Not logged inRybka Chess Community Forum
Forum RybkaChess Help Search Register Login
Up Topic The Rybka Lounge / Computer Chess / Toga ll
- - By AxeFever (*) [us] Date 2012-02-28 05:35
Where can I go to download the chess engine Toga ll 3.1.2 64 bit?  Thanks!
Parent - - By Graham Banks (****) [nz] Date 2012-02-28 07:28

> Where can I go to download the chess engine Toga ll 3.1.2 64 bit?  Thanks!


Was there a 64-bit version released and if it was, do you know if was any stronger than the 32-bit version?
Parent - - By Uly (Gold) [mx] Date 2012-02-28 19:32

> Was there a 64-bit version released


Yes. I could attach it to this post, but since it was released without source code, I wonder if it was in violation of the GPL?

The executable reads:

"Toga II 3.1.2SE UCI based on Fruit 2.1 by Thomas Gaksch and Fabien Letouzey. Settings by Dieter Eberle"

As Toga isn't Bitboard I doubt that it's faster than the 32bit version.
Parent - By Wrath of the Titans (****) [sa] Date 2012-02-28 19:46
quick question, Is there a cluster toga?
Parent - - By Graham Banks (****) [nz] Date 2012-02-29 04:56

> Yes. I could attach it to this post, but since it was released without source code, I wonder if it was in violation of the GPL?


Theoretically, I guess it was. However somebody said that it would only be in violation if the "author" failed to provide the source code upon request to anybody who asked?
Parent - - By Arrière Pensée (Gold) Date 2012-02-29 05:07
Well, here is the deal- CCRL give a link to WBEC Ridderkerk- Which in turn gives the link to - Superchessengine where one can download Toga-but, it is Toga 1.2.1a -never saw anything higher than that -would be interesting.
Parent - By Graham Banks (****) [nz] Date 2012-02-29 05:23
These are the Togas and Fruits in our lists. I can only see the 32-bit version of Toga II 3.1.2SE, which had source code provided.

CCRL 40/40 Rating List - Custom engine selection
381140 games played by 1136 programs, run by 18 testers
Ponder off, General books (up to 12 moves), 3-4-5 piece EGTB
Time control: Equivalent to 40 moves in 40 minutes on Athlon 64 X2 4600+ (2.4 GHz)
Computed on February 22, 2012 with Bayeselo based on 381'140 games
Tested by CCRL team, 2005-2012, http://computerchess.org.uk/ccrl/4040/

Rank                Engine                 Elo   +    -   Score  AvOp  Games
1 Toga II 1.4.1SE 4CPU                 3008  +13  -13  49.8%   +2.4  1981
  Toga II 1.4 beta5c 4CPU              2990  +18  -18  51.8%  -12.8   999
  Grapefruit 1.0 32-bit 2CPU           2977  +37  -37  42.8%  +44.8   222
  TheMadPrune 1.1.25 2CPU              2977  +39  -39  53.0%  -19.3   201
  Cyclone 3.4 2CPU                     2971  +38  -38  51.5%   -8.0   199
  Cyclone 2.3 2CPU                     2966  +40  -40  43.8%  +37.7   185
  Toga II 1.4 beta5c 2CPU              2965  +24  -24  53.2%  -21.7   546
  Toga II 1.4.1SE 2CPU                 2963  +33  -33  42.2%  +56.0   300
  Grapefruit 1.0 64-bit 2CPU           2962  +38  -38  47.1%  +19.4   208
  Toga II 1.4.1SE6 2CPU                2962  +38  -38  50.5%   -1.4   209
  Cyclone 2.2 2CPU                     2954  +39  -40  47.7%  +15.2   196
  Cyclone 3.1 2CPU                     2944  +41  -41  45.1%  +31.9   185
  Cyclone xTreme                       2927  +13  -13  44.0%  +38.9  1941
  Cyclone xTreme Fury                  2927  +19  -19  48.1%  +11.0   811
  Grapefruit 1.0 32-bit                2927   +9   -9  48.4%  +10.7  3991
  Toga II 1.4.1SE                      2927  +12  -12  49.8%   +3.8  2449
  Toga II 1.4.1SE6                     2927  +19  -19  50.2%   +8.8   825
  Toga II 3.1.2SE                      2927  +18  -18  55.2%  -33.5   941
  Cyclone 3.4                          2923  +21  -21  51.5%   -8.2   685
  Cyclone 2.2                          2920  +28  -28  49.1%   +5.1   382
  Cyclone xTreme Rage                  2920  +25  -25  48.4%   +9.7   476
  TheMadPrune 1.1.25                   2906  +26  -26  51.7%  -12.0   446
  Cyclone 1.0 2CPU                     2905  +33  -33  46.8%  +18.7   299
  Toga II 1.4 beta5c                   2905  +24  -24  53.4%  -21.9   532
  Toga II 1.3.1                        2891  +11  -11  49.1%   +4.9  2958
  Fruit 2.3.1                          2885  +14  -14  50.0%   -0.1  1734
  Grapefruit 1.0a3                     2884  +32  -32  47.7%  +14.3   309
  Cyclone 1.0                          2882  +32  -32  51.0%   -7.9   300
  Fruit 051103                         2869  +20  -20  49.9%   -1.1   761
  Toga II 1.3.4                        2864  +24  -24  47.9%  +14.4   550
  Toga II 1.2.1a                       2859  +11  -11  48.9%   +4.6  2792
  Toga II 1.2 32-bit                   2846  +27  -27  56.0%  -43.1   435
  Fruit 2.2.1                          2840  +13  -13  52.3%  -15.2  2037
  Toga II 1.1a                         2825  +18  -18  55.1%  -33.0  1058
  Fruit 2.1                            2792  +25  -25  51.5%  -13.6   507

