Sender: eric@ruckus.brouhaha.com From: Eric Smith Newsgroups: comp.arch,alt.folklore.computers Subject: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> Organization: Eric Conspiracy Secret Labs X-Eric-Conspiracy: There is no conspiracy. Date: 04 Sep 2001 16:27:11 -0700 Message-ID: Lines: 82 User-Agent: Gnus/5.0807 (Gnus v5.8.7) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii NNTP-Posting-Host: ruckus.brouhaha.com X-Trace: 4 Sep 2001 16:39:56 -0700, ruckus.brouhaha.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.mathworks.com!news.kjsl.com!news.spies.com!ruckus.brouhaha.com Xref: chonsp.franklin.ch alt.folklore.computers:89665 Joe Pfeiffer wrote: > third interrupt; if somebody wants to do a disk access we'll set the > timer for when we expect the disk seek to complete! Steve O'Hara-Smith writes: > The incredible thing is that this machine actually ran applications > succesfully. I had never realised they were so crude. It's not that it was crude, it's that it was a minimalistic design. It was intended to be a closed-box machine with no expansion other than through the external interface ports. Thus the internal wiring of interrupt lines, use of timers, etc. could be done in the lowest cost method. Since user software was NOT supposed to touch the hardware directly, but go through the supplied device drivers, this was an entirely reasonable design. Unlike, for instance, the IBM PC, in which the "supplied device drivers" (i.e., the BIOS) were a complete crock, and it was not possible to write sophisticated software without bypassing it. For instance, the BIOS does not support any form of asynchronous completion of I/O, nor does it have even the most basic means of arbitrating for use of devices (e.g., there is no close function for the printer interface). Note that when Apple introduced the Macintosh II, the first open-architecture Macintosh, properly written software for the original Mac still ran on it just fine. The floppy controller hardware was still strange, but that still didn't matter, because the driver still abstracted it. The new slots had a reasonably elegant design, with well defined electrical characteristics and a fairly good software frameword (the Slot Manager). Again, compare this to the IBM PC, which didn't get a reasonable, well-defined, open bus until five years after the Macintosh II. There is a fine art (and a mostly forgotten one at that) to designing minimalistic hardware and/or software that performs the desired function well. The Apple ][, Disk ][, and the original 128K Macintosh are three of the most well-known hardware examples of this art. Minimal design is *much* harder than doing things the obvious way, so it usually takes much longer. Which is, of course, one reason why it's so rarely done: I have only made this letter longer because I have not had the time to make it shorter. -- Blaise Pascal, Lettres Provinciales, letter 16 (1657) Today's version of minimalism is to throw together however many million lines of code or however many million gates one needs for a quick-and-dirty design, on the basis that memory and gates are cheap. The result works, but is often buggy due to unanticipated interactions: ...there are two ways of constructing a software design: One way is to make it so simple that there are _obviously_ no deficiencies and the other way is to make it so complicated that there are no _obvious_ deficiencies. The first method is far more difficult. It demands the same skill, devotion, insight, and even inspiration as the discovery of the simple physical laws which underlie the complex phenomena of nature. It also requires a willingness to accept objectives which are limited by physical, logical, and technological constraints, and to accept a compromise when conflicting objectives cannot be met. No committee will ever do this until it is too late. -- C.A.R. Hoare, "The Emperor's Old Clothes" (1980 ACM Turing Award Lecture), CACM vol. 24 no. 2, February 1981 Today's markets demand more and more features, at the expense of any concern for quality. People have the wrong metric for "perfection": Perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away. -- Antoine de Saint Exupery Minimalist design is most commonly found only in places where it was an absolute necessity: And in those days many a clever programmer derived an immense intellectual satisfaction from the cunning tricks by means of which he contrived to squeeze the impossible into the constraints of his equipment. -- Edsger W. Dijkstra, "The Humble Programmer", 1972 ACM Turing Award Lecture ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Thu, 06 Sep 01 08:23:30 GMT Organization: UltraNet Communications, Inc. Lines: 21 Message-ID: <9n7lkr$8p$4@bob.news.rcn.net> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> X-Trace: UmFuZG9tSVZ1RbtucZOHb44e5nNej9OhfJ7WZHvSUzo1Etxe+6oxYTmJufRSMFpV X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 6 Sep 2001 11:11:23 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!96651!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-102-46 Xref: chonsp.franklin.ch alt.folklore.computers:89710 In article <1bitex1pfy.fsf@cs.nmsu.edu>, Joe Pfeiffer wrote: >Steve O'Hara-Smith writes: >> ES> It's not that it was crude, it's that it was a minimalistic design. It >> ES> was intended to be a closed-box machine with no expansion other than >> >> OK, but using two out of three interrupts up on the mouse and not >> having one for the floppy - that's over the edge IMHO. > >That's my impression, as well. Yes, they did a very good job of >building a totally closed box and keeping the user from touching the >hardware. Was that really such a good goal? My impression was always >that the original Mac was pretty far over the edge. > Wait a minute. Are you saying that these systems can't deal with more than one interrupt to a level? How bizarre. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Thu, 06 Sep 01 08:24:38 GMT Organization: UltraNet Communications, Inc. Lines: 19 Message-ID: <9n7lmv$8p$5@bob.news.rcn.net> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n5vpo$qa9$1@news.panix.com> X-Trace: UmFuZG9tSVZyWYBm0N/EGp5hjZllxDEx6iec/PrqijV/pJw43yE+OctD6UB5+PX2 X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 6 Sep 2001 11:12:31 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.mathworks.com!news.voicenet.com!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-102-46 Xref: chonsp.franklin.ch alt.folklore.computers:89712 In article <9n5vpo$qa9$1@news.panix.com>, never+mail@panics.com.invalid (Michael Roach) wrote: > >In article , >Eric Smith wrote: >>If the objective was wrong, it may make the product less useful, but >>it doesn't make the design engineering any less impressive. > >I'm somehow reminded of the dialogue in the movie, "Short Circuit", that >goes something like this: > >"It's malfunctioning!" >"Yes, but what a malfunction!" I seem to recall JMF and TW admiring certain failures like that. /BAH Subtract a hundred and four for e-mail. ###### From: Steve O'Hara-Smith Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Wed, 5 Sep 2001 08:11:18 +0200 Organization: Wanadoo NL Lines: 26 Message-ID: <20010905081118.0a4e2a04.steveo@eircom.net> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> NNTP-Posting-Host: p1522.vcu.wanadoo.nl Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-Trace: scavenger.euro.net 999709211 33421 194.134.170.247 (5 Sep 2001 17:00:11 GMT) X-Complaints-To: abuse@wanadoo.nl NNTP-Posting-Date: Wed, 5 Sep 2001 17:00:11 +0000 (UTC) X-Newsreader: Sylpheed version 0.6.0 (GTK+ 1.2.10; i386--freebsd4.4) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!news-lei1.dfn.de!news-nue1.dfn.de!uni-erlangen.de!newsfeeds.belnet.be!news.belnet.be!newsgate.cistron.nl!news2.euro.net!news.euronet.nl!ams-gw.sohara.org!nobody Xref: chonsp.franklin.ch alt.folklore.computers:89636 On 04 Sep 2001 16:27:11 -0700 Eric Smith wrote: ES> Joe Pfeiffer wrote: ES> > third interrupt; if somebody wants to do a disk access we'll set the ES> > timer for when we expect the disk seek to complete! ES> ES> Steve O'Hara-Smith writes: ES> > The incredible thing is that this machine actually ran applications ES> > succesfully. I had never realised they were so crude. ES> ES> It's not that it was crude, it's that it was a minimalistic design. It ES> was intended to be a closed-box machine with no expansion other than OK, but using two out of three interrupts up on the mouse and not having one for the floppy - that's over the edge IMHO. ES> sophisticated software without bypassing it. For instance, the BIOS s/sophisticated/working/. No question the software layers of the Mac made the PC look like a toy, hence my surprise at the hardware. -- Directable Mirrors - A Better Way To Focus The Sun http://www.best.com/~sohara ###### From: Joe Pfeiffer Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 05 Sep 2001 12:07:45 -0600 Organization: NMSU Computer Science Lines: 17 Message-ID: <1bitex1pfy.fsf@cs.nmsu.edu> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> NNTP-Posting-Host: viper.cs.nmsu.edu X-Trace: bubba.NMSU.Edu 999713262 5076 128.123.64.113 (5 Sep 2001 18:07:42 GMT) X-Complaints-To: usenet@bubba.NMSU.Edu NNTP-Posting-Date: 5 Sep 2001 18:07:42 GMT X-Newsreader: Gnus v5.7/Emacs 20.5 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsxfer.eecs.umich.edu!hardy.tc.umn.edu!lynx.unm.edu!news.NMSU.Edu!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:89611 Steve O'Hara-Smith writes: > ES> It's not that it was crude, it's that it was a minimalistic design. It > ES> was intended to be a closed-box machine with no expansion other than > > OK, but using two out of three interrupts up on the mouse and not > having one for the floppy - that's over the edge IMHO. That's my impression, as well. Yes, they did a very good job of building a totally closed box and keeping the user from touching the hardware. Was that really such a good goal? My impression was always that the original Mac was pretty far over the edge. -- Joseph J. Pfeiffer, Jr., Ph.D. Phone -- (505) 646-1605 Department of Computer Science FAX -- (505) 646-1002 New Mexico State University http://www.cs.nmsu.edu/~pfeiffer SWNMRSEF: http://www.nmsu.edu/~scifair ###### From: never+mail@panics.com.invalid (Michael Roach) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 5 Sep 2001 19:52:24 GMT Organization: A small notepad underneath my in box Lines: 13 Message-ID: <9n5vpo$qa9$1@news.panix.com> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> NNTP-Posting-Host: panix3.panix.com X-Trace: news.panix.com 999719544 26953 166.84.1.3 (5 Sep 2001 19:52:24 GMT) X-Complaints-To: abuse@panix.com NNTP-Posting-Date: 5 Sep 2001 19:52:24 GMT X-Newsreader: trn 4.0-test74 (May 26, 2000) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!newscore.gigabell.net!newsfeeds.belnet.be!news.belnet.be!newsfeed.stanford.edu!logbridge.uoregon.edu!panix!news.panix.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:89705 In article , Eric Smith wrote: >If the objective was wrong, it may make the product less useful, but >it doesn't make the design engineering any less impressive. I'm somehow reminded of the dialogue in the movie, "Short Circuit", that goes something like this: "It's malfunctioning!" "Yes, but what a malfunction!" -- 10.0 times 0.1 is hardly ever 1.0. ###### From: name99@mac.com (Maynard Handley) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Wed, 05 Sep 2001 14:00:40 -0700 Organization: None Lines: 26 Message-ID: References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> NNTP-Posting-Host: handma2.apple.com X-Trace: news.apple.com 999723667 5059 17.202.32.152 (5 Sep 2001 21:01:07 GMT) X-Complaints-To: usenet@news.apple.com NNTP-Posting-Date: 5 Sep 2001 21:01:07 GMT X-Newsreader: MT-NewsWatcher 2.3.1 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.mathworks.com!nycmny1-snh1.gtei.net!paloalto-snf1.gtei.net!news.gtei.net!forum.apple.com!news.apple.com!handma2.apple.com!user Xref: chonsp.franklin.ch alt.folklore.computers:89733 In article <1bitex1pfy.fsf@cs.nmsu.edu>, Joe Pfeiffer wrote: > Steve O'Hara-Smith writes: > > ES> It's not that it was crude, it's that it was a minimalistic design. It > > ES> was intended to be a closed-box machine with no expansion other than > > > > OK, but using two out of three interrupts up on the mouse and not > > having one for the floppy - that's over the edge IMHO. > > That's my impression, as well. Yes, they did a very good job of > building a totally closed box and keeping the user from touching the > hardware. Was that really such a good goal? My impression was always > that the original Mac was pretty far over the edge. It was clearly ahead of its time---experience showed that with the state of HW at the time. people wanted and needed to expand their boxes. But the basic idea was not wrong---with better hardware (USB and Firewire) iMacs have sold well with no-one complaining about them being closed boxes with no-one touching the hardware. For that matter, the same is true of my Ti PowerBook---sure it has a PC Card slot, but I've never used it. Also Apple has realized that, at least for now, people do want different amounts of RAM, meaning that on iMacs RAM is the one HW thing you can pretty easily upgrade, a major mistake in the original Mac. Maynard ###### Message-ID: <3B971D19.E2071ADF@beagle-ears.com> Date: Wed, 05 Sep 2001 23:52:09 -0700 From: Lars Poulsen X-Mailer: Mozilla 4.75 [en]C-CCK-MCD (Win98; U) X-Accept-Language: da,en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 45 NNTP-Posting-Host: 207.154.106.6 X-Trace: azure.impulse.net 999759062 195 207.154.106.6 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!75768!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsxfer.interpacket.net!sienna.impulse.net!azure.impulse.net!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:89713 > ...there are two ways of constructing a software design: One way is > to make it so simple that there are _obviously_ no deficiencies and > the other way is to make it so complicated that there are no > _obvious_ deficiencies. > -- C.A.R. Hoare, "The Emperor's Old Clothes" (1980 ACM Turing > Award Lecture), CACM vol. 24 no. 2, February 1981 Eric Smith wrote: > Today's markets demand more and more features, at the expense of any > concern for quality. People have the wrong metric for "perfection": A current and concrete example from the software side: I am just going into a major discussion at my workplace (we build wireless network access systems). Someone in another division wants to build a "comprehensive, unified network administration and management platform", which they claim makes life easier for customers. I claim that this is going to be an endless sink for resources, and nobody will want to pay for it. I want to implement (a) user authentication and minimal accounting records using RADIUS in a way that makes the RADIUS server think we are a very slightly odd modem server box plus (b) a generic lowest common denominator SNMP MIB so that a comprehensive network management system will think we are "some sort of router, not quite sure what model" and finally (c) a custom GUI "query and configure" program, for Windows only, that can read and set all state and mode variables in the box, but which does not integrate with any "comprehensive platform". I know that I will lose this battle, but it is worth fighting, because it gets 95% of the job done at 25% of the cost, and I can predict the amount of effort. This is so obvious that I have a hard time explaining it to those who don't get it the first time. 5 years ago, the people I was working with abandoned all develop- ment of (originally portable) applications for MacIntosh, OS/2, Solaris and AIX because "it is obvious that in 5 years NT is going to be the only OS; not because it is better - it obviously isn't - but because Microsoft's monopolistic marketing practices will wipe everything else out". Lots of people reasoned that way, thus creating a self-fulfilling prophecy. -- / Lars Poulsen - http://www.beagle-ears.com - lars@beagle-ears.com 125 South Ontare Road, Santa Barbara, CA 93105 - +1-805-569-5277 ###### From: Joe Pfeiffer Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 06 Sep 2001 08:44:46 -0600 Organization: NMSU Computer Science Lines: 20 Message-ID: <1b66awidk1.fsf@cs.nmsu.edu> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> NNTP-Posting-Host: viper.cs.nmsu.edu X-Trace: bubba.NMSU.Edu 999787483 16358 128.123.64.113 (6 Sep 2001 14:44:43 GMT) X-Complaints-To: usenet@bubba.NMSU.Edu NNTP-Posting-Date: 6 Sep 2001 14:44:43 GMT X-Newsreader: Gnus v5.7/Emacs 20.5 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!64.245.249.19.MISMATCH!dfw3-feed1.news.digex.net!lax2-feed1.news.digex.net!intermedia!lynx.unm.edu!news.NMSU.Edu!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:89695 jmfbahciv@aol.com writes: > > > Wait a minute. Are you saying that these systems can't deal with > more than one interrupt to a level? How bizarre. One of the PC's screwups was that it was effectively impossible for it to deal with more than one interrupt at a level (it used assert-high interrupt requests). The Mac scheme was worse: it could only generate three interrupts in the first place, and any time a device requested an interrupt it could come in at four different levels depending on what else was interrupting at the same time. Bizarre is being kind. -- Joseph J. Pfeiffer, Jr., Ph.D. Phone -- (505) 646-1605 Department of Computer Science FAX -- (505) 646-1002 New Mexico State University http://www.cs.nmsu.edu/~pfeiffer SWNMRSEF: http://www.nmsu.edu/~scifair ###### Path: chonsp.franklin.ch!not-for-mail From: Neil Franklin Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 06 Sep 2001 20:55:16 +0200 Organization: My own Private Self Lines: 32 Message-ID: <6uy9nsqhd7.fsf@chonsp.franklin.ch> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> NNTP-Posting-Host: chonsp.franklin.ch X-Trace: chonsp.franklin.ch 999802520 432 10.0.3.2 (6 Sep 2001 18:55:20 GMT) X-Complaints-To: news@chonsp.franklin.ch NNTP-Posting-Date: 6 Sep 2001 18:55:20 GMT X-Newsreader: Gnus v5.7/Emacs 20.4 Xref: chonsp.franklin.ch alt.folklore.computers:89772 Joe Pfeiffer writes: > jmfbahciv@aol.com writes: > > > > > Wait a minute. Are you saying that these systems can't deal with > > more than one interrupt to a level? How bizarre. > > One of the PC's screwups was that it was effectively impossible for it > to deal with more than one interrupt at a level (it used assert-high > interrupt requests). Assert-high was not the problem (it is bad for catching interference). The problem was that the PC uses edge-triggered and not level-triggered interrupts. With edge-triggered The first device send an edge that sets an Flip-Flop. A hypothetical second device would have no effect and so be missed. With level-triggered the line is simply an wire-OR of all interrupt requests and keeps on interruptiong until it falls to zero as soon as everyone is satisfied. This sharing is possible. Worst is that the PIC chip can handle both types. The PC bus wiring simply demands edge (it lacks the resistors needed to pull wire-OR to the "no one" state). -- Neil Franklin, neil@franklin.ch.remove http://neil.franklin.ch/ Hacker, Unix Guru, El Eng HTL/BSc, Sysadmin, Archer, Roleplayer - Intellectual Property is Intellectual Robbery ###### From: never+mail@panics.com.invalid (Michael Roach) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 6 Sep 2001 18:07:40 GMT Organization: A small notepad underneath my in box Lines: 15 Message-ID: <9n8e1c$ohr$3@news.panix.com> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9n5vpo$qa9$1@news.panix.com> <9n7lmv$8p$5@bob.news.rcn.net> NNTP-Posting-Host: panix2.panix.com X-Trace: news.panix.com 999799660 25147 166.84.1.2 (6 Sep 2001 18:07:40 GMT) X-Complaints-To: abuse@panix.com NNTP-Posting-Date: 6 Sep 2001 18:07:40 GMT X-Newsreader: trn 4.0-test74 (May 26, 2000) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!nntp-out.monmouth.com!newspeer.monmouth.com!howland.erols.net!panix!news.panix.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:89777 In article <9n7lmv$8p$5@bob.news.rcn.net>, wrote: >In article <9n5vpo$qa9$1@news.panix.com>, > never+mail@panics.com.invalid (Michael Roach) wrote: >>"It's malfunctioning!" >>"Yes, but what a malfunction!" > >I seem to recall JMF and TW admiring certain failures like that. I remember a card punch that would occasionally send a stream of cards in a nice pretty arc right into the waste basket 8^) -- There's only one way to have a happy marriage and as soon as I learn what it is I'll get married again. -- Clint Eastwood ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Fri, 07 Sep 01 10:41:41 GMT Organization: UltraNet Communications, Inc. Lines: 28 Message-ID: <9nai45$f38$15@bob.news.rcn.net> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> X-Trace: UmFuZG9tSVa3nB+7WibFs/asAGJka+ZBNwqFah6IwKJ14v79zHtjZaY0rqu/FpoO X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 7 Sep 2001 13:29:41 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!newsfeed.r-kom.de!news0.de.colt.net!colt.net!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-255-184 Xref: chonsp.franklin.ch alt.folklore.computers:89831 In article , Eric Smith wrote: >Joe Pfeiffer writes: >> The Mac scheme was worse: it could only generate three interrupts in >> the first place, and any time a device requested an interrupt it could >> come in at four different levels depending on what else was >> interrupting at the same time. >> >> Bizarre is being kind. > >Bizarre, yet, but I dispute the characterization as "worse". Since it >was a bounded box, with no possibility for the addition of more >interrupt sources, and with the interrupt handlers carefully programmed >to understand the hardware, it was an entirely reasonable design. But that hogties the monitor! It would be nice if the software and hardware development can be mutually exclusive. That way pure software improvements can be done while the hardware guys are doing their thing. By tying a software development cycle to a hardware development cycle boggles my mind. It just ain't efficient. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Fri, 07 Sep 01 10:29:22 GMT Organization: UltraNet Communications, Inc. Lines: 20 Message-ID: <9nahd3$f38$13@bob.news.rcn.net> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9n5vpo$qa9$1@news.panix.com> <9n7lmv$8p$5@bob.news.rcn.net> <9n8e1c$ohr$3@news.panix.com> X-Trace: UmFuZG9tSVbhDWUwsKAtWpHiXHcK2SiAN924/wDOb2tOe0KYOf7Iff/wVTjk7Jgt X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 7 Sep 2001 13:17:23 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!7562!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-255-184 Xref: chonsp.franklin.ch alt.folklore.computers:89837 In article <9n8e1c$ohr$3@news.panix.com>, never+mail@panics.com.invalid (Michael Roach) wrote: > >In article <9n7lmv$8p$5@bob.news.rcn.net>, wrote: >>In article <9n5vpo$qa9$1@news.panix.com>, >> never+mail@panics.com.invalid (Michael Roach) wrote: >>>"It's malfunctioning!" >>>"Yes, but what a malfunction!" >> >>I seem to recall JMF and TW admiring certain failures like that. > >I remember a card punch that would occasionally send a stream of cards >in a nice pretty arc right into the waste basket 8^) Neat! I've never seen that! /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Fri, 07 Sep 01 10:28:30 GMT Organization: UltraNet Communications, Inc. Lines: 26 Message-ID: <9nahbf$f38$12@bob.news.rcn.net> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n5vpo$qa9$1@news.panix.com> <9n7lmv$8p$5@bob.news.rcn.net> <9n8c0n$fdm$2@plutonium.btinternet.com> X-Trace: UmFuZG9tSVaYycfoQy4MIWtkHgBktReu7I3EE485Ut0vJw7jPpDB1EKVY/C1/KtW X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 7 Sep 2001 13:16:31 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!7562!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-255-184 Xref: chonsp.franklin.ch alt.folklore.computers:89843 In article <9n8c0n$fdm$2@plutonium.btinternet.com>, nailed_barnacle@NOSPAMhotmail.com (Neil Barnes) wrote: >jmfbahciv@aol.com wrote in <9n7lmv$8p$5@bob.news.rcn.net>: > >> In article <9n5vpo$qa9$1@news.panix.com>, >> never+mail@panics.com.invalid (Michael Roach) wrote: > >>> "It's malfunctioning!" >>> "Yes, but what a malfunction!" >> >> I seem to recall JMF and TW admiring certain failures like that. >> > >Hah. Broadcast engineer, new base, standard line-up >of vision mixer (this is >before they were all digital). Observe fault in FX unit. Fix it. Two hours >later, a cry from the production gallery 'Who fixed that? Put that bloody >fault back!' Oh, yeah. Dependency on a bug or "how to design your way into a corner with no doors". /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Fri, 07 Sep 01 10:34:51 GMT Organization: UltraNet Communications, Inc. Lines: 36 Message-ID: <9nahnc$f38$14@bob.news.rcn.net> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> X-Trace: UmFuZG9tSVbMnFXKnF/mEto+Kwy2WS9MSn5NXHYvv1THNgZyqnuXctrUPNlq7MXM X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 7 Sep 2001 13:22:52 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!news-lei1.dfn.de!news-was.dfn.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-255-184 Xref: chonsp.franklin.ch alt.folklore.computers:89845 In article <1b66awidk1.fsf@cs.nmsu.edu>, Joe Pfeiffer wrote: >jmfbahciv@aol.com writes: >> > >> Wait a minute. Are you saying that these systems can't deal with >> more than one interrupt to a level? How bizarre. > >One of the PC's screwups was that it was effectively impossible for it >to deal with more than one interrupt at a level (it used assert-high >interrupt requests). But couldn't the code sort those out? When you've got time, maybe you ought to take a look at our OS code. > >The Mac scheme was worse: it could only generate three interrupts in >the first place, and any time a device requested an interrupt it could >come in at four different levels depending on what else was >interrupting at the same time. But how the hell can they make the interrupt system heirarchal(sp?). The reason a lot of thought went into assigning devices to interrupt levels was to make some interrupts more important than others. So if a PI0 interrupt comes in while the monitor is dealing with something at PI7, then the PI0 takes precedence over the PI7. (I think I've got the numbers right.) > >Bizarre is being kind. I withdraw my comment and declare shock. ;-) /BAH Subtract a hundred and four for e-mail. ###### Newsgroups: comp.arch,alt.folklore.computers From: Mike Schaeffer X-X-Sender: Subject: Re: Minimalist design (was Re: Parity - why even or odd) In-Reply-To: <6uy9nsqhd7.fsf@chonsp.franklin.ch> Message-ID: References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Lines: 14 NNTP-Posting-Date: Thu, 06 Sep 2001 17:02:01 CDT Organization: Giganews.Com - Premium News Outsourcing X-Trace: sv3-WJeGWj6UTxQhEg8PmRJlWDM6MOOgGah/N/OawdaZlXcSKya1w+mWvuj1XgTsnNPmDLjUm6tqQI58Wo/!F46wKQ8Uwm9RMmTrE1dZtuHc/QzdDusB9RoguuBlA+AS2hvpzBUrfc1CwO8+XvHKGQF28ufubPCZ!oIErpsV7NmigCNgo1pNzy03n X-Complaints-To: abuse@GigaNews.Com X-DMCA-Notifications: http://www.giganews.com/info/dmca.html X-Abuse-Info: Please be sure to forward a copy of ALL headers X-Abuse-Info: Otherwise we will be unable to process your complaint properly Date: Thu, 06 Sep 2001 22:02:01 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeeds.belnet.be!news.belnet.be!border1.nntp.aus1.giganews.com!nntp.giganews.com!nntp3.aus1.giganews.com!bin1.nnrp.aus1.giganews.com.POSTED!hagbard.io.com!mschaef Xref: chonsp.franklin.ch alt.folklore.computers:89891 On 6 Sep 2001, Neil Franklin wrote: > With level-triggered the line is simply an wire-OR of all interrupt > requests and keeps on interruptiong until it falls to zero as soon > as everyone is satisfied. This sharing is possible. > > Worst is that the PIC chip can handle both types. The PC bus wiring > simply demands edge (it lacks the resistors needed to pull wire-OR > to the "no one" state). So, if I understand your logic correctly, eight resistors, 20 years ago, would have made interrupts sharable??? -Mike ###### Sender: eric@ruckus.brouhaha.com From: Eric Smith Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> Organization: Eric Conspiracy Secret Labs X-Eric-Conspiracy: There is no conspiracy. Date: 06 Sep 2001 15:35:49 -0700 Message-ID: Lines: 22 User-Agent: Gnus/5.0807 (Gnus v5.8.7) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii NNTP-Posting-Host: ruckus.brouhaha.com X-Trace: 6 Sep 2001 15:48:53 -0700, ruckus.brouhaha.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!1531214!news.imp.ch!uni-erlangen.de!news-nue1.dfn.de!news-lei1.dfn.de!news-was.dfn.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed.stanford.edu!news.kjsl.com!news.spies.com!ruckus.brouhaha.com Xref: chonsp.franklin.ch alt.folklore.computers:89869 Joe Pfeiffer writes: > The Mac scheme was worse: it could only generate three interrupts in > the first place, and any time a device requested an interrupt it could > come in at four different levels depending on what else was > interrupting at the same time. > > Bizarre is being kind. Bizarre, yet, but I dispute the characterization as "worse". Since it was a bounded box, with no possibility for the addition of more interrupt sources, and with the interrupt handlers carefully programmed to understand the hardware, it was an entirely reasonable design. Part of the art of elegant, minimal design is knowing when NOT to add more stuff to increase the "theoretical elegance" without actually resulting in any practical improvement to the product. Of course, it is MUCH easier to build a closed hardware box and say that nothing will be added to it later (since you haven't designed in expansion connectors). Doing the same thing in software is much more prone to being a mistake, since software is generally much more malleable. ###### From: Joe Pfeiffer Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 06 Sep 2001 17:01:06 -0600 Organization: NMSU Computer Science Lines: 25 Message-ID: <1bzo87q5zh.fsf@cs.nmsu.edu> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> NNTP-Posting-Host: viper.cs.nmsu.edu X-Trace: bubba.NMSU.Edu 999817265 21544 128.123.64.113 (6 Sep 2001 23:01:05 GMT) X-Complaints-To: usenet@bubba.NMSU.Edu NNTP-Posting-Date: 6 Sep 2001 23:01:05 GMT X-Newsreader: Gnus v5.7/Emacs 20.5 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.mathworks.com!portc03.blue.aol.com!newsfeed.skycache.com.MISMATCH!newsfeed1.cidera.com!Cidera!dca6-feed2.news.digex.net!lax2-feed1.news.digex.net!intermedia!lynx.unm.edu!news.NMSU.Edu!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:89824 Neil Franklin writes: > > With level-triggered the line is simply an wire-OR of all interrupt > requests and keeps on interruptiong until it falls to zero as soon > as everyone is satisfied. This sharing is possible. > > Worst is that the PIC chip can handle both types. The PC bus wiring > simply demands edge (it lacks the resistors needed to pull wire-OR > to the "no one" state). While I can imagine a tri-state TTL wire-AND (that's what my digital design professor called it)(implementation that does assert-high, I've sure never seen it. In the case of the PC, a device's interrupt line was Hi-Z when it wasn't open; when opened, it would pull down. When it requested an interrupt, it would pull up. This is a problem with sharing interrupts that goes a whole lot deeper than mere edge- vs. level-triggered! Going to level-triggered could have been done in software. -- Joseph J. Pfeiffer, Jr., Ph.D. Phone -- (505) 646-1605 Department of Computer Science FAX -- (505) 646-1002 New Mexico State University http://www.cs.nmsu.edu/~pfeiffer SWNMRSEF: http://www.nmsu.edu/~scifair ###### From: rhn@nicholson.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 7 Sep 2001 02:50:30 GMT Organization: a2i network Lines: 13 Message-ID: <9n9clm$7ic$1@samba.rahul.net> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> NNTP-Posting-Host: waltz.rahul.net NNTP-Posting-User: rhn Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.mathworks.com!news.voicenet.com!news-out.spamkiller.net!propagator-la!news-in-la.newsfeeds.com!news-in.superfeed.net!feedwest.news.agis.net!us.telia.net!news.mainstreet.net!bug.rahul.net!samba.rahul.net!rahul.net!a2i!rhn.a2i!rhn Xref: chonsp.franklin.ch alt.folklore.computers:89830 In article <20010905081118.0a4e2a04.steveo@eircom.net>, Steve O'Hara-Smith wrote: > OK, but using two out of three interrupts up on the mouse and not >having one for the floppy - that's over the edge IMHO. Great UI design. The mouse tracked better on a '84 vintage Mac than on a contemporary 1+ GHz windows boxen. IMHO. YMMV. -- Ron Nicholson rhn@nicholson.com http://www.nicholson.com/rhn/ #include // only my own opinions, etc. ###### Message-ID: <3B986CCE.4852151B@beagle-ears.com> Date: Thu, 06 Sep 2001 23:44:30 -0700 From: Lars Poulsen X-Mailer: Mozilla 4.75 [en]C-CCK-MCD (Win98; U) X-Accept-Language: da,en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <1bzo87q5zh.fsf@cs.nmsu.edu> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 31 NNTP-Posting-Host: 207.154.106.6 X-Trace: azure.impulse.net 999845002 192 207.154.106.6 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!nntp-out.monmouth.com!newspeer.monmouth.com!newsfeed.wirehub.nl!sienna.impulse.net!azure.impulse.net!