Author: Don Dailey
Date: 12:10:08 05/24/98
Go up one level in this thread
On May 24, 1998 at 14:28:22, Georg Langrath wrote: >Sorry for the bad sentence structure. I posted wrong copy from Word. But >my asking is as you perhaps understood: Which year will there be a >pocketchess, that is as strong as Kasparov. I think there will be such a >pocketchess one day. One can only guess. Year 2005? > >Georg I'll be more than happy to speculate. Let's start with generally available micro's. I think we need a 10x over an Alpha to have the power, but this must be combined with further improvements in software which I'm guessing will come. Software has improved very slowly over the years, but not at the same pace as hardware. By the time the 10X comes, we might be about there. A purely subjective opinion I must admit. How far behind are pocket devices? I don't really know. What is the most powerful pocket device you can get right now (common consumer stuff that is?) There is the palm pilot, using 68000 type of technology with 1 or 2 meg of ram I think. There is 386 based stuff too which I think is more powerful. I think we tend to be at least a couple generations behind here. Someone else will correct me if I'm wrong. Are there 486 handheld machines? Probably. Is the windows CE thingy 486 based? What about power PC stuff that's handheld, is this available? I'll assume about 5 or 6 years behind for palmtop devices all things considered, compared to Alpha's. If anyone want to correct me on this, feel free, I'm just winging it and am not too aware of current developements in the market. What was the rule of thumb for processor advancment? Doubling every n months but I don't remember what n is, something like 18 months or so. If this also is correct we would need 3 doublings to be close. But I'm guessing 4 is safer! That gives us 6 years to acquire the necessary hardware and another 5 or 6 for palmtop technology to catch up. I think you might get your device within about 12 years! Of ocurse this is pretty wild speculation. If someone developed a special "chess chip" for us there could be a revival of computer specific chess devices like we used to have! This might knock a couple years or more off the time. I think there is also a slight trend for the small technology to catch up with the desktop technology in the same way micro's got closer to mainframes and that micro chess programs started competing with mainframe programs. Probably the most speculative part of my estimate is how much extra power do we need. Based on Deep Blue's result it would appear we need a hellishly greater amount that what I said. But I'm assuming we'll be writing better programs too within the next 10 years. Not only that, but I'm not assuming parallelism, and 10 or 20 X serially is a lot more than a 20 processor SMP (parallel shared memory machine.) Of course if we get parallelism this might also knock a little time off, can you imagine an 8 processor SMP in your hand! Pretty cool! My guess is that before too long chips will be built with multiple processors on them. - Don
This page took 0.01 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.