CCRL 40/4 Rating List - Custom engine selection
819677 games played by 915 programs, run by 17 testers
Ponder off, General books (up to 12 moves), 3-4-5 piece EGTB
Time control: Equivalent to 40 moves in 4 minutes on Athlon 64 X2 4600+ (2.4 GHz)
Computed on January 28, 2012 with Bayeselo based on 819'677 games
Tested by CCRL team, 2005-2012, http://computerchess.org.uk/ccrl/404/

Rank                 Engine                   Elo   +    -   Score  AvOp  Games
1 Toga II 1.4 beta5c 4CPU                 2998   +9   -9  36.8%  +94.7  4947
  Toga II 1.4.1SE 4CPU                    2994  +10  -10  44.2%  +42.9  3446
  Toga II 1.4.1SE 2CPU                    2973  +30  -31  41.7%  +53.2   342
  Toga II 1.4 beta5c 2CPU                 2944  +29  -29  45.4%  +30.5   369
  Toga II 1.4.1SE                         2922   +9  -10  50.2%   +0.4  4046
  Toga II 1.4 beta5c                      2918  +13  -14  48.9%   +6.3  1918
  Grapefruit 1.0 a2b                      2908  +22  -22  49.3%   +4.5   648
  Fruit 05/11/03                          2887  +11  -12  49.0%   +6.3  2730
  Fruit 2.3                               2881  +14  -14  62.0%  -82.0  1784
  Toga II 1.3.1                           2870   +9   -9  44.9%  +36.8  4937
  Fruit 2.3.1                             2864  +23  -23  57.7%  -52.1   614
  Toga II 1.2.1a 32-bit                   2861   +9   -9  49.2%   +1.6  5319
  Fruit 2.2.1                             2851  +18  -18  56.3%  -45.0  1076
  Toga II 1.1a 32-bit                     2836  +39  -37  76.7% -204.5   302
Parent - - By Graham Banks (****) [nz] Date 2012-02-29 05:33 Edited 2012-02-29 05:57

> Well, here is the deal- CCRL give a link to WBEC Ridderkerk- Which in turn gives the link to - Superchessengine where one can download Toga-but, it is Toga 1.2.1a -never saw anything higher than that -would be interesting.


http://talkchess.com/forum/viewtopic.php?t=18813&start=0

You might find this interesting too:

http://talkchess.com/forum/viewtopic.php?t=28044&highlight=toga++++2se

Only links I can find to Toga II 3.1.2SE 64-bit are these:

http://thesicilian.forumotion.net/t301-all-toga-engines
http://immortalchess.net/forum/showthread.php?t=263
Parent - - By Uly (Gold) [mx] Date 2012-02-29 06:18
The first site requires registration to download. The second site has dead links, so I'm attaching Toga 3.1.2SE 64bit to this post.
Attachment: TogaIIv3.1.2SE64.7z (166k)
Parent - - By Peter Grayson (***) [gb] Date 2012-02-29 08:09
But illegal to distribute because no source code or license agreement included. Are you sure these were actually released by the author because the zip file I have from the time it was released only includes a legal 32 bit version.

