Page 3 of 4

Re: Rodent III 0.273 - official release

Posted: Wed Sep 05, 2018 9:20 am
by Gabor Szots
tpoppins wrote: Tue Sep 04, 2018 9:58 pm There were 14 time losses out of 238 games after raising the TimeBuffer value to 100. Using today's fixed version there are no time losses after 79 games.

Thank you, Pawel, and better luck in the Championship!
What fixed version? The small pack is the same as what I downloaded 4 days ago.

Re: Rodent III 0.273 - official release

Posted: Wed Sep 05, 2018 9:45 am
by Guenther
Gabor Szots wrote: Wed Sep 05, 2018 9:20 am
tpoppins wrote: Tue Sep 04, 2018 9:58 pm There were 14 time losses out of 238 games after raising the TimeBuffer value to 100. Using today's fixed version there are no time losses after 79 games.

Thank you, Pawel, and better luck in the Championship!
What fixed version? The small pack is the same as what I downloaded 4 days ago.
Gabor, you can at least use the appveyor build.
The difference is this option now:

Code: Select all

option name TimeTricks type check default false

Re: Rodent III 0.273 - official release

Posted: Wed Sep 05, 2018 9:54 am
by Gabor Szots
Guenther wrote: Wed Sep 05, 2018 9:45 am Gabor, you can at least use the appveyor build.
The difference is this option now:

Code: Select all

option name TimeTricks type check default false
Thank you, Günther. I see it is already at version 0.275 now.

Re: Rodent III 0.273 - official release

Posted: Wed Sep 05, 2018 2:26 pm
by CMCanavessi
Gabor Szots wrote: Wed Sep 05, 2018 9:20 am
tpoppins wrote: Tue Sep 04, 2018 9:58 pm There were 14 time losses out of 238 games after raising the TimeBuffer value to 100. Using today's fixed version there are no time losses after 79 games.

Thank you, Pawel, and better luck in the Championship!
What fixed version? The small pack is the same as what I downloaded 4 days ago.
Yeah, the small pack hasn't been updated, but the full pack has a new binary from September 4.

Re: Rodent III 0.273 - official release

Posted: Sat Sep 08, 2018 11:21 am
by Norbert Raimund Leisner
Are Windows binaries of RodentIII 0.275 still available? cf.https://github.com/nescitus/Rodent_III/ ... c/main.cpp ?

Norbert

Re: Rodent III 0.273 - official release

Posted: Sat Sep 08, 2018 12:00 pm
by Guenther
Norbert Raimund Leisner wrote: Sat Sep 08, 2018 11:21 am Are Windows binaries of RodentIII 0.275 still available? cf.https://github.com/nescitus/Rodent_III/ ... c/main.cpp ?

Norbert
https://ci.appveyor.com/project/nescitu ... /artifacts

Re: Rodent III 0.273 - official release

Posted: Mon Sep 10, 2018 9:18 pm
by PK
No changes over the last few days, so please go directly to the AppVeyor builds. I'm insanely busy these days.

Re: Rodent III 0.273 - official release

Posted: Sat Sep 15, 2018 4:53 pm
by Colin-G
Image

I think this is the most symetrical endgame position, with so many pieces on the board, that I have ever see in my engine-engine matches. The game was played today.
[pgn] [Event "Arena tournament"] [Site "PC70a"] [Date "2018.09.15"] [Round "2"] [White "Rodent III"] [Black "Spark 1.0"] [Result "1/2-1/2"] [ECO "E29"] [TimeControl "40/240:40/240:40/240"] [PlyCount "100"] 1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.e3 O-O 5.Bd3 c5 6.Ne2 Nc6 7.a3 Bxc3+ 8.bxc3 b6 9.e4 d6 10.O-O Ba6 11.Bg5 h6 12.Bh4 g5 13.Bg3 Na5 14.Bxd6 Qxd6 15.e5 Qd8 16.exf6 cxd4 17.Nxd4 Qxf6 18.f4 gxf4 19.Qg4+ Qg5 20.Rxf4 Kh8 21.Qf3 Bb7 22.Be4 Bxe4 23.Qxe4 f5 24.Qf3 Rae8 25.Re1 Nb3 26.g3 Nc5 27.Re5 Rf6 28.Qe3 Rg8 29.Nxe6 Rxe6 30.Rxe6 Nxe6 31.Qxe6 Qxf4 32.Qxg8+ Kxg8 33.gxf4 Kf7 34.Kf2 Ke6 35.Ke3 Kd6 36.Kd4 h5 37.h4 a6 38.Ke3 Kc6 39.Kd2 a5 40.a4 Kd6 41.Ke3 Kc6 42.Ke2 Kd7 43.Kd3 Kc7 44.c5 bxc5 45.c4 Kd7 46.Ke3 Kd6 47.Kd3 Ke6 48.Ke3 Kd6 49.Kd3 Ke6 50.Ke3 Kd6 3-fold repetition 1/2-1/2 [/pgn]

I had some struggle compiling Windows executables of Rodent III v0.275 for some of my computers.
I had to use a linux computer to compile them since the supplied Makefile was incompatible with my Windows TDM-GCC compiler.
I had to use my Mint 18.1 computer since Mint 17 had an old version of gcc/g++ that had no support for the required c++14 standard that was used.
The windows executables were made by replacing "CXX = g++" in the Makefile with "CXX = x86_64-w64-mingw32-g++-posix" to make 64 bit or "CXX = i686-w64-mingw32-g++-posix" for 32 bit executables.
The posix version was required or it would not compile.
I also had to use the "build-static" option for the executables.
The 64 bit versions worked ok if the correct -march value was used for the target computer, but for the 32 bit versions, I had to use -march=i686, other values compiled without errors and appeared to work in Arena until the engine started calculating and immediately crashed.
The 32 bit versions were significantly slower than the 64 bit ones but were faster than the supplied 32 bit version.

