Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Fritz 5.32 - the truth

Author: blass uri

Date: 09:28:38 04/12/99

Go up one level in this thread



On April 12, 1999 at 02:58:00, Jouni Uski wrote:

>Despite much praised conversion from 16 to 32 bit, I don't found
>any speed up in Fritz analysis. NPS speed seems to be exactly
>as with previous 16 bit version and in endgame area there seems to be
>even some setbacks - e.g. these positions are solved fast with 5.16,
>but 5.32 takes very long time:
>
>3b4/5kp1/1p1p1p1p/pP1PpP1P/P1P1P3/3KN3/8/8 w - - bm c5;
>2r5/2pR1pk1/p1P3p1/P2K4/5P1p/5P1P/8/8 w - - bm Rxc7;
>
>Maybe 5.32 is optimised for BS2830 test were it is the best program, but
>in other testsuites it's not better than 5.16! And if You need tactical
>analysis speed Fritz4.01 is the fastest.
>
>Also the evaluation seems to be exact the same in many position
>with 5.16 and 5.32.

In some endgames there are setbacks in the evaluation:

8/4R2p/7k/3rN1p1/5K2/8/8/8 w - - 0 1

Fritz5.32 evaluated this position as advantage for white in the ssdf games when
Fritz5.16 understamds that it is a draw.

Uri





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.