Author: Bernhard Bauer
Date: 06:04:06 03/21/00
Go up one level in this thread
On March 20, 2000 at 12:28:57, blass uri wrote: >On March 20, 2000 at 03:49:50, Paulo Soares wrote: > >>On March 20, 2000 at 00:58:48, Côme wrote: >> >>>On March 19, 2000 at 22:22:34, Paulo Soares wrote: >>> >>>>On March 19, 2000 at 21:48:30, Aloisio Ponti Lopes wrote: >>>> >>>>>...maybe it will be available in the next 2 years. >>>>> >>>>>A. Ponti >>>> >>>>Which would be the largest number of pieces to be used in Nalimov >>>>Tablebases with 140Mgz? 7, 8, 9? >>>> >>>>Paulo Soares, from Brazil >>> >>>Hello, >>>IMHO only 6 pieces. >>>Best Regards >>>Alexandre Côme >> >>Only 6! How many memory for 7 pieces? >>Thanks, >>Paulo > >I think that if you want all of them you need about some hundreds times more >memory for adding one piece. > >The problem is that the number of possibility of 7 pieces is significantly >bigger than the number of possibility of 6 pieces and not only the fact that you >need about 60 times more memory for one 7 pieces tablebase relative to 6 pieces >tablebase. > >You need some Gbytes for 5 pieces tablebases >then you probably need some hundreds of thoushands of Gbytes for all the 7 >pieces tablebases > >You need sometimes hundreds of mbytes for one 5 pieces tablebases so you need >something some Gbytes to use one 6 men tablebases and you need something like >100-200 Gbytes to use only one 7 men tablebase. > >Uri As a rule of thumb the number of files for n+1 men tablebases is 4 times the number of n men files. The size is going up with a factor of ~60. So if you need 5 gigs for the 5-men class you will need about 1.2 tera bytes for the 6-men class. This is "only" 30 40-gig drives. IMHO it would make a good project for a university to generate and hold them on disk. Kind regards Bernhard
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.