Author: Dan Newman
Date: 23:49:05 02/11/00
Go up one level in this thread
On February 11, 2000 at 20:54:01, Tom Kerrigan wrote: >On February 11, 2000 at 20:07:56, Dan Newman wrote: > >> Skyblue3 (0x88) Shrike (bitboards) >>Capture gen: 1.2 M moves/s 2.0 M moves/s >>Non-cap gen: 28.6 M moves/s 30.7 M moves/s >>Full move gen: 14.9 M moves/s 21.7 M moves/s >>Make/unmake: 3.0 M/s 2.1 M/s >>WAC @1s 456 knps 530 knps > >Is it possible that you have some sort of systematic experimental error? > >If something seems too good to be true, it usually is... > >-Tom Well, I don't think I'm double counting or anything like that. The faster program gets to depth faster too. There is the possibility that my 0x88 program was badly implemented of course. But I have tried all sorts of things to speed the 0x88 program up--I'd been tweeking it for about a year before I went over to bitboards. (Over the last 5 years of so I've started maybe a dozen chess programs trying to find just the right basic data structures and so forth and have written about 5 that I actually got to point of playing chess. So I don't think I'm making any gross blunders.) Still, I can't say that the 0x88 program has the most efficient implementation possible... The thing that really helped was switching from the Watcom compiler to MSVC 5.0. That compiler made the 0x88 program 17% faster, but it boosted the bitboard program by 50%. Comparing the Watcom compiles, the bitboard program is slower... All in all, I wouldn't switch from 0x88 + piece-lists or whatever to bitboards, or vice-versa, for speed alone--the difference just isn't likely to be all that great. (Though eval calculations can obviously be affected by the choice of basic data structures too...and that may be where the largest differences come into effect.) To me one of the greatest benefits from switching to bitboards is the elimination of piece lists. They had always seemed to be a neccessity but had also been a lot of trouble too. I was forever figuring out how to find the captured piece in its list, whether to contract/expand them or not for captures, what order the pieces should be in, and so forth. So, at least in my case, bitboards are faster and easier... -Dan.
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.