For the moderators : the text in my .pgn file says "3-fold repetition". Its displaying as "3-fold retition"

Re: Rodent III 0.273 - official release

Posted: Sat Sep 15, 2018 10:44 pm
by tpoppins
Colin-G wrote: Sat Sep 15, 2018 4:53 pm I think this is the most symetrical endgame position, with so many pieces on the board, that I have ever see in my engine-engine matches. The game was played today.
A nice position! ;)
Colin-G wrote: Sat Sep 15, 2018 4:53 pm For the moderators : the text in my .pgn file says "3-fold repetition". Its displaying as "3-fold retition"
The viewer interprets the comment as a move. A comment must be enclosed in curly braces, viz:
[pgn] [Event "Arena tournament"] [Site "PC70a"] [Date "2018.09.15"] [Round "2"] [White "Rodent III"] [Black "Spark 1.0"] [Result "1/2-1/2"] [ECO "E29"] [TimeControl "40/240:40/240:40/240"] [PlyCount "100"] 1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.e3 O-O 5.Bd3 c5 6.Ne2 Nc6 7.a3 Bxc3+ 8.bxc3 b6 9.e4 d6 10.O-O Ba6 11.Bg5 h6 12.Bh4 g5 13.Bg3 Na5 14.Bxd6 Qxd6 15.e5 Qd8 16.exf6 cxd4 17.Nxd4 Qxf6 18.f4 gxf4 19.Qg4+ Qg5 20.Rxf4 Kh8 21.Qf3 Bb7 22.Be4 Bxe4 23.Qxe4 f5 24.Qf3 Rae8 25.Re1 Nb3 26.g3 Nc5 27.Re5 Rf6 28.Qe3 Rg8 29.Nxe6 Rxe6 30.Rxe6 Nxe6 31.Qxe6 Qxf4 32.Qxg8+ Kxg8 33.gxf4 Kf7 34.Kf2 Ke6 35.Ke3 Kd6 36.Kd4 h5 37.h4 a6 38.Ke3 Kc6 39.Kd2 a5 40.a4 Kd6 41.Ke3 Kc6 42.Ke2 Kd7 43.Kd3 Kc7 44.c5 bxc5 45.c4 Kd7 46.Ke3 Kd6 47.Kd3 Ke6 48.Ke3 Kd6 49.Kd3 Ke6 50.Ke3 Kd6 {3-fold repetition} 1/2-1/2 [/pgn]

I wonder why you need to build your own execs, with every build you could think of available via Appveyor.

Re: Rodent III 0.273 - official release

Posted: Sun Sep 16, 2018 10:03 am
by Colin-G
tpoppins wrote: Sat Sep 15, 2018 10:44 pm
Colin-G wrote: Sat Sep 15, 2018 4:53 pm I think this is the most symetrical endgame position, with so many pieces on the board, that I have ever see in my engine-engine matches. The game was played today.
A nice position! ;)
Colin-G wrote: Sat Sep 15, 2018 4:53 pm For the moderators : the text in my .pgn file says "3-fold repetition". Its displaying as "3-fold retition"
The viewer interprets the comment as a move. A comment must be enclosed in curly braces, viz:
[pgn] [Event "Arena tournament"] [Site "PC70a"] [Date "2018.09.15"] [Round "2"] [White "Rodent III"] [Black "Spark 1.0"] [Result "1/2-1/2"] [ECO "E29"] [TimeControl "40/240:40/240:40/240"] [PlyCount "100"] 1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.e3 O-O 5.Bd3 c5 6.Ne2 Nc6 7.a3 Bxc3+ 8.bxc3 b6 9.e4 d6 10.O-O Ba6 11.Bg5 h6 12.Bh4 g5 13.Bg3 Na5 14.Bxd6 Qxd6 15.e5 Qd8 16.exf6 cxd4 17.Nxd4 Qxf6 18.f4 gxf4 19.Qg4+ Qg5 20.Rxf4 Kh8 21.Qf3 Bb7 22.Be4 Bxe4 23.Qxe4 f5 24.Qf3 Rae8 25.Re1 Nb3 26.g3 Nc5 27.Re5 Rf6 28.Qe3 Rg8 29.Nxe6 Rxe6 30.Rxe6 Nxe6 31.Qxe6 Qxf4 32.Qxg8+ Kxg8 33.gxf4 Kf7 34.Kf2 Ke6 35.Ke3 Kd6 36.Kd4 h5 37.h4 a6 38.Ke3 Kc6 39.Kd2 a5 40.a4 Kd6 41.Ke3 Kc6 42.Ke2 Kd7 43.Kd3 Kc7 44.c5 bxc5 45.c4 Kd7 46.Ke3 Kd6 47.Kd3 Ke6 48.Ke3 Kd6 49.Kd3 Ke6 50.Ke3 Kd6 {3-fold repetition} 1/2-1/2 [/pgn]

I wonder why you need to build your own execs, with every build you could think of available via Appveyor.
Thanks for the reply.
I just copied the pgn direct from the Scid pgn window and did not spot the comment had the { and } missing.
In Scid, comments are just displayed in green in the window, without the brackets.

I compiled a new 32 bit version because the supplied 32 bit version (I only saw one offered) was very slow compared to the 64 bit versions and the 64 bit versions worked ok in Windows but not in Arena, using wine, on my linux computers.