PeterG
Parent - - By Uly (Gold) [mx] Date 2012-02-29 08:52

> But illegal to distribute because no source code or license agreement included.


Graham said it's not illegal, it's only illegal if the author refuses to provide source if asked.
Parent - - By Graham Banks (****) [nz] Date 2012-02-29 09:28

>> But illegal to distribute because no source code or license agreement included.
> Graham said it's not illegal, it's only illegal if the author refuses to provide source if asked.


No, I reported what somebody else said. No idea whether it's fact or not.   :wink:
Parent - - By Adam Hair (**) [us] Date 2012-02-29 17:23
From the GNU General Public License, version 2 (which is the license that was distributed with Fruit 2.1 and Toga II):

  3. You may copy and distribute the Program (or a work based on it,
under Section 2) in object code or executable form under the terms of
Sections 1 and 2 above provided that you also do one of the following:

    a) Accompany it with the complete corresponding machine-readable
    source code, which must be distributed under the terms of Sections
    1 and 2 above on a medium customarily used for software interchange; or,

    b) Accompany it with a written offer, valid for at least three
    years, to give any third party, for a charge no more than your
    cost of physically performing source distribution, a complete
    machine-readable copy of the corresponding source code, to be
    distributed under the terms of Sections 1 and 2 above on a medium
    customarily used for software interchange; or,

    c) Accompany it with the information you received as to the offer
    to distribute corresponding source code.  (This alternative is
    allowed only for noncommercial distribution and only if you
    received the program in object code or executable form with such
    an offer, in accord with Subsection b above.)

Denis Mendoza, who created the 64-bit compiles (which also can be found at CCRL public forum), failed to provide the source code or a written offer to provide the source code with his distribution of the 64-bit compiles. Most likely because he would readily provide the source code if asked. Simply adding the source and the GPL to the distribution would solve the problem.
Parent - - By Adam Hair (**) [us] Date 2012-02-29 17:37
By the way, the person who made the changes to Toga which resulted in Toga II 3.1.2 SE, is Chris Formula. This is in the readme:

ACKNOWLEDGMENT
--------------------------------

Huge thanks to Fabien Letouzey (France) for the great source code of the program Fruit 2.1.
Huge thanks to Thomas Gaksch (Germany) for the great source code of the program Toga II based on Fruit 2.1.

Special thanks to:
Dieter Eberle, Karl-Heinz Söntges, Shaun Brewer for testing Beta Versions.
A big thanks to Dieter Eberle for his setting.
Orlando Mouchel for bugfixes and new ideas.
Wilhelm Hudetz for the Logo.
Daniel Shawul for sharing his great endgame bitbases and his dll.
Walter Eigenmann for the EET testsuite.
Tony Hedlund for the SSDF Positional testsuite.
Martin Blume for the Arena software.
Denis Mendoza for the compile support for different platforms.

Without their help, the improved playing strength of Toga II based on Fruit 2.1 would not be possible.
Thank you! Maraming salamat!

Regards,
Thomas Gaksch (Germany)
Chris Formula (Philippines)

So, I think the author endorsed Denis' compiles (or would if asked). I do not think any of the authors associated with Toga II 3.1.2 SE would think Denis was doing anything illegal (other than in a limited technical sense). However, the FSF may would have some say also, since Fabian transferred the rights to his part of the source to the FSF.
Parent - - By Graham Banks (****) [nz] Date 2012-02-29 17:50

> By the way, the person who made the changes to Toga which resulted in Toga II 3.1.2 SE, is Chris Formula.


Chris put out the 32-bit version. I don't think that he was responsible for the 64-bit version which had no source code attached.
Parent - - By Uly (Gold) [mx] Date 2012-02-29 19:49
Dieter Eberle is probably responsible of the 64bit version.
Parent - - By Eelco de Groot (***) Date 2012-03-01 16:34 Edited 2012-03-01 16:41
Why the sudden interest in specifically this version? It is very old... There are dozens of later versions but all approximately the same strength as this one, only the playing style of S.E is probably more interesting I think. Not sure if this version is multithreaded, probably yes. The older 1.2.1a and related versions had a bit more aggessiveness but could only use one core. The multithreading in later versions always seemed a bit hackish although it still worked reasonably well. There was no real YBW multithreading though.

