Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: speeding up epd/fen generation from pgn

Author: Dann Corbit

Date: 15:39:58 01/20/06

Go up one level in this thread


On January 20, 2006 at 18:08:25, George Jempty wrote:

>I'm trying to create a database of epd/fen positions from a 2 million game PGN
>file.  I'm using a couple of Perl modules, Chess::PGN::Parse & Chess::PGN::EPD.
>I'm not totally unhappy with Parse, it starts off processing 250 games per
>second, so it should get through the 2 million games in a little over 2 hours.
>The EPD generation is another story altogether.  It's only processing 10 games
>per second, so this could literally take half a week.  I also looked at pgn2fen
>but at the rate it took to get through Fischer's 60 memorable games, it would
>probably take a  month to get through 2 million games.
>
>So the Perl modules seem to be a better bet, but actually to tell you the truth
>I am also unhappy with Chess::PGN::Parse.  I mean, how many thousands of
>positions do chess engines evaluate per second, but only 250 PGN scores can be
>parsed in the same period of time.  Can anybody point me to PGN parsers/FEN
>generators that perform better, the latter in particular.
>
>Thanks in advance

There are several solutions that will turn standard PGN games into standard EPD
records.

Examples are:
PGN-Extract by Barnes
PGNPosition by Veluri
Pgn2Epd by Coulom

I have parsed 4 million games in a shot with them.  Probably, they are faster
than perl.



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.