Author: Dann Corbit
Date: 10:35:23 02/01/01
Go up one level in this thread
On February 01, 2001 at 02:43:44, David Blackman wrote: >On January 31, 2001 at 19:14:19, Dann Corbit wrote: > >>On January 31, 2001 at 11:36:01, Robert Hyatt wrote: >>[snip] >>>The idea seems pretty good for 3-4 piece files, and even for 5 piece files >>>although the memory to hold them becomes prodigous. But 6's are hopeless >>>as todays machines are no better at probing a 1gb file than they are at >>>probing a 1 terrabyte file. >> >>It might be worthwhile to store them in a real database with hashed index. >> >>Modern database systems will cache database requests very efficiently, and so I >>think it might be doable. The database model would be the hard part (finding a >>representation which is still highly compact like the tablebase files) > >Modern database systems are pretty hopeless for raw speed and space efficiency. >A custom built data structure is almost always faster and smaller, and usually >much simpler as well. > >The good things about modern database systems, are safe multiuser access and >update, robustness in the face of various disasters including hardware failure, >reliable backups that can be done even while transactions are in progress, nice >system administration tools, nice programming tools for building simple GUI >programs for end users, a highly generalised and flexible query mechanism, and >so on. > >All of these good things are irrelevant for endgame tablebases, and the >performance cost would really hurt. We have machines here which have more than 500 MB/sec IO bandwidth. They can read tablebase files in the time a PC can open one. I am speaking of a real database system, not a toy. You are thinking of a PC. I am thinking of an IBM 3090, RS/6000 or multiple CPU Alpha machine (we have all of these in house).
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.