And here is the backside of Moore's Law: Lile's Upgrade Law. It works like this: Lile will not upgrade until his new computer can kick his old computer's a** by 10X. Many consumers think the same way, consciously or unconsciously. So my first computer had 10MB hard drives, 4MHZ processor, and the next one had 100MB hard drive and 40 MHZ processor, about 10X as much memory, and brand new features like sound cards. The law has held pretty well, give or take, until now. I would not upgrade until I truly needed a faster computer and could get a whopping boost in performance. This has kept me on the dying tail edge of computer technology for 25 years, and also has saved a bundle of cash. I haven't suffered a bit, either. I finally am retiring my 1998 computer. It does everything I need to do, and I have no need to upgrade for any reason of speed. My Hard drive is half empty despite a huge collection of MP3's, my processor seems to handle anything I can throw at it, and the only reason I have to upgrade is to get a more reliable box. The old box surfs just as fast as a new box, since the Internet is usually the bottleneck, it plays nice sound, displays movies and flash files, I can't think of a single feature that I don't already have that I would want in a new machine, except better reliability and maybe a 50% larger screen. Neither of those is controlled by Moore's Law, and both could have been had on the old machine for a few bucks. Egad's, I remember the jump from a monochrome DOS machine to a VGA windows machine, this was a real tangible upgrade! Wow! I won't see anything like this going from a WIN98 pentium to a WINXP Pentium. Likewise the jump from a 386 to a 486 running Autocad - screen refreshes went from 30 second slogs to instant, there was a visible increase in productivity. So for the first time I am settling for a hard drive only about 2X as big, but am going for reliability - a RAID array instead of a single drive. I bought a relatively slow computer by today's standards, deeming a 2.6 gig processor as good as a 3.x gig processor, what is the diff? I would have bought a slower one if they were offered. If I am any indication of what the market wants (quite a stretch, my needs are oddball) then the days of exponential growth in computer power are numbered. The car industry, if it experienced Moore's law growth, would now be sporting 2000MPH cars that got 5000 miles per gallon and carried 453 passengers in a space the size of a bicycle. Obviously, I am riding around in a four passenger car just like my parents did, a bit more comfy but well within an order of magnitude on all performance criteria. Smaller, in fact. And safer. The PC industry is soon to go the same way, who needs a 1000GHZ machine when a 200MHZ Pentium will surf the web just the same? Nobody will pay for the huge new Fabs that are required for such progress, the PC industry has already seen the death of Moore's law, but they haven't realized it yet. -- Lawrence Lile, P.E. Electrical and Electronic Solutions Project Solutions Companies www.projsolco.com > -----Original Message----- > From: Russell McMahon [mailto:apptech@paradise.net.nz] > Sent: Monday, December 13, 2004 4:58 PM > To: PIC List > Subject: [EE] Moore's law failing ? > > Moore's law informally predicts that "microprocessor" processing power > will > double every 18 months. (Originally it was 12 months and related to > transistors in an IC but it was extended long ago to the currently > accepted > meaning.) For many years people have been saying that it can't last, but > Moore's law just kept on being true. Each 3 years you could expect a 4 > times > increase in processing power. The same figures *roughly* also applied to > disk capacity, RAM size on a typical system and CD rom speed fwiw. > However, > > A year ago if I'd bough the latest and greatest PC it would have had a > Pentium IV with a 3 GHZ processor. This Christmas if you want the latest > and > greatest it may be a Pentium IV with a 3.3 GHz processor. 3 GHz is still > so > close to the leading edge that it's hardly worthwhile. The Pentium IV 4 > GHZ > which was originally to be introduced late 2004, and then early 2005 has > been cancelled. The Pentium IV 3.6 GHz will be introduced sometime on > 2005. > Instead of the 4 GHz part they are going to introduce their "dual core" > processor. > > Now the new "Prescott" processor does have an 800 MHz FSB as opposed to > the > 400 or ?533? MHz of previous models so there will be some extra speed > there, > but not a vast amount. But overall the advances seem to have stumbled. > Dual > core is NOT an advance along the Moore's law curve - it's desperate > (albeit > possibly effectual) cheating. And the competition do not seem to be taking > the opportunity to leapfrog over Intel and take the lead in processor > speed. > Something fundamental seems to have happened. There was nothing sacrosanct > about Moore's law. There was no reason that it had to work and no reason > why > we should be disturbed when it doesn't anymore. After all it was just an > observation that seemed to fit the facts without any known basis. Just > like > Quantum Mechanics is :-) > > > > Russell McMahon > > > > _______________________________________________ > http://www.piclist.com PIC/SX FAQ & list archive > View/change your membership options at > http://mailman.mit.edu/mailman/listinfo/piclist > > --- > Incoming mail is certified Virus Free. > Checked by AVG anti-virus system (http://www.grisoft.com). > Version: 6.0.762 / Virus Database: 510 - Release Date: 9/13/2004 > --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.762 / Virus Database: 510 - Release Date: 9/13/2004 _______________________________________________ http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist