reston.ans.net!ix.netcom.com!netcomsv!netcomsv!annwfn!annwfn!merlin From: merlin@annwfn.com (Fred McCall) Subject: Re: What's Real-Time? (was Re: Widespread C++ Competency Gap?) Date: Wed, 04 Jan 1995 16:05:03 -0600 Newsgroups: comp.lang.c++,comp.object,comp.arch,comp.multimedia,comp.lang.ada Message-ID: <86203738B3A@annwfn.com> References: Organization: Is For People Who Don't Have Real Work X-Newsreader: Newsy 0.23 Xref: nntp.gmd.de comp.lang.c++:85203 comp.object:19214 comp.arch:27440 comp.multimedia:25874 comp.lang.ada:17766 Date: 1995-01-04T16:05:03-06:00 List-Id: In hbaker@netcom.com Henry Baker writes: >In the RT community there are the terms '_hard_ real-time' and >'_soft_ real-time'. 'Hard' RT means that the consequences of missing >a deadline are catastrophic, while 'soft' RT means that the consequences >are expensive, but not catastrophic. This doesn't quite jibe with the way I learned it (although reflection seems to show that we're talking about different aspects of the same thing). Real-time (in general) means that there are time-driven constraints on the system which *must* be met. 'Hard' real-time implies that the required timeline must be met *exactly*; in other words, you need a deterministic machine where you know the timing of all operations and combinations thereof in order to do this correctly. If you blow the timeline (either early or late, within a very narrow window), the system will fail and cannot be recovered. 'Soft' real-time sets upper limits on how long things can take, but if you get there early the system will run along fine. Blown timelines can generally be recovered if other later operations can be performed fast enough to 'make up' the blown timeline. >Unfortunately, nearly _all_ of the progress in HW performance in the last >20 years has been in _mean_ performance, _not_ worst-case performance. >In fact, worst-case performance has only improved modestly over this >period. The _ratio_ of mean to worst-case performance has increased >dramatically, so that it may be 100:1 or more today. Not to mention that most of the things that we do to make the hardware faster also tend to make it less deterministic (super-scalar, super-pipelined designs, etc.) and hence less suitable for 'hard' real-time unless you design with extraordinary margins (so that you can still make the timeline if the pipeline stalls and has to be flushed, for example). > -- "Insisting on perfect safety is for people who don't have the balls to live in the real world." -- Mary Shafer, NASA Ames Dryden --------------------------------------------------------------------------- merlin@annwfn.com -- I don't speak for others and they don't speak for me.