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:89832 Neil Franklin writes: > > With level-triggered the line is simply an wire-OR of all interrupt > > requests and keeps on interruptiong until it falls to zero as soon > > as everyone is satisfied. This sharing is possible. > > Worst is that the PIC chip can handle both types. The PC bus wiring > > simply demands edge (it lacks the resistors needed to pull wire-OR > > to the "no one" state). Joe Pfeiffer wrote: > In the case of the PC, a device's interrupt line was Hi-Z when it > wasn't open; when opened, it would pull down. When it requested > an interrupt, it would pull up. This is a problem with sharing > interrupts that goes a whole lot deeper than mere edge- vs. level- > triggered! Going to level-triggered could have been done in > software. There is a bit of oversimplification happening here. Yes, the (8051?) PIC can do either edge- or level-triggered inter- rupts. The bus designer specified edge-triggered. Consequently, many of the board designers designed the boards to signal inter- rupts by PULSING the IRQ line instead of HOLDING it. Bad ... because now reprogramming the PIC to level-trigger will cause all sorts of problems. Bad design decision. Probably, they did this because of the one good thing about edge- triggered interrupts: An interrupt from an unknown device is harm- less: You get a single spurious interrupt, and after you dismiss it, it will not bother you again. -- / Lars Poulsen - http://www.beagle-ears.com - lars@beagle-ears.com 125 South Ontare Road, Santa Barbara, CA 93105 - +1-805-569-5277 ###### From: Jan Vorbrueggen Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 07 Sep 2001 09:18:33 +0200 Organization: Institut fuer Neuroinformatik, Ruhr-Universitaet Bochum, Germany Lines: 12 Message-ID: References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <9n9clm$7ic$1@samba.rahul.net> NNTP-Posting-Host: luda.neuroinformatik.ruhr-uni-bochum.de X-Trace: sunu789.rz.ruhr-uni-bochum.de 999847114 14865 134.147.176.178 (7 Sep 2001 07:18:34 GMT) NNTP-Posting-Date: 7 Sep 2001 07:18:34 GMT X-Newsreader: Gnus v5.7/Emacs 20.5 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!uni-erlangen.de!news-nue1.dfn.de!news-han1.dfn.de!news-koe1.dfn.de!news.ruhr-uni-bochum.de!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:89884 rhn@nicholson.com writes: > > OK, but using two out of three interrupts up on the mouse and not > >having one for the floppy - that's over the edge IMHO. > Great UI design. The mouse tracked better on a '84 vintage Mac than on > a contemporary 1+ GHz windows boxen. I'm pretty sure that's a question of OS design, not hardware or UI design (the latter probably to a small extent). A 68030 running NeXTStep was doing much better in this respect than even current processors running X11, for instance. Jan ###### Message-ID: <3B98F97A.6D1714FD@ev1.net> From: Charles Richmond Reply-To: richmond@ev1.net Organization: Cannine Computer Center X-Mailer: Mozilla 4.7C-CCK-MCD {C-UDP; EBM-APPLE} (Macintosh; I; PPC) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9n5vpo$qa9$1@news.panix.com> <9n7lmv$8p$5@bob.news.rcn.net> <9n8e1c$ohr$3@news.panix.com> <9nahd3$f38$13@bob.news.rcn.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 25 Date: Fri, 07 Sep 2001 14:47:06 GMT NNTP-Posting-Host: 24.179.111.125 X-Complaints-To: abuse@home.net X-Trace: news2.rdc2.tx.home.com 999874026 24.179.111.125 (Fri, 07 Sep 2001 07:47:06 PDT) NNTP-Posting-Date: Fri, 07 Sep 2001 07:47:06 PDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed.direct.ca!look.ca!newshub2.rdc1.sfba.home.com!news.home.com!news2.rdc2.tx.home.com.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:89903 jmfbahciv@aol.com wrote: > > In article <9n8e1c$ohr$3@news.panix.com>, > never+mail@panics.com.invalid (Michael Roach) wrote: > > > >In article <9n7lmv$8p$5@bob.news.rcn.net>, wrote: > >>In article <9n5vpo$qa9$1@news.panix.com>, > >> never+mail@panics.com.invalid (Michael Roach) wrote: > >>>"It's malfunctioning!" > >>>"Yes, but what a malfunction!" > >> > >>I seem to recall JMF and TW admiring certain failures like that. > > > >I remember a card punch that would occasionally send a stream of cards > >in a nice pretty arc right into the waste basket 8^) > > Neat! I've never seen that! > Perhaps it was the punch's comment on someone's code that was being punched... -- +-------------------------------------------------------------+ | Charles and Francis Richmond | +-------------------------------------------------------------+ ###### Path: chonsp.franklin.ch!not-for-mail From: Neil Franklin Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 07 Sep 2001 22:08:44 +0200 Organization: My own Private Self Lines: 67 Message-ID: <6upu92ix0z.fsf@chonsp.franklin.ch> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <1bzo87q5zh.fsf@cs.nmsu.edu> <3B986CCE.4852151B@beagle-ears.com> NNTP-Posting-Host: chonsp.franklin.ch X-Trace: chonsp.franklin.ch 999893325 865 10.0.3.2 (7 Sep 2001 20:08:45 GMT) X-Complaints-To: news@chonsp.franklin.ch NNTP-Posting-Date: 7 Sep 2001 20:08:45 GMT X-Newsreader: Gnus v5.7/Emacs 20.4 Xref: chonsp.franklin.ch alt.folklore.computers:89916 Lars Poulsen writes: > Neil Franklin writes: > > > With level-triggered the line is simply an wire-OR of all interrupt > > > requests and keeps on interruptiong until it falls to zero as soon > > Joe Pfeiffer wrote: > > While I can imagine a tri-state TTL wire-AND (that's what my digital > > design professor called it) He seems to have been confused. Here the details on wiring: Wire-OR and Wire-AND happen when you have an resistor holding the output to one state and then one transistor per gate pulling to the other state. You can hook multiple outs to one in. If any or multiple are asserted the line reacts. Used for shared interrupts. Totem-Pole is when each gate has 2 transistors, one for each state, one of these allways active. You can only hook one out to an in (or multiple ins), else you get dead transistors as soon as the outs want different states. Used for non-shared interrupts. Tri-State is a variant of Totem-Pole where an additional tri-state control line switches _both_ transistors off (=Hi-Z). You can have multiple outs on one in, so long _never_ more than one control line is active at one time. Requires some form of decoder to drive the control lines. Used for data buses to select driving device. > > In the case of the PC, a device's interrupt line was Hi-Z when it > > wasn't open; when opened, it would pull down. When it requested > > an interrupt, it would pull up. That text looks self contradicting. Hi-Z and 3 states describes an tri-state system. But interrupts have no controller that goes selecting who can now drive the line. The PIC recieves multiple lines passively and then selects who the CPU shall service. So they can not be a tri-state system. > > interrupts that goes a whole lot deeper than mere edge- vs. level- > > triggered! Going to level-triggered could have been done in > > software. > > Yes, the (8051?) PIC 8259 PIC. 8051 is an 8bit microcontroller. > rupts. The bus designer specified edge-triggered. Consequently, > many of the board designers designed the boards to signal inter- > rupts by PULSING the IRQ line instead of HOLDING it. Bad ... Definitely bad. > because now reprogramming the PIC to level-trigger will cause > all sorts of problems. Bad design decision. Such as lost interrupts. -- Neil Franklin, neil@franklin.ch.remove http://neil.franklin.ch/ Hacker, Unix Guru, El Eng HTL/BSc, Sysadmin, Archer, Roleplayer - Intellectual Property is Intellectual Robbery ###### From: never+mail@panics.com.invalid (Michael Roach) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 7 Sep 2001 19:15:39 GMT Organization: A small notepad underneath my in box Lines: 22 Message-ID: <9nb6cr$t81$1@news.panix.com> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9n7lmv$8p$5@bob.news.rcn.net> <9n8e1c$ohr$3@news.panix.com> <9nahd3$f38$13@bob.news.rcn.net> NNTP-Posting-Host: panix3.panix.com X-Trace: news.panix.com 999890139 29953 166.84.1.3 (7 Sep 2001 19:15:39 GMT) X-Complaints-To: abuse@panix.com NNTP-Posting-Date: 7 Sep 2001 19:15:39 GMT X-Newsreader: trn 4.0-test74 (May 26, 2000) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!newsfeed01.sul.t-online.de!t-online.de!news.csl-gmbh.net!news.stealth.net!jfk3-feed1.news.digex.net!dca6-feed2.news.digex.net!intermedia!newsfeed1.cidera.com!Cidera!cpk-news-hub1.bbnplanet.com!nycmny1-snh1.gtei.net!news.gtei.net!panix!news.panix.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:89917 In article <9nahd3$f38$13@bob.news.rcn.net>, wrote: >In article <9n8e1c$ohr$3@news.panix.com>, > never+mail@panics.com.invalid (Michael Roach) wrote: >> >>In article <9n7lmv$8p$5@bob.news.rcn.net>, wrote: >>>In article <9n5vpo$qa9$1@news.panix.com>, >>> never+mail@panics.com.invalid (Michael Roach) wrote: >>>>"It's malfunctioning!" >>>>"Yes, but what a malfunction!" >>> >>>I seem to recall JMF and TW admiring certain failures like that. >> >>I remember a card punch that would occasionally send a stream of cards >>in a nice pretty arc right into the waste basket 8^) > >Neat! I've never seen that! It was a Honeywell cardpunch, I don't know which model. -- According to the latest official figures, 43% of all statistics are totally worthless. ###### From: nmm1@cus.cam.ac.uk (Nick Maclaren) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 7 Sep 2001 19:39:29 GMT Organization: University of Cambridge, England Lines: 36 Message-ID: <9nb7ph$d7g$1@pegasus.csx.cam.ac.uk> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9n8e1c$ohr$3@news.panix.com> <9nahd3$f38$13@bob.news.rcn.net> <9nb6cr$t81$1@news.panix.com> NNTP-Posting-Host: draco.cus.cam.ac.uk Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!53168!news.imp.ch!fu-berlin.de!server1.netnews.ja.net!pegasus.csx.cam.ac.uk!nmm1 Xref: chonsp.franklin.ch alt.folklore.computers:89922 In article <9nb6cr$t81$1@news.panix.com>, Michael Roach wrote: > >In article <9nahd3$f38$13@bob.news.rcn.net>, wrote: >>In article <9n8e1c$ohr$3@news.panix.com>, >> never+mail@panics.com.invalid (Michael Roach) wrote: >>> >>>In article <9n7lmv$8p$5@bob.news.rcn.net>, wrote: >>>>In article <9n5vpo$qa9$1@news.panix.com>, >>>> never+mail@panics.com.invalid (Michael Roach) wrote: >>>>>"It's malfunctioning!" >>>>>"Yes, but what a malfunction!" >>>> >>>>I seem to recall JMF and TW admiring certain failures like that. >>> >>>I remember a card punch that would occasionally send a stream of cards >>>in a nice pretty arc right into the waste basket 8^) >> >>Neat! I've never seen that! > >It was a Honeywell cardpunch, I don't know which model. There was an ICL card reader that did much better - it sent a 3' stack across the room, because someone left the weight off the output hopper. That was the one that had the built in shuffle and deal feature, and used to have 2-card wrecks, where the first few were turned back into wood pulp. But it was faster than anything IBM was marketing :-) Regards, Nick Maclaren, University of Cambridge Computing Service, New Museums Site, Pembroke Street, Cambridge CB2 3QG, England. Email: nmm1@cam.ac.uk Tel.: +44 1223 334761 Fax: +44 1223 334679 ###### From: mwilson@the-wire.com (Mel Wilson) Newsgroups: alt.folklore.computers Subject: Re: Minimalist design Message-ID: References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> Lines: 15 X-Newsreader: VSoup v1.2.9.37Beta [95/NT] Date: Sun, 09 Sep 2001 08:32:41 -0400 NNTP-Posting-Host: 205.206.39.64 X-Trace: nnrp1.uunet.ca 1000044208 205.206.39.64 (Sun, 09 Sep 2001 10:03:28 EDT) NNTP-Posting-Date: Sun, 09 Sep 2001 10:03:28 EDT Organization: UUNET Canada News Reader Service Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeeds.belnet.be!news.belnet.be!feed2.onemain.com!feed1.onemain.com!feeder.qis.net!sunqbc.risq.qc.ca!news.uunet.ca!nnrp1.uunet.ca.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:89991 In article <9nd4ic$i2m$9@bob.news.rcn.net>, jmfbahciv@aol.com wrote: >In article <6uy9nsqhd7.fsf@chonsp.franklin.ch>, > Neil Franklin wrote: >>With edge-triggered The first device send an edge that sets an >>Flip-Flop. A hypothetical second device would have no effect and >>so be missed. >Wouldn't the second device cause a Flop-Flip? Or did I just >demonstrate how ignorant I am about how all of that hardware >works? I think the hardware-correct word you're thinking of is 'toggle'. When you set a flip-flop that's set it stays set. Regards. Mel. ###### From: mwilson@the-wire.com (Mel Wilson) Newsgroups: alt.folklore.computers Subject: Re: Minimalist design Message-ID: References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <9nahnc$f38$14@bob.news.rcn.net> Lines: 10 X-Newsreader: VSoup v1.2.9.37Beta [95/NT] Date: Sun, 09 Sep 2001 08:37:00 -0400 NNTP-Posting-Host: 205.206.39.64 X-Trace: nnrp1.uunet.ca 1000044209 205.206.39.64 (Sun, 09 Sep 2001 10:03:29 EDT) NNTP-Posting-Date: Sun, 09 Sep 2001 10:03:29 EDT Organization: UUNET Canada News Reader Service Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!newsfeed01.sul.t-online.de!t-online.de!feed.news.nacamar.de!feed2.news.rcn.net!rcn!cpk-news-hub1.bbnplanet.com!news.gtei.net!sunqbc.risq.qc.ca!news.uunet.ca!nnrp1.uunet.ca.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:89994 In article , "no-spam" wrote: >I can still buy stock in Intel, IBM, Microsoft, Apple. How's DEC stock >doing these days ? Irrelevant. See George "Adam Smith" Goodmans's books for demonstrations of how a great company can be a lousy stock pick. Regards. Mel. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Sat, 08 Sep 01 10:02:43 GMT Organization: UltraNet Communications, Inc. Lines: 29 Message-ID: <9nd47e$i2m$7@bob.news.rcn.net> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9n5vpo$qa9$1@news.panix.com> <9n7lmv$8p$5@bob.news.rcn.net> <9n8e1c$ohr$3@news.panix.com> <9nahd3$f38$13@bob.news.rcn.net> <3B98F97A.6D1714FD@ev1.net> X-Trace: UmFuZG9tSVa5+7l4k2bUiW5WVGm5Q/SEBZTIoJtLRfOZoFWZghBY2DMjf9joYI5C X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 8 Sep 2001 12:50:54 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!2282808!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-189 Xref: chonsp.franklin.ch alt.folklore.computers:89958 In article <3B98F97A.6D1714FD@ev1.net>, Charles Richmond wrote: >jmfbahciv@aol.com wrote: >> >> In article <9n8e1c$ohr$3@news.panix.com>, >> never+mail@panics.com.invalid (Michael Roach) wrote: >> > >> >In article <9n7lmv$8p$5@bob.news.rcn.net>, wrote: >> >>In article <9n5vpo$qa9$1@news.panix.com>, >> >> never+mail@panics.com.invalid (Michael Roach) wrote: >> >>>"It's malfunctioning!" >> >>>"Yes, but what a malfunction!" >> >> >> >>I seem to recall JMF and TW admiring certain failures like that. >> > >> >I remember a card punch that would occasionally send a stream of cards >> >in a nice pretty arc right into the waste basket 8^) >> >> Neat! I've never seen that! >> >Perhaps it was the punch's comment on someone's code that was >being punched... > Yup. Puts a whole new slant on the meaning of bit bucket. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Sat, 08 Sep 01 10:05:22 GMT Organization: UltraNet Communications, Inc. Lines: 43 Message-ID: <9nd4cc$i2m$8@bob.news.rcn.net> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9n8e1c$ohr$3@news.panix.com> <9nahd3$f38$13@bob.news.rcn.net> <9nb6cr$t81$1@news.panix.com> <9nb7ph$d7g$1@pegasus.csx.cam.ac.uk> X-Trace: UmFuZG9tSVYqpGJUGZZelt4+xxzWEWAZ8RvGArgRlxdM5yCA5g8kIlmmi65iSrP/ X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 8 Sep 2001 12:53:32 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!223936!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-189 Xref: chonsp.franklin.ch alt.folklore.computers:89961 In article <9nb7ph$d7g$1@pegasus.csx.cam.ac.uk>, nmm1@cus.cam.ac.uk (Nick Maclaren) wrote: >In article <9nb6cr$t81$1@news.panix.com>, >Michael Roach wrote: >> >>In article <9nahd3$f38$13@bob.news.rcn.net>, wrote: >>>In article <9n8e1c$ohr$3@news.panix.com>, >>> never+mail@panics.com.invalid (Michael Roach) wrote: >>>> >>>>In article <9n7lmv$8p$5@bob.news.rcn.net>, wrote: >>>>>In article <9n5vpo$qa9$1@news.panix.com>, >>>>> never+mail@panics.com.invalid (Michael Roach) wrote: >>>>>>"It's malfunctioning!" >>>>>>"Yes, but what a malfunction!" >>>>> >>>>>I seem to recall JMF and TW admiring certain failures like that. >>>> >>>>I remember a card punch that would occasionally send a stream of cards >>>>in a nice pretty arc right into the waste basket 8^) >>> >>>Neat! I've never seen that! >> >>It was a Honeywell cardpunch, I don't know which model. > >There was an ICL card reader that did much better - it sent a 3' >stack across the room, because someone left the weight off the >output hopper. That's preferable than one card at a time. If the whole stack went, reconstructing the card order would be a snap. ;-) > ...That was the one that had the built in shuffle >and deal feature, and used to have 2-card wrecks, where the first >few were turned back into wood pulp. But it was faster than >anything IBM was marketing :-) Hmm....[emoticon puts on its slogan hat and types] Our readers munch faster. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Sat, 08 Sep 01 10:08:33 GMT Organization: UltraNet Communications, Inc. Lines: 39 Message-ID: <9nd4ic$i2m$9@bob.news.rcn.net> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> X-Trace: UmFuZG9tSVYUwFnHaeDcK5Z37i4tNBAiFg0QENj4gFJn7cj3/S+IIrELcmOKUqXY X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 8 Sep 2001 12:56:44 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-189 Xref: chonsp.franklin.ch alt.folklore.computers:89964 In article <6uy9nsqhd7.fsf@chonsp.franklin.ch>, Neil Franklin wrote: >Joe Pfeiffer writes: > >> jmfbahciv@aol.com writes: >> > > >> > Wait a minute. Are you saying that these systems can't deal with >> > more than one interrupt to a level? How bizarre. >> >> One of the PC's screwups was that it was effectively impossible for it >> to deal with more than one interrupt at a level (it used assert-high >> interrupt requests). > >Assert-high was not the problem (it is bad for catching interference). >The problem was that the PC uses edge-triggered and not >level-triggered interrupts. > >With edge-triggered The first device send an edge that sets an >Flip-Flop. A hypothetical second device would have no effect and >so be missed. Wouldn't the second device cause a Flop-Flip? Or did I just demonstrate how ignorant I am about how all of that hardware works? > >With level-triggered the line is simply an wire-OR of all interrupt >requests and keeps on interruptiong until it falls to zero as soon >as everyone is satisfied. This sharing is possible. But interruptions won't ever fall to zero...the clock will continue. > /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Sun, 09 Sep 01 09:35:51 GMT Organization: UltraNet Communications, Inc. Lines: 38 Message-ID: <9nfn1c$cse$4@bob.news.rcn.net> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <9nahnc$f38$14@bob.news.rcn.net> X-Trace: UmFuZG9tSVb2SiSy0WvPT+O+CGPWxkm7TxCDCMygO3pJT6hkABSSzC5lWavcKKSz X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 9 Sep 2001 12:24:12 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.mathworks.com!news.voicenet.com!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-219 Xref: chonsp.franklin.ch alt.folklore.computers:89975 In article , "no-spam" wrote: >/BAH --- > >Somehow it seems that the Mac and IBM PCs were just a >bit more successful in >the marketplace than the PDP-10 systems. They must have done something >right for that to be the case. How long did the PDP-10 architecture and >software survive and remain a viable offering ? My DECUS token has it beginning in 1964 (including the PDP-6). PDP-10 CPUs are still getting manufactured and sold under another product id. There is also the opportunity of running a TOPS-10 or TOPS-20 operating system on top of a choice of emulators which run on top of Unix..and simultaneously, too! There have been verified rumors that the software functions on top of Windows. TW and JMF are either laughing their asses off or spinning in their graves. > ... How much revenue did it >generate ? We never counted. The books didn't attribute PDP-10 income to the -10 product line. Most of it "counted" against other product lines. This was one of the rationales to cancel the -10. > >I can still buy stock in Intel, IBM, Microsoft, Apple. How's DEC stock >doing these days ? > >Case closed. Quality is not determined by stock price. You need to learn a little bit more about life in the bit lanes. /BAH ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Sun, 09 Sep 01 09:28:11 GMT Organization: UltraNet Communications, Inc. Lines: 37 Message-ID: <9nfmj0$cse$3@bob.news.rcn.net> References: <1bitex1pfy.fsf@cs.nmsu.edu> <9n5vpo$qa9$1@news.panix.com> <9n7lmv$8p$5@bob.news.rcn.net> <9n8c0n$fdm$2@plutonium.btinternet.com> <9nahbf$f38$12@bob.news.rcn.net> <930.650T165T7153626@nowhere.in.particular> X-Trace: UmFuZG9tSVZ+rJxprQW+2hQ8Ef9ly/SscZEzzfGqlf0d2OEQaE9N7iW0L6VEBmA1 X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 9 Sep 2001 12:16:32 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!39618283!news.imp.ch!sunqbc.risq.qc.ca!cpk-news-hub1.bbnplanet.com!news.gtei.net!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-219 Xref: chonsp.franklin.ch alt.folklore.computers:89977 In article <930.650T165T7153626@nowhere.in.particular>, "Charlie Gibbs" wrote: >In article <9nahbf$f38$12@bob.news.rcn.net> jmfbahciv@aol.com >(jmfbahciv) writes: > >>In article <9n8c0n$fdm$2@plutonium.btinternet.com>, >>nailed_barnacle@NOSPAMhotmail.com (Neil Barnes) wrote: >> >>>Hah. Broadcast engineer, new base, standard line-up of vision mixer >>>(this is before they were all digital). Observe fault in FX unit. >>>Fix it. Two hours later, a cry from the production gallery 'Who >>>fixed that? Put that bloody fault back!' >> >> Oh, yeah. Dependency on a bug or "how to design your way >>into a corner with no doors". > >This is a practice which has been enshrined in the Wintel world - >both in hardware and software - from the beginning. It shows. :-) yea. No plans beyond tomorrow's ship. We spent tons of time just thinking about future and past ship implications. > >I've heard it described as making future releases "bug-compatible", >and it just proves the old saying: "If at first you don't succeed, >you might as well forget it." Oh, but Misoft has had to drop the bug-compatible constraint. They're so pathetic that they can't think of new bugs, so they have to resort to recycling the old bugs that had been "fixed" with the next release. /BAH Subtract a hundred and four for e-mail. ###### From: "Charlie Gibbs" Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 07 Sep 01 11:55:03 -0800 Organization: http://extra.newsguy.com Lines: 25 Message-ID: <930.650T165T7153626@nowhere.in.particular> References: <1bitex1pfy.fsf@cs.nmsu.edu> <9n5vpo$qa9$1@news.panix.com> <9n7lmv$8p$5@bob.news.rcn.net> <9n8c0n$fdm$2@plutonium.btinternet.com> <9nahbf$f38$12@bob.news.rcn.net> NNTP-Posting-Host: p-383.newsdawg.com X-Newsreader: THOR 2.5a (Amiga;TCP/IP) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.mathworks.com!pln-e!spln!dex!extra.newsguy.com!newsp.newsguy.com!news1 Xref: chonsp.franklin.ch alt.folklore.computers:90035 In article <9nahbf$f38$12@bob.news.rcn.net> jmfbahciv@aol.com (jmfbahciv) writes: >In article <9n8c0n$fdm$2@plutonium.btinternet.com>, >nailed_barnacle@NOSPAMhotmail.com (Neil Barnes) wrote: > >>Hah. Broadcast engineer, new base, standard line-up of vision mixer >>(this is before they were all digital). Observe fault in FX unit. >>Fix it. Two hours later, a cry from the production gallery 'Who >>fixed that? Put that bloody fault back!' > > Oh, yeah. Dependency on a bug or "how to design your way >into a corner with no doors". This is a practice which has been enshrined in the Wintel world - both in hardware and software - from the beginning. It shows. :-) I've heard it described as making future releases "bug-compatible", and it just proves the old saying: "If at first you don't succeed, you might as well forget it." -- cgibbs@nowhere.in.particular (Charlie Gibbs) I'm switching ISPs - watch this space. ###### From: Joe Pfeiffer Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 07 Sep 2001 17:04:04 -0600 Organization: NMSU Computer Science Lines: 47 Message-ID: <1bofomppqz.fsf@cs.nmsu.edu> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <1bzo87q5zh.fsf@cs.nmsu.edu> <3B986CCE.4852151B@beagle-ears.com> <6upu92ix0z.fsf@chonsp.franklin.ch> NNTP-Posting-Host: viper.cs.nmsu.edu X-Trace: bubba.NMSU.Edu 999903842 15974 128.123.64.113 (7 Sep 2001 23:04:02 GMT) X-Complaints-To: usenet@bubba.NMSU.Edu NNTP-Posting-Date: 7 Sep 2001 23:04:02 GMT X-Newsreader: Gnus v5.7/Emacs 20.5 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!newsfeed01.sul.t-online.de!t-online.de!newsfeed.online.be!logbridge.uoregon.edu!lax2-feed1.news.digex.net!intermedia!lynx.unm.edu!news.NMSU.Edu!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:89972 Neil Franklin writes: > Lars Poulsen writes: > > > Neil Franklin writes: > > > > With level-triggered the line is simply an wire-OR of all interrupt > > > > requests and keeps on interruptiong until it falls to zero as soon > > > > Joe Pfeiffer wrote: > > > While I can imagine a tri-state TTL wire-AND (that's what my digital > > > design professor called it) > > He seems to have been confused. Here the details on wiring: > > Wire-OR and Wire-AND happen when you have an resistor holding the > output to one state and then one transistor per gate pulling to the > other state. You can hook multiple outs to one in. If any or multiple > are asserted the line reacts. Used for shared interrupts. The only TTL variant I've ever seen has been open-collector, using a passive pullup and an active pulldown. This is wire-AND, since if any of the gates pull down the result is 0. I've frequently seen it called wire-OR, since it's always used as an OR in assert-low logic. > > > > In the case of the PC, a device's interrupt line was Hi-Z when it > > > wasn't open; when opened, it would pull down. When it requested > > > an interrupt, it would pull up. > > That text looks self contradicting. Probably because I was using the software term in a hardware description. When the software opens the device, the device leaves the Hi-Z state, and (if it not requesting an interrupt at the moment) pulls down. When it does want an interrupt, it pulls up. > Hi-Z and 3 states describes an tri-state system. But interrupts have > no controller that goes selecting who can now drive the line. The PIC > recieves multiple lines passively and then selects who the CPU shall > service. So they can not be a tri-state system. Remember, this design doesn't share interrupts. It doesn't use tri-state in any sane way, but it does use them... -- Joseph J. Pfeiffer, Jr., Ph.D. Phone -- (505) 646-1605 Department of Computer Science FAX -- (505) 646-1002 New Mexico State University http://www.cs.nmsu.edu/~pfeiffer SWNMRSEF: http://www.nmsu.edu/~scifair ###### Sender: eric@ruckus.brouhaha.com From: Eric Smith Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <9nai45$f38$15@bob.news.rcn.net> Organization: Eric Conspiracy Secret Labs X-Eric-Conspiracy: There is no conspiracy. Date: 07 Sep 2001 17:21:18 -0700 Message-ID: Lines: 25 User-Agent: Gnus/5.0807 (Gnus v5.8.7) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii NNTP-Posting-Host: ruckus.brouhaha.com X-Trace: 7 Sep 2001 17:34:35 -0700, ruckus.brouhaha.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!newsfeed01.sul.t-online.de!t-online.de!newsfeed.online.be!news-hog.berkeley.edu!ucberkeley!newsfeed.stanford.edu!news.kjsl.com!news.spies.com!ruckus.brouhaha.com Xref: chonsp.franklin.ch alt.folklore.computers:90043 I wrote: > Bizarre, yet, but I dispute the characterization as "worse". Since it > was a bounded box, with no possibility for the addition of more > interrupt sources, and with the interrupt handlers carefully programmed > to understand the hardware, it was an entirely reasonable design. jmfbahciv@aol.com writes: > But that hogties the monitor! It would be nice if the software > and hardware development can be mutually exclusive. That way > pure software improvements can be done while the hardware guys > are doing their thing. By tying a software development cycle to > a hardware development cycle boggles my mind. It just ain't > efficient. They did the interrupts differently on the expandable Macintoshes, starting with the Macintosh II. And the OS was updated to understand it. So there's still no justification for saying that there was anything "wrong" with the original design. If they'd set up the interrupts strangely, AND required that users deal with it in their software (rather than having the OS do it), that would have been attrocious. That's what the IBM PC did, and we have the legacy of it to this day. By comparison, the Mac developers were utter geniuses. ###### From: Paul DeMone Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Fri, 07 Sep 2001 20:39:03 -0400 Organization: IGS - Information Gateway Services Lines: 19 Message-ID: <3B9968A7.43086976@igs.net> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> NNTP-Posting-Host: i216-58-49-161.igs.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: news.igs.net 999909066 20970 216.58.49.161 (8 Sep 2001 00:31:06 GMT) X-Complaints-To: abuse@igs.net NNTP-Posting-Date: 8 Sep 2001 00:31:06 GMT X-Mailer: Mozilla 4.61 [en] (Win98; I) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!news0.de.colt.net!colt.net!news.maxwell.syr.edu!newsfeed.mathworks.com!wn3feed!worldnet.att.net!206.191.82.231!rockie.attcanada.net!newsfeed.attcanada.net!216.58.1.11!nntp.igs.net!news.igs.net!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90076 Eric Smith wrote: > You have illuminated some interesting points about the hardware and software design of early PCs and Macs. I appreciate your comments because I haven't heard of or given much thought to a few of the points you raised. But ease up on the quotes please. http://www.winternet.com/~mikelr/flame53.html :-) -- Paul W. DeMone The 801 experiment SPARCed an ARMs race of EPIC Kanata, Ontario proportions to put more PRECISION and POWER into demone@mosaid.com architectures with MIPSed results but ALPHA's well pdemone@igs.net that ends well. ###### From: pechter@i4got.pechter.dyndns.org (Bill Pechter) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 8 Sep 2001 08:09:49 -0400 Organization: Unknown Lines: 32 Message-ID: <9nd1qd$1q0m$1@i4got.pechter.dyndns.org> References: <1bitex1pfy.fsf@cs.nmsu.edu> <9n8c0n$fdm$2@plutonium.btinternet.com> <9nahbf$f38$12@bob.news.rcn.net> <930.650T165T7153626@nowhere.in.particular> NNTP-Posting-Host: bg-tc-ppp1323.monmouth.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.tele.dk!small.news.tele.dk!205.231.236.10!newspeer.monmouth.com!news.monmouth.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90051 In article <930.650T165T7153626@nowhere.in.particular>, Charlie Gibbs wrote: >In article <9nahbf$f38$12@bob.news.rcn.net> jmfbahciv@aol.com >(jmfbahciv) writes: > >>In article <9n8c0n$fdm$2@plutonium.btinternet.com>, >>nailed_barnacle@NOSPAMhotmail.com (Neil Barnes) wrote: >> >> Oh, yeah. Dependency on a bug or "how to design your way >>into a corner with no doors". > >This is a practice which has been enshrined in the Wintel world - >both in hardware and software - from the beginning. It shows. :-) Wintel hell... Ciscoworks won't install on Solaris unless you back out a recommended Sun security patch involving swapping a directory for a symlink because their f****ing install program relies on this mis-feature to work correctly. I'm not too surprized at the code level that would do crap like this and I know I wouldn't have done this myself... and I'm not a programmer. The damned Cisco folks have maintained this misfeature through the past two sets of CD's for Cisco works -- which is one reason why I figure they're right up there with Microsoft. Bill -- -- d|i|g|i|t|a|l had it THEN. Don't you wish you could still buy it now! bpechter@shell.monmouth.com|pechter@pechter.dyndns.org ###### From: "no-spam" Newsgroups: comp.arch,alt.folklore.computers References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <9nahnc$f38$14@bob.news.rcn.net> Subject: Re: Minimalist design (was Re: Parity - why even or odd) Lines: 54 X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 5.50.4133.2400 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400 Message-ID: Date: Sun, 09 Sep 2001 02:04:01 GMT NNTP-Posting-Host: 12.89.171.129 X-Complaints-To: abuse@worldnet.att.net X-Trace: bgtnsc05-news.ops.worldnet.att.net 1000001041 12.89.171.129 (Sun, 09 Sep 2001 02:04:01 GMT) NNTP-Posting-Date: Sun, 09 Sep 2001 02:04:01 GMT Organization: AT&T Worldnet Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!uni-erlangen.de!news-nue1.dfn.de!news-han1.dfn.de!news-koe1.dfn.de!do.de.uu.net!newsfeed.esat.net!logbridge.uoregon.edu!arclight.uoregon.edu!wn4feed!worldnet.att.net!135.173.83.71!wnfilter1!worldnet-localpost!bgtnsc05-news.ops.worldnet.att.