Author: Robert Hyatt
Date: 09:02:43 04/04/03
Go up one level in this thread
On April 04, 2003 at 00:22:40, Russell Reagan wrote: >On April 03, 2003 at 23:10:09, Robert Hyatt wrote: > >>6 pieces are a challenge. 7 will be a bigger challenge. What happens beyond >>there depends on what happens to storage technology. > >Could you give some examples of what makes 6 pieces challenging, and 7 pieces >moreso? IE Does "challenge" == "not enough storage"? Or is there more to it than >that? First, 6 piece files require roughly 6 bits per piece, or 36 bits to address the entries in the file. Eugene's indexing reduces this but it still leaves a huge number of entries to deal with. Second, so far, we are scraping by with 8 bit entries, although I don't know if Eugene has blown this in the more recent tables. But it will be blown somewhere in the 6-piece files and that will instantly double the size. Third, you would need all the 6 piece files to produce the 7 piece files. And a 7 piece file will require a 2^42 address space, where 2^32 is four billion. So we are talking about 4 trillion positions in a single file, and if they need 2 bytes per position, we are talking a paltry 8 terabytes for a _single_ 7 piece file. _that_ is a challenge. :) The 6's are challenging today's technology as we are likely going to end up with over a terabyte as is. And while a terabyte is doable, building all those files requires a 64 bit machine to have the virtual address space for Eugene's code.
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.