Author: Greg Lindahl
Date: 21:40:04 12/20/99
Go up one level in this thread
On December 20, 1999 at 21:49:12, Robert Hyatt wrote: >That's where you are wrong. A simple analytical approach: pick any part >of the engine you want, and assume the time for that part can be driven to >0.00 by hardware. Amdahl's law steps in quickly for the parts you can't >drive to zero, and gives you a limit. Right. So on my example engine which spends 90% of its time in eval, you get at most a 10x speedup. That's why I brought up that particular example. However, you can then use a stronger eval, and get even more than 10x quality. > The search is at _least_ 10% of any >program I know of. That means that there is _no way_ to speed any program >up by more than a factor of 10, If you ignore quality. >I understand alpha/beta searching _perfectly_. If that is what you mean. If >you are asking "would you like to have an eval in hardware so you can keep >adding more and more knowledge without slowing the program down?" then I would >say "heck yes". Wow, so we finally agree on something. >But that is _not_ going to produce a revolutionarily strong >program in a year. Did I say that anything would produce a revolutionarily strong program in a year? Nope. Perhaps this kind of statement is part of your conversational style, but I am wondering where you got the 1 year figure, why you think it's important, and so forth. You sure aren't answering a question I asked. >The ASIC is the _right_ way to solve this problem. I'm glad that you still know all the answers. -- g
This page took 0 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.