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:89979 /BAH --- Somehow it seems that the Mac and IBM PCs were just a bit more successful in the marketplace than the PDP-10 systems. They must have done something right for that to be the case. How long did the PDP-10 architecture and software survive and remain a viable offering ? How much revenue did it generate ? I can still buy stock in Intel, IBM, Microsoft, Apple. How's DEC stock doing these days ? Case closed. ============= wrote in message news:9nahnc$f38$14@bob.news.rcn.net... > In article <1b66awidk1.fsf@cs.nmsu.edu>, > Joe Pfeiffer wrote: > >jmfbahciv@aol.com writes: > >> > > >> Wait a minute. Are you saying that these systems can't deal with > >> more than one interrupt to a level? How bizarre. > > > >One of the PC's screwups was that it was effectively impossible for it > >to deal with more than one interrupt at a level (it used assert-high > >interrupt requests). > > But couldn't the code sort those out? When you've got time, maybe > you ought to take a look at our OS code. > > > > >The Mac scheme was worse: it could only generate three interrupts in > >the first place, and any time a device requested an interrupt it could > >come in at four different levels depending on what else was > >interrupting at the same time. > > But how the hell can they make the interrupt system heirarchal(sp?). > The reason a lot of thought went into assigning devices to interrupt > levels was to make some interrupts more important than others. > So if a PI0 interrupt comes in while the monitor is dealing > with something at PI7, then the PI0 takes precedence over the PI7. > (I think I've got the numbers right.) > > > > > >Bizarre is being kind. > > I withdraw my comment and declare shock. ;-) > > /BAH > > Subtract a hundred and four for e-mail. ###### From: Elliott Roper Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Sun, 09 Sep 2001 12:20:34 +0100 Organization: Posted via Supernews, http://www.supernews.com Message-ID: <090920011220344692%elliott@yrl.co.uk> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <9nahnc$f38$14@bob.news.rcn.net> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-transfer-encoding: 8bit User-Agent: YA-NewsWatcher/5.0.1 X-Complaints-To: abuse@xo.supernews.co.uk Lines: 53 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feeder.qis.net!sn-xit-02!sn-uk-post-01!supernews.com!xo.supernews.co.uk!elliott Xref: chonsp.franklin.ch alt.folklore.computers:90001 In article , no-spam wrote: > /BAH --- > > Somehow it seems that the Mac and IBM PCs were just a bit more successful in > the marketplace than the PDP-10 systems. They must have done something > right for that to be the case. How long did the PDP-10 architecture and > software survive and remain a viable offering ? How much revenue did it > generate ? > > I can still buy stock in Intel, IBM, Microsoft, Apple. How's DEC stock > doing these days ? > > Case closed. > > ============= You jest. Digital Equipment Corp destroyed the DEC-10 as part of its early training for the mis-management Olympics. You could still buy DEC stock 20 years later. Digital (cough, spit) management were openly saying 'their product was their share price' as Robert Palmer sold the seed corn. I remember the displays on the elevator monitors at Digital's 'Enterprise House' in London. They had to clean the chunder off the floor after my visits. It is a long long line of poor logic to get from interrupt line pull-ups to short term share price 30 years after the event. There is a huge difference between good engineering and bad business. These days, the short term stoopids running funds and tech businesses wouldn't know a decently engineered product if it jumped up and bit 'em on the bum. Or a decent team of engineers either. How could they put a history degree in charge of Hewlett Packard? Back to your question. DEC-10 and TOPS-10 survive in ways you can't begin to imagine. It's up near the top of the 20th century's greatest achievements. Paul Allen and your mate BillG built on what they learned on one of them as did thousands of others who have shaped this place. What were the hosts on the original ARPANET? What did the airborne data acquistion in Concorde during its development? Hmm. Maybe a history degree is not such a bad idea. Now if someone could teach Carly distributed lock manager and word-tearing...? ###### From: lysse Subject: Re: Minimalist design (was Re: Parity - why even or odd) Newsgroups: comp.arch,alt.folklore.computers References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <9nahnc$f38$14@bob.news.rcn.net> Organization: lysse's domain Reply-To: lysse.news@blueyonder.co.uk User-Agent: tin/pre-1.4-980226 (UNIX) (Linux/2.2.6 (i486)) Message-ID: <28ifn9.6en.ln@setter.lysse.co.uk> Lines: 32 Date: Sun, 09 Sep 2001 11:32:47 GMT NNTP-Posting-Host: 62.31.9.47 X-Complaints-To: http://www.blueyonder.co.uk/abuse X-Trace: news1.cableinet.net 1000035167 62.31.9.47 (Sun, 09 Sep 2001 12:32:47 BST) NNTP-Posting-Date: Sun, 09 Sep 2001 12:32:47 BST Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!news0.de.colt.net!colt.net!newspeer.clara.net!news.clara.net!news-hub.cableinet.net!blueyonder!internal-news-hub.cableinet.net!news1.cableinet.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90022 no-spam wrote: > Somehow it seems that the Mac and IBM PCs were just a bit more successful in > the marketplace than the PDP-10 systems. They must have done something > right for that to be the case. How long did the PDP-10 architecture and > software survive and remain a viable offering ? How much revenue did it > generate ? PCs and Macs were commodity hardware, cheap to build, cheaply designed, and frankly it shows - but to most people, the fact that a computer is cheaper than a holiday these days will be what counts. PDP-10s were not commodity hardware, and I don't think I can recall seeing any room-sized architecture succeed in the supermarket. But it's worth pointing out that the PDP-10 architecture originated in 1963 with the PDP-6, and a PDP-10 is still available from XKL, and there's at least one project on the go to create a PDP-10 processor in a <$10 FPGA. Sure, the PC and Mac may have market share, but will the PC be so adored that they're being rebuilt from scratch in 2019? Will copies of MS-DOS 6.22 be lovingly preserved until the hardware exists to run it again? Somehow I think not... > I can still buy stock in Intel, IBM, Microsoft, Apple. How's DEC stock > doing these days ? > Case closed. This is ridiculous. Do you pick your favourite music by only looking at the top 5? -- lysse at lysse dot co dot uk "Why are your problems always so much bigger than everyone else's?" "Because they're mine." -- Ally McBeal ###### From: Pete Fenelon Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Sun, 09 Sep 2001 13:41:43 -0000 Organization: Posted via Supernews, http://www.supernews.com Message-ID: Sender: Pete Fenelon References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <9nahnc$f38$14@bob.news.rcn.net> <28ifn9.6en.ln@setter.lysse.co.uk> User-Agent: tin/1.5.8-20010221 ("Blue Water") (UNIX) (FreeBSD/4.3-STABLE (i386)) X-Complaints-To: newsabuse@supernews.com Lines: 13 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!sn-xit-03!sn-post-01!supernews.com!corp.supernews.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90046 In alt.folklore.computers lysse wrote: > in a <$10 FPGA. Sure, the PC and Mac may have market share, but will > the PC be so adored that they're being rebuilt from scratch in 2019? > Will copies of MS-DOS 6.22 be lovingly preserved until the hardware > exists to run it again? Somehow I think not... I have this horrible feeling that PCs that are still suffering from Intel brain damage will still be if not on sale then at least in widespread use by 2019. After all, we've had 20 years of x86 hell things already... presumably IA64 is intended for a fairly long design life, too? :) pete ###### From: pechter@i4got.pechter.dyndns.org (Bill Pechter) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 9 Sep 2001 12:55:02 -0400 Organization: Unknown Lines: 28 Message-ID: <9ng6t6$1gg$1@i4got.pechter.dyndns.org> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9nahnc$f38$14@bob.news.rcn.net> <090920011220344692%elliott@yrl.co.uk> NNTP-Posting-Host: bg-tc-ppp1669.monmouth.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!nntp-out.monmouth.com!newspeer.monmouth.com!news.monmouth.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90038 In article <090920011220344692%elliott@yrl.co.uk>, Elliott Roper wrote: > >How could they put a history degree in charge of Hewlett Packard? > >Hmm. Maybe a history degree is not such a bad idea. >Now if someone could teach Carly distributed lock manager and >word-tearing...? As a computer geek and ex-DEC Field Service guy who has a history degree... I volunteer to take the DEC operations back from Compaq/HP and run them at a profit. As far as the distributed lock manager goes it's time to move Digital Unix to a *BSD (FreeBSD preferred) or Linux base with cluster add-ons and port VMS to IA64. As far as the Alpha goes -- I'd try to get them produced at commodity prices and get them out there in numbers and prices that would compete with the Pentium III line. Bill -- -- -- d|i|g|i|t|a|l had it THEN. Don't you wish you could still buy it now! bpechter@shell.monmouth.com|pechter@pechter.dyndns.org ###### From: lysse Subject: Re: Minimalist design (was Re: Parity - why even or odd) Newsgroups: comp.arch,alt.folklore.computers References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <9nahnc$f38$14@bob.news.rcn.net> <28ifn9.6en.ln@setter.lysse.co.uk> Organization: lysse's domain Reply-To: lysse.news@blueyonder.co.uk User-Agent: tin/pre-1.4-980226 (UNIX) (Linux/2.2.6 (i486)) Message-ID: <6b5gn9.qkn.ln@setter.lysse.co.uk> Lines: 46 Date: Sun, 09 Sep 2001 17:32:47 GMT NNTP-Posting-Host: 62.31.9.47 X-Complaints-To: http://www.blueyonder.co.uk/abuse X-Trace: news1.cableinet.net 1000056767 62.31.9.47 (Sun, 09 Sep 2001 18:32:47 BST) NNTP-Posting-Date: Sun, 09 Sep 2001 18:32:47 BST Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeeds.belnet.be!news.belnet.be!colt.net!dispose.news.demon.net!demon!btnet-peer0!btnet-peer!btnet!news-hub.cableinet.net!blueyonder!internal-news-hub.cableinet.net!news1.cableinet.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90025 Pete Fenelon wrote: > I have this horrible feeling that PCs that are still suffering from Intel > brain damage will still be if not on sale then at least in widespread > use by 2019. After all, we've had 20 years of x86 hell things > already... presumably IA64 is intended for a fairly long design > life, too? :) Yes, but as I said in another thread, if M$ topples, the x86 will probably die fairly quickly afterwards. I have the IA64 instruction manual; it doesn't look too bad, but then it was co-designed by HP. Although I do wonder if HP have stiffed Intel with this one; lock a major competitor into co-designing a ridiculously overblown 64-bit architecture that's perfect for what you've been doing for the last fifteen years, safe in the knowledge that they will stuff it up by including a hardware x86 emulation, and then two years down the line release an IA64-only chip with a software emulation not unlike that of Transmeta. For about half the price. It's worth remembering that HP have done a good deal of work on dynamic retranslation in the past. They developed a translator that translated PA-RISC code into PA-RISC code; yes, the same architecture, but the code ran about 30% faster post-translation, mainly because compile-time unknowns had been resolved by that point, enabling a constant-folding orgy. Here's a reference: http://www.hpl.hp.com/cambridge/projects/Dynamo and it seems to have spawned a commercial successor: http://www.hp.com/products1/itanium/advantage/aries.html In fact, in idler moments I wonder if this won't be the future of commercial CPU design; everyone will go back to making the best architectures they can, and code will be generated to run on one of a number of common architectural platforms, which are dynamically translated into target code. It'd even be the best position for the processor manufacturers, as they wouldn't be required to implement the same instruction set for every revision. Meanwhile, the rest of us can make do with FPGAs and downloadable CPU definitions. And maybe even build our own... ;-) -- lysse at lysse dot co dot uk "Why are your problems always so much bigger than everyone else's?" "Because they're mine." -- Ally McBeal ###### From: "del cecchi" Newsgroups: comp.arch,alt.folklore.computers References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <9nahnc$f38$14@bob.news.rcn.net> <28ifn9.6en.ln@setter.lysse.co.uk> Subject: Re: Minimalist design (was Re: Parity - why even or odd) Lines: 48 X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 5.50.4133.2400 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400 Message-ID: Date: Sun, 9 Sep 2001 20:37:32 -0500 NNTP-Posting-Host: 63.38.178.127 X-Trace: eagle.america.net 1000085868 63.38.178.127 (Sun, 09 Sep 2001 21:37:48 EDT) NNTP-Posting-Date: Sun, 09 Sep 2001 21:37:48 EDT Organization: 24hoursupport.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!sunqbc.risq.qc.ca!falcon.america.net!eagle.america.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90019 "lysse" wrote in message news:28ifn9.6en.ln@setter.lysse.co.uk... > no-spam wrote: > > Somehow it seems that the Mac and IBM PCs were just a bit more successful in > > the marketplace than the PDP-10 systems. They must have done something > > right for that to be the case. How long did the PDP-10 architecture and > > software survive and remain a viable offering ? How much revenue did it > > generate ? > > PCs and Macs were commodity hardware, cheap to build, cheaply designed, > and frankly it shows - but to most people, the fact that a computer is > cheaper than a holiday these days will be what counts. > > PDP-10s were not commodity hardware, and I don't think I can recall > seeing any room-sized architecture succeed in the supermarket. But > it's worth pointing out that the PDP-10 architecture originated in > 1963 with the PDP-6, and a PDP-10 is still available from XKL, and > there's at least one project on the go to create a PDP-10 processor > in a <$10 FPGA. Sure, the PC and Mac may have market share, but will > the PC be so adored that they're being rebuilt from scratch in 2019? > Will copies of MS-DOS 6.22 be lovingly preserved until the hardware > exists to run it again? Somehow I think not... > > > I can still buy stock in Intel, IBM, Microsoft, Apple. How's DEC stock > > doing these days ? > > > Case closed. > > This is ridiculous. Do you pick your favourite music by only > looking at the top 5? > -- > lysse at lysse dot co dot uk > "Why are your problems always so much bigger than everyone else's?" > "Because they're mine." -- Ally McBeal Ask the guys in Endicott how well desk side small 370 machines went over in the market. Oops you can't. They are gone. del cecchi ###### Path: chonsp.franklin.ch!not-for-mail From: Neil Franklin Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 10 Sep 2001 14:55:10 +0200 Organization: My own Private Self Lines: 39 Message-ID: <6ubskj434h.fsf@chonsp.franklin.ch> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <9nahnc$f38$14@bob.news.rcn.net> <28ifn9.6en.ln@setter.lysse.co.uk> NNTP-Posting-Host: chonsp.franklin.ch X-Trace: chonsp.franklin.ch 1000126512 460 10.0.3.2 (10 Sep 2001 12:55:12 GMT) X-Complaints-To: news@chonsp.franklin.ch NNTP-Posting-Date: 10 Sep 2001 12:55:12 GMT X-Newsreader: Gnus v5.7/Emacs 20.4 Xref: chonsp.franklin.ch alt.folklore.computers:90100 lysse writes: > it's worth pointing out that the PDP-10 architecture originated in > 1963 with the PDP-6, and a PDP-10 is still available from XKL, and > there's at least one project on the go to create a PDP-10 processor > in a <$10 FPGA. $45 FPGA. That is what a XC2S200 costs presently. I don't think that anything $10 has enough room for an 10. Them 36bits and 512 instructions take up space. > the PC be so adored that they're being rebuilt from scratch in 2019? > Will copies of MS-DOS 6.22 be lovingly preserved until the hardware > exists to run it again? Somehow I think not... :-) Actually I expect old some PCs and OSes to be saved to then. Simply for the sentimental "I started with one of these" type reason. It may only be 0.0001%, but with 100s of millions made that will be enough to ensure immortality. > > I can still buy stock in Intel, IBM, Microsoft, Apple. How's DEC stock > > doing these days ? I can still download the PDP-10 reference manual and build one. Engineering outlives business. Just like the pyramids outlived the pharaoes. Perhaps we are the last one to have the laugh, after all. Neil "cloning PDP-10" Franklin ObURL: http://neil.franklin.ch/Projects/PDP-10/ -- Neil Franklin, neil@franklin.ch.remove http://neil.franklin.ch/ Hacker, Unix Guru, El Eng HTL/BSc, Sysadmin, Archer, Roleplayer - Intellectual Property is Intellectual Robbery ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Mon, 10 Sep 01 08:27:01 GMT Organization: UltraNet Communications, Inc. Lines: 15 Message-ID: <9ni7ck$c78$2@bob.news.rcn.net> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9nahnc$f38$14@bob.news.rcn.net> <090920011220344692%elliott@yrl.co.uk> <9ng6t6$1gg$1@i4got.pechter.dyndns.org> X-Trace: UmFuZG9tSVbyEjhHuYWwVrahDTBgEYPBIi/9rViQfYfQHqX6s5xxY6oASmE8PBnk X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 10 Sep 2001 11:15:32 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-127 Xref: chonsp.franklin.ch alt.folklore.computers:90107 In article <9ng6t6$1gg$1@i4got.pechter.dyndns.org>, pechter@i4got.pechter.dyndns.org (Bill Pechter) wrote: >As far as the Alpha goes -- I'd try to get them produced at commodity >prices and get them out there in numbers and prices that would compete >with the Pentium III line. That's something I never understood. Why didn't they do that? Buying an Alpha was essentially barred from retail. //BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Tue, 11 Sep 01 08:37:20 GMT Organization: UltraNet Communications, Inc. Lines: 55 Message-ID: <9nksc9$ast$9@bob.news.rcn.net> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9nahnc$f38$14@bob.news.rcn.net> <090920011220344692%elliott@yrl.co.uk> <9ng6t6$1gg$1@i4got.pechter.dyndns.org> <9ni7ck$c78$2@bob.news.rcn.net> <_Ccn7.24433$bl4.4553314@typhoon.ne.mediaone.net> X-Trace: UmFuZG9tSVbkaOb5JS6iEbq7UnZt+p60t/J8ts0/akCn40Tf+kV3YePMzYo5scrG X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 11 Sep 2001 11:26:01 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-156 Xref: chonsp.franklin.ch alt.folklore.computers:90165 In article <_Ccn7.24433$bl4.4553314@typhoon.ne.mediaone.net>, Ric Werme wrote: >jmfbahciv@aol.com writes: > >>In article <9ng6t6$1gg$1@i4got.pechter.dyndns.org>, >> pechter@i4got.pechter.dyndns.org (Bill Pechter) wrote: > >> > >>>As far as the Alpha goes -- I'd try to get them produced at commodity >>>prices and get them out there in numbers and prices that would compete >>>with the Pentium III line. > >>That's something I never understood. Why didn't they do that? >>Buying an Alpha was essentially barred from retail. > >Why didn't DEC try to come up with a lowest possible cost PDP-11 ala >Heathkit's H-11 and put RT-11 on it? They could've called it a personal >computer and sold it to individuals. (Hmm, "Personal Data Processor?") > >Catch is, it would have competed with the "minicomputer" PDP-11s. No, it wouldn't have. The minis weren't off-the-shelf installations either...especially the -11s. Companies bought them and either hired somebody or had DEC doing the "adjustments" to do the task. > >There was an "Alpha PC" processor. Handled by >Digital's Stealth marketing, I never learned much >about it until it was over. I think they were concerned >about it competing with workstations. But this doesn't make any sense. You want the future equipment procurers to be familiar with your stuff. It isn't competition in the outside world that was the problem; rather it was the arcane way the books were kept that made two products competitive _inside_ the company. > >When I left DEC in 1978, it occurred to me that DEC's >concern about competing against itself was misplaced, >as everyone else was quite willing to. And did. But that blurb about competition had nothing to do with the field. It had everything to do with product lines "competing" within the company. Do you realize that the PDP-10 line never got mini sales put on the -10's books? Do you realize that we didn't get evaluated on the money coming in for documentation or support if it had anything to do with a non-PDP-10? It's not clear to me that the -10 line got credit for the field service or maintenance income. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Tue, 11 Sep 01 08:54:19 GMT Organization: UltraNet Communications, Inc. Lines: 36 Message-ID: <9nktc3$ast$12@bob.news.rcn.net> References: <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> X-Trace: UmFuZG9tSVaWPgMgly6gSEJgGrCkWOb19gC/grOrATMFOSsw12VALm0/xzljJXJ8 X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 11 Sep 2001 11:42:59 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-156 Xref: chonsp.franklin.ch alt.folklore.computers:90166 In article <3B9DE34F.8DBAB95@hda.hydro.com>, Terje Mathisen wrote: >CBFalconer wrote: >> >> Terje Mathisen wrote: >> > It is sufficient that each interrupt driver will chain to the previous >> > one (at least when there's no obvious casue for the current interrupt), >> > and that the drivers are willing and capable of handling what looks (to >> > them) as a spurious interrupt, i.e. with no real cause. >> > >> > This way each driver doesn't need to know anything about any other hw >> > that happens to share the same interrupt line. >> >> Not quite. If you have multiple priority levels of interrupts, >> you also have to re-enable the interrupt controller level, and >> this shouldn't be done until all the possibilities have been >> examined. Thus you either have to have a master service routine, >> or ensure that only the last one in the chain does this job. > >If you re-enable it prematurely, you just suffer an immediate nested >interrupt, but beside needing some extra stack resources, this should be >OK. > >I.e. only if you re-enable interrupts would it be OK to return >immediately, instead of first chaining to the next driver in the list. But a driver doesn't get involved during an interrupt...does it? I thought the role of a driver is to drive a hardware device. It's only when that device has to say "OK, I'm done; what do I do next?" or "Something's wrong!" that an interrupt occurs. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Tue, 11 Sep 01 08:50:17 GMT Organization: UltraNet Communications, Inc. Lines: 42 Message-ID: <9nkt4i$ast$11@bob.news.rcn.net> References: <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> X-Trace: UmFuZG9tSVYfc8YNUcsReBLRBgRpfTgUkD90ngM+fB/RTbgmMqPG8K426tchgZZT X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 11 Sep 2001 11:38:58 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-156 Xref: chonsp.franklin.ch alt.folklore.computers:90170 In article <3B9DD935.583B7141@yahoo.com>, CBFalconer wrote: >Terje Mathisen wrote: >> >> CBFalconer wrote: >> > The interrupt module (assuming level sensitive) simply >> > interrogates the (limited) possible sources of an interrupt, and >> > resets those that it services. At completion either the interrupt >> > level has reverted to the default or a new interrupt has occured. >> > Not quite as efficient as one event per interrupt, but far better >> > than a monstrous polling cycle. The rub is that the software has >> > to know about the hardware configuration. >> >> Not really: >> >> It is sufficient that each interrupt driver will chain to the previous >> one (at least when there's no obvious casue for the current interrupt), >> and that the drivers are willing and capable of handling what looks (to >> them) as a spurious interrupt, i.e. with no real cause. >> >> This way each driver doesn't need to know anything about any other hw >> that happens to share the same interrupt line. > >Not quite. If you have multiple priority levels of interrupts, >you also have to re-enable the interrupt controller level, and >this shouldn't be done until all the possibilities have been >examined. Thus you either have to have a master service routine, >or ensure that only the last one in the chain does this job. > I can't tell you the details about how JMF or TW did it. Your description of how it has to be done just doesn't sound like the way they did it. Magic incantations to look for in the listings are EPT (exec proccess table). On the last page of my cheat sheet ("decsystem10 SYSTEM Reference Card including the DECsystem-1070)", words 42-57 of the EPT are defined to Standard Priority Interrupt Instructions. I think I've copied it correctly; my eyes can't seem to see the little teensy tiny printing anymore. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Tue, 11 Sep 01 08:40:35 GMT Organization: UltraNet Communications, Inc. Lines: 42 Message-ID: <9nksib$ast$10@bob.news.rcn.net> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9nahnc$f38$14@bob.news.rcn.net> <090920011220344692%elliott@yrl.co.uk> <9ng6t6$1gg$1@i4got.pechter.dyndns.org> <9ni7ck$c78$2@bob.news.rcn.net> <_Ccn7.24433$bl4.4553314@typhoon.ne.mediaone.net> <1bae02s9ub.fsf@cs.nmsu.edu> X-Trace: UmFuZG9tSVbIynhjAptRWzHjW/Ouca9DJk4u2r7v4PwjRuTB0NCEbcMXPhyV9XVM X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 11 Sep 2001 11:29:15 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.mathworks.com!news.voicenet.com!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-156 Xref: chonsp.franklin.ch alt.folklore.computers:90177 In article <1bae02s9ub.fsf@cs.nmsu.edu>, Joe Pfeiffer wrote: >Ric Werme writes: >> >> Catch is, it would have competed with the "minicomputer" PDP-11s. > >That's the best one-line description of what went wrong with DEC in >the 1980s that I've ever seen. They didn't understand that *somebody* >was going to build PCs that were going to compete with their >minicomputers, and that it was long-term inevitable that the PCs were >going to win. > >> There was an "Alpha PC" processor. Handled by Digital's Stealth marketing, >> I never learned much about it until it was over. I think they were concerned >> about it competing with workstations. >> >> When I left DEC in 1978, it occurred to me that DEC's concern about competing >> against itself was misplaced, as everyone else was quite willing to. And did. > >Yep. Meanwhile, their hardware turned into crap. I knew guys who >told me stories about DEC tape drives on board ships in the 1970s, out >on deck, in salt spray, surviving forever. That's because they weren't our tape drives (unless you're talking about DECtape which was the best magnetic media ever). > ... Meanwhile every time El >Paso Electric flickered our lights (which happens pretty frequently) >another VT-220 terminal would pop its power supply. We never made a decent terminal or magtape drive. If it was designed and manufactured within DEC, it sucked. I never did figure that one out since we did make DECtapes, disks, and CPUs that were super. /BAH Subtract a hundred and four for e-mail. ###### From: "Charlie Gibbs" Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 10 Sep 01 12:07:56 -0800 Organization: http://extra.newsguy.com Lines: 20 Message-ID: <543.653T2811T7275748@nowhere.in.particular> References: <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <9nahnc$f38$14@bob.news.rcn.net> <28ifn9.6en.ln@setter.lysse.co.uk> NNTP-Posting-Host: p-993.newsdawg.com X-Newsreader: THOR 2.5a (Amiga;TCP/IP) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!newsfeed.r-kom.de!news-nue1.dfn.de!news-han1.dfn.de!news-koe1.dfn.de!do.de.uu.net!pln-e!spln!dex!extra.newsguy.com!newsp.newsguy.com!news3 Xref: chonsp.franklin.ch alt.folklore.computers:90199 In article <28ifn9.6en.ln@setter.lysse.co.uk> lysse@retriever.lysse.co.uk (lysse) writes: >no-spam wrote: > >> I can still buy stock in Intel, IBM, Microsoft, Apple. How's DEC >> stock doing these days ? >> >> Case closed. > >This is ridiculous. Do you pick your favourite music by only >looking at the top 5? He probably doesn't go to gourmet restaurants because McDonald's, with their billions served, is obviously better. -- cgibbs@nowhere.in.particular (Charlie Gibbs) I'm switching ISPs - watch this space. ###### From: lysse Subject: Re: Minimalist design (was Re: Parity - why even or odd) Newsgroups: comp.arch,alt.folklore.computers References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <9nahnc$f38$14@bob.news.rcn.net> <28ifn9.6en.ln@setter.lysse.co.uk> <6ubskj434h.fsf@chonsp.franklin.ch> Organization: lysse's domain Reply-To: lysse.news@blueyonder.co.uk User-Agent: tin/pre-1.4-980226 (UNIX) (Linux/2.2.6 (i486)) Message-ID: Lines: 20 Date: Mon, 10 Sep 2001 19:32:53 GMT NNTP-Posting-Host: 62.31.9.47 X-Complaints-To: http://www.blueyonder.co.uk/abuse X-Trace: news1.cableinet.net 1000150373 62.31.9.47 (Mon, 10 Sep 2001 20:32:53 BST) NNTP-Posting-Date: Mon, 10 Sep 2001 20:32:53 BST Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!fr.clara.net!heighliner.fr.clara.net!diablo.netcom.net.uk!netcom.net.uk!news-hub.cableinet.net!blueyonder!internal-news-hub.cableinet.net!news1.cableinet.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90189 Neil Franklin wrote: > $45 FPGA. That is what a XC2S200 costs presently. > I don't think that anything $10 has enough room for an 10. Them > 36bits and 512 instructions take up space. Thanks for the correction. However, a -10 isn't the only cool architecture that needs to be revived, and a $10 FPGA might have plenty of room for a more minimalistic architecture, eg the P21. > I can still download the PDP-10 reference manual and build one. > Engineering outlives business. Just like the pyramids outlived the > pharaoes. Perhaps we are the last one to have the laugh, after all. I never thought I'd see the day of samizdat CPUs. ;-) -- lysse at lysse dot co dot uk "Why are your problems always so much bigger than everyone else's?" "Because they're mine." -- Ally McBeal ###### Sender: eric@ruckus.brouhaha.com From: Eric Smith Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <9nahnc$f38$14@bob.news.rcn.net> <28ifn9.6en.ln@setter.lysse.co.uk> Organization: Eric Conspiracy Secret Labs X-Eric-Conspiracy: There is no conspiracy. Date: 10 Sep 2001 12:41:47 -0700 Message-ID: Lines: 4 User-Agent: Gnus/5.0807 (Gnus v5.8.7) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii NNTP-Posting-Host: ruckus.brouhaha.com X-Trace: 10 Sep 2001 12:55:35 -0700, ruckus.brouhaha.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.mathworks.com!news.kjsl.com!news.spies.com!ruckus.brouhaha.com Xref: chonsp.franklin.ch alt.folklore.computers:90201 lysse writes: > and a PDP-10 is still available from XKL, Unfortunately not. ###### From: hawk@fac13.ds.psu.edu (Prof. Richard E. Hawkins) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 10 Sep 2001 20:03:15 GMT Organization: Penn State University, Center for Academic Computing Lines: 20 Message-ID: <9nj6a3$156c@r02n01.cac.psu.edu> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <28ifn9.6en.ln@setter.lysse.co.uk> <6b5gn9.qkn.ln@setter.lysse.co.uk> NNTP-Posting-Host: fac13.ds.psu.edu X-Newsreader: trn 4.0-test76 (Apr 2, 2001) Originator: hawk@fac13.ds.psu.edu (Prof. Richard E. Hawkins) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed.stanford.edu!news.ems.psu.edu!news3.cac.psu.edu!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90148 In article <6b5gn9.qkn.ln@setter.lysse.co.uk>, lysse wrote: >Yes, but as I said in another thread, if M$ topples, the x86 will >probably die fairly quickly afterwards. I have the IA64 instruction >manual; it doesn't look too bad, but then it was co-designed by HP. But lets be fair here: the x86 was *never* meant to go on this long. It was supposed to tide people over until the 860 was established. Unfortunately, people bought more, an dmore, and more, and the 8 bit hangover continues . . . hawk -- Prof. Richard E. Hawkins, Esq. /"\ ASCII ribbon campaign dochawk@psu.edu Smeal 178 (814) 375-4700 \ / against HTML mail These opinions will not be those of X and postings Penn State until it pays my retainer. / \ ###### Sender: eric@ruckus.brouhaha.com From: Eric Smith Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <28ifn9.6en.ln@setter.lysse.co.uk> <6b5gn9.qkn.ln@setter.lysse.co.uk> <9nj6a3$156c@r02n01.cac.psu.edu> Organization: Eric Conspiracy Secret Labs X-Eric-Conspiracy: There is no conspiracy. Date: 10 Sep 2001 14:19:12 -0700 Message-ID: Lines: 6 User-Agent: Gnus/5.0807 (Gnus v5.8.7) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii NNTP-Posting-Host: ruckus.brouhaha.com X-Trace: 10 Sep 2001 14:33:01 -0700, ruckus.brouhaha.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news-out.nuthinbutnews.com!propagator2-austin!propagator!feed2.newsfeeds.com!newsfeeds.com!news-in-austin.nuthinbutnews.com!news.exile.org!news.kjsl.com!news.spies.com!ruckus.brouhaha.com Xref: chonsp.franklin.ch alt.folklore.computers:90206 hawk@fac13.ds.psu.edu (Prof. Richard E. Hawkins) writes: > But lets be fair here: the x86 was *never* meant to go on this long. > It was supposed to tide people over until the 860 was established. No, until the iAPX 432 was established. The i860 wasn't even dreamt of back in 1977; design work on the i860 started after the i386. ###### Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9nahnc$f38$14@bob.news.rcn.net> <090920011220344692%elliott@yrl.co.uk> <9ng6t6$1gg$1@i4got.pechter.dyndns.org> <9ni7ck$c78$2@bob.news.rcn.net> From: Ric Werme X-Newsreader: NN version 6.5.0 CURRENT #119 Lines: 32 Message-ID: <_Ccn7.24433$bl4.4553314@typhoon.ne.mediaone.net> Date: Tue, 11 Sep 2001 00:09:30 GMT NNTP-Posting-Host: 24.128.105.166 X-Complaints-To: abuse@mediaone.net X-Trace: typhoon.ne.mediaone.net 1000166970 24.128.105.166 (Mon, 10 Sep 2001 20:09:30 EDT) NNTP-Posting-Date: Mon, 10 Sep 2001 20:09:30 EDT Organization: Road Runner Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feed2.news.rcn.net!rcn!chnws02.mediaone.net!chnws06.ne.mediaone.net!24.128.8.70!typhoon.ne.mediaone.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90224 jmfbahciv@aol.com writes: >In article <9ng6t6$1gg$1@i4got.pechter.dyndns.org>, > pechter@i4got.pechter.dyndns.org (Bill Pechter) wrote: > >>As far as the Alpha goes -- I'd try to get them produced at commodity >>prices and get them out there in numbers and prices that would compete >>with the Pentium III line. >That's something I never understood. Why didn't they do that? >Buying an Alpha was essentially barred from retail. Why didn't DEC try to come up with a lowest possible cost PDP-11 ala Heathkit's H-11 and put RT-11 on it? They could've called it a personal computer and sold it to individuals. (Hmm, "Personal Data Processor?") Catch is, it would have competed with the "minicomputer" PDP-11s. There was an "Alpha PC" processor. Handled by Digital's Stealth marketing, I never learned much about it until it was over. I think they were concerned about it competing with workstations. When I left DEC in 1978, it occurred to me that DEC's concern about competing against itself was misplaced, as everyone else was quite willing to. And did. -Ric Werme -- Ric Werme | werme@nospam.mediaone.net http://people.ne.mediaone.net/werme | ^^^^^^^ delete ###### From: Joe Pfeiffer Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 10 Sep 2001 21:08:28 -0600 Organization: NMSU Computer Science Lines: 27 Message-ID: <1bae02s9ub.fsf@cs.nmsu.edu> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9nahnc$f38$14@bob.news.rcn.net> <090920011220344692%elliott@yrl.co.uk> <9ng6t6$1gg$1@i4got.pechter.dyndns.org> <9ni7ck$c78$2@bob.news.rcn.net> <_Ccn7.24433$bl4.4553314@typhoon.ne.mediaone.net> NNTP-Posting-Host: viper.cs.nmsu.edu X-Trace: bubba.NMSU.Edu 1000177707 22137 128.123.64.113 (11 Sep 2001 03:08:27 GMT) X-Complaints-To: usenet@bubba.NMSU.Edu NNTP-Posting-Date: 11 Sep 2001 03:08:27 GMT X-Newsreader: Gnus v5.7/Emacs 20.5 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!upp1.onvoy!msc1.onvoy!onvoy.com!hardy.tc.umn.edu!lynx.unm.edu!news.NMSU.Edu!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90146 Ric Werme writes: > > Catch is, it would have competed with the "minicomputer" PDP-11s. That's the best one-line description of what went wrong with DEC in the 1980s that I've ever seen. They didn't understand that *somebody* was going to build PCs that were going to compete with their minicomputers, and that it was long-term inevitable that the PCs were going to win. > There was an "Alpha PC" processor. Handled by Digital's Stealth marketing, > I never learned much about it until it was over. I think they were concerned > about it competing with workstations. > > When I left DEC in 1978, it occurred to me that DEC's concern about competing > against itself was misplaced, as everyone else was quite willing to. And did. Yep. Meanwhile, their hardware turned into crap. I knew guys who told me stories about DEC tape drives on board ships in the 1970s, out on deck, in salt spray, surviving forever. Meanwhile every time El Paso Electric flickered our lights (which happens pretty frequently) another VT-220 terminal would pop its power supply. -- Joseph J. Pfeiffer, Jr., Ph.D. Phone -- (505) 646-1605 Department of Computer Science FAX -- (505) 646-1002 New Mexico State University http://www.cs.nmsu.edu/~pfeiffer SWNMRSEF: http://www.nmsu.edu/~scifair ###### Message-ID: <3B9DB262.78744BC6@ev1.net> From: Charles Richmond Reply-To: richmond@ev1.net Organization: Cannine Computer Center X-Mailer: Mozilla 4.7C-CCK-MCD {C-UDP; EBM-APPLE} (Macintosh; I; PPC) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <9nahnc$f38$14@bob.news.rcn.net> <28ifn9.6en.ln@setter.lysse.co.uk> <6ubskj434h.fsf@chonsp.franklin.ch> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 25 Date: Tue, 11 Sep 2001 04:45:09 GMT NNTP-Posting-Host: 24.179.111.125 X-Complaints-To: abuse@home.net X-Trace: news2.rdc2.tx.home.com 1000183509 24.179.111.125 (Mon, 10 Sep 2001 21:45:09 PDT) NNTP-Posting-Date: Mon, 10 Sep 2001 21:45:09 PDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!newsfeed.hanau.net!fr.clara.net!heighliner.fr.clara.net!feed2.onemain.com!feed1.onemain.com!newsfeed.direct.ca!look.ca!newshub2.rdc1.sfba.home.com!news.home.com!news2.rdc2.tx.home.com.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90218 lysse wrote: > > Neil Franklin wrote: > > > $45 FPGA. That is what a XC2S200 costs presently. > > > I don't think that anything $10 has enough room for an 10. Them > > 36bits and 512 instructions take up space. > > Thanks for the correction. However, a -10 isn't the only cool > architecture that needs to be revived, and a $10 FPGA might have > plenty of room for a more minimalistic architecture, eg the P21. > > > I can still download the PDP-10 reference manual and build one. > > Engineering outlives business. Just like the pyramids outlived the > > pharaoes. Perhaps we are the last one to have the laugh, after all. > > I never thought I'd see the day of samizdat CPUs. ;-) > You mean...this CPU is samizdat CPU??? -- +-------------------------------------------------------------+ | Charles and Francis Richmond | +-------------------------------------------------------------+ ###### Message-ID: <3B9D2E53.3D5C720F@yahoo.com> From: CBFalconer Reply-To: cbfalconer@worldnet.att.net Organization: Ched Research X-Mailer: Mozilla 4.75 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 51 Date: Tue, 11 Sep 2001 05:03:01 GMT NNTP-Posting-Host: 12.90.167.60 X-Complaints-To: abuse@worldnet.att.net X-Trace: bgtnsc06-news.ops.worldnet.att.net 1000184581 12.90.167.60 (Tue, 11 Sep 2001 05:03:01 GMT) NNTP-Posting-Date: Tue, 11 Sep 2001 05:03:01 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!howland.erols.net!news-out.worldnet.att.net.MISMATCH!wn3feed!worldnet.att.net!135.173.83.71!wnfilter1!worldnet-localpost!bgtnsc06-news.ops.worldnet.att.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90129 jmfbahciv@aol.com wrote: > > In article <6uy9nsqhd7.fsf@chonsp.franklin.ch>, > Neil Franklin wrote: > >Joe Pfeiffer writes: > > > >> jmfbahciv@aol.com writes: > >> > > > >> > Wait a minute. Are you saying that these systems can't deal with > >> > more than one interrupt to a level? How bizarre. > >> > >> One of the PC's screwups was that it was effectively impossible for it > >> to deal with more than one interrupt at a level (it used assert-high > >> interrupt requests). > > > >Assert-high was not the problem (it is bad for catching interference). > >The problem was that the PC uses edge-triggered and not > >level-triggered interrupts. > > > >With edge-triggered The first device send an edge that sets an > >Flip-Flop. A hypothetical second device would have no effect and > >so be missed. > > Wouldn't the second device cause a Flop-Flip? Or did I just > demonstrate how ignorant I am about how all of that hardware > works? > > > > >With level-triggered the line is simply an wire-OR of all interrupt > >requests and keeps on interruptiong until it falls to zero as soon > >as everyone is satisfied. This sharing is possible. > > But interruptions won't ever fall to zero...the clock will > continue. This is imnpossible to snip and maintain coherence :-) The interrupt module (assuming level sensitive) simply interrogates the (limited) possible sources of an interrupt, and resets those that it services. At completion either the interrupt level has reverted to the default or a new interrupt has occured. Not quite as efficient as one event per interrupt, but far better than a monstrous polling cycle. The rub is that the software has to know about the hardware configuration. -- Chuck F (cbfalconer@yahoo.com) (cbfalconer@XXXXworldnet.att.net) (Remove "XXXX" from reply address. yahoo works unmodified) mailto:uce@ftc.gov (for spambots to harvest) ###### From: Terje Mathisen Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Tue, 11 Sep 2001 09:36:29 +0200 Organization: Hydro Lines: 25 Message-ID: <3B9DBEFD.BAAE6C8A@hda.hydro.com> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> NNTP-Posting-Host: no115350c8.hda.hydro.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.77 [en] (Windows NT 5.0; U) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!fr.clara.net!heighliner.fr.clara.net!news.tele.dk!small.news.tele.dk!148.122.208.68!news2.oke.nextra.no!nextra.com!hydro.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90152 CBFalconer wrote: > The interrupt module (assuming level sensitive) simply > interrogates the (limited) possible sources of an interrupt, and > resets those that it services. At completion either the interrupt > level has reverted to the default or a new interrupt has occured. > Not quite as efficient as one event per interrupt, but far better > than a monstrous polling cycle. The rub is that the software has > to know about the hardware configuration. Not really: It is sufficient that each interrupt driver will chain to the previous one (at least when there's no obvious casue for the current interrupt), and that the drivers are willing and capable of handling what looks (to them) as a spurious interrupt, i.e. with no real cause. This way each driver doesn't need to know anything about any other hw that happens to share the same interrupt line. Terje -- - Using self-discipline, see http://www.eiffel.com/discipline "almost all programming can be viewed as an exercise in caching" ###### Message-ID: <3B9DD935.583B7141@yahoo.com> From: CBFalconer Reply-To: cbfalconer@worldnet.att.net Organization: Ched Research X-Mailer: Mozilla 4.75 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 32 Date: Tue, 11 Sep 2001 09:34:03 GMT NNTP-Posting-Host: 12.90.168.89 X-Complaints-To: abuse@worldnet.att.net X-Trace: bgtnsc05-news.ops.worldnet.att.net 1000200843 12.90.168.89 (Tue, 11 Sep 2001 09:34:03 GMT) NNTP-Posting-Date: Tue, 11 Sep 2001 09:34:03 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!newsfeed01.sul.t-online.de!t-online.de!newsfeed.hanau.net!news-fra1.dfn.de!news0.de.colt.net!colt.net!dispose.news.demon.net!demon!howland.erols.net!news-out.worldnet.att.net.MISMATCH!wn3feed!worldnet.att.net!135.173.83.71!wnfilter1!worldnet-localpost!bgtnsc05-news.ops.worldnet.att.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90240 Terje Mathisen wrote: > > CBFalconer wrote: > > The interrupt module (assuming level sensitive) simply > > interrogates the (limited) possible sources of an interrupt, and > > resets those that it services. At completion either the interrupt > > level has reverted to the default or a new interrupt has occured. > > Not quite as efficient as one event per interrupt, but far better > > than a monstrous polling cycle. The rub is that the software has > > to know about the hardware configuration. > > Not really: > > It is sufficient that each interrupt driver will chain to the previous > one (at least when there's no obvious casue for the current interrupt), > and that the drivers are willing and capable of handling what looks (to > them) as a spurious interrupt, i.e. with no real cause. > > This way each driver doesn't need to know anything about any other hw > that happens to share the same interrupt line. Not quite. If you have multiple priority levels of interrupts, you also have to re-enable the interrupt controller level, and this shouldn't be done until all the possibilities have been examined. Thus you either have to have a master service routine, or ensure that only the last one in the chain does this job. -- Chuck F (cbfalconer@yahoo.com) (cbfalconer@XXXXworldnet.att.net) (Remove "XXXX" from reply address. yahoo works unmodified) mailto:uce@ftc.gov (for spambots to harvest) ###### From: Terje Mathisen Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Tue, 11 Sep 2001 12:11:27 +0200 Organization: Hydro Lines: 32 Message-ID: <3B9DE34F.8DBAB95@hda.hydro.com> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> NNTP-Posting-Host: no115350c8.hda.hydro.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.77 [en] (Windows NT 5.0; U) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.tele.dk!small.news.tele.dk!148.122.208.68!news2.oke.nextra.no!nextra.com!hydro.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90141 CBFalconer wrote: > > Terje Mathisen wrote: > > It is sufficient that each interrupt driver will chain to the previous > > one (at least when there's no obvious casue for the current interrupt), > > and that the drivers are willing and capable of handling what looks (to > > them) as a spurious interrupt, i.e. with no real cause. > > > > This way each driver doesn't need to know anything about any other hw > > that happens to share the same interrupt line. > > Not quite. If you have multiple priority levels of interrupts, > you also have to re-enable the interrupt controller level, and > this shouldn't be done until all the possibilities have been > examined. Thus you either have to have a master service routine, > or ensure that only the last one in the chain does this job. If you re-enable it prematurely, you just suffer an immediate nested interrupt, but beside needing some extra stack resources, this should be OK. I.e. only if you re-enable interrupts would it be OK to return immediately, instead of first chaining to the next driver in the list. OK? Terje -- - Using self-discipline, see http://www.eiffel.com/discipline "almost all programming can be viewed as an exercise in caching" ###### Message-ID: <3B9DEBA0.A72EDEFC@mikron.de> Date: Tue, 11 Sep 2001 12:46:56 +0200 From: Bernd Paysan X-Mailer: Mozilla 4.77 [de] (X11; U; Linux 2.4.4-4GB i686) X-Accept-Language: de, en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <28ifn9.6en.ln@setter.lysse.co.uk> <6b5gn9.qkn.ln@setter.lysse.co.uk> <9nj6a3$156c@r02n01.cac.psu.edu> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 18 NNTP-Posting-Host: 194.139.17.38 X-Trace: 1000206004 read.news.de.uu.net 186 194.139.17.38 X-Complaints-To: abuse@de.uu.net Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!news-lei1.dfn.de!news-fra1.dfn.de!news.tele.dk!small.news.tele.dk!195.158.233.21!news1.ebone.net!news.ebone.net!lnewspeer00.lnd.ops.eu.uu.net!lnewsifeed00.lnd.ops.eu.uu.net!lnewsifeed01.lnd.ops.eu.uu.net!bnewspost00.bru.ops.eu.uu.net!emea.uu.net!read.news.de.uu.net!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90135 Eric Smith wrote: > > hawk@fac13.ds.psu.edu (Prof. Richard E. Hawkins) writes: > > But lets be fair here: the x86 was *never* meant to go on this long. > > It was supposed to tide people over until the 860 was established. > > No, until the iAPX 432 was established. The i860 wasn't even dreamt > of back in 1977; design work on the i860 started after the i386. Other way round. When Intel saw that there was no light at the end of the i432 tunnel, they ordered a few engineers to design a quick and dirty 16 bit extended ISA, source-level backward compatible the 8085 (to fill the gap): the 8086. -- Bernd Paysan "If you want it done right, you have to do it yourself" http://www.jwdt.com/~paysan/ ###### Newsgroups: alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9ni7ck$c78$2@bob.news.rcn.net> <_Ccn7.24433$bl4.4553314@typhoon.ne.mediaone.net> <1bae02s9ub.fsf@cs.nmsu.edu> Organization: Daedalus Consulting X-Newsreader: trn 4.0-test72 (19 April 1999) From: don@news.daedalus.co.nz (Don Stokes) Lines: 35 Message-ID: Date: Tue, 11 Sep 2001 11:09:58 GMT NNTP-Posting-Host: 203.96.144.16 X-Complaints-To: abuse@tsnz.net X-Trace: news02.tsnz.net 1000206598 203.96.144.16 (Tue, 11 Sep 2001 23:09:58 NZST) NNTP-Posting-Date: Tue, 11 Sep 2001 23:09:58 NZST Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!newsfeed01.sul.t-online.de!t-online.de!fr.clara.net!heighliner.fr.clara.net!skynet.be!skynet.be!newsfeed01.tsnz.net!news02.tsnz.net!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90185 In article <1bae02s9ub.fsf@cs.nmsu.edu>, Joe Pfeiffer wrote: >Ric Werme writes: >> There was an "Alpha PC" processor. Handled by Digital's Stealth marketing, >> I never learned much about it until it was over. I think they were concerned >> about it competing with workstations. I recall a DECUS presentation, by a Digit, about the Alpha motherboard to go into a PC case. This was when the Alphas were only just hitting the market, and it wasn't entirely clear what one was supposed to run on the thing once you got it, much less how you were supposed to get one in the first place. But they did exist. Coupla years later I got to see an Alpha board, at of all things an embedded systems trade show (which happened to be co-located with a networking conference, and had a much more interesting show). >Yep. Meanwhile, their hardware turned into crap. I knew guys who >told me stories about DEC tape drives on board ships in the 1970s, out >on deck, in salt spray, surviving forever. Meanwhile every time El >Paso Electric flickered our lights (which happens pretty frequently) >another VT-220 terminal would pop its power supply. Hmm, the only time I ever saw a VT-220 go pop was when the folks renovating out building decided to pump water into the heating radiators without bothering to connect up the plumbing first. VT-320s on the other hand ... "hmm, funny place for a LED ... hang on, that's a capacitor ..." -- don (comp.arch trimmed) ###### From: Joe Pfeiffer Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 11 Sep 2001 08:13:16 -0600 Organization: NMSU Computer Science Lines: 25 Message-ID: <1br8tdiznn.fsf@cs.nmsu.edu> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9nahnc$f38$14@bob.news.rcn.net> <090920011220344692%elliott@yrl.co.uk> <9ng6t6$1gg$1@i4got.pechter.dyndns.org> <9ni7ck$c78$2@bob.news.rcn.net> <_Ccn7.24433$bl4.4553314@typhoon.ne.mediaone.net> <1bae02s9ub.fsf@cs.nmsu.edu> <9nksib$ast$10@bob.news.rcn.net> NNTP-Posting-Host: viper.cs.nmsu.edu X-Trace: bubba.NMSU.Edu 1000217594 19780 128.123.64.113 (11 Sep 2001 14:13:14 GMT) X-Complaints-To: usenet@bubba.NMSU.Edu NNTP-Posting-Date: 11 Sep 2001 14:13:14 GMT X-Newsreader: Gnus v5.7/Emacs 20.5 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!news0.de.colt.net!colt.net!news.maxwell.syr.edu!feed2.news.rcn.net!rcn!dca6-feed2.news.digex.net!lax2-feed1.news.digex.net!intermedia!lynx.unm.edu!news.NMSU.Edu!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90158 jmfbahciv@aol.com writes: > >Yep. Meanwhile, their hardware turned into crap. I knew guys who > >told me stories about DEC tape drives on board ships in the 1970s, out > >on deck, in salt spray, surviving forever. > > That's because they weren't our tape drives (unless you're > talking about DECtape which was the best magnetic media ever). Yep, I do mean DECtapes. > > > ... Meanwhile every time El > >Paso Electric flickered our lights (which happens pretty frequently) > >another VT-220 terminal would pop its power supply. > > We never made a decent terminal or magtape drive. If it was > designed and manufactured within DEC, it sucked. I never > did figure that one out since we did make DECtapes, disks, > and CPUs that were super. I remember VT-100s as being very solid. -- Joseph J. Pfeiffer, Jr., Ph.D. Phone -- (505) 646-1605 Department of Computer Science FAX -- (505) 646-1002 New Mexico State University http://www.cs.nmsu.edu/~pfeiffer SWNMRSEF: http://www.nmsu.edu/~scifair ###### From: Joe Pfeiffer Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 11 Sep 2001 08:17:07 -0600 Organization: NMSU Computer Science Lines: 13 Message-ID: <1bofohizh8.fsf@cs.nmsu.edu> References: <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <9nktc3$ast$12@bob.news.rcn.net> NNTP-Posting-Host: viper.cs.nmsu.edu X-Trace: bubba.NMSU.Edu 1000217824 19780 128.123.64.113 (11 Sep 2001 14:17:04 GMT) X-Complaints-To: usenet@bubba.NMSU.Edu NNTP-Posting-Date: 11 Sep 2001 14:17:04 GMT X-Newsreader: Gnus v5.7/Emacs 20.5 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feed2.news.rcn.net!rcn!dca6-feed2.news.digex.net!lax2-feed1.news.digex.net!intermedia!lynx.unm.edu!news.NMSU.Edu!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90156 jmfbahciv@aol.com writes: > > But a driver doesn't get involved during an interrupt...does it? > I thought the role of a driver is to drive a hardware device. > It's only when that device has to say "OK, I'm done; what do I > do next?" or "Something's wrong!" that an interrupt occurs. Yes -- the device's interrupt handler is part of its driver. -- Joseph J. Pfeiffer, Jr., Ph.D. Phone -- (505) 646-1605 Department of Computer Science FAX -- (505) 646-1002 New Mexico State University http://www.cs.nmsu.edu/~pfeiffer SWNMRSEF: http://www.nmsu.edu/~scifair ###### From: Terje Mathisen Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Tue, 11 Sep 2001 18:58:08 +0200 Organization: Hydro Lines: 78 Message-ID: <3B9E42A0.97ABB794@hda.hydro.com> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> NNTP-Posting-Host: iabrahmspc.nho.hydro.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.77 [en] (Windows NT 5.0; U) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.tele.dk!small.news.tele.dk!148.122.208.68!news2.oke.nextra.no!nextra.com!hydro.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90143 Konrad Schwarz wrote: > > Terje Mathisen schrieb: > > If you re-enable it prematurely, you just suffer an immediate nested > > interrupt, but beside needing some extra stack resources, this should be > > OK. > > No. > > The previous interrupt level must become active only at the completion of the > complete interrupt handler. Why? The reason is the immediate nested interrupt. I.e., just as I stated above. > > Say we have two devices A and B attached to an interrupt. (Actually the problem > is more general: A and B can be attached to different IRQs with different priorities. > The example also works with only one device A.) > > A posts an interrupt and the CPU fields it by calling A's interrupt handler (aka ISR). > A's interrupt handler enables interrupts prematurely and B posts an interrupt. > The CPU starts executing B's interrupt handler. > > B's interrupt handler enables interrupts prematurely, > A interrupts, CPU calls A's interrupt handler. > > We now have two copies of A's interrupt handler running. This means that A's > interrupt handler needs to be reentrant. Since A's interrupt handler is > updating global state to communicate with > base level code, this is fairly difficult to achieve. It might be difficult, and I have made a few mistakes here, but it wasn't really that hard to write a serial port driver that allows multiple (4) different interrupt sources from each port, as well as having multiple ports sharing the same interrupt. Yes, the parts of the code that can run reentrantly obviously needs to be reentrant, but the only really critical parts are the code sequences that reads some hardware and tranfers that info to some software structure. > Even if A's interrupt handler > were reentrant, we could now have the situation that, since the nested call to A's > interrupt routine completes before the outer invocation of A's interrupt > routine, that the data read from/written to A is reversed, i.e., the nested > call's data arrive first, the outer call's data arrives second. E.g., > A is a UART and suddenly two adjoining characters are reversed. (In general, > arbitrary permutations of the data can be achieved using additional nested > invocations.) This was not a real problem. > > An additional problem is the additional stack space. Allowing premature As I also noted. > nested interrupts means that the space required by interrupt handlers > is unbounded. Thus the probability of the (interrupt) stack overflowing is > greater than zero. If interrupt handler nest strictly, the stack is bounded > by the sum of the requirements of each interrupt handler. This is correct, and the reason why I always wrote my code to put absolute limits to how much stack could ever be used in a worst-case situation. > > People who believe that this problem can be solved by making interrupt handlers > as short as possible deserve the reliability of MS Windows. Terje -- - Using self-discipline, see http://www.eiffel.com/discipline "almost all programming can be viewed as an exercise in caching" ###### From: Terje Mathisen Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Tue, 11 Sep 2001 19:00:28 +0200 Organization: Hydro Lines: 23 Message-ID: <3B9E432C.AE2A3FA5@hda.hydro.com> References: <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <9nktc3$ast$12@bob.news.rcn.net> <1bofohizh8.fsf@cs.nmsu.edu> NNTP-Posting-Host: iabrahmspc.nho.hydro.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.77 [en] (Windows NT 5.0; U) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeeds.belnet.be!news.belnet.be!news.tele.dk!small.news.tele.dk!148.122.208.68!news2.oke.nextra.no!nextra.com!hydro.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90144 Joe Pfeiffer wrote: > > jmfbahciv@aol.com writes: > > > > But a driver doesn't get involved during an interrupt...does it? > > I thought the role of a driver is to drive a hardware device. > > It's only when that device has to say "OK, I'm done; what do I > > do next?" or "Something's wrong!" that an interrupt occurs. > > Yes -- the device's interrupt handler is part of its driver. On a small system the hw interrupt handler can be more or less all of the driver. Terje PS. We've just turned on the news here, my condolences to anyone with friends/family involved in the horrible terrorist attacks. -- - Using self-discipline, see http://www.eiffel.com/discipline "almost all programming can be viewed as an exercise in caching" ###### Sender: eric@ruckus.brouhaha.com From: Eric Smith Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <28ifn9.6en.ln@setter.lysse.co.uk> <6b5gn9.qkn.ln@setter.lysse.co.uk> <9nj6a3$156c@r02n01.cac.psu.edu> <3B9DEBA0.A72EDEFC@mikron.de> Organization: Eric Conspiracy Secret Labs X-Eric-Conspiracy: There is no conspiracy. Date: 11 Sep 2001 12:15:38 -0700 Message-ID: Lines: 29 User-Agent: Gnus/5.0807 (Gnus v5.8.7) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii NNTP-Posting-Host: ruckus.brouhaha.com X-Trace: 11 Sep 2001 12:29:37 -0700, ruckus.brouhaha.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.tele.dk!small.news.tele.dk!216.218.192.242!news.he.net!news.kjsl.com!news.spies.com!ruckus.brouhaha.com Xref: chonsp.franklin.ch alt.folklore.computers:90203 hawk@fac13.ds.psu.edu (Prof. Richard E. Hawkins) writes: > But lets be fair here: the x86 was *never* meant to go on this long. > It was supposed to tide people over until the 860 was established. Eric Smith wrote: > No, until the iAPX 432 was established. The i860 wasn't even dreamt > of back in 1977; design work on the i860 started after the i386. Bernd Paysan writes: > Other way round. When Intel saw that there was no light at the end of > the i432 tunnel, they ordered a few engineers to design a quick and > dirty 16 bit extended ISA, source-level backward compatible the 8085 (to > fill the gap): the 8086. From personal conversations with some of the engineers and managers involved, it is definitely the case that some of the company thought that the 432 was either not going to succeed, or not going to be competitive at the low-end to midrange market, and those people expected the 8086 to fill that market. Others thought that the 432 would succeed, but that it would take a while, and they needed to stay competitive until the 432 displaced things, so they viewed the 8086 as a stopgap measure. But NONE of those people, at that time, thought the 8086 was just to "tide people over until the 860 was established", since the i860 hadn't even been thought up. The i860 came AFTER the company had written off the 432. ###### Sender: lynn@LYNNPC Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <28ifn9.6en.ln@setter.lysse.co.uk> <6b5gn9.qkn.ln@setter.lysse.co.uk> <9nj6a3$156c@r02n01.cac.psu.edu> <3B9DEBA0.A72EDEFC@mikron.de> Reply-To: Anne & Lynn Wheeler From: Anne & Lynn Wheeler Message-ID: Organization: Wheeler&Wheeler Lines: 26 User-Agent: Gnus/5.0808 (Gnus v5.8.8) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Tue, 11 Sep 2001 19:34:01 GMT NNTP-Posting-Host: 199.174.226.45 X-Complaints-To: abuse@earthlink.net X-Trace: newsread2.prod.itd.earthlink.net 1000236841 199.174.226.45 (Tue, 11 Sep 2001 12:34:01 PDT) NNTP-Posting-Date: Tue, 11 Sep 2001 12:34:01 PDT X-Received-Date: Tue, 11 Sep 2001 12:30:15 PDT (newsmaster1.prod.itd.earthlink.net) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!2402646!news.imp.ch!sunqbc.risq.qc.ca!newsfeed.direct.ca!look.ca!newsfeed1.earthlink.net!newsfeed.earthlink.net!newsmaster1.prod.itd.earthlink.net!newsread2.prod.itd.earthlink.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90178 Eric Smith writes: > From personal conversations with some of the engineers and managers > involved, it is definitely the case that some of the company thought > that the 432 was either not going to succeed, or not going to be > competitive at the low-end to midrange market, and those people expected > the 8086 to fill that market. Others thought that the 432 would > succeed, but that it would take a while, and they needed to stay > competitive until the 432 displaced things, so they viewed the 8086 as > a stopgap measure. a late '70s? sigops at asilomar there was talk on 432 ... big unresolved issue was how to apply fixes ... the "code/algorithms/ideas that had been sunk into silicon was significant ... as were the associated bugs that were being found. unresolved 432 issue was how to fix silicon (software downloads didn't work) ... random past 432 threads http://www.garlic.com/~lynn/2000d.html#57 iAPX-432 (was: 36 to 32 bit transition http://www.garlic.com/~lynn/2000d.html#62 iAPX-432 (was: 36 to 32 bit transition http://www.garlic.com/~lynn/2000e.html#6 Ridiculous http://www.garlic.com/~lynn/2000f.html#48 Famous Machines and Software that didn't http://www.garlic.com/~lynn/2001g.html#36 What was object oriented in iAPX432? -- Anne & Lynn Wheeler | lynn@garlic.com - http://www.garlic.com/~lynn/ ###### Message-ID: <3B9E639B.EF5A1AA2@yahoo.com> From: CBFalconer Reply-To: cbfalconer@worldnet.att.net Organization: Ched Research X-Mailer: Mozilla 4.75 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 43 Date: Tue, 11 Sep 2001 20:18:24 GMT NNTP-Posting-Host: 12.90.167.112 X-Complaints-To: abuse@worldnet.att.net X-Trace: bgtnsc05-news.ops.worldnet.att.net 1000239504 12.90.167.112 (Tue, 11 Sep 2001 20:18:24 GMT) NNTP-Posting-Date: Tue, 11 Sep 2001 20:18:24 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed.mathworks.com!wn3feed!worldnet.att.net!135.173.83.71!wnfilter1!worldnet-localpost!bgtnsc05-news.ops.worldnet.att.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90133 Terje Mathisen wrote: > > CBFalconer wrote: > > > > Terje Mathisen wrote: > > > It is sufficient that each interrupt driver will chain to the previous > > > one (at least when there's no obvious casue for the current interrupt), > > > and that the drivers are willing and capable of handling what looks (to > > > them) as a spurious interrupt, i.e. with no real cause. > > > > > > This way each driver doesn't need to know anything about any other hw > > > that happens to share the same interrupt line. > > > > Not quite. If you have multiple priority levels of interrupts, > > you also have to re-enable the interrupt controller level, and > > this shouldn't be done until all the possibilities have been > > examined. Thus you either have to have a master service routine, > > or ensure that only the last one in the chain does this job. > > If you re-enable it prematurely, you just suffer an immediate nested > interrupt, but beside needing some extra stack resources, this should be > OK. > > I.e. only if you re-enable interrupts would it be OK to return > immediately, instead of first chaining to the next driver in the list. To my mind, horrible inefficiencies at the most critical points. That would require completely restoring the CPU state before the next interrupt happens. Interrupt stack space is VERY likely to be a scarce resource. I once found I could create an almost total denial of service on an HP3000 by sending it a continuous stream of s at 2400 baud, each of which created a CPU interrupt (there was no i/o processor on that version). I rapidly changed the communicating application to supply the 'keep-alive' operation at 10 second intervals. -- Chuck F (cbfalconer@yahoo.com) (cbfalconer@XXXXworldnet.att.net) (Remove "XXXX" from reply address. yahoo works unmodified) mailto:uce@ftc.gov (for spambots to harvest) ###### Message-ID: <3B9E6481.B7DC2B27@yahoo.com> From: CBFalconer Reply-To: cbfalconer@worldnet.att.net Organization: Ched Research X-Mailer: Mozilla 4.75 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <9nktc3$ast$12@bob.news.rcn.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 22 Date: Tue, 11 Sep 2001 20:18:26 GMT NNTP-Posting-Host: 12.90.167.112 X-Complaints-To: abuse@worldnet.att.net X-Trace: bgtnsc05-news.ops.worldnet.att.net 1000239506 12.90.167.112 (Tue, 11 Sep 2001 20:18:26 GMT) NNTP-Posting-Date: Tue, 11 Sep 2001 20:18:26 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!2761720!news.imp.ch!sunqbc.risq.qc.ca!howland.erols.net!news-out.worldnet.att.net.MISMATCH!wn3feed!worldnet.att.net!135.173.83.71!wnfilter1!worldnet-localpost!bgtnsc05-news.ops.worldnet.att.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90130 jmfbahciv@aol.com wrote: > ... snip ... > > > >I.e. only if you re-enable interrupts would it be OK to return > >immediately, instead of first chaining to the next driver in the list. > > But a driver doesn't get involved during an interrupt...does it? > I thought the role of a driver is to drive a hardware device. > It's only when that device has to say "OK, I'm done; what do I > do next?" or "Something's wrong!" that an interrupt occurs. Yes it does. Something has to accept the partial i/o operation and buffer it. Even if you have hardware for this, at some point it will get full. -- Chuck F (cbfalconer@yahoo.com) (cbfalconer@XXXXworldnet.att.net) (Remove "XXXX" from reply address. yahoo works unmodified) mailto:uce@ftc.gov (for spambots to harvest) ###### Message-ID: <3B9E6C3F.1D45BA0D@yahoo.com> From: CBFalconer Reply-To: cbfalconer@worldnet.att.net Organization: Ched Research X-Mailer: Mozilla 4.75 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 78 Date: Tue, 11 Sep 2001 20:18:28 GMT NNTP-Posting-Host: 12.90.167.112 X-Complaints-To: abuse@worldnet.att.net X-Trace: bgtnsc05-news.ops.worldnet.att.net 1000239508 12.90.167.112 (Tue, 11 Sep 2001 20:18:28 GMT) NNTP-Posting-Date: Tue, 11 Sep 2001 20:18:28 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!howland.erols.net!news-out.worldnet.att.net.MISMATCH!wn3feed!worldnet.att.net!135.173.83.71!wnfilter1!worldnet-localpost!bgtnsc05-news.ops.worldnet.att.