Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!julius.cs.uiuc.edu!coolidge From: coolidge@cs.uiuc.edu (John Coolidge) Newsgroups: comp.arch Subject: Re: Sun's Competitive Strategy (Was: Re: P1754 Message-ID: <1990Dec7.030303.20157@julius.cs.uiuc.edu> Date: 7 Dec 90 03:03:03 GMT References: <1990Nov16.225515.494@zoo.toronto.edu> <1990Nov25.194404.3376@dircon.uucp> <1635@unix386.Convergent.COM> <1990Dec2.014554.3491@Neon.Stanford.EDU> <2760@cirrusl.UUCP> <2764@cirrusl.UUCP> <76095@iuvax.cs.indiana.edu> Sender: news@julius.cs.uiuc.edu (USENet News) Reply-To: coolidge@cs.uiuc.edu Distribution: na Organization: U of Illinois, Dept. of Computer Science, Systems Research Group Lines: 85 kls30@duts.ccc.amdahl.com (Kent L Shephard) writes: >In article <1990Dec6.005211.7490@julius.cs.uiuc.edu> coolidge@cs.uiuc.edu (John Coolidge) writes: >>386sx's are pretty much the same, operationally, as 68020's. I've got >I don't agree because you have to get a PMMU on a 68020 which adds wait >states to memory. But besides that they are equivalent. I was comparing after the PMMU, of course. I'm running Unix on my '020 most of the time. >>You're confusing resolution with pixels. The resolution of a Macintosh >>monitor (any Macintosh monitor from Apple and almost all third party >>monitors) is 72+/- 1 dpi. >No I'm not, resolution is the number of pixels, dot pitch determines >dpi. Look at any number of publications and manufacturer's literature >and resolution is listed as 800x600, 1024x768, etc. Manufacturers who don't understand English, either conversational or technical. Resolution has a very clear and precise meaning, and it isn't the number of dots on your (arbitrarily big) monitor. Resolution, in monitors as in printers, is the number of dots placed in a given area. Dot pitch and dpi are both measures of resolution, dots on the screen is not. The term is based on the idea of resolving a point --- how precisely can you display images of a given size? Clearly the number of dots on your display impacts this --- more dots on a display of the same size affects resolution. Adding more dots, but making the screen bigger, doesn't change resolution. It just gives you the ability to put more things on your screen at the same resolution. I'll concede, however, that this is a common mistake, similar to the misuse of hacker as cracker, MIPs as a measure of anything useful, and so forth. I'll also drop the point, since I don't want this to become a terminology flame. Others can pick up the discussion from here. >>The 680x0 is a much cleaner architecture than the 80x86, if for >>no other reason than that the backward compatibility is limited to >>a reasonable processor (8086! gimme a break!). >If I'm not mistaken the 68k is circa 8086. BTW what is wrong w/ backward >compatibility in a processor. If you don't have backward compatibility >you've got a different processor family. Backward compatibility is great and wonderful, as long as it doesn't cost you anything. Staying compatible with a 1980 (or earlier) design (the 8086) is clearly straining the 80x86 family of processors. The 680x0 series is based on a 1982-3 design (the 68000) (BTW, the Lisa, which uses the 68000, came out in 1983 I believe. Possibly 1982). Those two years are one of the reasons why a 68030 is only about twice as slow as a 80486 at the same clock speed and why a 68040 is about half again as fast as the 80486. The 680x0 series maintains backward compatibility with the 68000, which was two years later and based on a number of lessons learned in those two years (as well as borrowing a few more good ideas from the mainframes of yore, from whence quite a lot of microcomputer 'advances' come). >The 68040 is a kluge just like the i486 both use a RISC core wrapped up >in a CISC instruction set. The 68k arch. is not better that the 80x86 >they're just different. I agree; they're both 'kludges' in the sense you describe. They're both processor families that are popular but aging. There are clearly better processors around then either of them by a number of criteria. But, comparing only the two, I believe the 680x0 series to be a better architecture than the 80x86 series, at least from a programmer's point of view (I'm not a hardware designer, I'm an operating systems person. I know which series I'm happier working with). >Before you start blasting away at arch. and screens check you sources. >Mine are industry standard for what is called resolution and dot pitch. I have checked my sources. They indicate (you guessed it) that, on the subject of resolution, I'm correct. On the comparison between 80x86 and 680x0, there's clearly a lot of opinion involved in the comparison. On the factual questions I stand by my sources. --John -------------------------------------------------------------------------- John L. Coolidge Internet:coolidge@cs.uiuc.edu UUCP:uiucdcs!coolidge Of course I don't speak for the U of I (or anyone else except myself) Copyright 1990 John L. Coolidge. Copying allowed if (and only if) attributed. You may redistribute this article if and only if your recipients may as well. Brought to you by Super Global Mega Corp .com