Author: Tony Werten
Date: 23:52:15 03/02/06
Go up one level in this thread
On March 02, 2006 at 12:15:05, Dezhi Zhao wrote: >On March 02, 2006 at 11:43:20, Robert Hyatt wrote: > >>On March 02, 2006 at 01:08:47, Dezhi Zhao wrote: >> >>>If you want to try adding another dimension to the history table, I think the >>>ply number is the best candidate. You just replace the side index with the ply >>>number. This will greatly localize the histroy table and probally make it more >>>useful. We can call it history killer table by doing so:) >> >>Huge table. becomes 4096 * 64 for my program. = 2^18 or 262K entries... >> > >you may use [ply][piece][to], 64*12*64 = 48K entries. and only the very first >quarter of the table is really active most of the time. So, this looks not that >bad. Actually, it's only 24K. If you have the plynumber, you don't need to distinguish between black and white pieces. Tony > >>I'll add this to my list of things to try however, although I suspect it will >>only be useful for the reduction test, since the point of global history >>counters is to emulate a sort of global killer move list... >> > >Yes. I understand history table is in fact a set of global killers. However, >[ply][piece][to] or [ply][from][to] obviously dominate your original >[side][from][to]. If you still need the globalness, just add all the history >values of a move from odd plys and even plys respectively. > > >> >>> >>>On March 01, 2006 at 16:05:06, Robert Hyatt wrote: >>> >>>>I am still using my old 12-bit history index <to><from> to index into these new >>>>values. I also still maintain one for white and one for black. I once tried a >>>>17 bit index <piece><to><from> but it offered no improvement to the basic >>>>history heuristic, but I have not tested that with this reduction stuff. >>>> >>>>Any ideas what others are doing here? I can think of several possibilities: >>>> >>>><piece><to> >>>> >>>><piece><from> >>>> >>>><to><from> (I am doing this now) >>>> >>>><piece><to><from> >>>> >>>>to name at least 4. First two seem too simplistic. Last one turns the history >>>>tables into pretty good sized arrays (2^17 for white, ditto for black). >>>> >>>>Note that I am essentially factoring in wtm already since I have two separate >>>>sets of history tables for black and white.
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.