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90132 Konrad Schwarz wrote: > > Terje Mathisen schrieb: > > > > CBFalconer wrote: > > > > > > Terje Mathisen wrote: > > > > It is sufficient that each interrupt driver will chain to the previous > > > > one (at least when there's no obvious casue for the current interrupt), > > > > and that the drivers are willing and capable of handling what looks (to > > > > them) as a spurious interrupt, i.e. with no real cause. > > > > > > > > This way each driver doesn't need to know anything about any other hw > > > > that happens to share the same interrupt line. > > > > > > Not quite. If you have multiple priority levels of interrupts, > > > you also have to re-enable the interrupt controller level, and > > > this shouldn't be done until all the possibilities have been > > > examined. Thus you either have to have a master service routine, > > > or ensure that only the last one in the chain does this job. > > > > If you re-enable it prematurely, you just suffer an immediate nested > > interrupt, but beside needing some extra stack resources, this should be > > OK. > > No. > > The previous interrupt level must become active only at the completion of the > complete interrupt handler. Why? The reason is the immediate nested interrupt. > > Say we have two devices A and B attached to an interrupt. (Actually the problem > is more general: A and B can be attached to different IRQs with different priorities. > The example also works with only one device A.) > > A posts an interrupt and the CPU fields it by calling A's interrupt handler (aka ISR). > A's interrupt handler enables interrupts prematurely and B posts an interrupt. > The CPU starts executing B's interrupt handler. > > B's interrupt handler enables interrupts prematurely, > A interrupts, CPU calls A's interrupt handler. > > We now have two copies of A's interrupt handler running. This means that A's > interrupt handler needs to be reentrant. Since A's interrupt handler is > updating global state to communicate with > base level code, this is fairly difficult to achieve. > > Even if A's interrupt handler > were reentrant, we could now have the situation that, since the nested call to A's > interrupt routine completes before the outer invocation of A's interrupt > routine, that the data read from/written to A is reversed, i.e., the nested > call's data arrive first, the outer call's data arrives second. E.g., > A is a UART and suddenly two adjoining characters are reversed. (In general, > arbitrary permutations of the data can be achieved using additional nested > invocations.) > > An additional problem is the additional stack space. Allowing premature > nested interrupts means that the space required by interrupt handlers > is unbounded. Thus the probability of the (interrupt) stack overflowing is > greater than zero. If interrupt handler nest strictly, the stack is bounded > by the sum of the requirements of each interrupt handler. ... snip ... You can fix all this (except the stack space) by letting each routine only re-enable interrupts just before exit. Most systems have provision to re-enable only after the interrupted context is restored, which fixes the stack space problem. So with proper design you CAN get a workable system with individual service routines. However you can't fix the inherent overhead of multiple (unnecessary) context switches, which are likely to be the major portion of the service time. -- Chuck F (cbfalconer@yahoo.com) (cbfalconer@XXXXworldnet.att.net) (Remove "XXXX" from reply address. yahoo works unmodified) mailto:uce@ftc.gov (for spambots to harvest) ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Wed, 12 Sep 01 09:35:00 GMT Organization: UltraNet Communications, Inc. Lines: 21 Message-ID: <9nnk4k$r4v$8@bob.news.rcn.net> References: <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <3B9E42A0.97ABB794@hda.hydro.com> <3B9E4CDC.198D6274@mchpDOTsiemens.de> X-Trace: UmFuZG9tSVbstO4YClLNxnOc62EV0pkpCVu5ZJzwkYC20T+os9udMgSCBVMwdXij X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 12 Sep 2001 12:23:48 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-255-128 Xref: chonsp.franklin.ch alt.folklore.computers:90251 In article <3B9E4CDC.198D6274@mchpDOTsiemens.de>, Konrad Schwarz wrote: >NB I hope I'm not coming across as too brusque. I've thought about >this problem for some time and I think I have understood it pretty >completely. I have also seen a lot of suspect code and I think >many people who do this type of work haven't understood all >the ramifications (premature optimization the root of all evil, >etc.). However, there is always the chance that I've overlooked >something. There is something you've overlooked. Take a look at any TOPS-10 SMP implementation; it will be version 7.nn. Actually take a look at any PDP-10 implementation especially TOPS-10 or TOPS-20. They knew how to take prioritized interrupts, do the minimal necessary and be ready for the next one. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Wed, 12 Sep 01 09:22:46 GMT Organization: UltraNet Communications, Inc. Lines: 36 Message-ID: <9nnjdl$r4v$4@bob.news.rcn.net> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9nahnc$f38$14@bob.news.rcn.net> <090920011220344692%elliott@yrl.co.uk> <9ng6t6$1gg$1@i4got.pechter.dyndns.org> <9ni7ck$c78$2@bob.news.rcn.net> <_Ccn7.24433$bl4.4553314@typhoon.ne.mediaone.net> <1bae02s9ub.fsf@cs.nmsu.edu> <9nksib$ast$10@bob.news.rcn.net> <1br8tdiznn.fsf@cs.nmsu.edu> X-Trace: UmFuZG9tSVbbA4CcJQzgKJNp2Ayr0f8jwQO8fxhHYFSZM6C9uUpH6m2DxmHKWmBh X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 12 Sep 2001 12:11:33 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.mailgate.org!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-255-128 Xref: chonsp.franklin.ch alt.folklore.computers:90255 In article <1br8tdiznn.fsf@cs.nmsu.edu>, Joe Pfeiffer wrote: >jmfbahciv@aol.com writes: >> >Yep. Meanwhile, their hardware turned into crap. I knew guys who >> >told me stories about DEC tape drives on board ships in the 1970s, out >> >on deck, in salt spray, surviving forever. >> >> That's because they weren't our tape drives (unless you're >> talking about DECtape which was the best magnetic media ever). > >Yep, I do mean DECtapes. Oh, then of course it ran and read anywhere. There are legends based on DECtapes. >> >> > ... Meanwhile every time El >> >Paso Electric flickered our lights (which happens pretty frequently) >> >another VT-220 terminal would pop its power supply. >> >> We never made a decent terminal or magtape drive. If it was >> designed and manufactured within DEC, it sucked. I never >> did figure that one out since we did make DECtapes, disks, >> and CPUs that were super. > >I remember VT-100s as being very solid. They were a bad ergonomic design. And the keyboard layout still sucked. I should have whacked Tom over the head in the early 70s when he produced the VT05 layout. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Wed, 12 Sep 01 09:26:03 GMT Organization: UltraNet Communications, Inc. Lines: 23 Message-ID: <9nnjjq$r4v$5@bob.news.rcn.net> References: <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <9nktc3$ast$12@bob.news.rcn.net> <1bofohizh8.fsf@cs.nmsu.edu> <3B9E432C.AE2A3FA5@hda.hydro.com> X-Trace: UmFuZG9tSVasLodwBcFLne9uHXKt8Dkqn806uVX9wx1sdWXXwX5FkcXyDb8b3HhM X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 12 Sep 2001 12:14:50 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-255-128 Xref: chonsp.franklin.ch alt.folklore.computers:90257 In article <3B9E432C.AE2A3FA5@hda.hydro.com>, Terje Mathisen wrote: >Joe Pfeiffer wrote: >> >> jmfbahciv@aol.com writes: >> > >> > But a driver doesn't get involved during an interrupt...does it? >> > I thought the role of a driver is to drive a hardware device. >> > It's only when that device has to say "OK, I'm done; what do I >> > do next?" or "Something's wrong!" that an interrupt occurs. >> >> Yes -- the device's interrupt handler is part of its driver. > >On a small system the hw interrupt handler can be more or less all of >the driver. No. That's an aspect of small computer thinking. The interrupt handler should be as short as possible, leaving the device driving business to the core of the driver. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Wed, 12 Sep 01 09:31:15 GMT Organization: UltraNet Communications, Inc. Lines: 68 Message-ID: <9nnjtj$r4v$7@bob.news.rcn.net> References: <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> X-Trace: UmFuZG9tSVYtd9wueyQuwPnNZ7vFFCTcsAkfMExqCOHmemd4cD3DjvFPiSI1LE8A X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 12 Sep 2001 12:20:03 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-255-128 Xref: chonsp.franklin.ch alt.folklore.computers:90259 In article <3B9E06A0.F021007C@mchpDOTsiemens.de>, Konrad Schwarz wrote: >Terje Mathisen schrieb: >> >> CBFalconer wrote: >> > >> > Terje Mathisen wrote: >> > > It is sufficient that each interrupt driver will chain to the previous >> > > one (at least when there's no obvious casue for the current interrupt), >> > > and that the drivers are willing and capable of handling what looks (to >> > > them) as a spurious interrupt, i.e. with no real cause. >> > > >> > > This way each driver doesn't need to know anything about any other hw >> > > that happens to share the same interrupt line. >> > >> > Not quite. If you have multiple priority levels of interrupts, >> > you also have to re-enable the interrupt controller level, and >> > this shouldn't be done until all the possibilities have been >> > examined. Thus you either have to have a master service routine, >> > or ensure that only the last one in the chain does this job. >> >> If you re-enable it prematurely, you just suffer an immediate nested >> interrupt, but beside needing some extra stack resources, this should be >> OK. > >No. > >The previous interrupt level must become active only at the completion of the >complete interrupt handler. Why? The reason is the immediate nested interrupt. > >Say we have two devices A and B attached to an interrupt. (Actually the problem >is more general: A and B can be attached to different IRQs with different priorities. >The example also works with only one device A.) > >A posts an interrupt and the CPU fields it by calling A's interrupt handler (aka ISR). >A's interrupt handler enables interrupts prematurely and B posts an interrupt. >The CPU starts executing B's interrupt handler. > >B's interrupt handler enables interrupts prematurely, >A interrupts, CPU calls A's interrupt handler. > >We now have two copies of A's interrupt handler running. This means that A's >interrupt handler needs to be reentrant. Since A's interrupt handler is >updating global state to communicate with >base level code, this is fairly difficult to achieve. No it's not. IIRC, it's difficult to do under the Unix design. JMF and TW worked this stuff out when they designed SMP. Fix you line lenght. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Wed, 12 Sep 01 09:28:13 GMT Organization: UltraNet Communications, Inc. Lines: 25 Message-ID: <9nnjns$r4v$6@bob.news.rcn.net> References: <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <9nktc3$ast$12@bob.news.rcn.net> <3B9E6481.B7DC2B27@yahoo.com> X-Trace: UmFuZG9tSVaA/rwT7tzrVK9fMIz3uWKdNUvQHTUBoArNlMm7IR52p3LtJnf+SpY7 X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 12 Sep 2001 12:17:00 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-255-128 Xref: chonsp.franklin.ch alt.folklore.computers:90261 In article <3B9E6481.B7DC2B27@yahoo.com>, CBFalconer wrote: >jmfbahciv@aol.com wrote: >> >.... snip ... >> > >> >I.e. only if you re-enable interrupts would it be OK to return >> >immediately, instead of first chaining to the next driver in the list. >> >> But a driver doesn't get involved during an interrupt...does it? >> I thought the role of a driver is to drive a hardware device. >> It's only when that device has to say "OK, I'm done; what do I >> do next?" or "Something's wrong!" that an interrupt occurs. > >Yes it does. Something has to accept the partial i/o operation >and buffer it. Even if you have hardware for this, at some point >it will get full. You don't have to do it all at interrupt level. All that needs to be done is save the info so that the processor needs when coming along to manage the oops that caused the interrupt (like I/O done). /BAH Subtract a hundred and four for e-mail. ###### Message-ID: <3B9EE907.F4F6EB78@ev1.net> From: Charles Richmond Reply-To: richmond@ev1.net Organization: Cannine Computer Center X-Mailer: Mozilla 4.7C-CCK-MCD {C-UDP; EBM-APPLE} (Macintosh; I; PPC) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <9nktc3$ast$12@bob.news.rcn.net> <3B9E6481.B7DC2B27@yahoo.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 37 Date: Wed, 12 Sep 2001 02:50:35 GMT NNTP-Posting-Host: 24.179.111.125 X-Complaints-To: abuse@home.net X-Trace: news2.rdc2.tx.home.com 1000263035 24.179.111.125 (Tue, 11 Sep 2001 19:50:35 PDT) NNTP-Posting-Date: Tue, 11 Sep 2001 19:50:35 PDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!enews.sgi.com!newshub2.rdc1.sfba.home.com!news.home.com!news2.rdc2.tx.home.com.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90273 CBFalconer wrote: > > jmfbahciv@aol.com wrote: > > > ... snip ... > > > > > >I.e. only if you re-enable interrupts would it be OK to return > > >immediately, instead of first chaining to the next driver in the list. > > > > But a driver doesn't get involved during an interrupt...does it? > > I thought the role of a driver is to drive a hardware device. > > It's only when that device has to say "OK, I'm done; what do I > > do next?" or "Something's wrong!" that an interrupt occurs. > > Yes it does. Something has to accept the partial i/o operation > and buffer it. Even if you have hardware for this, at some point > it will get full. > At least the maskable interrupt on the Motorola MC6800 caused *all* the registers to be pushed into the stack (well, maybe *not* the stack pointer). IIRC, with the 8080, the interrupt service routine (ISR) could begin by pushing whatever registers that might change... To return from the ISR, you used a RTI (return from interrupt) instruction, which restored the registers from the stack. I have the source (on paper) to a Tiny BASIC interpreter that used this fact to implement the usr() function (that calls a machine language routine). usr() took parameters which were to be used for register contents. Tiny BASIC pushed these on the stack and pushed the address of the machine language routine...then did an RTI to pop all this junk into the registers and branch to the machine language routine... -- +-------------------------------------------------------------+ | Charles and Francis Richmond | +-------------------------------------------------------------+ ###### From: Terje Mathisen Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Wed, 12 Sep 2001 08:33:52 +0200 Organization: Hydro Lines: 131 Message-ID: <3B9F01D0.CF028AB6@hda.hydro.com> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <3B9E42A0.97ABB794@hda.hydro.com> <3B9E4CDC.198D6274@mchpDOTsiemens.de> NNTP-Posting-Host: 136.164.110.166 Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.77 [en] (Windows NT 5.0; U) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.tele.dk!small.news.tele.dk!148.122.208.68!news2.oke.nextra.no!nextra.com!hydro.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90246 Konrad Schwarz wrote: > > Terje Mathisen schrieb: > > > We now have two copies of A's interrupt handler running. This means that A's > > > interrupt handler needs to be reentrant. Since A's interrupt handler is > > > updating global state to communicate with > > > base level code, this is fairly difficult to achieve. > > > > It might be difficult, and I have made a few mistakes here, but it > > wasn't really that hard to write a serial port driver that allows > > multiple (4) different interrupt sources from each port, as well as > > having multiple ports sharing the same interrupt. > > What benefit was gained by reenabling interrupts early? Oh, I see that we've been talking past each other! Most of my comments, particularly the one about non-cooperating drivers sharing an interrupt signal, was _not_ based on x86 code I've written. As I (not clearly enough) stated elsewhere, the only drivers I've written which shared interrupts between multiple sources used a single driver to handle all of them. Since x86 didn't have sharable interrupts, my speculations were partly theoretical, but I'll get back to that below. > > Yes, the parts of the code that can run reentrantly obviously needs to > > be reentrant, but the only really critical parts are the code sequences > > that reads some hardware and tranfers that info to some software > > structure. > > How did you synchronize access to the software structure? Did you using > non-blocking methods and the help of atomic instructions like > Compare & Swap or did you raise the interrupt priority level > of the processor while modifying the software structure? The latter. > > This was not a real problem. > > Testing can show only the presence of bugs, never their absence. OK, 'not a real problem' as in: It was possible to prove (to my own satisfaction) that there was no possible way for a nested interrupt to occur within the critical windows, since the relevant interrupts were blocked for the duration. > What would have happened if additional, higher-priority interrupts > were being handled in your system? And if the higher-priority interrupt's > handler took a long time to complete...say so > long that the serial port caused a second interrupt? On an unbuffered serial port (i.e. original PC) obviously no higher-priority interrupts can ever be allowed to take longer than the gap between two incoming characters. When this sometimes did happen (i.e. badly written code that chained into the clock interrupt), data was indeed lost. :-( > > This is correct, and the reason why I always wrote my code to put > > absolute limits to how much stack could ever be used in a worst-case > > situation. > > And what did you do then? Drop the character? The only absolute requirement (as in hard real time) is (of course) that the sum of all possible interrupt sources must take less time to handle than what's actually available. It this is false, then the only thing you can do is to make sure that your own code is part of the solution (i.e. as efficient as possible) instead of a part of the problem. :-( I did try to rotate the priority levels, making serial port interrupts the the highest, but since this left the clock interrupt near the bottom of the list, it could cause other/more serious problems. On an AT which nests the 8 secondary interrupt lines on top of IRQ 2, the IRQ 3 & 4 which is allocated to serial ports is even further down the list. > NB I hope I'm not coming across as too brusque. I've thought about > this problem for some time and I think I have understood it pretty > completely. I have also seen a lot of suspect code and I think > many people who do this type of work haven't understood all > the ramifications (premature optimization the root of all evil, > etc.). However, there is always the chance that I've overlooked > something. I believe we (mostly) agree, it is just that I believe that it would indeed be possible to share interrupts even if one or some of the drivers wasn't written totally according to the rules. I.e. Assume source A and B share a signal, with A first in line to handle the interrupt signal: I have the following alternatives: 1. A signals an event, but not B: In this case driver-A handles the event, and can either chain to the driver-B which finds nothing to do (before re-enabling interrupts), or driver-A can re-enable and return immediately. Both is obviously OK. 2. B signals, but not A: Driver-A determines that it has no even to handle and must therefore chain to B. B handles the event and re-enables interrupts. 3. Both A and B: Worst case: Driver-A handles A and re-enables interrupts prior to returning. A nested interrupt will reinvoke driver-A, which now finds nothing to do, and therefore must chan to B. Driver-B handles B, re-enables interrupts and returns, whereupon the first invocation finally returns to the interrupted code. Different variations like A being signalled after driver-B has started is also OK, as long as the total processing time for all possible interrupt sources is less than the minimum time between interrupts from the fastest generator. Yes, I do agree that it would be better (and cleaner) for all the drivers to cooperate! Terje -- - Using self-discipline, see http://www.eiffel.com/discipline "almost all programming can be viewed as an exercise in caching" ###### From: Terje Mathisen Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Wed, 12 Sep 2001 08:38:22 +0200 Organization: Hydro Lines: 31 Message-ID: <3B9F02DE.4128D006@hda.hydro.com> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E639B.EF5A1AA2@yahoo.com> NNTP-Posting-Host: 136.164.110.166 Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.77 [en] (Windows NT 5.0; U) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.tele.dk!small.news.tele.dk!148.122.208.68!news2.oke.nextra.no!nextra.com!hydro.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90245 CBFalconer wrote: > > Terje Mathisen wrote: > > I.e. only if you re-enable interrupts would it be OK to return > > immediately, instead of first chaining to the next driver in the list. > > To my mind, horrible inefficiencies at the most critical points. > That would require completely restoring the CPU state before the > next interrupt happens. Interrupt stack space is VERY likely to > be a scarce resource. Relax, we agree! I am just pointing out that in some situations it is possible to have the code 'work' even if some parts of it isn't written with cooperation in mind. > I once found I could create an almost total denial of service on > an HP3000 by sending it a continuous stream of s at 2400 baud, > each of which created a CPU interrupt (there was no i/o processor > on that version). I rapidly changed the communicating application > to supply the 'keep-alive' operation at 10 second intervals. BTDT. A Norsk Data ND machine likewise had unbuffered serial cards, so a single typist could overflow it when repainting at the end of the screen. Terje -- - Using self-discipline, see http://www.eiffel.com/discipline "almost all programming can be viewed as an exercise in caching" ###### Sender: lynn@LYNNPC Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <3B9E42A0.97ABB794@hda.hydro.com> <3B9E4CDC.198D6274@mchpDOTsiemens.de> <9nnk4k$r4v$8@bob.news.rcn.net> Reply-To: Anne & Lynn Wheeler From: Anne & Lynn Wheeler Message-ID: Organization: Wheeler&Wheeler Lines: 22 User-Agent: Gnus/5.0808 (Gnus v5.8.8) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Wed, 12 Sep 2001 14:17:05 GMT NNTP-Posting-Host: 199.174.230.119 X-Complaints-To: abuse@earthlink.net X-Trace: newsread2.prod.itd.earthlink.net 1000304225 199.174.230.119 (Wed, 12 Sep 2001 07:17:05 PDT) NNTP-Posting-Date: Wed, 12 Sep 2001 07:17:05 PDT X-Received-Date: Wed, 12 Sep 2001 07:13:19 PDT (newsmaster1.prod.itd.earthlink.net) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!1164084!news.imp.ch!sunqbc.risq.qc.ca!cpk-news-hub1.bbnplanet.com!lsanca1-snf1!news.gtei.net!newsfeed2.earthlink.net!newsfeed.earthlink.net!newsmaster1.prod.itd.earthlink.net!newsread2.prod.itd.earthlink.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90265 jmfbahciv@aol.com writes: > > There is something you've overlooked. Take a look at any TOPS-10 > SMP implementation; it will be version 7.nn. Actually take > a look at any PDP-10 implementation especially TOPS-10 or > TOPS-20. They knew how to take prioritized interrupts, do > the minimal necessary and be ready for the next one. There was a trade-off in CP/67 and the revision of the "CP" kernel for VM/370. In CP/67, it would handle I/O interrupt up-thru "channel level" processing (i.e. things specific to channel level operation, i/o operations tended to be viewed as hierarchy: channel, control unit, device) and then flag things clear and be ready for next interrupt (and/or another processor could do some things). In VM/370, processing went thru until device-level processing was complete ... there were some increased CPU pathlength overhead issues in CP/67 allowing the earlier processing (namely the processing emboddied in the "CHFREE" macro ... aka channel free ... which disappeared in VM/370). -- Anne & Lynn Wheeler | lynn@garlic.com - http://www.garlic.com/~lynn/ ###### From: mirian@trantor.cosmic.com (Mirian Crzig Lennox) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> Message-ID: User-Agent: slrn/0.9.6.4 (Linux) Lines: 20 Date: Wed, 12 Sep 2001 14:23:32 GMT NNTP-Posting-Host: 209.58.189.187 X-Complaints-To: abuse@shore.net X-Trace: news.shore.net 1000304612 209.58.189.187 (Wed, 12 Sep 2001 10:23:32 EDT) NNTP-Posting-Date: Wed, 12 Sep 2001 10:23:32 EDT Organization: Shore.Net, a PRIMUS Company (info@shore.net) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news.shore.net!mirian Xref: chonsp.franklin.ch alt.folklore.computers:90275 On Wed, 12 Sep 01 09:31:15 GMT, jmfbahciv@aol.com wrote: >> >>We now have two copies of A's interrupt handler running. This means that >A's >>interrupt handler needs to be reentrant. Since A's interrupt handler is >>updating global state to communicate with >>base level code, this is fairly difficult to achieve. > >No it's not. IIRC, it's difficult to do under the Unix >design. JMF and TW worked this stuff out when they designed >SMP. And, in fact, Linux deviates from the traditional Unix design for just this reason (SMP). Under Linux, an interrupt handler does the bare minimum to shove the necessary information about the request onto a task queue, and then immediately returns. All the I/O and whatnot related to the interrupt is done in a thread-safe context, as it should. --Mirian ###### Message-ID: <3B9F74E0.34A77C3@yahoo.com> From: CBFalconer Reply-To: cbfalconer@worldnet.att.net Organization: Ched Research X-Mailer: Mozilla 4.75 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <9nktc3$ast$12@bob.news.rcn.net> <3B9E6481.B7DC2B27@yahoo.com> <3B9EE907.F4F6EB78@ev1.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 25 Date: Wed, 12 Sep 2001 15:32:58 GMT NNTP-Posting-Host: 12.90.168.80 X-Complaints-To: abuse@worldnet.att.net X-Trace: bgtnsc04-news.ops.worldnet.att.net 1000308778 12.90.168.80 (Wed, 12 Sep 2001 15:32:58 GMT) NNTP-Posting-Date: Wed, 12 Sep 2001 15:32:58 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeeds.belnet.be!news.belnet.be!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!howland.erols.net!news-out.worldnet.att.net.MISMATCH!wn3feed!worldnet.att.net!135.173.83.71!wnfilter1!worldnet-localpost!bgtnsc04-news.ops.worldnet.att.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90276 Charles Richmond wrote: > > > ... snip ... > > > At least the maskable interrupt on the Motorola MC6800 caused *all* > the registers to be pushed into the stack (well, maybe *not* the stack > pointer). IIRC, with the 8080, the interrupt service routine (ISR) could > begin by pushing whatever registers that might change... To return > from the ISR, you used a RTI (return from interrupt) instruction, which > restored the registers from the stack. No, that was one of the nuisances with the 8080 - there was no RTI. Instead EI delayed execution for one instruction, so the sequence EI RET could exit completely. A major failing in the 8080 instruction set was that there was no way to record whether interrupts were enabled, and this often required duplicating routines, or nesting them. The Z80 fixed this AFAIR, but the 8086/8 was better in having RTI auto-pop the flags from the stack. -- Chuck F (cbfalconer@yahoo.com) (cbfalconer@XXXXworldnet.att.net) (Remove "XXXX" from reply address. yahoo works unmodified) mailto:uce@ftc.gov (for spambots to harvest) ###### Message-ID: <3B9F75F9.F3F844B0@yahoo.com> From: CBFalconer Reply-To: cbfalconer@worldnet.att.net Organization: Ched Research X-Mailer: Mozilla 4.75 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <3B9E42A0.97ABB794@hda.hydro.com> <3B9E4CDC.198D6274@mchpDOTsiemens.de> <3B9F01D0.CF028AB6@hda.hydro.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 18 Date: Wed, 12 Sep 2001 15:32:59 GMT NNTP-Posting-Host: 12.90.168.80 X-Complaints-To: abuse@worldnet.att.net X-Trace: bgtnsc04-news.ops.worldnet.att.net 1000308779 12.90.168.80 (Wed, 12 Sep 2001 15:32:59 GMT) NNTP-Posting-Date: Wed, 12 Sep 2001 15:32:59 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!howland.erols.net!news-out.worldnet.att.net.MISMATCH!wn3feed!worldnet.att.net!135.173.83.71!wnfilter1!worldnet-localpost!bgtnsc04-news.ops.worldnet.att.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90243 Terje Mathisen wrote: > ... snip ... > > As I (not clearly enough) stated elsewhere, the only drivers I've > written which shared interrupts between multiple sources used a single > driver to handle all of them. Since x86 didn't have sharable interrupts, > my speculations were partly theoretical, but I'll get back to that > below. The X86 *has* shareable interrupts, the ISA bus does not. -- Chuck F (cbfalconer@yahoo.com) (cbfalconer@XXXXworldnet.att.net) (Remove "XXXX" from reply address. yahoo works unmodified) mailto:uce@ftc.gov (for spambots to harvest) ###### Message-ID: <3B9F787B.C06809B0@yahoo.com> From: CBFalconer Reply-To: cbfalconer@worldnet.att.net Organization: Ched Research X-Mailer: Mozilla 4.75 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <3B9E6C3F.1D45BA0D@yahoo.com> <3B9F3764.A3BF626@mchpDOTsiemens.de> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 37 Date: Wed, 12 Sep 2001 15:33:01 GMT NNTP-Posting-Host: 12.90.168.80 X-Complaints-To: abuse@worldnet.att.net X-Trace: bgtnsc04-news.ops.worldnet.att.net 1000308781 12.90.168.80 (Wed, 12 Sep 2001 15:33:01 GMT) NNTP-Posting-Date: Wed, 12 Sep 2001 15:33:01 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.mathworks.com!wn3feed!worldnet.att.net!135.173.83.71!wnfilter1!worldnet-localpost!bgtnsc04-news.ops.worldnet.att.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90242 Konrad Schwarz wrote: > > CBFalconer schrieb: > > You can fix all this (except the stack space) by letting each > > routine only re-enable interrupts just before exit. Most systems > > have provision to re-enable only after the interrupted context is > > restored, which fixes the stack space problem. > > The interresting thing about RFE is that it both returns > to the previous PC *and* unmasks interrupts in the CPU. > > A CPU that does not have this cannot deal with interrupts > at all, irrespective of interrupt chaining, because of > incorrect stack nesting. You have two things to control (which may not both exist on any particular processor) - the master interrupt enable, and the priority level. You can use the master to ensure proper handling of the priority. The sequence is: 1. DI (often automatic) 2. Save current priority level 3. Set priority level to allow higher priorities 4. EI 5. Body of service 6. DI 7. Restore priority level 8. EI (may be part of the RET) 9. RET You can look at all this from a higher level as critical sections. -- Chuck F (cbfalconer@yahoo.com) (cbfalconer@XXXXworldnet.att.net) (Remove "XXXX" from reply address. yahoo works unmodified) mailto:uce@ftc.gov (for spambots to harvest) ###### From: Greg Menke Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 12 Sep 2001 12:34:09 -0400 Organization: MindSpring Enterprises Lines: 23 Sender: gregm@europa.pienet Message-ID: References: <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <3B9F82B0.92BDFA34@mchpDOTsiemens.de> NNTP-Posting-Host: a5.f7.5b.ec X-Server-Date: 12 Sep 2001 16:34:12 GMT X-Newsreader: Gnus v5.7/Emacs 20.7 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!newsfeed.r-kom.de!news0.de.colt.net!colt.net!nycmny1-snf1.gtei.net!nycmny1-snh1.gtei.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!feed2.news.rcn.net!rcn!newsfeed1.earthlink.net!newsfeed2.earthlink.net!newsfeed.earthlink.net!news.mindspring.net!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90252 > > And, in fact, Linux deviates from the traditional Unix design for just > > this reason (SMP). Under Linux, an interrupt handler does the bare > > minimum to shove the necessary information about the request onto a task > > queue, and then immediately returns. All the I/O and whatnot related to > > the interrupt is done in a thread-safe context, as it should. > > To me, this seems like a design-trade off, at least for a uni-processor: > Using a thread-safe context to handle the interrupt may be simpler to > program for (since all calls are available) but undoubtedly increases > the interrupt latency and the CPU cycles required to handle the > interrupt. It depends a lot on the OS archtecture. In a priority scheduling regime, if the ISR wakes up a high priority "service" thread, it can start immediately after the ISR code returns. The overhead can be small and the gain from doing the I/O service in a less constrained context can be considerable. In this case, extended interrupt processing can be handled within the normal task architecture of the system. Gregm ###### Message-ID: <3B9FA31B.7630B371@thinkage.ca> From: "Alan T. Bowler" Organization: Thinkage Ltd. X-Mailer: Mozilla 4.5 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <3B9E42A0.97ABB794@hda.hydro.com> <3B9E4CDC.198D6274@mchpDOTsiemens.de> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 8 Date: Wed, 12 Sep 2001 14:02:03 -0400 NNTP-Posting-Host: 192.102.11.4 X-Trace: nnrp1.uunet.ca 1000317704 192.102.11.4 (Wed, 12 Sep 2001 14:01:44 EDT) NNTP-Posting-Date: Wed, 12 Sep 2001 14:01:44 EDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!sunqbc.risq.qc.ca!news.uunet.ca!nnrp1.uunet.ca.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90244 Konrad Schwarz wrote: > > Testing can show only the presence of bugs, never their absence. It can show the absence of specific historical bugs. I.e. an evolving test suite can show that you have not reintroduced an old bug. Unfortunately, it does not guarentee that you have not created a close varient. ###### From: Terje Mathisen Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Wed, 12 Sep 2001 21:35:53 +0200 Organization: Hydro Lines: 23 Message-ID: <3B9FB919.48EA7CD9@hda.hydro.com> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <20010902204807.26de8fd5.steveo@eircom.net> <3B9280CC.A39F9E14@jetnet.ab.ca> <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <3B9E42A0.97ABB794@hda.hydro.com> <3B9E4CDC.198D6274@mchpDOTsiemens.de> <3B9F01D0.CF028AB6@hda.hydro.com> <3B9F75F9.F3F844B0@yahoo.com> NNTP-Posting-Host: iabrahmspc.nho.hydro.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.77 [en] (Windows NT 5.0; U) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!newscore.gigabell.net!news0.de.colt.net!colt.net!pop-news-1.colt-telecom.nl!newsgate.cistron.nl!news.tele.dk!small.news.tele.dk!148.122.208.68!news2.oke.nextra.no!nextra.com!hydro.