No, Dieter had made some new settings for Toga but when he reported his results to Thomas Gaksch, it turned out that testing under the Chessbase interface had not actually used those settings, because Chessbase reverts to the defaults unless you make a specific .eng file with the new settings. So the settings were not actually tested, only later by the CCRL tests themselves. The 64 bits builds by Denis did not require codechanges I think so one can assume the 32 bit code is valid for this build too. There is no speed up to be expected from Toga 64 bit compared to 32 bit. The only thing is I think we can't ask Denis about it because he has not been on the Toga Developers Discussion Board since New Year and I am a bit worried if he is okay :sad:

Eelco
Parent - - By Adam Hair (**) [us] Date 2012-03-01 21:28
Denis is okay. He has had some blood pressure problems, and his doctor advised that he not spend as much time on the computer. But he still pops in from time to time at CCRL.
Parent - By Eelco de Groot (***) Date 2012-03-01 22:09
Thanks for letting me know Adam! That sure is a relief to know, Denis already said that he had to give up on computerchess lifestyle for a while and also Denis is not living the most safe region when it comes to the weather and huricaneforce storms. But maybe better than working in Russia for him :eek: with the very bad cold spell they had at least in countries like Romenia and Bulgaria. I do suppose computerchess especially professional 'have to pay the rent with it' computerchess can be not so good for your bloodpressure, too much time sitting behind a monitor and especially the frustration if another magnificent idea is doing the opposite of what you thought it would, you keep on working till you get it right, forgetting to eat your two days old pizza slice that the cat found first, or even forgetting to sleep or that it's your turn bringing your kids to pre-school maybe and the deadline for your program already three months behind... Not sure if Nelson is right calling that the perfect job Vas has chosen for himself.

It's been a while since I have been there at the CCRL forum so I would not have found Denis postings there I think.

Thanks, Eelco
Parent - By Uly (Gold) [mx] Date 2012-03-01 22:38

> only the playing style of S.E is probably more interesting I think.


Yeah, I stopped testing Toga versions at some point, but it was clear 3.1.2 was the one with the best style, specially with the Chekov settings :wink:, future versions of Toga were way too materialistic for my taste, even though they performed stronger than oldie 3.1.2 on the elo department. But for analysis of games you don't use engines for their strength (one uses top engines for that), you use them for their style and the ideas they provide on the game, ideas that the top engine would just prune.

So unless someone comes to tell that a future Toga version has a better playing style, 3.1.2SE is THE Toga version I'd suggest to use for analysis of games, also, because Fruit 2.3.1 is the Fruit version I suggest for analysis (good at showing defensive resources top engines might prune) and their move choices are dissimilar enough that they can be used together on the same game.

Unfortunately engines have gone a long way, and I stopped using "weak" engines because a refuter engine like Critter may kill all their original ideas that would have been playable in times with less elo and hardware, but the best moves against them are just evident today, while Critter may also come with its own original ideas that other top engines don't consider, so offshoot ideas are somewhat deprecated because main ideas have become offshoot themselves.

>The 64 bits builds by Denis did not require codechanges I think so one can assume the 32 bit code is valid for this build too


That explains it. It still doesn't explain why isn't Chris Formula in the credits? I thought he was the main person responsible for the SE version, did he choose to not include himself?

>There is no speed up to be expected from Toga 64 bit compared to 32 bit.


Yeah, one wonders what's the fuzz about the 64 bit version then :eek:
Parent - By Adam Hair (**) [us] Date 2012-03-01 21:35
As shown in the link I provided ( here it is again: http://kirill-kryukov.com/chess/discussion-board/viewtopic.php?f=7&t=2992 ), Denis created the 64-bit compiles of Toga II 3.1.2 SE (though there may be other compiles out there).

I do not believe there was any code differences between the 32-bit and 64-bit compiles.
Parent - By Peter Grayson (***) [gb] Date 2012-02-29 08:04
I checked my engines database and Toga 3.1.2 SE was only released as a 32 bit engine. The files confirm it included a zipped source code folder too as well as the GNU license agreement so it was all perfectly legal.

PeterG
Up Topic The Rybka Lounge / Computer Chess / Toga ll

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill