Author: Dann Corbit
Date: 13:01:57 05/21/04
Go up one level in this thread
On May 21, 2004 at 14:30:51, Dieter Buerssner wrote: >On May 20, 2004 at 21:48:25, Dann Corbit wrote: > >>This is also too low level for the most part, but well worth reading anyway: >>http://www.azillionmonkeys.com/qed/optimize.html > >I had a fast look at this one. Certainly many useful suggestions, but some also >seem doubtful. There is some emphasis on microoptimizations, that probably many >modern compilers will do as well, and perhaps better from the normal code, than >from the already microoptimized code. I also use often do while loops instead of >for (when it is clear that the first comparision in the "end-condition" will >always be true). In the example given in the document, only one out of 100 >comparisions will be saved. The prize is, that the data will be accessed from >back to front. I guess on many computers, this will make the code slower. >Suggestions about some floating point optimizations seem dangerous. I agree that some of the stuff might be a little dubious, and I think the "tweaking" angle is way too strong. But when it comes to making something fast, Paul is very good at it. Algorithms are the sensible way to optimize. But if you can't find a better one (anyone figured out how to make chess polynomial time yet instead of exponential?) then some of his ideas are rather useful.
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.