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90280 CBFalconer wrote: > > Terje Mathisen wrote: > > > ... snip ... > > > > As I (not clearly enough) stated elsewhere, the only drivers I've > > written which shared interrupts between multiple sources used a single > > driver to handle all of them. Since x86 didn't have sharable interrupts, > > my speculations were partly theoretical, but I'll get back to that > > below. > > The X86 *has* shareable interrupts, the ISA bus does not. I'll spell it out completely: Since all available (to me) x86-based machines (i.e. IBM-compatible PCs) used edge-triggered interrupts they weren't shareable. OK? Terje -- - Using self-discipline, see http://www.eiffel.com/discipline "almost all programming can be viewed as an exercise in caching" ###### Message-ID: <3B9FBEE0.B24B0161@yahoo.com> From: CBFalconer Reply-To: cbfalconer@worldnet.att.net Organization: Ched Research X-Mailer: Mozilla 4.75 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <9nktc3$ast$12@bob.news.rcn.net> <3B9E6481.B7DC2B27@yahoo.com> <9nnjns$r4v$6@bob.news.rcn.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 55 Date: Wed, 12 Sep 2001 20:26:26 GMT NNTP-Posting-Host: 12.90.173.80 X-Complaints-To: abuse@worldnet.att.net X-Trace: bgtnsc06-news.ops.worldnet.att.net 1000326386 12.90.173.80 (Wed, 12 Sep 2001 20:26:26 GMT) NNTP-Posting-Date: Wed, 12 Sep 2001 20:26:26 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!3454976!news.imp.ch!sunqbc.risq.qc.ca!cpk-news-hub1.bbnplanet.com!news.gtei.net!howland.erols.net!news-out.worldnet.att.net.MISMATCH!wn3feed!worldnet.att.net!135.173.83.71!wnfilter1!worldnet-localpost!bgtnsc06-news.ops.worldnet.att.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90277 jmfbahciv@aol.com wrote: > > In article <3B9E6481.B7DC2B27@yahoo.com>, > CBFalconer wrote: > >jmfbahciv@aol.com wrote: > >> > >.... snip ... > >> > > >> >I.e. only if you re-enable interrupts would it be OK to return > >> >immediately, instead of first chaining to the next driver in the list. > >> > >> But a driver doesn't get involved during an interrupt...does it? > >> I thought the role of a driver is to drive a hardware device. > >> It's only when that device has to say "OK, I'm done; what do I > >> do next?" or "Something's wrong!" that an interrupt occurs. > > > >Yes it does. Something has to accept the partial i/o operation > >and buffer it. Even if you have hardware for this, at some point > >it will get full. > > You don't have to do it all at interrupt level. All that needs to > be done is save the info so that the processor needs when coming > along to manage the oops that caused the interrupt (like I/O done). Yes you do, up to a point. Otherwise the service process can get delayed and the buffer overwritten. For example, an incoming UART interrupt will store a char in a buffer, using an index, advance the index, and exit. If the char is eoln, it will also signal the service routine and switch to a free buffer. Now the service process has line time, not char time, to handle the servicing. We don't have to worry about overruns unless we run out of buffers, or the hardware overrun itself appears (maybe due to tie ups in a higher priority interrupt). The delay of the service process doesn't even imply multi-tasking. A single thread program just has to get busy and fail to poll the buffer often enough. A single char buffer is no better than a hardware poll, it just postpones the problem for one char time. So you have to think about throughput, using the buffers to iron out the bumps. If you try to use a single circular buffer you have to worry about concurrent processes connected by a buffer. You have to worry about critical sections and synchronization. If you don't worry about these things, you probably work for some firm with an M in its name. Big iron or little iron, the problems are the same. -- Chuck F (cbfalconer@yahoo.com) (cbfalconer@XXXXworldnet.att.net) (Remove "XXXX" from reply address. yahoo works unmodified) mailto:uce@ftc.gov (for spambots to harvest) ###### Message-ID: <3B9FC02F.E13FB7CA@yahoo.com> From: CBFalconer Reply-To: cbfalconer@worldnet.att.net Organization: Ched Research X-Mailer: Mozilla 4.75 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 30 Date: Wed, 12 Sep 2001 20:26:29 GMT NNTP-Posting-Host: 12.90.173.80 X-Complaints-To: abuse@worldnet.att.net X-Trace: bgtnsc06-news.ops.worldnet.att.net 1000326389 12.90.173.80 (Wed, 12 Sep 2001 20:26:29 GMT) NNTP-Posting-Date: Wed, 12 Sep 2001 20:26:29 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!3635287!news.imp.ch!sunqbc.risq.qc.ca!cpk-news-hub1.bbnplanet.com!news.gtei.net!newsfeed1.cidera.com!Cidera!cyclone1.gnilink.net!wn3feed!worldnet.att.net!135.173.83.71!wnfilter1!worldnet-localpost!bgtnsc06-news.ops.worldnet.att.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90278 Mirian Crzig Lennox wrote: > > On Wed, 12 Sep 01 09:31:15 GMT, jmfbahciv@aol.com wrote: > >> > >>We now have two copies of A's interrupt handler running. This means that > >A's > >>interrupt handler needs to be reentrant. Since A's interrupt handler is > >>updating global state to communicate with > >>base level code, this is fairly difficult to achieve. > > > >No it's not. IIRC, it's difficult to do under the Unix > >design. JMF and TW worked this stuff out when they designed > >SMP. > > And, in fact, Linux deviates from the traditional Unix design for just > this reason (SMP). Under Linux, an interrupt handler does the bare > minimum to shove the necessary information about the request onto a task > queue, and then immediately returns. All the I/O and whatnot related to > the interrupt is done in a thread-safe context, as it should. As I understand it Linux can postpone a portion of the interrupt service to what is effectively a super high priority process (or set of processes) that are executed between completion of interrupt service and return to the current task. -- Chuck F (cbfalconer@yahoo.com) (cbfalconer@XXXXworldnet.att.net) (Remove "XXXX" from reply address. yahoo works unmodified) mailto:uce@ftc.gov (for spambots to harvest) ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Thu, 13 Sep 01 13:07:53 GMT Organization: UltraNet Communications, Inc. Lines: 29 Message-ID: <9nql04$1hl$16@bob.news.rcn.net> References: <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <3B9F82B0.92BDFA34@mchpDOTsiemens.de> <3BA071A8.CA2DBB0E@mchpDOTsiemens.de> X-Trace: UmFuZG9tSVbThowlCnwodox5W7PolRj5/hfwg7PvYRo9aNitWuLtuKGx3kBY5N9c X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 13 Sep 2001 15:56:52 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-131 Xref: chonsp.franklin.ch alt.folklore.computers:90299 In article <3BA071A8.CA2DBB0E@mchpDOTsiemens.de>, Konrad Schwarz wrote: > > >Greg Menke schrieb: >> It depends a lot on the OS archtecture. In a priority scheduling >> regime, if the ISR wakes up a high priority "service" thread, it can >> start immediately after the ISR code returns. The overhead can be >> small and the gain from doing the I/O service in a less constrained >> context can be considerable. In this case, extended interrupt >> processing can be handled within the normal task architecture of the >> system. > >A quick return to base-level code is obviously desirable: the full complement >of OS facilities is available only there (esp. the ability to block). > >However, I see no compelling reason to architect specialized interrupt >service "threads" into the OS (that can do much more than interrupt handlers >but perhaps not everything that ordinary threads can do), except when the >scheduler for ordinary threads is so slow that a quick path is required. Think. You've got 500 users, all demanding attention; then the operator hits the halt button on CPU0 (or the boot CPU). /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Thu, 13 Sep 01 13:14:20 GMT Organization: UltraNet Communications, Inc. Lines: 47 Message-ID: <9nqlc7$1hl$17@bob.news.rcn.net> References: <_Ccn7.24433$bl4.4553314@typhoon.ne.mediaone.net> <1bae02s9ub.fsf@cs.nmsu.edu> <9nksib$ast$10@bob.news.rcn.net> <1br8tdiznn.fsf@cs.nmsu.edu> <9nnjdl$r4v$4@bob.news.rcn.net> <3B9FF71E.8B8ED2E3@ev1.net> <1192.655T2092T10255577@nowhere.in.particular> X-Trace: UmFuZG9tSVYlx1n4k4KZ0xRHs4Dye1dCIFF35FW32QrG1fZQORkkQRItibrViCdh X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 13 Sep 2001 16:03:19 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!57018!news.imp.ch!uni-erlangen.de!news-nue1.dfn.de!news-han1.dfn.de!news-koe1.dfn.de!do.de.uu.net!lnewspeer01.lnd.ops.eu.uu.net!emea.uu.net!newsfeed.icl.net!netnews.com!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-131 Xref: chonsp.franklin.ch alt.folklore.computers:90302 In article <1192.655T2092T10255577@nowhere.in.particular>, "Charlie Gibbs" wrote: >In article <3B9FF71E.8B8ED2E3@ev1.net> richmond@ev1.net >(Charles Richmond) writes: > >>jmfbahciv@aol.com wrote: >> >>> In article <1br8tdiznn.fsf@cs.nmsu.edu>, >>> Joe Pfeiffer wrote: >>> > >>> > [snip...] [snip...] [snip...] >>> > >>> >I remember VT-100s as being very solid. >>> >>> They were a bad ergonomic design. And the keyboard layout still >>> sucked. I should have whacked Tom over the head in the early >>> 70s when he produced the VT05 layout. >>> >>But the VT-100 had that beautiful 10" green screen... (;-)) >> >>I *never* used a VT-100 much, but friends who did complained >>most about the keyboard being hard to type on. They said you >>had to push each key *straight* down to get it to work...and >>the key travel was a bit long... > >Still, I'd be willing to forgive that when you consider what they >did with that abominable VT-220 layout. That was my last terminal. I hated those. It seemed like we started off bad and just got worse. For typing, I guess the best we made was the VT52....at least you could lay your listing, coffee cup and ashtray on the top without shorting dangers. > ... The whole world was going >crazy with respect to keyboard layouts in those days: first the IBM >PC, then the VT-220, and Sperry->Unisys did some things that can >only be described as obscene. Keyboard design was set back 10 years. > The best terminal keyboard design that I've ever used was something we called a VT06. Not only could one type at it, one could look at it without eye strain even when florescent lights were flickering all over. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Thu, 13 Sep 01 12:59:29 GMT Organization: UltraNet Communications, Inc. Lines: 35 Message-ID: <9nqkgb$1hl$14@bob.news.rcn.net> References: <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> X-Trace: UmFuZG9tSVY6a7a0fEKhA0JtiPEYTiQ3kM7Nm1vyQRp+75kdHI8CyYCyduMqE11k X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 13 Sep 2001 15:48:27 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-131 Xref: chonsp.franklin.ch alt.folklore.computers:90308 In article , mirian@trantor.cosmic.com (Mirian Crzig Lennox) wrote: >On Wed, 12 Sep 01 09:31:15 GMT, jmfbahciv@aol.com wrote: >>> >>>We now have two copies of A's interrupt handler running. This means that >>A's >>>interrupt handler needs to be reentrant. Since A's interrupt handler is >>>updating global state to communicate with >>>base level code, this is fairly difficult to achieve. >> >>No it's not. IIRC, it's difficult to do under the Unix >>design. JMF and TW worked this stuff out when they designed >>SMP. > >And, in fact, Linux deviates from the traditional Unix design for just >this reason (SMP). Right. But it seems to me (based on JMF's gag reflex) that there's a fudge required that isn't very nice. > ...Under Linux, an interrupt handler does the bare >minimum to shove the necessary information about the request onto a task >queue, and then immediately returns. All the I/O and whatnot related to >the interrupt is done in a thread-safe context, as it should. But (big BUT here), I don't think it was easy or possible to make that non-interrupt I/O code reentrant. One of these days, I might even find JMF's and other guys' paper they wrote about their Unix SMP experience. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Thu, 13 Sep 01 13:03:49 GMT Organization: UltraNet Communications, Inc. Lines: 46 Message-ID: <9nqkog$1hl$15@bob.news.rcn.net> References: <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <3B9F82B0.92BDFA34@mchpDOTsiemens.de> X-Trace: UmFuZG9tSVZ3iwHfbwtgh7R9cVEQ3wuHArFdlnzyHtDdAtWWTUzsJUE7E1gjzjKq X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 13 Sep 2001 15:52:48 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!9938998!news.imp.ch!sunqbc.risq.qc.ca!cpk-news-hub1.bbnplanet.com!news.gtei.net!feed2.news.rcn.net!feed1.news.rcn.net!rcn!207-172-245-131 Xref: chonsp.franklin.ch alt.folklore.computers:90309 In article <3B9F82B0.92BDFA34@mchpDOTsiemens.de>, Konrad Schwarz wrote: >Mirian Crzig Lennox schrieb: >> And, in fact, Linux deviates from the traditional Unix design for just >> this reason (SMP). Under Linux, an interrupt handler does the bare >> minimum to shove the necessary information about the request onto a task >> queue, and then immediately returns. All the I/O and whatnot related to >> the interrupt is done in a thread-safe context, as it should. > >To me, this seems like a design-trade off, at least for a uni-processor: Most of this usually is. A lot of the trade-off also depends on the philosophy of the OS. >Using a thread-safe context to handle the interrupt may be simpler to >program for (since all calls are available) But they aren't. This can get into a mess. > ... but undoubtedly increases >the interrupt latency and the CPU cycles required to handle the >interrupt. It's nice if the hardware and software cooperate with each other. > >An old IBM book about the RS/6000 (sorry, no reference at hand) describes >their interrupt scheme that allows any processor in a SMP to handle >an interrupt from any device, using processor priorities: the "symmetric" >interrupt controllers dynamically routed the interrupt to the >processor running at the lowest interrupt priority. > >I'm not sure how much is gained by handling interrupts symmetrically >in a SMP: I don't expect the time spent in interrupt routines to >unbalance a machine. I.e., I would just handle the interrupts on >the CPU where it is convenient to do so (e.g., the interrupt line >is routed to that CPU). That's the goal. However, the hardware configuration may not be completely symmetrical. JMF and TW had to deal with that because of hardware config contraints, e.g, disk drives only having two ports so only two CPUs could start their I/O. /BAH Subtract a hundred and four for e-mail. ###### From: "Charlie Gibbs" Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 12 Sep 01 17:05:41 -0800 Organization: http://extra.newsguy.com Lines: 33 Message-ID: <1192.655T2092T10255577@nowhere.in.particular> References: <_Ccn7.24433$bl4.4553314@typhoon.ne.mediaone.net> <1bae02s9ub.fsf@cs.nmsu.edu> <9nksib$ast$10@bob.news.rcn.net> <1br8tdiznn.fsf@cs.nmsu.edu> <9nnjdl$r4v$4@bob.news.rcn.net> <3B9FF71E.8B8ED2E3@ev1.net> NNTP-Posting-Host: p-426.newsdawg.com X-Newsreader: THOR 2.5a (Amiga;TCP/IP) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.mathworks.com!pln-e!spln!dex!extra.newsguy.com!newsp.newsguy.com!news3 Xref: chonsp.franklin.ch alt.folklore.computers:90320 In article <3B9FF71E.8B8ED2E3@ev1.net> richmond@ev1.net (Charles Richmond) writes: >jmfbahciv@aol.com wrote: > >> In article <1br8tdiznn.fsf@cs.nmsu.edu>, >> Joe Pfeiffer wrote: >> > >> > [snip...] [snip...] [snip...] >> > >> >I remember VT-100s as being very solid. >> >> They were a bad ergonomic design. And the keyboard layout still >> sucked. I should have whacked Tom over the head in the early >> 70s when he produced the VT05 layout. >> >But the VT-100 had that beautiful 10" green screen... (;-)) > >I *never* used a VT-100 much, but friends who did complained >most about the keyboard being hard to type on. They said you >had to push each key *straight* down to get it to work...and >the key travel was a bit long... Still, I'd be willing to forgive that when you consider what they did with that abominable VT-220 layout. The whole world was going crazy with respect to keyboard layouts in those days: first the IBM PC, then the VT-220, and Sperry->Unisys did some things that can only be described as obscene. Keyboard design was set back 10 years. -- cgibbs@nowhere.in.particular (Charlie Gibbs) I'm switching ISPs - watch this space. ###### Message-ID: <3B9FF71E.8B8ED2E3@ev1.net> From: Charles Richmond Reply-To: richmond@ev1.net Organization: Cannine Computer Center X-Mailer: Mozilla 4.7C-CCK-MCD {C-UDP; EBM-APPLE} (Macintosh; I; PPC) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9nahnc$f38$14@bob.news.rcn.net> <090920011220344692%elliott@yrl.co.uk> <9ng6t6$1gg$1@i4got.pechter.dyndns.org> <9ni7ck$c78$2@bob.news.rcn.net> <_Ccn7.24433$bl4.4553314@typhoon.ne.mediaone.net> <1bae02s9ub.fsf@cs.nmsu.edu> <9nksib$ast$10@bob.news.rcn.net> <1br8tdiznn.fsf@cs.nmsu.edu> <9nnjdl$r4v$4@bob.news.rcn.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 24 Date: Wed, 12 Sep 2001 22:02:58 GMT NNTP-Posting-Host: 24.179.111.125 X-Complaints-To: abuse@home.net X-Trace: news2.rdc2.tx.home.com 1000332178 24.179.111.125 (Wed, 12 Sep 2001 15:02:58 PDT) NNTP-Posting-Date: Wed, 12 Sep 2001 15:02:58 PDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!1193625!news.imp.ch!sunqbc.risq.qc.ca!newsfeed.direct.ca!look.ca!newshub2.rdc1.sfba.home.com!news.home.com!news2.rdc2.tx.home.com.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90331 jmfbahciv@aol.com wrote: > > In article <1br8tdiznn.fsf@cs.nmsu.edu>, > Joe Pfeiffer wrote: > > > > [snip...] [snip...] [snip...] > > > >I remember VT-100s as being very solid. > > They were a bad ergonomic design. And the keyboard layout still > sucked. I should have whacked Tom over the head in the early > 70s when he produced the VT05 layout. > But the VT-100 had that beautiful 10" green screen... (;-)) I *never* used a VT-100 much, but friends who did complained most about the keyboard being hard to type on. They said you had to push each key *straight* down to get it to work...and the key travel was a bit long... -- +-------------------------------------------------------------+ | Charles and Francis Richmond | +-------------------------------------------------------------+ ###### From: stanb@dial.pipex.com (Stan Barr) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <_Ccn7.24433$bl4.4553314@typhoon.ne.mediaone.net> <1bae02s9ub.fsf@cs.nmsu.edu> <9nksib$ast$10@bob.news.rcn.net> <1br8tdiznn.fsf@cs.nmsu.edu> <9nnjdl$r4v$4@bob.news.rcn.net> <3B9FF71E.8B8ED2E3@ev1.net> <1192.655T2092T10255577@nowhere.in.particular> Organization: Metropolis Grafix Reply-To: stanb@dial.pipex.com Message-ID: X-Newsreader: slrn (0.9.4.3 UNIX) Date: 13 Sep 2001 17:02:59 GMT Lines: 12 NNTP-Posting-Host: usercr24.uk.uudial.com X-Trace: 1000400579 news.dial.pipex.com 8512 62.188.156.251 X-Complaints-To: abuse@uk.uu.net Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!newsfeed01.sul.t-online.de!t-online.de!lnewspeer00.lnd.ops.eu.uu.net!lnewspost00.lnd.ops.eu.uu.net!emea.uu.net!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90292 On 12 Sep 01 17:05:41 -0800, Charlie Gibbs wrote: >Sperry->Unisys did some things that can >only be described as obscene. Keyboard design was set back 10 years. God...I remember them...unfortunately ;-( -- Cheers, Stan Barr stanb@dial.pipex.com The future was never like this! ###### From: mirian@trantor.cosmic.com (Mirian Crzig Lennox) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> Message-ID: User-Agent: slrn/0.9.6.4 (Linux) Lines: 34 Date: Thu, 13 Sep 2001 18:04:56 GMT NNTP-Posting-Host: 209.58.189.187 X-Complaints-To: abuse@shore.net X-Trace: news.shore.net 1000404296 209.58.189.187 (Thu, 13 Sep 2001 14:04:56 EDT) NNTP-Posting-Date: Thu, 13 Sep 2001 14:04:56 EDT Organization: Shore.Net, a PRIMUS Company (info@shore.net) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news.shore.net!mirian Xref: chonsp.franklin.ch alt.folklore.computers:90336 On Thu, 13 Sep 01 12:59:29 GMT, jmfbahciv@aol.com wrote: >> >>And, in fact, Linux deviates from the traditional Unix design for just >>this reason (SMP). > >Right. But it seems to me (based on JMF's gag reflex) that >there's a fudge required that isn't very nice. Absolutely, and the gag reflex I would expect from any self-respecting DEC engineer is all to do with the fact that you're running on a major lose hardware architecture with no cross-processor atomicity support. >> ...Under Linux, an interrupt handler does the bare >>minimum to shove the necessary information about the request onto a task >>queue, and then immediately returns. All the I/O and whatnot related to >>the interrupt is done in a thread-safe context, as it should. > >But (big BUT here), I don't think it was easy or possible to >make that non-interrupt I/O code reentrant. Correct. This is because you have to still need to synchronize critical sections of I/O (or paging, or whatever) code at non-interrupt level. All DEC computers which supported SMP had instructions which were guaranteed to be atomic across all processors. With Linux, you have to fake it. >One of these days, >I might even find JMF's and other guys' paper they wrote about >their Unix SMP experience. It would be great if you could get that on line somehow. --Mirian ###### From: Terje Mathisen Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Fri, 14 Sep 2001 09:08:28 +0200 Organization: Hydro Lines: 19 Message-ID: <3BA1ACEC.8752E3BD@hda.hydro.com> References: <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> NNTP-Posting-Host: no115350c8.hda.hydro.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.77 [en] (Windows NT 5.0; U) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.tele.dk!small.news.tele.dk!148.122.208.68!news2.oke.nextra.no!nextra.com!hydro.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90353 Mirian Crzig Lennox wrote: > Correct. This is because you have to still need to synchronize critical > sections of I/O (or paging, or whatever) code at non-interrupt level. > All DEC computers which supported SMP had instructions which were > guaranteed to be atomic across all processors. With Linux, you have to > fake it. Excuse me, but isn't that what all the cache-bypassing, memory-locking read-modify-write opcodes on x86 do? I.e. XCHG, CMPXCHG, XADD etc. Are you talking about some magic opcode that doesn't touch memory, but still does an atomic update, as seen from all cpus??? Terje -- - Using self-discipline, see http://www.eiffel.com/discipline "almost all programming can be viewed as an exercise in caching" ###### From: mirian@trantor.cosmic.com (Mirian Crzig Lennox) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> <3BA1ACEC.8752E3BD@hda.hydro.com> Message-ID: User-Agent: slrn/0.9.6.4 (Linux) Lines: 28 Date: Fri, 14 Sep 2001 16:41:57 GMT NNTP-Posting-Host: 209.58.189.187 X-Complaints-To: abuse@shore.net X-Trace: news.shore.net 1000485717 209.58.189.187 (Fri, 14 Sep 2001 12:41:57 EDT) NNTP-Posting-Date: Fri, 14 Sep 2001 12:41:57 EDT Organization: Shore.Net, a PRIMUS Company (info@shore.net) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news.shore.net!mirian Xref: chonsp.franklin.ch alt.folklore.computers:90424 On Fri, 14 Sep 2001 09:08:28 +0200, Terje Mathisen wrote: >Mirian Crzig Lennox wrote: >> Correct. This is because you have to still need to synchronize critical >> sections of I/O (or paging, or whatever) code at non-interrupt level. >> All DEC computers which supported SMP had instructions which were >> guaranteed to be atomic across all processors. With Linux, you have to >> fake it. > >Excuse me, but isn't that what all the cache-bypassing, memory-locking >read-modify-write opcodes on x86 do? I.e. XCHG, CMPXCHG, XADD etc. None of those instructions is MP-atomic per se. What I think you may be referring to is the LOCK prefix. It's a kludge from a hardware design standpoint, because the x86 was not really architected with MP in mind, and the LOCK prefix (which came about as late as the 486, IIRC), just engages an external mechanism not actually specified in the x86 architecture. Also, they make your code non-portable to 386-and-before, which means realistically you need separate binaries for MP and non-MP. In contrast, the VAX interlocked instructions have rigidly specified behavior as part of the CPU architecture, which is guaranteed to work consistenly in both UP and MP modes, and on any VAX. So, "you have to fake it" is not strictly correct, since SMP does work on current x86 hardware. But it sure ain't pretty. --Mirian ###### Sender: lynn@LYNNPC Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> <3BA1ACEC.8752E3BD@hda.hydro.com> Reply-To: Anne & Lynn Wheeler From: Anne & Lynn Wheeler Message-ID: Organization: Wheeler&Wheeler Lines: 80 User-Agent: Gnus/5.0808 (Gnus v5.8.8) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Fri, 14 Sep 2001 17:14:44 GMT NNTP-Posting-Host: 199.174.230.245 X-Complaints-To: abuse@earthlink.net X-Trace: newsread1.prod.itd.earthlink.net 1000487684 199.174.230.245 (Fri, 14 Sep 2001 10:14:44 PDT) NNTP-Posting-Date: Fri, 14 Sep 2001 10:14:44 PDT X-Received-Date: Fri, 14 Sep 2001 10:11:30 PDT (newsmaster1.prod.itd.earthlink.net) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.direct.ca!look.ca!newsfeed1.earthlink.net!newsfeed.earthlink.net!newsmaster1.prod.itd.earthlink.net!newsread1.prod.itd.earthlink.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90393 mirian@trantor.cosmic.com (Mirian Crzig Lennox) writes: > > None of those instructions is MP-atomic per se. What I think you may be > referring to is the LOCK prefix. It's a kludge from a hardware design > standpoint, because the x86 was not really architected with MP in mind, > and the LOCK prefix (which came about as late as the 486, IIRC), just > engages an external mechanism not actually specified in the x86 > architecture. Also, they make your code non-portable to 386-and-before, > which means realistically you need separate binaries for MP and non-MP. > In contrast, the VAX interlocked instructions have rigidly specified > behavior as part of the CPU architecture, which is guaranteed to work > consistenly in both UP and MP modes, and on any VAX. as an outgrowth of the 360 test&set (TS) atomic instruction for MP operation, Charlie invented atomic Compare&Swap (aka CAS are charlie's inititals). The architecture/POP group in POK (namely Padegs & Ron Smith) said to get it into 370 architecture, it needed to have a non-MP programming paradigm defined for it ... giving rise to the POP programming notes for C&S operation in non-MP world (serializing multi-threaded, not necessarily MP, applications). 370 also had some privileged instructions that were also defined for serialized operation in MP environment ... namely PTLB, IPTE, ISTE, & ISTO (for managing virtual memory tables). When 370m165 engineers said that it would take an extra six months to design/build support for IPTE, ISTE, & ISTO for virtual memory hardware retro-fit to 165, all but PTLB was dropped. PTLB - purge table lookaside buffer (on all processors) IPTE - invalidate page table entry (and any associated TLB entries) ISTE - invalidate segment table entry (and any associate TLB entries) ISTO - invalidate segment table (origin) (and any associate TLB entries) later for the 3033, the IPTE selective invalidate was (re-)introduced. for aix (rios/power, non-mp & power/pc) defined a C&S macro for uniprocessor operation was defined ... however, this just generated an svc interrupt and executed some "disabled" code in the FLIH that simulated a C&S instruction. This was to support non-kernal, multi-threaded application serizlization operation (i.e. the original stuff that was done in cambridge for the POP programming notes to get C&S accepted for 370). misc ref: http://www.garlic.com/~lynn/subtopic.html#smp random refs: http://www.garlic.com/~lynn/93.html#0 360/67, was Re: IBM's Project F/S ? http://www.garlic.com/~lynn/93.html#14 S/360 addressing http://www.garlic.com/~lynn/94.html#28 370 ECPS VM microcode assist http://www.garlic.com/~lynn/99.html#70 Series/1 as NCP (was: Re: System/1 ?) http://www.garlic.com/~lynn/99.html#204 Core (word usage) was anti-equipment etc. http://www.garlic.com/~lynn/2000g.html#16 360/370 instruction cycle time http://www.garlic.com/~lynn/2000g.html#32 Multitasking and resource sharing http://www.garlic.com/~lynn/2001.html#63 Are the L1 and L2 caches flushed on a page fault ? http://www.garlic.com/~lynn/2001.html#64 Are the L1 and L2 caches flushed on a page fault ? http://www.garlic.com/~lynn/2001b.html#37 John Mashey's greatest hits http://www.garlic.com/~lynn/2001c.html#7 LINUS for S/390 http://www.garlic.com/~lynn/2001c.html#87 "Bootstrap" http://www.garlic.com/~lynn/2001d.html#42 IBM was/is: Imitation... http://www.garlic.com/~lynn/2001e.html#73 CS instruction, when introducted ? http://www.garlic.com/~lynn/2001f.html#41 Test and Set (TS) vs Compare and Swap (CS) http://www.garlic.com/~lynn/2001f.html#61 Test and Set (TS) vs Compare and Swap (CS) http://www.garlic.com/~lynn/2001f.html#69 Test and Set (TS) vs Compare and Swap (CS) http://www.garlic.com/~lynn/2001f.html#70 Test and Set (TS) vs Compare and Swap (CS) http://www.garlic.com/~lynn/2001f.html#73 Test and Set (TS) vs Compare and Swap (CS) http://www.garlic.com/~lynn/2001f.html#74 Test and Set (TS) vs Compare and Swap (CS) http://www.garlic.com/~lynn/2001f.html#75 Test and Set (TS) vs Compare and Swap (CS) http://www.garlic.com/~lynn/2001f.html#76 Test and Set (TS) vs Compare and Swap (CS) http://www.garlic.com/~lynn/2001g.html#4 Extended memory error recovery http://www.garlic.com/~lynn/2001g.html#8 Test and Set (TS) vs Compare and Swap (CS) http://www.garlic.com/~lynn/2001g.html#9 Test and Set (TS) vs Compare and Swap (CS) http://www.garlic.com/~lynn/2001h.html#17 IBM 9020 FAA/ATC Systems from 1960's http://www.garlic.com/~lynn/2001i.html#2 Most complex instructions (was Re: IBM 9020 FAA/ATC Systems from 1960's) http://www.garlic.com/~lynn/2001i.html#34 IBM OS Timeline? -- Anne & Lynn Wheeler | lynn@garlic.com - http://www.garlic.com/~lynn/ ###### From: hawk@fac13.ds.psu.edu (Prof. Richard E. Hawkins) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 14 Sep 2001 17:56:27 GMT Organization: Penn State University, Center for Academic Computing Lines: 27 Message-ID: <9ntgcb$13ua@r02n01.cac.psu.edu> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <6b5gn9.qkn.ln@setter.lysse.co.uk> <9nj6a3$156c@r02n01.cac.psu.edu> NNTP-Posting-Host: fac13.ds.psu.edu X-Newsreader: trn 4.0-test76 (Apr 2, 2001) Originator: hawk@fac13.ds.psu.edu (Prof. Richard E. Hawkins) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!newsfeed.r-kom.de!news0.de.colt.net!colt.net!diablo.netcom.net.uk!netcom.net.uk!cpk-news-hub1.bbnplanet.com!news.gtei.net!news.maxwell.syr.edu!news.cis.ohio-state.edu!news.ems.psu.edu!news3.cac.psu.edu!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90391 In article , Eric Smith wrote: >hawk@fac13.ds.psu.edu (Prof. Richard E. Hawkins) writes: >> But lets be fair here: the x86 was *never* meant to go on this long. >> It was supposed to tide people over until the 860 was established. >No, until the iAPX 432 was established. The i860 wasn't even dreamt >of back in 1977; design work on the i860 started after the i386. so many numbers, so many years :) ANyway, the 8086 was the stopgap until the "real" new chip took over. And VW kept poducing the bug waiting for the Type 3 (squareback/notchback/fastvack) to replace it. And still while waiting for the 411 and then the 412 to replace it. Then with the rabbit/golf they made sure by yanking the plug (though the superbeetle lasted another couple of years) product planning . . . hawk -- Prof. Richard E. Hawkins, Esq. /"\ ASCII ribbon campaign dochawk@psu.edu Smeal 178 (814) 375-4700 \ / against HTML mail These opinions will not be those of X and postings Penn State until it pays my retainer. / \ ###### From: Arargh! Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Fri, 14 Sep 2001 16:16:22 -0500 Organization: Arargh!! Lines: 35 Message-ID: References: <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> <3BA1ACEC.8752E3BD@hda.hydro.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: UmFuZG9tSVaTEZf1EYR3kThbVadjxxw9KaPm0NCZObiWeO+TYzZtJPBilWtz5Qqg X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 14 Sep 2001 21:16:29 GMT X-Newsreader: Forte Agent 1.8/32.548 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!feed2.news.rcn.net!feed1.news.rcn.net!rcn!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90422 On Fri, 14 Sep 2001 16:41:57 GMT, mirian@trantor.cosmic.com (Mirian Crzig Lennox) wrote: >On Fri, 14 Sep 2001 09:08:28 +0200, >Terje Mathisen wrote: >>Mirian Crzig Lennox wrote: >>> Correct. This is because you have to still need to synchronize critical >>> sections of I/O (or paging, or whatever) code at non-interrupt level. >>> All DEC computers which supported SMP had instructions which were >>> guaranteed to be atomic across all processors. With Linux, you have to >>> fake it. >> >>Excuse me, but isn't that what all the cache-bypassing, memory-locking >>read-modify-write opcodes on x86 do? I.e. XCHG, CMPXCHG, XADD etc. > >None of those instructions is MP-atomic per se. What I think you may be >referring to is the LOCK prefix. It's a kludge from a hardware design >standpoint, because the x86 was not really architected with MP in mind, >and the LOCK prefix (which came about as late as the 486, IIRC), just The LOCK prefix is described on page 3-112 of my Intel 8086 book. >engages an external mechanism not actually specified in the x86 >architecture. Also, they make your code non-portable to 386-and-before, >which means realistically you need separate binaries for MP and non-MP. >In contrast, the VAX interlocked instructions have rigidly specified >behavior as part of the CPU architecture, which is guaranteed to work >consistenly in both UP and MP modes, and on any VAX. > >So, "you have to fake it" is not strictly correct, since SMP does work >on current x86 hardware. But it sure ain't pretty. > >--Mirian -- Arargh (at enteract dot com) http://www.arargh.com ###### Message-ID: <3BA2A128.79F17B15@yahoo.com> From: CBFalconer Reply-To: cbfalconer@worldnet.att.net Organization: Ched Research X-Mailer: Mozilla 4.75 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <6b5gn9.qkn.ln@setter.lysse.co.uk> <9nj6a3$156c@r02n01.cac.psu.edu> <9ntgcb$13ua@r02n01.cac.psu.edu> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 18 Date: Sat, 15 Sep 2001 00:50:11 GMT NNTP-Posting-Host: 12.90.172.211 X-Complaints-To: abuse@worldnet.att.net X-Trace: bgtnsc05-news.ops.worldnet.att.net 1000515011 12.90.172.211 (Sat, 15 Sep 2001 00:50:11 GMT) NNTP-Posting-Date: Sat, 15 Sep 2001 00:50:11 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.mathworks.com!wn3feed!worldnet.att.net!135.173.83.72!wnfilter2!worldnet-localpost!bgtnsc05-news.ops.worldnet.att.