Author: Uri Blass
Date: 01:30:17 04/26/02
Go up one level in this thread
On April 26, 2002 at 04:11:26, Odd Gunnar Malin wrote: >On April 26, 2002 at 03:49:27, Uri Blass wrote: > >>On April 26, 2002 at 03:20:36, Odd Gunnar Malin wrote: >> >>>On April 25, 2002 at 12:01:12, Uri Blass wrote: >>>>My program in the 5th division of the winboard programs >>>>has 9.5 out of 11 >>>>inspite of not having opening book,hash tables,pondering >>>>and even does not use the null move pruning that >>>>seems to be a significant improvement relative to >>>>the previous version. >>>> >>>>Uri >>> >>>As I read before so did you start with an 4 div. engine (tscp) and apply changes >>>to this that would gain speed etc. >>>When you compare with the engines in 5 div. you have to remember that these are >>>engines that start from scratch, with reading ideas and pseudo code and not >>>source codes. >>> >>>Odd Gunnar >> >> >>I did not start by modifying tscp. >> >>I do not know if to say that I started from scratch because I copied part of the >>data structures from tscp. >> >>I started by move generator of only legal moves. >>My code may has may have today 10% in common with tscp. >> >>I copied part of the data structure of tscp but most of my varaible are not from >>tscp and I never used a lot of variables that are needed for tscp. >> >>Here are list of arrays that were never used by my program: >> >>int mailbox[120] >>int mailbox64[64] >>BOOL slide[6] >>int offsets[6] >>int offset[6][8] >>char piece_char[6] >> >>My program today also does not use the following arrays but it used them in the >>past. >> >>int color[64]; >>int piece[64]; >> >>some arrays are also used today >> >>gen_t gen_dat[GEN_STACK]; >>int first_move[MAX_PLY]; >>int history[64][64]; >>hist_t hist_dat[HIST_STACK]; >>move pv[MAX_PLY][MAX_PLY]; >>int pv_length[MAX_PLY]; >>int castle_mask[64]; >> >>Uri > >Hi. > >I don't want to start a discussion on this but my reflection was: > >In the node/move debate you argued that you did no pruning so all generated >moves is also a node. >TCPS do alphabeta pruning so either you had copyed this without testing or you >do a minimax/negamax search. If it is the later than you must have a very fast >movegenerator. I guess that there is a misunderstanding. I do not remember that I said that I do no pruning(I will try to find the relevant post and the reason for this misunderstanding). I did not consider the alpha beta algorithm as pruning but I use other rules that I consider as pruning so I do not think that I do no pruning. I generate only legal moves and I generate 150000-300000 nodes per second. I do not consider every move that is gnerated as a node but only every move that my program makes. > >A beginner engine start on 1500 (in real below 1000), at least did my do and to >quote a 'wise' man, I could beat it easy blindfolded :) > >Odd Gunnar I simply did not do an engine before being satisfied with the speed of my move generator. There are examples for engines like resp or genesis when the first public version is higher than 2000 so I do not understand the claim "a beginner engine start on 1500(in real below 1000)". Uri
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.