Dirty ACCA 4 32-bit Gauntlets (CCRL 40/40)

[Aser Huerga]

Dirty ACCA 4 32-bit Gauntlets

Hardware: Intel(R) Core(TM)2 Duo CPU E8400 @ 3.00GHzDual 3000 MHz with 4.088 MB Memory
Operating system: Microsoft Windows XP 64 Bit Professional Service Pack 2 (Build 3790)
All engines 1 CPU
Arena 2.0.1 GUI
128 Hash Tables
3-4-5 EGTBs
Ponder off
Book GrandPQRS-3moves (2.360 starting selected positions, from Norm Pollock collections), each pair plays the same position with switched colours.
40 moves in 25 minutes repeating (adapted to the CCRL 40/40 standard)

Gauntlet I:

Code: Select all

1   Dirty ACCA 4 32-bit -2625- 54.0/120
2   Amyan 1.72           2678  16.5/24
3   Arasan 11.6 64-bit   2690  14.0/24
4   Rotor 0.5            2666  13.0/24
5   Ufim 8.02            2643  11.5/24
6   DanaSah 4.24         2630  11.0/24

(All ELOs referred to the last Chess Computer Rating List when tournament starts, except -XXXX- one, wich is referred to performance).

Statistics:

Code: Select all

Dirty ACCA 4 32-bit           : 120 (+ 37,= 34,- 49), 45.0 %

Arasan 11.6 64-bit            :  24 (+  7,=  6,- 11), 41.7 %
Amyan 1.72                    :  24 (+  5,=  5,- 14), 31.2 %
Rotor 0.5                     :  24 (+  8,=  6,- 10), 45.8 %
Ufim 8.02                     :  24 (+  8,=  9,-  7), 52.1 %
DanaSah 4.24                  :  24 (+  9,=  8,-  7), 54.2 %

(All statistic results are from the engine tested view).

Games: http://kirill-kryukov.com/chess/discuss ... f=7&t=5418

[Aser Huerga]

Gauntlet II:

Code: Select all

1   Dirty ACCA 4 32-bit        -2647- 63.5/120
2   Brutus 8.05 32-bit          2673  15.0/24
3   SOS 5.1                     2661  13.0/24
4   Gaviota 0.74.41 64-bit      2590  11.5/24
5   Tao 5.6                     2610  10.5/24
6   LittleThought 1.052 64-bit  2609  6.5/24

(All ELOs referred to the last Chess Computer Rating List when tournament starts, except -XXXX- one, wich is referred to performance).

Statistics:

Code: Select all

Dirty ACCA 4 32-bit           : 120 (+ 45,= 37,- 38), 52.9 %

Brutus 8.05 32-bit            :  24 (+  6,=  6,- 12), 37.5 %
SOS 5.1                       :  24 (+  6,= 10,-  8), 45.8 %
Tao 5.6                       :  24 (+ 10,=  7,-  7), 56.2 %
LittleThought 1.052 64-bit    :  24 (+ 15,=  5,-  4), 72.9 %
Gaviota 0.74.41 64-bit        :  24 (+  8,=  9,-  7), 52.1 %

(All statistic results are from the engine tested view).

Games: http://kirill-kryukov.com/chess/discuss ... 242#p57242

[Andres Valverde]

Gauntlet II:

Code: Select all

1   Dirty ACCA 4 32-bit        -2647- 63.5/120
2   Brutus 8.05 32-bit          2673  15.0/24
3   SOS 5.1                     2661  13.0/24
4   Gaviota 0.74.41 64-bit      2590  11.5/24
5   Tao 5.6                     2610  10.5/24
6   LittleThought 1.052 64-bit  2609  6.5/24

(All ELOs referred to the last Chess Computer Rating List when tournament starts, except -XXXX- one, wich is referred to performance).

Statistics:

Code: Select all

Dirty ACCA 4 32-bit           : 120 (+ 45,= 37,- 38), 52.9 %

Brutus 8.05 32-bit            :  24 (+  6,=  6,- 12), 37.5 %
SOS 5.1                       :  24 (+  6,= 10,-  8), 45.8 %
Tao 5.6                       :  24 (+ 10,=  7,-  7), 56.2 %
LittleThought 1.052 64-bit    :  24 (+ 15,=  5,-  4), 72.9 %
Gaviota 0.74.41 64-bit        :  24 (+  8,=  9,-  7), 52.1 %

(All statistic results are from the engine tested view).

Games: http://kirill-kryukov.com/chess/discuss ... 242#p57242

Thx Aser, it is getting closer to 6J-22 but I guess it will be worse finally.
6J-22 has a heavier eval, and it was a little worse at fast TC (at which I test) but seems to be a little better at long TC. As Uri (iirc) said, given two versions with similar behaviour at short TC, choose that with a more complex eval....

[Aser Huerga]

Yes, I hope this means some progress for Dirty in the near future. the 64-bit with supposedly no crashes is in test right now. Tomorrow I could tell you something about the estability. Let's see if the 64-bit compile it's a step up in Dirty or not.

Regards.

[Aser Huerga]

More games are needed, but it seems Dirty ACCA4 32-bit, with 2638 ELO points, is not an improvement regarding previous versions (Dirty 6J-22 2669).

I wish a better success to the next Engine version!

http://www.computerchess.org.uk/ccrl/40 ... ons_only=1

BTW: (new 64-bit version shows no crashes by the moment...)