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90430 "Prof. Richard E. Hawkins" wrote: > ... snip ... > > And VW kept poducing the bug waiting for the Type 3 > (squareback/notchback/fastvack) to replace it. And still while waiting > for the 411 and then the 412 to replace it. Then with the rabbit/golf > they made sure by yanking the plug (though the superbeetle lasted > another couple of years) They're still building beetles in Brazil. -- Chuck F (cbfalconer@yahoo.com) (cbfalconer@XXXXworldnet.att.net) (Remove "XXXX" from reply address. yahoo works unmodified) mailto:uce@ftc.gov (for spambots to harvest) ###### From: lysse Subject: Re: Minimalist design (was Re: Parity - why even or odd) Newsgroups: comp.arch,alt.folklore.computers References: <9n7lkr$8p$4@bob.news.rcn.net> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> <3BA1ACEC.8752E3BD@hda.hydro.com> Organization: lysse's domain Reply-To: lysse.news@blueyonder.co.uk User-Agent: tin/pre-1.4-980226 (UNIX) (Linux/2.2.6 (i486)) Message-ID: <4dcun9.29s.ln@setter.lysse.co.uk> Lines: 12 Date: Sat, 15 Sep 2001 03:33:11 GMT NNTP-Posting-Host: 62.31.9.47 X-Complaints-To: http://www.blueyonder.co.uk/abuse X-Trace: news1.cableinet.net 1000524791 62.31.9.47 (Sat, 15 Sep 2001 04:33:11 BST) NNTP-Posting-Date: Sat, 15 Sep 2001 04:33:11 BST Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.tele.dk!small.news.tele.dk!212.74.64.35!colt.net!newspeer.clara.net!news.clara.net!news-hub.cableinet.net!blueyonder!internal-news-hub.cableinet.net!news1.cableinet.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90398 Mirian Crzig Lennox wrote: > standpoint, because the x86 was not really architected with MP in mind, > and the LOCK prefix (which came about as late as the 486, IIRC), just I have an Osborne book dating from 1979 which details a number of 16-bit processors, including the 8086 and 8088. It lists the LOCK instruction, and points out that you would use LOCK XCHG to do an atomic update in a multiprocessor system. -- lysse at lysse dot co dot uk "Why are your problems always so much bigger than everyone else's?" "Because they're mine." -- Ally McBeal ###### Message-ID: <3BA2D989.51AC4E0A@yahoo.com> From: CBFalconer Reply-To: cbfalconer@worldnet.att.net Organization: Ched Research X-Mailer: Mozilla 4.75 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9n7lkr$8p$4@bob.news.rcn.net> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> <3BA1ACEC.8752E3BD@hda.hydro.com> <4dcun9.29s.ln@setter.lysse.co.uk> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 22 Date: Sat, 15 Sep 2001 04:46:30 GMT NNTP-Posting-Host: 12.90.169.7 X-Complaints-To: abuse@worldnet.att.net X-Trace: bgtnsc06-news.ops.worldnet.att.net 1000529190 12.90.169.7 (Sat, 15 Sep 2001 04:46:30 GMT) NNTP-Posting-Date: Sat, 15 Sep 2001 04:46:30 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!7533453!news.imp.ch!sunqbc.risq.qc.ca!howland.erols.net!news-out.worldnet.att.net.MISMATCH!wn3feed!worldnet.att.net!135.173.83.71!wnfilter1!worldnet-localpost!bgtnsc06-news.ops.worldnet.att.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90376 lysse wrote: > > Mirian Crzig Lennox wrote: > > standpoint, because the x86 was not really architected with MP in mind, > > and the LOCK prefix (which came about as late as the 486, IIRC), just > > I have an Osborne book dating from 1979 which details a number of > 16-bit processors, including the 8086 and 8088. It lists the LOCK > instruction, and points out that you would use LOCK XCHG to do an > atomic update in a multiprocessor system. And what is wrong with that sort of approach? It is not as if a system is about to spend the majority of its time testing semaphores. There are much better things on which to spend valuable opcodes. -- Chuck F (cbfalconer@yahoo.com) (cbfalconer@XXXXworldnet.att.net) (Remove "XXXX" from reply address. yahoo works unmodified) mailto:uce@ftc.gov (for spambots to harvest) ###### From: Terje Mathisen Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Sat, 15 Sep 2001 11:24:22 +0200 Organization: Hydro Lines: 40 Message-ID: <3BA31E46.8AB11EC1@hda.hydro.com> References: <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> <3BA1ACEC.8752E3BD@hda.hydro.com> NNTP-Posting-Host: iabrahmspc.nho.hydro.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.77 [en] (Windows NT 5.0; U) X-Accept-Language: en Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.tele.dk!small.news.tele.dk!148.122.208.68!news2.oke.nextra.no!nextra.com!hydro.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90380 "Arargh!" wrote: > > On Fri, 14 Sep 2001 16:41:57 GMT, mirian@trantor.cosmic.com (Mirian > Crzig Lennox) wrote: > > >On Fri, 14 Sep 2001 09:08:28 +0200, > >Terje Mathisen wrote: > >>Mirian Crzig Lennox wrote: > >>> Correct. This is because you have to still need to synchronize critical > >>> sections of I/O (or paging, or whatever) code at non-interrupt level. > >>> All DEC computers which supported SMP had instructions which were > >>> guaranteed to be atomic across all processors. With Linux, you have to > >>> fake it. > >> > >>Excuse me, but isn't that what all the cache-bypassing, memory-locking > >>read-modify-write opcodes on x86 do? I.e. XCHG, CMPXCHG, XADD etc. > > > >None of those instructions is MP-atomic per se. What I think you may be Rather the opposite, see below! > >referring to is the LOCK prefix. It's a kludge from a hardware design > >standpoint, because the x86 was not really architected with MP in mind, > >and the LOCK prefix (which came about as late as the 486, IIRC), just > The LOCK prefix is described on page 3-112 of my Intel 8086 book. Not only was LOCK part of the original x86 architecture, the XCHG reg,[mem] opcode was specified to have an implied LOCK prefix, i.e. there was no way to use that opcode in a non-atomic manner, even on cpus that didn't support SMP. It could still be useful in a system which had DMA or some other form of assymetric MP. Terje -- - Using self-discipline, see http://www.eiffel.com/discipline "almost all programming can be viewed as an exercise in caching" ###### Sender: lynn@LYNNPC Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> <3BA1ACEC.8752E3BD@hda.hydro.com> <3BA31E46.8AB11EC1@hda.hydro.com> Reply-To: Anne & Lynn Wheeler From: Anne & Lynn Wheeler Message-ID: Organization: Wheeler&Wheeler Lines: 25 User-Agent: Gnus/5.0808 (Gnus v5.8.8) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Sat, 15 Sep 2001 19:44:23 GMT NNTP-Posting-Host: 199.174.228.51 X-Complaints-To: abuse@earthlink.net X-Trace: newsread2.prod.itd.earthlink.net 1000583063 199.174.228.51 (Sat, 15 Sep 2001 12:44:23 PDT) NNTP-Posting-Date: Sat, 15 Sep 2001 12:44:23 PDT X-Received-Date: Sat, 15 Sep 2001 12:41:07 PDT (newsmaster1.prod.itd.earthlink.net) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!newsfeed01.sul.t-online.de!t-online.de!newsfeed.hanau.net!news-fra1.dfn.de!news0.de.colt.net!colt.net!newsfeed.icl.net!netnews.com!xfer02.netnews.com!newsfeed2.earthlink.net!newsfeed.earthlink.net!newsmaster1.prod.itd.earthlink.net!newsread2.prod.itd.earthlink.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90445 Terje Mathisen writes: > > Not only was LOCK part of the original x86 architecture, the XCHG > reg,[mem] opcode was specified to have an implied LOCK prefix, i.e. > there was no way to use that opcode in a non-atomic manner, even on cpus > that didn't support SMP. > > It could still be useful in a system which had DMA or some other form of > assymetric MP. note that one of the uses of the atomic compare & swap work (from the late '60s and early '70s resulting in CAS going into system/370) was to support multi-threaded, pre-emptable application code (i.e. enabled for interrupts, non-kernal, etc). the MP barrier semantics from the '60s (and earlier?) with test & set type instruction had non-atomic type operations (set barrier, do whatever, clear barrier) which implied non-pre-emptable execution (unless you were very, very, very careful). random refs: http://www.garlic.com/~lynn/subtopic.html#smp -- Anne & Lynn Wheeler | lynn@garlic.com - http://www.garlic.com/~lynn/ ###### From: Bernd Paysan Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Sat, 15 Sep 2001 22:25:41 +0200 Organization: Bernd Paysan, 81477 Muenchen, Germany Lines: 17 Message-ID: <3BA3B945.D06E42ED@gmx.de> References: <9n7lkr$8p$4@bob.news.rcn.net> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> <3BA1ACEC.8752E3BD@hda.hydro.com> <4dcun9.29s.ln@setter.lysse.co.uk> <3BA2D989.51AC4E0A@yahoo.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Complaints-To: news@cohen.paysan.nom NNTP-Posting-Date: 15 Sep 2001 20:25:41 GMT X-Mailer: Mozilla 4.77 [de] (X11; U; Linux 2.4.9 i686) X-Accept-Language: de, en NNTP-Posting-Host: dial021.mucweb.de X-Original-NNTP-Posting-Host: dial021.mucweb.de X-Trace: 15 Sep 2001 22:54:44 +0200, dial021.mucweb.de X-Original-Trace: 15 Sep 2001 22:54:44 +0200, dial021.mucweb.de Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!newsfeed.freenet.de!news.space.net!news.touch.net!cohen.paysan.nom!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90467 CBFalconer wrote: > And what is wrong with that sort of approach? It is not as if a > system is about to spend the majority of its time testing > semaphores. There are much better things on which to spend > valuable opcodes. Actually, this prefix saves valuabe opcode space. How many ways of using atomic operations can you imagine? xchg reg, mem; inc mem; dec mem; add mem, reg at least? Others? By using a lock prefix, you can use these operations atomic if you like to, so you don't have to pollute your opcode space with atomic and non-atomic versions of the same instruction. -- Bernd Paysan "If you want it done right, you have to do it yourself" http://www.jwdt.com/~paysan/ ###### Message-ID: <3BA3EF15.9D2B0632@yahoo.com> From: CBFalconer Reply-To: cbfalconer@worldnet.att.net Organization: Ched Research X-Mailer: Mozilla 4.75 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9n7lkr$8p$4@bob.news.rcn.net> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> <3BA1ACEC.8752E3BD@hda.hydro.com> <4dcun9.29s.ln@setter.lysse.co.uk> <3BA2D989.51AC4E0A@yahoo.com> <3BA3B945.D06E42ED@gmx.de> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 23 Date: Sun, 16 Sep 2001 00:40:33 GMT NNTP-Posting-Host: 12.90.175.121 X-Complaints-To: abuse@worldnet.att.net X-Trace: bgtnsc04-news.ops.worldnet.att.net 1000600833 12.90.175.121 (Sun, 16 Sep 2001 00:40:33 GMT) NNTP-Posting-Date: Sun, 16 Sep 2001 00:40:33 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!enews.sgi.com!arclight.uoregon.edu!wn4feed!worldnet.att.net!135.173.83.71!wnfilter1!worldnet-localpost!bgtnsc04-news.ops.worldnet.att.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90431 Bernd Paysan wrote: > > CBFalconer wrote: > > And what is wrong with that sort of approach? It is not as if a > > system is about to spend the majority of its time testing > > semaphores. There are much better things on which to spend > > valuable opcodes. > > Actually, this prefix saves valuabe opcode space. How many ways of using > atomic operations can you imagine? xchg reg, mem; inc mem; dec mem; > add mem, reg at least? Others? By using a lock prefix, you can use these > operations atomic if you like to, so you don't have to pollute your > opcode space with atomic and non-atomic versions of the same > instruction. Which is what I intended to imply. -- Chuck F (cbfalconer@yahoo.com) (cbfalconer@XXXXworldnet.att.net) (Remove "XXXX" from reply address. yahoo works unmodified) mailto:uce@ftc.gov (for spambots to harvest) ###### From: Jan Vorbrueggen Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 17 Sep 2001 09:31:52 +0200 Organization: Institut fuer Neuroinformatik, Ruhr-Universitaet Bochum, Germany Lines: 16 Message-ID: References: <9n7lkr$8p$4@bob.news.rcn.net> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> <3BA1ACEC.8752E3BD@hda.hydro.com> <4dcun9.29s.ln@setter.lysse.co.uk> <3BA2D989.51AC4E0A@yahoo.com> <3BA3B945.D06E42ED@gmx.de> NNTP-Posting-Host: luda.neuroinformatik.ruhr-uni-bochum.de X-Trace: sunu789.rz.ruhr-uni-bochum.de 1000711913 15829 134.147.176.178 (17 Sep 2001 07:31:53 GMT) NNTP-Posting-Date: 17 Sep 2001 07:31:53 GMT X-Newsreader: Gnus v5.7/Emacs 20.5 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!news.rwth-aachen.de!news.ruhr-uni-bochum.de!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90504 Bernd Paysan writes: > Actually, this prefix saves valuabe opcode space. How many ways of using > atomic operations can you imagine? xchg reg, mem; inc mem; dec mem; > add mem, reg at least? Others? I do think so - there was this machine somewhere from New York which did these kinds of operations in the memory subsystem (the UltraComputer? memory is failing...) Load-locked/store-conditional has the advantage that only the basic building blocks for an atomic memory operation is architecturally specified; most of the contents of a traditional critical section can go between these two instructions. Jan ###### Sender: lynn@LYNNPC Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <6uy9nsqhd7.fsf@chonsp.franklin.ch> <9nd4ic$i2m$9@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> <3BA1ACEC.8752E3BD@hda.hydro.com> <3BA5A0C3.B7EA848E@mchpDOTsiemens.de> Reply-To: Anne & Lynn Wheeler From: Anne & Lynn Wheeler Organization: Wheeler&Wheeler User-Agent: Gnus/5.0808 (Gnus v5.8.8) Emacs/20.7 Message-ID: Lines: 87 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Mon, 17 Sep 2001 14:42:11 GMT NNTP-Posting-Host: 199.174.226.37 X-Complaints-To: abuse@earthlink.net X-Trace: newsread1.prod.itd.earthlink.net 1000737731 199.174.226.37 (Mon, 17 Sep 2001 07:42:11 PDT) NNTP-Posting-Date: Mon, 17 Sep 2001 07:42:11 PDT X-Received-Date: Mon, 17 Sep 2001 07:38:54 PDT (newsmaster1.prod.itd.earthlink.net) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!feed2.onemain.com!feed1.onemain.com!newsfeed2.earthlink.net!newsfeed.earthlink.net!newsmaster1.prod.itd.earthlink.net!newsread1.prod.itd.earthlink.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90494 Konrad Schwarz writes: > > Are the 370 Principles of Operation still available on the net? > (I know this has been discussed before...but I can't find > it via IBM's search or via Google.) now "390" pop ... http://publibz.boulder.ibm.com:80/cgi-bin/bookmgr_OS390/BOOKS/DZ9AR004/CCONTENTS appendix has differences: D.0 Appendix D. Comparison between ESA/370 and ESA/390 D.1 New Facilities in ESA/390 D.1.1 Access-List-Controlled Protection D.1.2 Branch and Set Authority D.1.3 Called-Space Identification D.1.4 Checksum D.1.5 Compare and Move Extended D.1.6 Concurrent Sense D.1.7 Immediate and Relative Instruction D.1.8 Move-Page Facility 2 D.1.9 PER 2 D.1.10 Perform Locked Operation D.1.11 Set Address Space Control Fast D.1.12 Square Root D.1.13 Storage-Protection Override D.1.14 String Instruction D.1.15 Subspace Group D.1.16 Suppression on Protection D.2 Comparison of Facilities E.0 Appendix E. Comparison between 370-XA and ESA/370 E.1 New Facilities in ESA/370 E.1.1 Access Registers E.1.2 Compare until Substring Equal E.1.3 Home Address Space E.1.4 Linkage Stack E.1.5 Load and Store Using Real Address E.1.6 Move Page Facility 1 E.1.7 Move with Source or Destination Key E.1.8 Private Space E.2 Comparison of Facilities E.3 Summary of Changes E.3.1 New Instructions Provided E.3.2 Comparison of PSW Formats E.3.3 New Control-Register Assignments E.3.4 New Assigned Storage Locations E.3.5 New Exceptions E.3.6 Change to Secondary-Space Mode E.3.7 Changes to ASN-Second-Table Entry and ASN Translation E.3.8 Changes to Entry-Table Entry and PC-Number Translation E.3.9 Changes to PROGRAM CALL E.3.10 Changes to SET ADDRESS SPACE CONTROL E.4 Effects in New Translation Modes E.4.1 Effects on Interlocks for Virtual-Storage References E.4.2 Effect on INSERT ADDRESS SPACE CONTROL E.4.3 Effect on LOAD REAL ADDRESS E.4.4 Effect on TEST PENDING INTERRUPTION E.4.5 Effect on TEST PROTECTION F.0 Appendix F. Comparison between System/370 and 370-XA F.1 New Facilities in 370-XA F.1.1 Bimodal Addressing F.1.2 31-Bit Logical Addressing F.1.3 31-Bit Real and Absolute Addressing F.1.4 Page Protection F.1.5 Tracing F.1.6 Incorrect-Length-Indication Suppression F.1.7 Status Verification F.2 Comparison of Facilities F.3 Summary of Changes F.3.1 Changes in Instructions Provided F.3.2 Input/Output Comparison F.3.3 Comparison of PSW Formats F.3.4 Changes in Control-Register Assignments F.3.5 Changes in Assigned Storage Locations F.3.6 Changes to SIGNAL PROCESSOR F.3.7 Machine-Check Changes F.3.8 Changes to Addressing Wraparound F.3.9 Changes to LOAD REAL ADDRESS F.3.10 Changes to 31-Bit Real Operand Addresses -- Anne & Lynn Wheeler | lynn@garlic.com - http://www.garlic.com/~lynn/ ###### From: cecchi@signa.rchland.ibm.com (Del Cecchi) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 17 Sep 2001 20:49:05 GMT Organization: IBM Corp. Lines: 53 Message-ID: <9o5nk1$tf4$1@news.rchland.ibm.com> References: <9n7lkr$8p$4@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> <3BA1ACEC.8752E3BD@hda.hydro.com> <3BA5A0C3.B7EA848E@mchpDOTsiemens.de> Reply-To: dcecchi@vnet.ibm.com NNTP-Posting-Host: signa.rchland.ibm.com X-Trace: news.rchland.ibm.com 1000759745 30180 9.5.54.183 (17 Sep 2001 20:49:05 GMT) X-Complaints-To: usenet@news.rchland.ibm.com NNTP-Posting-Date: 17 Sep 2001 20:49:05 GMT X-Newsreader: xrn 9.01-beta-3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.mathworks.com!nycmny1-snh1.gtei.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!znr.news.ans.net!news.chips.ibm.com!newsfeed.btv.ibm.com!news!signa.rchland.ibm.com!cecchi Xref: chonsp.franklin.ch alt.folklore.computers:90546 In article <3BA5A0C3.B7EA848E@mchpDOTsiemens.de>, Konrad Schwarz writes: |> Anne & Lynn Wheeler schrieb: |> > as an outgrowth of the 360 test&set (TS) atomic instruction for MP |> > operation, Charlie invented atomic Compare&Swap (aka CAS are charlie's |> > inititals). The architecture/POP group in POK (namely Padegs & Ron |> > Smith) said to get it into 370 architecture, it needed to have a |> > non-MP programming paradigm defined for it ... giving rise to the POP |> > programming notes for C&S operation in non-MP world (serializing |> > multi-threaded, not necessarily MP, applications). |> |> Are the 370 Principles of Operation still available on the net? |> (I know this has been discussed before...but I can't find |> it via IBM's search or via Google.) |> |> Thx, |> Konrad Search results Search argument: principles of operation Matches 1 to 10 of 25 For online publications and bookshelves click the Online icon to view or download. For hardcopy publications select the check box next to the required publication(s) and click the Add to shopping cart button to order. Or click help for more information about the following results. Publication title Publication no. Order Online 1.S/360 PRINCIPLES OF OPERATION GA22-6821-08 2. S/370 PRINCIPLES OF OPERATION GA22-7000-10 6. 370/XA PRINCIPLES OF OPERATION SA22-7085-01 7. ESA/370 PRINCIPLES OF OPERATION SA22-7200-00 8. ESA/390 Principles of Operation SA22-7201-07 z/Architecture Principles of Operation SA22-7832-00 The later ones are on line for free. the hardcopy costs. go to http://ehone.ibm.com/public/applications/publications/cgibin/pbi.cgi?CTY=US for the US. -- Del Cecchi cecchi@rchland ###### From: cecchi@signa.rchland.ibm.com (Del Cecchi) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 17 Sep 2001 20:50:12 GMT Organization: IBM Corp. Lines: 6 Message-ID: <9o5nm4$tf4$2@news.rchland.ibm.com> References: <9n7lkr$8p$4@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> <3BA1ACEC.8752E3BD@hda.hydro.com> <3BA5A0C3.B7EA848E@mchpDOTsiemens.de> Reply-To: dcecchi@vnet.ibm.com NNTP-Posting-Host: signa.rchland.ibm.com X-Trace: news.rchland.ibm.com 1000759812 30180 9.5.54.183 (17 Sep 2001 20:50:12 GMT) X-Complaints-To: usenet@news.rchland.ibm.com NNTP-Posting-Date: 17 Sep 2001 20:50:12 GMT X-Newsreader: xrn 9.01-beta-3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!news-lei1.dfn.de!news-was.dfn.de!znr.news.ans.net!news.chips.ibm.com!newsfeed.btv.ibm.com!news!signa.rchland.ibm.com!cecchi Xref: chonsp.franklin.ch alt.folklore.computers:90547 oops, that address might only work internally. let me know and I will look further. -- Del Cecchi cecchi@rchland ###### From: cecchi@signa.rchland.ibm.com (Del Cecchi) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 17 Sep 2001 20:51:43 GMT Organization: IBM Corp. Lines: 23 Message-ID: <9o5nov$tf4$3@news.rchland.ibm.com> References: <9n7lkr$8p$4@bob.news.rcn.net> <3B9D2E53.3D5C720F@yahoo.com> <3B9DBEFD.BAAE6C8A@hda.hydro.com> <3B9DD935.583B7141@yahoo.com> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> <3BA1ACEC.8752E3BD@hda.hydro.com> <3BA5A0C3.B7EA848E@mchpDOTsiemens.de> Reply-To: dcecchi@vnet.ibm.com NNTP-Posting-Host: signa.rchland.ibm.com X-Trace: news.rchland.ibm.com 1000759903 30180 9.5.54.183 (17 Sep 2001 20:51:43 GMT) X-Complaints-To: usenet@news.rchland.ibm.com NNTP-Posting-Date: 17 Sep 2001 20:51:43 GMT X-Newsreader: xrn 9.01-beta-3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.mathworks.com!nycmny1-snh1.gtei.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!znr.news.ans.net!news.chips.ibm.com!newsfeed.btv.ibm.com!news!signa.rchland.ibm.com!cecchi Xref: chonsp.franklin.ch alt.folklore.computers:90552 In article <3BA5A0C3.B7EA848E@mchpDOTsiemens.de>, Konrad Schwarz writes: |> Anne & Lynn Wheeler schrieb: |> > as an outgrowth of the 360 test&set (TS) atomic instruction for MP |> > operation, Charlie invented atomic Compare&Swap (aka CAS are charlie's |> > inititals). The architecture/POP group in POK (namely Padegs & Ron |> > Smith) said to get it into 370 architecture, it needed to have a |> > non-MP programming paradigm defined for it ... giving rise to the POP |> > programming notes for C&S operation in non-MP world (serializing |> > multi-threaded, not necessarily MP, applications). |> |> Are the 370 Principles of Operation still available on the net? |> (I know this has been discussed before...but I can't find |> it via IBM's search or via Google.) |> |> Thx, |> Konrad I see you are from siemens in germany. call the IBM branch office and give them that web address (it lets you change country). I think the book costs 5 dollars. -- Del Cecchi cecchi@rchland ###### From: cecchi@signa.rchland.ibm.com (Del Cecchi) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 18 Sep 2001 12:29:51 GMT Organization: IBM Corp. Lines: 37 Message-ID: <9o7env$13hu$1@news.rchland.ibm.com> References: <9n7lkr$8p$4@bob.news.rcn.net> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> <3BA1ACEC.8752E3BD@hda.hydro.com> <3BA5A0C3.B7EA848E@mchpDOTsiemens.de> <9o5nk1$tf4$1@news.rchland.ibm.com> <3BA6B54D.D9FFF44A@yahoo.com.NOSPAM> Reply-To: dcecchi@vnet.ibm.com NNTP-Posting-Host: signa.rchland.ibm.com X-Trace: news.rchland.ibm.com 1000816191 36414 9.5.54.183 (18 Sep 2001 12:29:51 GMT) X-Complaints-To: usenet@news.rchland.ibm.com NNTP-Posting-Date: 18 Sep 2001 12:29:51 GMT X-Newsreader: xrn 9.01-beta-3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!newsfeed.r-kom.de!feedme.news.mediaways.net!abq.news.ans.net!news.chips.ibm.com!newsfeed.btv.ibm.com!news!signa.rchland.ibm.com!cecchi Xref: chonsp.franklin.ch alt.folklore.computers:90554 In article <3BA6B54D.D9FFF44A@yahoo.com.NOSPAM>, CBFalconer writes: |> Del Cecchi wrote: |> > |> ... snip ... |> > Online |> > 1.S/360 PRINCIPLES OF OPERATION GA22-6821-08 |> > 2. S/370 PRINCIPLES OF OPERATION GA22-7000-10 |> > 6. 370/XA PRINCIPLES OF OPERATION SA22-7085-01 |> > 7. ESA/370 PRINCIPLES OF OPERATION SA22-7200-00 |> > 8. ESA/390 Principles of Operation SA22-7201-07 |> > |> > z/Architecture Principles of Operation SA22-7832-00 |> > |> > The later ones are on line for free. the hardcopy costs. |> > |> > go to http://ehone.ibm.com/public/applications/publications/cgibin/pbi.cgi?CTY=US |> > |> > for the US. |> |> URL not found. |> |> -- |> Chuck F (cbfalconer@yahoo.com) (cbfalconer@XXXXworldnet.att.net) |> (Remove "XXXX" from reply address. yahoo works unmodified) |> mailto:uce@ftc.gov (for spambots to harvest) |> Sorry, as I noted in an earlier followup, this is apparently an internal URL. The earlier publications are only available in hardcopy anyway. But they are available. The later ones are available online, as posted by Lynn Wheeler. Again, sorry for the internal URL. -- Del Cecchi cecchi@rchland ###### From: hawk@fac13.ds.psu.edu (Prof. Richard E. Hawkins) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 18 Sep 2001 18:00:43 GMT Organization: Penn State University, Center for Academic Computing Lines: 25 Message-ID: <9o824b$22lo@r02n01.cac.psu.edu> References: <9mhjn3$7nh$1@nnrp2.phx.gblx.net> <9ntgcb$13ua@r02n01.cac.psu.edu> <3BA2A128.79F17B15@yahoo.com> NNTP-Posting-Host: fac13.ds.psu.edu X-Newsreader: trn 4.0-test76 (Apr 2, 2001) Originator: hawk@fac13.ds.psu.edu (Prof. Richard E. Hawkins) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeeds.belnet.be!news.belnet.be!newsfeed.stanford.edu!news.ems.psu.edu!news3.cac.psu.edu!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:90532 In article <3BA2A128.79F17B15@yahoo.com>, CBFalconer wrote: >"Prof. Richard E. Hawkins" wrote: >... snip ... >> And VW kept poducing the bug waiting for the Type 3 >> (squareback/notchback/fastvack) to replace it. And still while waiting >> for the 411 and then the 412 to replace it. Then with the rabbit/golf >> they made sure by yanking the plug (though the superbeetle lasted >> another couple of years) >They're still building beetles in Brazil. Yes, and mexico too--at the same plant that builds the new bug-like contraptions. But they're not available in the industrialized countries. hawk -- rm -rf /bin/laden Prof. Richard E. Hawkins, Esq. /"\ ASCII ribbon campaign dochawk@psu.edu Smeal 178 (814) 375-4700 \ / against HTML mail These opinions will not be those of X and postings. Penn State until it pays my retainer. / \ ###### From: gah@ugcs.caltech.edu (glen herrmannsfeldt) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 18 Sep 2001 23:34:02 GMT Organization: California Institute of Technology, Pasadena Lines: 25 Message-ID: <9o8lla$rr7@gap.cco.caltech.edu> References: <9n7lkr$8p$4@bob.news.rcn.net> <3B9DE34F.8DBAB95@hda.hydro.com> <3B9E06A0.F021007C@mchpDOTsiemens.de> <9nnjtj$r4v$7@bob.news.rcn.net> <9nqkgb$1hl$14@bob.news.rcn.net> <3BA1ACEC.8752E3BD@hda.hydro.com> <3BA5A0C3.B7EA848E@mchpDOTsiemens.de> <9o5nk1$tf4$1@news.rchland.ibm.com> <3BA6B54D.D9FFF44A@yahoo.com.NOSPAM> <9o7env$13hu$1@news.rchland. NNTP-Posting-Host: yak.ugcs.caltech.edu X-Newsreader: NN version 6.5.0 #1 (NOV) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!270611!news.imp.ch!sunqbc.risq.qc.ca!newsfeed.direct.ca!look.ca!sfo2-feed1.news.digex.net!intermedia!news-out.spamkiller.net!propagator-la!news-in-la.newsfeeds.com!news-in.superfeed.net!news.exit.com!nntp-server.caltech.edu!gah Xref: chonsp.franklin.ch alt.folklore.computers:90605 cecchi@signa.rchland.ibm.com (Del Cecchi) writes: (snip) >|> > 1.S/360 PRINCIPLES OF OPERATION GA22-6821-08 Available until last year >|> > 2. S/370 PRINCIPLES OF OPERATION GA22-7000-10 Still available, $11.25+tax, but includes UPS ground shipping >|> > 6. 370/XA PRINCIPLES OF OPERATION SA22-7085-01 Still available, about $33.50+tax >|> > 7. ESA/370 PRINCIPLES OF OPERATION SA22-7200-00 Still available, $41.06+tax >|> > 8. ESA/390 Principles of Operation SA22-7201-07 Still available, the -6 was $14.75+tax, this is a very big book, nicely bound, too. >|> > z/Architecture Principles of Operation SA22-7832-00 Available, $21.75+tax, probably bigger than above. http://www.elink.ibmlink.ibm.com/public/applications/publications/cgibin/pbi.cgi? -- glen ###### From: Steve O'Hara-Smith Newsgroups: alt.folklore.computers,ed.general,alt.lang.teco Subject: Re: Newbie TOPS-10 7.03 question Date: Mon, 15 Oct 2001 19:51:39 +0200 Organization: Wanadoo NL Lines: 33 Message-ID: <20011015195139.7e261279.steveo@eircom.net> References: <9nont9$9cq@dispatch.concentric.net> <9nspfu$gqv$2@bob.news.rcn.net> <3BA3A9DF.58F2C0CB@virgin.net> <9o0sjb$2n9s$1@pc-news.cogsci.ed.ac.uk> <3BA5EBE2.DFE78360@owl.co.uk> <3bc7e261@news.sierratel.com> NNTP-Posting-Host: p1367.vcu.wanadoo.nl Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-Trace: scavenger.euro.net 1003170509 62540 194.134.170.92 (15 Oct 2001 18:28:29 GMT) X-Complaints-To: abuse@wanadoo.nl NNTP-Posting-Date: Mon, 15 Oct 2001 18:28:29 +0000 (UTC) X-Newsreader: Sylpheed version 0.6.3 (GTK+ 1.2.10; i386--freebsd4.4) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeeds.belnet.be!news.belnet.be!news2.euro.net!news.euronet.nl!ams-gw.sohara.org!nobody Xref: chonsp.franklin.ch alt.folklore.computers:92871 On Sun, 14 Oct 2001 21:01:58 -0600 Brian Inglis wrote: BI> On Fri, 12 Oct 2001 20:18:29 -0700, "Walter Rottenkolber" BI> wrote: BI> >Incredible! A text processing program with a script more obscure and BI> >cryptic than Perl. Is there nothing new under the sun? BI> BI> It's nice to see Perlites feel threatened by TECO's obvious BI> essence of hackishness. ;^> It's somewhat surprising that anyone expected cryptic(1), powerful and terse to be new with perl. It is clearly part of a long tradition. (1) I'm not sure I agree with this assesment, but I think I know what it means here. BI> And not a lot new these days, unfortunately, although as PHP and There was the fun with aspect weavers which seems to have died down. It looked to be quite a promising approach. BI> Python seem to be generating interest, I still have hopes of a BI> resurgence of interest in creating useful weird and wonderful BI> little languages. Weird and wonderful languages turn up in all sorts of places, mostly specialised IME. Bonus points to anyone who can identify Cheshire. -- Directable Mirrors - A Better Way To Focus The Sun http://www.best.com/~sohara ###### From: michael.wojcik@microfocus.com (Michael Wojcik) Newsgroups: alt.folklore.computers Subject: Re: Newbie TOPS-10 7.03 question Date: 15 Oct 2001 19:27:14 GMT Organization: http://extra.newsguy.com Lines: 25 Message-ID: <9qfdai0u9e@enews3.newsguy.com> References: <9nont9$9cq@dispatch.concentric.net> <3BA3A9DF.58F2C0CB@virgin.net> <9o0sjb$2n9s$1@pc-news.cogsci.ed.ac.uk> <3BA5EBE2.DFE78360@owl.co.uk> <3bc7e261@news.sierratel.com> Organizat Reply-To: michael.wojcik@microfocus.com NNTP-Posting-Host: p-075.newsdawg.com X-Newsreader: xrn 9.00 ion: MERANT Inc. Originator: mww@lorelei Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!pln-e!spln!dex!extra.newsguy.com!newsp.newsguy.com!mww Xref: chonsp.franklin.ch alt.folklore.computers:92851 In article , Brian Inglis writes: > And not a lot new these days, unfortunately, although as PHP and > Python seem to be generating interest, I still have hopes of a > resurgence of interest in creating useful weird and wonderful > little languages. Ruby seems to have some momentum, and it has some interesting characteristics, particularly its blocks and the "yield" operator. I haven't had a chance to play with it myself, but there was an intriguing article about it in _Dr Dobb's_ a while back.[1] 1. -- Michael Wojcik michael.wojcik@microfocus.com Comms Development, Micro Focus Department of English, Miami University This is a "rubbering action game," a 2D platformer where you control a girl equipped with an elastic rope with a fishing hook at the end. -- review of _Umihara Kawase Shun_ for the Sony Playstation ###### From: John F. Eldredge Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Sat, 24 Nov 2001 23:41:15 -0600 Lines: 37 Message-ID: References: <1bk7zh2hr3.fsf@cs.nmsu.edu> <9n1ju8$t5n$1@mpoli.fi> <1bu1yj8olm.fsf@cs.nmsu.edu> <20010904211234.1d1be108.steveo@eircom.net> <20010905081118.0a4e2a04.steveo@eircom.net> <1bitex1pfy.fsf@cs.nmsu.edu> <9n7lkr$8p$4@bob.news.rcn.net> <1b66awidk1.fsf@cs.nmsu.edu> <9nahnc$f38$14@bob.news.rcn.net> <28ifn9.6en.ln@setter.lysse.co.uk> <6ubskj434h.fsf@chonsp.franklin.ch> NNTP-Posting-Host: user-38lcdig.dialup.mindspring.com (209.86.54.80) Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: fu-berlin.de 1006666856 4184284 209.86.54.80 (16 [61967]) X-Newsreader: Forte Agent 1.8/32.548 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsmi-eu.news.garr.it!NewsITBone-GARR!fu-berlin.de!uni-berlin.de!user-38lcdig.dialup.mindspring.COM!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:95070 -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 10 Sep 2001 14:55:10 +0200, Neil Franklin wrote: >I can still download the PDP-10 reference manual and build one. >Engineering outlives business. Just like the pyramids outlived the >pharaoes. Perhaps we are the last one to have the laugh, after all. One of the "Seven Wonders of the World" in classical times was a lighthouse in Alexandria, Egypt, called the Pharos. When it was built, a plaque was placed in its base, honoring the current ruler, Ptolemy II. However, the architect had played a clever trick. He had a stone carved with an inscription honoring _him_, and had it set slightly recessed into the lighthouse. This was then covered with a layer of soft plaster, into which the official inscription was carved. Within a few years, the plaster had weathered away, leaving the more durable inscription bearing the architect's name. -----BEGIN PGP SIGNATURE----- Version: PGPfreeware 7.0.3 for non-commercial use iQA/AwUBPACBujMYPge5L34aEQJ8OQCcD4Y6CT0ci+HgmAK3tu+sIy1d2l0AoMo5 FO0cT0wHIc0kspgzcHQ7TMtb =G7Hv -----END PGP SIGNATURE----- -- John F. Eldredge -- eldredge@earthlink.net, eldredge@poboxes.com PGP key available from http://pgpkeys.mit.edu:11371 "There must be, not a balance of power, but a community of power; not organized rivalries, but an organized common peace." Woodrow Wilson ###### From: bhk@dsl.co.uk (Brian {Hamilton Kelly}) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Mon, 26 Nov 2001 01:54:37 GMT Organization: Dragonhill Systems Ltd Message-ID: <1006739677snz@dsl.co.uk> References: X-Trace: mail2news.demon.co.uk 1006759647 mail2news:2741 mail2news mail2news.demon.co.uk X-Complaints-To: abuse@demon.net X-Mail2News-Path: news.demon.net!dsl.demon.co.uk X-Newsreader: Demon Internet Simple News v1.31 Lines: 31 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!news0.de.colt.net!colt.net!dispose.news.demon.net!news.demon.co.uk!demon!mail2news.demon.co.uk!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:95152 In article eldredge@earthlink.net "John F. Eldredge" writes: > One of the "Seven Wonders of the World" in classical times was a > lighthouse in Alexandria, Egypt, called the Pharos. When it was > built, a plaque was placed in its base, honoring the current ruler, > Ptolemy II. However, the architect had played a clever trick. He > had a stone carved with an inscription honoring _him_, and had it set > slightly recessed into the lighthouse. This was then covered with a > layer of soft plaster, into which the official inscription was > carved. Within a few years, the plaster had weathered away, leaving > the more durable inscription bearing the architect's name. An example of the Pharos outliving the Pharoe? Who was the architect? Is his name preserved for posterity? (Bearing in mind that the remains of the Pharos, and other buildings of the period, have recently been rediscovered.) On a related note, what WAS the name of the idiot that burnt down the greatest library of its age, also at Alexandria? He did so in order that he should be remembered for all time as having committed such a heinous act. Buggered if I can remember his name though, and none of my reference books seem to be able to assist me in my search. -- Brian {Hamilton Kelly} bhk@dsl.co.uk "We have gone from a world of concentrated knowledge and wisdom to one of distributed ignorance. And we know and understand less while being incr- easingly capable." Prof. Peter Cochrane, formerly of BT Labs ###### From: Ketil Z Malde Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 26 Nov 2001 09:32:30 +0100 Organization: University of Bergen Lines: 14 Message-ID: References: <1006739677snz@dsl.co.uk> NNTP-Posting-Host: apal.ii.uib.no X-Trace: toralf.uib.no 1006763553 314 129.177.16.7 (26 Nov 2001 08:32:33 GMT) X-Complaints-To: abuse@uib.no NNTP-Posting-Date: 26 Nov 2001 08:32:33 GMT X-Newsreader: Gnus v5.7/Emacs 20.7 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!skynet.be!skynet.be!news.tele.dk!small.news.tele.dk!128.39.3.168!uninett.no!nntp.uib.no!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:95119 bhk@dsl.co.uk (Brian {Hamilton Kelly}) writes: > On a related note, what WAS the name of the idiot that burnt down the > greatest library of its age, also at Alexandria? He did so in order that > he should be remembered for all time as having committed such a heinous > act. Buggered if I can remember his name though, Apparently, it failed to work. (Anyway, ISTR the library was burnt down three times or so, each time with scholars rushing to save what they could of their dwindling treasures.) -kzm ###### From: "Walter Rottenkolber" Newsgroups: comp.arch,alt.folklore.computers References: <1006739677snz@dsl.co.uk> Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Mon, 26 Nov 2001 02:25:44 -0800 Lines: 32 X-Newsreader: Microsoft Outlook Express 4.72.2106.4 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4 NNTP-Posting-Host: 209.155.205.73 X-Original-NNTP-Posting-Host: 209.155.205.73 Message-ID: <3c0236c0$1@news.sierratel.com> X-Trace: news.sierratel.com 1006778048 209.155.205.73 (26 Nov 2001 04:34:08 -0800) Organization: news.sierratel.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!skynet.be!skynet.be!europa.netcrusader.net!usenetserver.com!208.184.7.66!newsfeed1.cidera.com!Cidera!news.sierratel.com Xref: chonsp.franklin.ch alt.folklore.computers:95114 Ketil Z Malde wrote in message ... >bhk@dsl.co.uk (Brian {Hamilton Kelly}) writes: > >> On a related note, what WAS the name of the idiot that burnt down the >> greatest library of its age, also at Alexandria? He did so in order that >> he should be remembered for all time as having committed such a heinous >> act. Buggered if I can remember his name though, > >Apparently, it failed to work. > >(Anyway, ISTR the library was burnt down three times or so, each >time with scholars rushing to save what they could of their dwindling >treasures.) > >-kzm IIRC the Romans accidentally set fire to it during their conquest of Egypt, but much of it was saved. The real destruction came later as Christian and Moslem fanatics in turn cleansed the library of its pagan and infidel scrolls as they swept through. In time, more enlightened Moslem scholars were able to save and translate the few scraps of the ancient manuscripts that came to light here and there. Actually, the library was just part of a larger complex, the equivalent of a university today. Scholars came there from all parts of the Greek and, later, Roman world. Walter Rottenkolber ###### From: Sander Vesik Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Mon, 26 Nov 2001 10:40:38 +0000 (UTC) Lines: 21 Message-ID: <1006771236.321258@haldjas.folklore.ee> References: <1006739677snz@dsl.co.uk> NNTP-Posting-Host: haldjas.folklore.ee X-Trace: kadri.ut.ee 1006771238 28003 193.40.6.121 (26 Nov 2001 10:40:38 GMT) X-Complaints-To: usenet@kadri.ut.ee NNTP-Posting-Date: Mon, 26 Nov 2001 10:40:38 +0000 (UTC) User-Agent: tin/pre-1.4-19990413 ("Endemoniada") (UNIX) (FreeBSD/4.3-RELEASE (i386)) Cache-Post-Path: haldjas.folklore.ee!unknown@localhost X-Cache: nntpcache 2.3.2 (see http://www.nntpcache.org/) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!sunqbc.risq.qc.ca!teleglobe.net!teleglobe.net!news.estpak.ee!news.infonet.ee!news.ut.ee!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:95124 In comp.arch Brian {Hamilton Kelly} wrote: > On a related note, what WAS the name of the idiot that burnt down the > greatest library of its age, also at Alexandria? He did so in order that > he should be remembered for all time as having committed such a heinous > act. Buggered if I can remember his name though, and none of my > reference books seem to be able to assist me in my search. Uhh.. no - he didn't burn down the library, he burnt down an "unrelated" temple > -- > Brian {Hamilton Kelly} bhk@dsl.co.uk > "We have gone from a world of concentrated knowledge and wisdom to one of > distributed ignorance. And we know and understand less while being incr- > easingly capable." Prof. Peter Cochrane, formerly of BT Labs -- Sander +++ Out of cheese error +++ ###### From: pw@panix.com (Paul Wallich) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Mon, 26 Nov 2001 12:08:57 -0500 Organization: PANIX Public Access Internet and UNIX, NYC Lines: 18 Message-ID: References: <1006739677snz@dsl.co.uk> NNTP-Posting-Host: vt-montpelier1a-442.bur.adelphia.net X-Trace: news.panix.com 1006794532 7627 24.48.165.186 (26 Nov 2001 17:08:52 GMT) X-Complaints-To: abuse@panix.com NNTP-Posting-Date: 26 Nov 2001 17:08:52 GMT X-Newsreader: MT-NewsWatcher 2.4.4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!skynet.be!skynet.be!howland.erols.net!newsfeed.mathworks.com!panix!news.panix.com!pw Xref: chonsp.franklin.ch alt.folklore.computers:95153 In article , Ketil Z Malde wrote: >bhk@dsl.co.uk (Brian {Hamilton Kelly}) writes: > >> On a related note, what WAS the name of the idiot that burnt down the >> greatest library of its age, also at Alexandria? He did so in order that >> he should be remembered for all time as having committed such a heinous >> act. Buggered if I can remember his name though, > >Apparently, it failed to work. That's because it wasn't the guy who burned the library of Alexandria, unless there were two such crazies. I think you're thinking of the early- dynasty chinese emperor who burned books because he wanted his age to be known for its flowering of new knowledge. paul ###### Newsgroups: comp.arch,alt.folklore.computers From: alexc@world.std.com (Alex Colvin) Subject: Re: Minimalist design (was Re: Parity - why even or odd) Organization: The World Public Access UNIX, Brookline, MA Message-ID: References: <1006739677snz@dsl.co.uk> <1006771236.321258@haldjas.folklore.ee> Date: Mon, 26 Nov 2001 17:36:32 GMT Lines: 11 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!out.nntp.be!propagator-SanJose!in.nntp.be!micro-heart-of-gold.mit.edu!bloom-beacon.mit.edu!world!alexc Xref: chonsp.franklin.ch alt.folklore.computers:95104 > On a related note, what WAS the name of the idiot that burnt down the > greatest library of its age, also at Alexandria? He did so in order that > he should be remembered for all time as having committed such a heinous > act. Buggered if I can remember his name though, and none of my > reference books seem to be able to assist me in my search. Erostratus burned the Temple of Artemis http://www.skidmore.edu/~j_cotton/ephesus/artemis.html -- mac the naïf ###### From: bhk@dsl.co.uk (Brian {Hamilton Kelly}) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Mon, 26 Nov 2001 22:08:08 GMT Organization: Dragonhill Systems Ltd Message-ID: <1006812488snz@dsl.co.uk> References: <1006739677snz@dsl.co.uk> X-Trace: mail2news.demon.co.uk 1006821911 mail2news:13086 mail2news mail2news.demon.co.uk X-Complaints-To: abuse@demon.net X-Mail2News-Path: news.demon.net!dsl.demon.co.uk X-Newsreader: Demon Internet Simple News v1.31 Lines: 14 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!newsfeed.gamma.ru!Gamma.RU!dispose.news.demon.net!news.demon.co.uk!demon!mail2news.demon.co.uk!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:95305 In article <1006739677snz@dsl.co.uk> bhk@dsl.co.uk "Brian {Hamilton Kelly}" writes: > An example of the Pharos outliving the Pharoe? Where the heck did that spelling come from? I meant to write, of course, "Pharoah". -- Brian {Hamilton Kelly} bhk@dsl.co.uk "We have gone from a world of concentrated knowledge and wisdom to one of distributed ignorance. And we know and understand less while being incr- easingly capable." Prof. Peter Cochrane, formerly of BT Labs ###### From: bhk@dsl.co.uk (Brian {Hamilton Kelly}) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Mon, 26 Nov 2001 22:13:31 GMT Organization: Dragonhill Systems Ltd Message-ID: <1006812811snz@dsl.co.uk> References: <1006739677snz@dsl.co.uk> <1006771236.321258@haldjas.folklore.ee> X-Trace: mail2news.demon.co.uk 1006821911 mail2news:13090 mail2news mail2news.demon.co.uk X-Complaints-To: abuse@demon.net X-Mail2News-Path: news.demon.net!dsl.demon.co.uk X-Newsreader: Demon Internet Simple News v1.31 Lines: 26 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!newsfeeder.inwind.it!inwind.it!teaser.fr!proxad.net!news-hub.cableinet.net!blueyonder!shale.ftech.net!news.ftech.net!dispose.news.demon.net!news.demon.co.uk!demon!mail2news.demon.co.uk!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:95302 In article alexc@world.std.com "Alex Colvin" writes: > > On a related note, what WAS the name of the idiot that burnt down the > > greatest library of its age, also at Alexandria? He did so in order that > > he should be remembered for all time as having committed such a heinous > > act. Buggered if I can remember his name though, and none of my > > reference books seem to be able to assist me in my search. > > Erostratus burned the Temple of Artemis > http://www.skidmore.edu/~j_cotton/ephesus/artemis.html Indeed; however, Ephesus is quite a long way from Alexandria, and it was /definitely/ the Library at Alexandria that was burnt down as a deliberate (?) act. (It may be that Sander Vesik is right in that he was burning down a temple which just happened to be nearby to the library, but I was always under the impression that his intention was to destroy this seat of learning.) -- Brian {Hamilton Kelly} bhk@dsl.co.uk "We have gone from a world of concentrated knowledge and wisdom to one of distributed ignorance. And we know and understand less while being incr- easingly capable." Prof. Peter Cochrane, formerly of BT Labs ###### From: Jeff Teunissen Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Organization: Dusk To Dawn Computing Lines: 17 Message-ID: <3C02F731.DA2607D@d2dc.net> References: <1006739677snz@dsl.co.uk> <1006812488snz@dsl.co.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.4.9 i686) X-Accept-Language: en-US, en Date: Tue, 27 Nov 2001 02:20:02 GMT NNTP-Posting-Host: 24.13.35.106 X-Complaints-To: abuse@home.net X-Trace: news1.rdc1.mi.home.com 1006827602 24.13.35.106 (Mon, 26 Nov 2001 18:20:02 PST) NNTP-Posting-Date: Mon, 26 Nov 2001 18:20:02 PST Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!wn4feed!wn2feed!worldnet.att.net!24.0.0.38!newshub2.rdc1.sfba.home.com!news.home.com!news1.rdc1.mi.home.com.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:95293 Brian {Hamilton Kelly} wrote: > > In article <1006739677snz@dsl.co.uk> > bhk@dsl.co.uk "Brian {Hamilton Kelly}" writes: > > > An example of the Pharos outliving the Pharoe? > > Where the heck did that spelling come from? I meant to write, of > course, "Pharoah". Or "Pharaoh", even. -- | Jeff Teunissen -=- Pres., Dusk To Dawn Computing -=- deek @ d2dc.net | GPG: 1024D/9840105A 7102 808A 7733 C2F3 097B 161B 9222 DAB8 9840 105A | Core developer, The QuakeForge Project http://www.quakeforge.net/ | Specializing in Debian GNU/Linux http://www.d2dc.net/~deek/ ###### From: John F. Eldredge Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Mon, 26 Nov 2001 20:54:12 -0600 Lines: 50 Message-ID: References: <1006739677snz@dsl.co.uk> NNTP-Posting-Host: dialup-153-107.tnnas2.usit.net (216.80.153.107) Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: fu-berlin.de 1006829646 5394184 216.80.153.107 (16 [61967]) X-Newsreader: Forte Agent 1.8/32.548 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!newsfeeder.inwind.it!inwind.it!nntp.infostrada.it!fu-berlin.de!uni-berlin.de!dialup-153-107.tnnas2.usit.NET!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:95277 -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Mon, 26 Nov 2001 01:54:37 GMT, bhk@dsl.co.uk (Brian {Hamilton Kelly}) wrote: >In article > eldredge@earthlink.net "John F. Eldredge" writes: > >> One of the "Seven Wonders of the World" in classical times was a >> lighthouse in Alexandria, Egypt, called the Pharos. When it was >> built, a plaque was placed in its base, honoring the current >> ruler, Ptolemy II. However, the architect had played a clever >> trick. He had a stone carved with an inscription honoring _him_, >> and had it set slightly recessed into the lighthouse. This was >> then covered with a layer of soft plaster, into which the official >> inscription was >> carved. Within a few years, the plaster had weathered away, >> leaving the more durable inscription bearing the architect's name. > >An example of the Pharos outliving the Pharoe? > >Who was the architect? Is his name preserved for posterity? >(Bearing in mind that the remains of the Pharos, and other buildings >of the period, have recently been rediscovered.) "Sostratus, the son of Dexiphanes, the Cnidian" is how he described himself on the plaque. The account about the plaster layer, with the official description, is from "The Ancient Engineers", an excellent popular history of civil engineering in the ancient world by L. Sprague de Camp. One of the web sites describing the Pharos is http://ce.eng.usf.edu/pharos/Wonders/pharos.html. -----BEGIN PGP SIGNATURE----- Version: PGPfreeware 7.0.3 for non-commercial use iQA/AwUBPAMAODMYPge5L34aEQIfiACgm70v8u05Fin1sPRHTxsnHlCtCioAoOEF FYUdoszzHBZpjRBuOThLmUqm =l387 -----END PGP SIGNATURE----- -- John F. Eldredge -- eldredge@earthlink.net, eldredge@poboxes.com PGP key available from http://pgpkeys.mit.edu:11371 "There must be, not a balance of power, but a community of power; not organized rivalries, but an organized common peace." Woodrow Wilson ###### From: Sander Vesik Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Tue, 27 Nov 2001 10:03:38 +0000 (UTC) Lines: 36 Message-ID: <1006855415.996728@haldjas.folklore.ee> References: <1006739677snz@dsl.co.uk> <1006771236.321258@haldjas.folklore.ee> <1006812811snz@dsl.co.uk> NNTP-Posting-Host: haldjas.folklore.ee X-Trace: kadri.ut.ee 1006855418 20458 193.40.6.121 (27 Nov 2001 10:03:38 GMT) X-Complaints-To: usenet@kadri.ut.ee NNTP-Posting-Date: Tue, 27 Nov 2001 10:03:38 +0000 (UTC) User-Agent: tin/pre-1.4-19990413 ("Endemoniada") (UNIX) (FreeBSD/4.3-RELEASE (i386)) Cache-Post-Path: haldjas.folklore.ee!unknown@localhost X-Cache: nntpcache 2.3.2 (see http://www.nntpcache.org/) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!skynet.be!skynet.be!news1.ebone.net!news.ebone.net!newsfeed.freenet.de!teleglobe.net!news.estpak.ee!news.infonet.ee!news.ut.ee!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:95246 In comp.arch Brian {Hamilton Kelly} wrote: > In article alexc@world.std.com "Alex Colvin" writes: >> > On a related note, what WAS the name of the idiot that burnt down the >> > greatest library of its age, also at Alexandria? He did so in order that >> > he should be remembered for all time as having committed such a heinous >> > act. Buggered if I can remember his name though, and none of my >> > reference books seem to be able to assist me in my search. >> >> Erostratus burned the Temple of Artemis >> http://www.skidmore.edu/~j_cotton/ephesus/artemis.html > Indeed; however, Ephesus is quite a long way from Alexandria, and it was > /definitely/ the Library at Alexandria that was burnt down as a > deliberate (?) act. > (It may be that Sander Vesik is right in > that he was burning down a temple which just happened to be nearby to the > library, but I was always under the impression that his intention was to > destroy this seat of learning.) I definately said *unrelated* temple. The Library of Alexandria was not ever burnt down for the purpose of immortalising ones name. I'm quite sure of that. > -- > Brian {Hamilton Kelly} bhk@dsl.co.uk > "We have gone from a world of concentrated knowledge and wisdom to one of > distributed ignorance. And we know and understand less while being incr- > easingly capable." Prof. Peter Cochrane, formerly of BT Labs -- Sander +++ Out of cheese error +++ ###### From: eal@etek.chalmers.se (Andreas Långström) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 27 Nov 2001 10:16:22 GMT Organization: Chalmers University of Technology Lines: 25 Message-ID: <9tvp5m$562$1@nyheter.chalmers.se> References: <1006771236.321258@haldjas.folklore.ee> <1006812811snz@dsl.co.uk> NNTP-Posting-Host: hortensia.etek.chalmers.se X-Trace: nyheter.chalmers.se 1006856182 5314 129.16.32.92 (27 Nov 2001 10:16:22 GMT) X-Complaints-To: abuse@chalmers.se NNTP-Posting-Date: 27 Nov 2001 10:16:22 GMT X-Newsreader: trn 4.0-test72 (19 April 1999) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.tele.dk!small.news.tele.dk!195.54.122.107!newsfeed1.bredband.com!bredband!newsfeed1.telenordia.se!algonet!newsfeed.sunet.se!news01.sunet.se!news.chalmers.se!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:95241 Brian {Hamilton Kelly} wrote: > >Indeed; however, Ephesus is quite a long way from Alexandria, and it was >/definitely/ the Library at Alexandria that was burnt down as a >deliberate (?) act. The Alexandria library was burnt 47 bc by Julius Caesar, when he set fire to the Alexandrian fleet in the harbour, near the library. Cleopatra, with help from Marcus Antonius rebuilt it. In 389, Theodosius I ordered all pagan monuments destroyed (he was christian), and the library had unfortunately been housed in an temple. Those who tried to save some books were slaughtered by the mob. The last remains that had survived the christian onslaught were destroyed in 652 by the moslem kalif Omar. Asked what to do with the library, he allegedly said that the books not agreeing with the Koran was to be destroyed as heretic, and those that did agree were superfluos and should be destroyed. And then the dark ages began... /Andreas Facts stolen from "Fermats riddle", Simon Singh ###### Message-ID: <3C0440A4.F542B836@ev1.net> From: Charles Richmond Reply-To: richmond@ev1.net Organization: Cannine Computer Center X-Mailer: Mozilla 4.7C-CCK-MCD {C-UDP; EBM-APPLE} (Macintosh; I; PPC) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <1006771236.321258@haldjas.folklore.ee> <1006812811snz@dsl.co.uk> <9tvp5m$562$1@nyheter.chalmers.se> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Lines: 34 Date: Tue, 27 Nov 2001 23:44:22 GMT NNTP-Posting-Host: 24.179.111.125 X-Complaints-To: abuse@home.net X-Trace: news2.rdc2.tx.home.com 1006904662 24.179.111.125 (Tue, 27 Nov 2001 15:44:22 PST) NNTP-Posting-Date: Tue, 27 Nov 2001 15:44:22 PST Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed00.sul.t-online.de!t-online.de!nntp-relay.ihug.net!ihug.co.nz!logbridge.uoregon.edu!newsfeed.direct.ca!look.ca!newshub2.rdc1.sfba.home.com!news.home.com!news2.rdc2.tx.home.com.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:95290 "Andreas Långström" wrote: > > Brian {Hamilton Kelly} wrote: > > > >Indeed; however, Ephesus is quite a long way from Alexandria, and it was > >/definitely/ the Library at Alexandria that was burnt down as a > >deliberate (?) act. > > The Alexandria library was burnt 47 bc by Julius Caesar, when he > set fire to the Alexandrian fleet in the harbour, near the library. > Cleopatra, with help from Marcus Antonius rebuilt it. > Quite a trick, since Julius Caesar was assassinated in 50 bc... > > In 389, Theodosius I ordered all pagan monuments destroyed (he was > christian), and the library had unfortunately been housed in an temple. > Those who tried to save some books were slaughtered by the mob. > > The last remains that had survived the christian onslaught were destroyed > in 652 by the moslem kalif Omar. Asked what to do with the library, he > allegedly said that the books not agreeing with the Koran was to be > destroyed as heretic, and those that did agree were superfluos and > should be destroyed. > > And then the dark ages began... > Actually, from the 600's A.D. to at least 1000 A.D., Arab scholarship flourished. In fact, the way Europe got out of the Dark Ages was by conquest of the Moors in Spain, and the discovery of the Alhambra. -- +-------------------------------------------------------------+ | Charles and Francis Richmond | +-------------------------------------------------------------+ ###### From: spam+@orion-com.com (Joe Thompson) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 28 Nov 2001 07:27:12 GMT Organization: MindSpring Enterprises Lines: 14 Message-ID: References: <1006771236.321258@haldjas.folklore.ee> <1006812811snz@dsl.co.uk> <9tvp5m$562$1@nyheter.chalmers.se> <3C0440A4.F542B836@ev1.net> NNTP-Posting-Host: 3f.1a.c6.50 Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit User-Agent: slrn/0.9.6.3 (Linux) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed1.earthlink.net!newsfeed2.earthlink.net!newsfeed.earthlink.net!news.mindspring.net!spam+ Xref: chonsp.franklin.ch alt.folklore.computers:95310 On Tue, 27 Nov 2001 23:44:22 GMT, Charles Richmond wrote: > "Andreas Långström" wrote: > > The Alexandria library was burnt 47 bc by Julius Caesar, when he > > set fire to the Alexandrian fleet in the harbour, near the library. > > Cleopatra, with help from Marcus Antonius rebuilt it. > > Quite a trick, since Julius Caesar was assassinated in 50 bc... ITYM "44 BC". HTH. HAND. -- Joe -- Joe Thompson | http://www.orion-com.com/~kensey/ spam+@orion-com.com | PGP key: Finger joe-jobs@mindspring.com Yeah, it's crazy, but I just ate a giant pixy stick, and I'm excited because my new computer is coming today. -- Apreche on Slashdot ###### From: bhk@dsl.co.uk (Brian {Hamilton Kelly}) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Wed, 28 Nov 2001 08:08:50 GMT Organization: Dragonhill Systems Ltd Message-ID: <1006934930snz@dsl.co.uk> References: <1006771236.321258@haldjas.folklore.ee> <1006812811snz@dsl.co.uk> <9tvp5m$562$1@nyheter.chalmers.se> <3C0440A4.F542B836@ev1.net> X-Trace: mail2news.demon.co.uk 1006973925 mail2news:10477 mail2news mail2news.demon.co.uk X-Complaints-To: abuse@demon.net X-Mail2News-Path: news.demon.net!dsl.demon.co.uk X-Newsreader: Demon Internet Simple News v1.31 Lines: 16 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!dispose.news.demon.net!news.demon.co.uk!demon!mail2news.demon.co.uk!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:95369 In article <3C0440A4.F542B836@ev1.net> richmond@ev1.net "Charles Richmond" writes: > Actually, from the 600's A.D. to at least 1000 A.D., Arab scholarship > flourished. In fact, the way Europe got out of the Dark Ages was by > conquest of the Moors in Spain, and the discovery of the Alhambra. The oldest university in the world is not Padua or one of those other early European seats of learning, but Fez in Morocco. -- Brian {Hamilton Kelly} bhk@dsl.co.uk "We have gone from a world of concentrated knowledge and wisdom to one of distributed ignorance. And we know and understand less while being incr- easingly capable." Prof. Peter Cochrane, formerly of BT Labs ###### Message-ID: <3C04C9BB.DC2A20DB@yahoo.com> From: CBFalconer Reply-To: cbfalconer@worldnet.att.net Organization: Ched Research X-Mailer: Mozilla 4.75 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) References: <1006771236.321258@haldjas.folklore.ee> <1006812811snz@dsl.co.uk> <9tvp5m$562$1@nyheter.chalmers.se> <3C0440A4.F542B836@ev1.net> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Lines: 24 Date: Wed, 28 Nov 2001 16:41:50 GMT NNTP-Posting-Host: 12.90.178.193 X-Complaints-To: abuse@worldnet.att.net X-Trace: bgtnsc06-news.ops.worldnet.att.net 1006965710 12.90.178.193 (Wed, 28 Nov 2001 16:41:50 GMT) NNTP-Posting-Date: Wed, 28 Nov 2001 16:41:50 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed.stanford.edu!logbridge.uoregon.edu!arclight.uoregon.edu!wn4feed!worldnet.att.net!135.173.83.71!wnfilter1!worldnet-localpost!bgtnsc06-news.ops.worldnet.att.net.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:95330 Charles Richmond wrote: > > "Andreas Långström" wrote: > > ... snip ... > > > > And then the dark ages began... > > > Actually, from the 600's A.D. to at least 1000 A.D., Arab scholarship > flourished. In fact, the way Europe got out of the Dark Ages was by > conquest of the Moors in Spain, and the discovery of the Alhambra. I understand it more to be the conquest of Constantinople, and the arrival in the West of expatriate Byzantines. Latter half of the 15th century for both events. At any rate, cross-cultural fertilization. -- Chuck F (cbfalconer@yahoo.com) (cbfalconer@XXXXworldnet.att.net) Available for consulting/temporary embedded and systems. (Remove "XXXX" from reply address. yahoo works unmodified) mailto:uce@ftc.gov (for spambots to harvest) ###### From: eal@etek.chalmers.se (Andreas Långström) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 28 Nov 2001 17:15:16 GMT Organization: Chalmers University of Technology Lines: 36 Message-ID: <9u3634$d69$1@nyheter.chalmers.se> References: <1006812811snz@dsl.co.uk> <9tvp5m$562$1@nyheter.chalmers.se> <3C0440A4.F542B836@ev1.net> NNTP-Posting-Host: hortensia.etek.chalmers.se X-Trace: nyheter.chalmers.se 1006967716 13513 129.16.32.92 (28 Nov 2001 17:15:16 GMT) X-Complaints-To: abuse@chalmers.se NNTP-Posting-Date: 28 Nov 2001 17:15:16 GMT X-Newsreader: trn 4.0-test72 (19 April 1999) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!uni-erlangen.de!newsfeed1.telenordia.se!algonet!newsfeed.sunet.se!news01.sunet.se!news.chalmers.se!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:95343 Charles Richmond wrote: >"Andreas Långström" wrote: >> Brian {Hamilton Kelly} wrote: >> >> The Alexandria library was burnt 47 bc by Julius Caesar, when he >> set fire to the Alexandrian fleet in the harbour, near the library. >> Cleopatra, with help from Marcus Antonius rebuilt it. >> >Quite a trick, since Julius Caesar was assassinated in 50 bc... http://www.incwell.com/Biographies/Caesar.html claims 44 bc http://www.vroma.org/~bmcmanus/caesar.html claims March 15 44 bc (and speaks of the library burning incident) http://www.livius.org/caa-can/caesar/caesar01.html claims March 15 44 bc etc. etc. etc. <> >> And then the dark ages began... >> >Actually, from the 600's A.D. to at least 1000 A.D., Arab scholarship >flourished. In fact, the way Europe got out of the Dark Ages was by >conquest of the Moors in Spain, and the discovery of the Alhambra. Gerbert d'Aurillac learnt the "new" (current) way of representing and calculating numbers, in the 900's, from the Moors. And he became a pope in 999 (but lets pretend this is a.f.u and just say BoR BoR BoR instead of pursuing the link between the church and the Dark Ages). Next major improvement of European math came in 1435 when Bysantine scholars fled west to escape the Turks invading Constantinopel. /Andreas ###### From: spam+@orion-com.com (Joe Thompson) Newsgroups: comp.arch,alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 28 Nov 2001 18:27:14 GMT Organization: MindSpring Enterprises Lines: 27 Message-ID: References: <1006771236.321258@haldjas.folklore.ee> <1006812811snz@dsl.co.uk> <9tvp5m$562$1@nyheter.chalmers.se> <3C0440A4.F542B836@ev1.net> <3C04C9BB.DC2A20DB@yahoo.com> NNTP-Posting-Host: 3f.17.7d.66 User-Agent: slrn/0.9.6.3 (Linux) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!netnews.com!xfer02.netnews.com!newsfeed2.earthlink.net!newsfeed.earthlink.net!news.mindspring.net!spam+ Xref: chonsp.franklin.ch alt.folklore.computers:95372 On Wed, 28 Nov 2001 16:41:50 GMT, CBFalconer wrote: > Charles Richmond wrote: > > Actually, from the 600's A.D. to at least 1000 A.D., Arab scholarship > > flourished. In fact, the way Europe got out of the Dark Ages was by > > conquest of the Moors in Spain, and the discovery of the Alhambra. > > I understand it more to be the conquest of Constantinople, and the > arrival in the West of expatriate Byzantines. Latter half of the > 15th century for both events. At any rate, cross-cultural > fertilization. There were reservoirs of learning that helped things along, of course. Celtic monasteries, Arabic and other Muslim scholars, the Moors, the Copts... Weren't a lot of older Greek and Roman works unearthed around that time too? The other question is how much of the science of the Renaissance was dependent on those (re)discoveries. For example, the line from Brahe through Kepler to Newton pretty much starts out of thin air with Brahe's naked-eye observations. IIRC the Arabs systematized astronomy (which is why so many stars bear Arabic-derived names) but didn't really develop it much. -- Joe -- Joe Thompson | http://www.orion-com.com/~kensey/ spam+@orion-com.com | PGP key: Finger joe-jobs@mindspring.com Yeah, it's crazy, but I just ate a giant pixy stick, and I'm excited because my new computer is coming today. -- Apreche on Slashdot ###### From: ab528@FreeNet.Carleton.CA (Heinz W. Wiggeshoff) Newsgroups: alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: 28 Nov 2001 20:37:05 GMT Organization: The National Capital FreeNet Lines: 7 Message-ID: <9u3hth$i2a$1@freenet9.carleton.ca> References: <1006771236.321258@haldjas.folklore.ee> <1006812811snz@dsl.co.uk> <9tvp5m$562$1@nyheter.chalmers.se> <3C0440A4.F542B836@ev1.net> <1006934930snz@dsl.co.uk> Reply-To: ab528@FreeNet.Carleton.CA (Heinz W. Wiggeshoff) NNTP-Posting-Host: freenet10 X-Trace: freenet9.carleton.ca 1006979825 18506 134.117.136.30 (28 Nov 2001 20:37:05 GMT) X-Complaints-To: complaints@ncf.ca NNTP-Posting-Date: 28 Nov 2001 20:37:05 GMT X-Given-Sender: ab528@freenet10.carleton.ca (Heinz W. Wiggeshoff) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newscore.univie.ac.at!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed.stanford.edu!news.kjsl.com!xcski.com!freenet-news!FreeNet.Carleton.CA!ab528 Xref: chonsp.franklin.ch alt.folklore.computers:95347 Brian {Hamilton Kelly} (bhk@dsl.co.uk) writes: > > The oldest university in the world is not Padua or one of those other > early European seats of learning, but Fez in Morocco. This goes a long way to explaining the funny looking hats the grads wear! ###### From: David Powell Newsgroups: alt.folklore.computers Subject: Re: Minimalist design (was Re: Parity - why even or odd) Date: Fri, 30 Nov 2001 21:12:16 +0000 Reply-To: ddotpowell@netscapeonline.co.uk Message-ID: <79tf0ug7d446s3gn1n8bj2s88akv0q36ni@4ax.com> References: <1006812811snz@dsl.co.uk> <9tvp5m$562$1@nyheter.chalmers.se> <3C0440A4.F542B836@ev1.net> <9u3634$d69$1@nyheter.chalmers.se> <9u6jah04ju@enews2.newsguy.com> X-Newsreader: Forte Agent 1.8/32.548 MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Lines: 22 X-Report: Report abuse to abuse@netscapeonline.co.uk X-Original-NNTP-Posting-Host: 212.137.251.154 NNTP-Posting-Host: 10.20.1.166 X-Trace: 30 Nov 2001 20:16:16 GMT, 10.20.1.166 X-Report: Report abuse to abuse@netscapeonline.co.uk Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newscore.univie.ac.at!194.25.134.126.MISMATCH!newsfeed01.sul.t-online.de!t-online.de!fr.clara.net!heighliner.fr.clara.net!opentransit.net!newsfeed.icl.net!iclnet!plato.netscapeonline.co.uk!10.20.1.166 Xref: chonsp.franklin.ch alt.folklore.computers:95568 In article <9u6jah04ju@enews2.newsguy.com>, mwojcik@newsguy.com (Michael Wojcik) in alt.folklore.computers wrote: > >In article <9u3634$d69$1@nyheter.chalmers.se>, eal@etek.chalmers.se (Andreas Långström) writes: >> Charles Richmond wrote: >> >Quite a trick, since Julius Caesar was assassinated in 50 bc... >> >> http://www.incwell.com/Biographies/Caesar.html claims 44 bc > >Perhaps Charles was using base 8.8? > >(Frankly, the only date I can ever remember is 1066. What's the other >one?) Yup, Sellar and Yeatman were right. Regards, David P.