Author: Dann Corbit
Date: 14:40:18 09/19/02
Go up one level in this thread
On September 19, 2002 at 17:05:42, David Hanley wrote: >On September 19, 2002 at 16:53:54, Uri Blass wrote: >>I think that speed is the most important thing that can help you to improve your >>program. >> >>Doing your program more than 10 times faster is more important than everything >>that you can add to your evaluation in the near future. >> >>Tscp is not optimized for speed and if tscp is clearly faster than your program >>then it mean that you can get significant improvement by speed improvement. > >You are right. I have not considered speed much at all while programming. My >idea is to get the program entirely correct with all the features i want, then >to worry about speed if it's still not playing as well as i hope. But, i am not >intending it to compete against the likes of crafty and yace, but just to play >well ( 2200-2400 ) while being short and clear. I think i can do that in under >64K of non-obfuscated source code, but i'll see.. How to get the speed is the key thing, though. You might add hash tables to TSCP and it will double the speed. You might add null move to TSCP, again, probably doubling the speed. You might fix the iteration in eval.c to use a move list instead of iteration over the whole board. If you go to individual piece lists, it will be about 4x faster in the early game, and 10x faster in the endgame. Combine them all together and it will be quite a speedup. Also, you should profile to find where the program bottlenecks are. Sometimes it is not the routine that takes the most time that is the bottleneck, because a function can act like a gate. The Intel profiler has a really nice hot spot locator. >>If you think about adding evaluation then my opinion is that mobility is the >>most important thing to add. > >I need to try that--i currently don't consider it at all, but i know i should.. > >dave
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.