Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: perft records

Author: Sune Fischer

Date: 05:18:37 09/07/04

Go up one level in this thread


On September 07, 2004 at 07:19:44, Reinhard Scharnagl wrote:

Ok, but I'm not going to strip down my makemove for perft optimization, so just
running it as is with the incremental eval stuff (it would only be a few %
faster I think).

>Hi Sune,
>
>it would be interesting to know your detailed data on:
>
>FEN: rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1
>
>  +-*--b--c--d--*--f--g--*-+ MS Visual Studio C++ Version 13.10
>8 |[r][n][b][q][k][b][n][r]|
>7 |[p][p][p][p][p][p][p][p]| Perft Testseries
>6 |   :::   :::   :::   :::|
>5 |:::   :::   :::   :::   | (With TT Caching 512 MB / 5-fold)
>4 |   :::   :::   :::   :::|
>3 |:::   :::   :::   :::   | Test #:      00
>2 |<P><P><P><P><P><P><P><P>|
>1 |<R><N><B><Q><K><B><N><R>| Break Time 50.0 Sec.
>=>+-*--b--c--d--*--f--g--*-+
>
>Ply          Nodes         all (x)         all (+)   Seconds
>------------------------------------------------------------
>01              20               0               0         0
>02             400               0               0         0
>03            8902              34              12         0
>04          197281            1576             469         0
>05         4865609           82719           27351       0.1
>06       119060324         2812008          809099       1.7
>07      3195901860       108329926        33103848      19.9
>08     84998978956      3523740106       968981593     285.5
>------------------------------------------------------------

You used 512 MB hash?
Gosh, I only have 512 total in the machine.

Well default is 50 MB hash as is also used in the latest public version.

ply -- seconds

(50 MB)
05 -- 0.16
06 -- 1.98
07 -- 27.70
08 -- 531.19

(200 MB)
05 -- 0.19
06 -- 2.08
07 -- 24.86
08 -- 389.23


>and on:
>
>FEN: r3k2r/p1ppqpb1/bn2pnp1/3PN3/1p2P3/2N2Q1p/PPPBBPPP/R3K2R w KQkq - 0 25
>
>  +-*--b--c--d--*--f--g--*-+ MS Visual Studio C++ Version 13.10
>8 |[r]:::   :::[k]:::   [r]|
>7 |[p]   [p][p][q][p][b]   | Perft Testseries
>6 |[b][n]   :::[p][n][p]:::|
>5 |:::   :::<P><N>   :::   | (With TT Caching 512 MB / 5-fold)
>4 |   [p]   :::<P>:::   :::|
>3 |:::   <N>   :::<Q>:::[p]| Test #:      01
>2 |<P><P><P><B><B><P><P><P>|
>1 |<R>   :::   <K>   :::<R>| Break Time 50.0 Sec.
>=>+-*--b--c--d--*--f--g--*-+
>
>Ply          Nodes         all (x)         all (+)   Seconds
>------------------------------------------------------------
>01              48               8               0         0
>02            2039             351               3         0
>03           97862           17102             993         0
>04         4085603          757163           25523       0.2
>05       193690690        35043416         3309887       3.9
>06      8031647685      1558445089        92238050      90.3
>------------------------------------------------------------

50 MB:
04 -- 0.16
05 -- 3.64
06 -- 115.12

200 MB:
04 -- 0.19
05 -- 3.80
06 -- 108.92

>because it would be very interesting for me to know the ply-timings,
>for to estimate and to compare the groth of dure from ply to ply,
>when using a transposition table.

The replacement scheme is not unimportant here either, the hash will saturate
badly when going deep.

And of course the machines are not identical.

-S.
>Regards, Reinhard.



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.