Author: Dave Kuntzsch
Date: 09:03:18 09/10/02
Go up one level in this thread
On September 10, 2002 at 08:33:48, Eli Liang wrote: >In 1957, Alex Bernstein created a computer chess program that handled all the >rules of chess and ran on an IBM 704 with a _total_ of 7K x 36bits of valve >storage. His program completed a 4-ply selective search in 8 minutes (but the >processor speed of those IBM 704 was 0.042Mhz). > >In 1983, David Horne created a legendary chess program in 672 bytes of Z80 >assembly language and Basic, but his program didn't handle castling or >en-passant, had a fixed opening move, only did a 1-ply search, and was unable to >actually finish a game. > >In comparison, last month, Douglas Bagnall created a chess program with a static >evaluator and a full-width 3-ply search which seems pretty hefty at 4.4K bytes >of JavaScript since not only was it coded in a HLA, but it is not limited in RAM >as Horne's and Bernstein's programs were. (Not to mention that I've tried it >and unfortunately, it seems buggy.) > >Questions: Was Bernstein's the smallest/tightest/most-memory-efficient _full_ >chess program ever written? What do you think is the smallest possible full >chess program with alpha-beta search that could ever be written in Intel x86 >native code, if you disregard how to get moves in and out of it (no UI)? I know >the latter question is rather imprecise because of issues concerning complexity >of static evaluation, depth of search, etc., but I am just looking for some >imprecise opinions. In the late 70s and early 80s, I wrote an 18K Z80 assembler program. It followed all rules including castling on both sides, en passant, and pawn promotion. It had a small imbedded opening book, all keyboard and display routines, alpha-beta, mini-max, move ordering and killer move heuristic. The search depth could be set to any value by the user. It was inspired by an article in Popular Science describing game theory. I still have the printed source, but in looking for it the past month, I haven't been able to find it yet. BTW, the Z80 was an 8 bit machine and ran at 4Mhz. When it finally deveped a memory bug that I was never able to pinpoint, I bought one of the first IBM PCs (floppy disk, 8080 chip), but never converted the program. The Z80 is still sitting in my basement. Dave
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.