Floyd 0.5 released

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Harvey Williamson, bob

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
mvk
Posts: 589
Joined: Tue Jun 04, 2013 8:15 pm

Floyd 0.5 released

Post by mvk » Sun Oct 11, 2015 12:28 pm

I have created a mini UCI engine for study purposes. Yesterday it played in HGM's monthly and I decided to make a release out of that. Download page is here: https://marcelk.net/floyd/

From the release notes:

Version 0.5

- Initial release. Support for Windows, Linux and Mac.
- Very bare-bones chess engine with source and permissive license.
- (Almost) full UCI support, except "go ponder", "go nodes" and "go mate"
- Just PVS + transposition table + null move and a very simplistic eval.
- Evaluation: No king safety. Some of the eval is done in floating point. No pawn table, no material table, no piece/square tables.
- Search: No reductions, no killers, no history table, single-core only.
- Everything is recalculated again and again in every node for simplicity.
- Move generation is a far descendent of MSCP's. All other code is new.
- Comes with built-in CPU benchmark. Type "bench" on the console interface.
- Not super strong. Finished 18 out of 30 in the October 2015 HGM's monthly on-line blitz tourney.
- Win64 exe courtesy J. Buijs.
[Account deleted]

Ferdy
Posts: 3607
Joined: Sun Aug 10, 2008 1:15 pm
Location: Philippines

Re: Floyd 0.5 released

Post by Ferdy » Sun Oct 11, 2015 3:20 pm

This is not right.

Code: Select all

option name Hash type spin default 0 min 0

Code: Select all

Floyd Chess Engine - Version 0.5
Copyright (C) 1998-2015, Marcel van Kervinck
All rights reserved

Type "help" for more information, or "quit" to leave.

uci
id name Floyd 0.5
id author Marcel van Kervinck
option name Hash type spin default 0 min 0
option name Clear Hash type button
uciok

mvk
Posts: 589
Joined: Tue Jun 04, 2013 8:15 pm

Re: Floyd 0.5 released

Post by mvk » Sun Oct 11, 2015 6:12 pm

Ferdy wrote:This is not right.

Code: Select all

option name Hash type spin default 0 min 0
Indeed, default should say 128. I will fix that in the next version. It should still play and be changeable. The minimum is 0, there is no max imposed by the engine.
[Account deleted]

User avatar
MikeB
Posts: 2522
Joined: Thu Mar 09, 2006 5:34 am
Location: Pen Argyl, Pennsylvania

Re: Floyd 0.5 released

Post by MikeB » Sun Oct 11, 2015 6:41 pm

mvk wrote:I have created a mini UCI engine for study purposes. Yesterday it played in HGM's monthly and I decided to make a release out of that. Download page is here: https://marcelk.net/floyd/

From the release notes:

Version 0.5

- Initial release. Support for Windows, Linux and Mac.
- Very bare-bones chess engine with source and permissive license.
- (Almost) full UCI support, except "go ponder", "go nodes" and "go mate"
- Just PVS + transposition table + null move and a very simplistic eval.
- Evaluation: No king safety. Some of the eval is done in floating point. No pawn table, no material table, no piece/square tables.
- Search: No reductions, no killers, no history table, single-core only.
- Everything is recalculated again and again in every node for simplicity.
- Move generation is a far descendent of MSCP's. All other code is new.
- Comes with built-in CPU benchmark. Type "bench" on the console interface.
- Not super strong. Finished 18 out of 30 in the October 2015 HGM's monthly on-line blitz tourney.
- Win64 exe courtesy J. Buijs.
Very nice - you might want to put the mac exe in a zip file, it gets mangled during the download. Was able to get it from the "big" zipped file.

mac bench :

Code: Select all

iMac:floyd0-2.5 michaelbyrne$ ./floyd0.5.osx

Floyd Chess Engine - Version 0.5
Copyright (C) 1998-2015, Marcel van Kervinck
All rights reserved

Type "help" for more information, or "quit" to leave.

uci
id name Floyd 0.5
id author Marcel van Kervinck
option name Hash type spin default 0 min 0
option name Clear Hash type button
uciok
bench
info time 1001 nps 689291 fen r1bqk1nr/pp3ppp/2nb4/1B1p4/8/1N3N2/PPP2PPP/R1BQ1RK1 b kq -
info time 1002 nps 820137 fen 4rrk1/ppqb1ppn/3p2np/2pP4/2P1P3/2PBN1B1/P5PP/1R1Q1RK1 w - -
info time 1003 nps 734361 fen 3rk2r/1pq1bpp1/p1n1p1n1/3pP3/3P2P1/1PPQNNB1/5P2/R3R1K1 b k -
info time 1003 nps 778789 fen 7k/8/1p1Q2pp/1p3b2/3p3P/P3q3/BP4PK/8 b - -
info time 1000 nps 672242 fen 2r1rnk1/4qppp/p3b3/1p4P1/4p2P/2N1PN2/PP1Q1P2/1KR4R w - -
info time 1004 nps 754296 fen 2rk3r/p2n1pp1/bqn1p2p/1p1pP3/3P1N1P/bP1BBN2/P3QPP1/2R1R1K1 w - -
info time 1004 nps 908488 fen 1nr2rk1/p3Qppp/1p2p1n1/8/3P4/2N2N2/PP3PPP/2R2RK1 b - -
info time 1001 nps 1068648 fen 3nr1k1/1p1r3p/pR1pNpp1/3P4/4P3/3P3P/6P1/1R5K b - -
info time 1005 nps 713639 fen 4r1k1/pp3ppp/3n4/2n1r3/2P2NP1/4KB2/PP2P2P/R5R1 w - -
info time 1000 nps 1312800 fen 8/p5pk/1p6/5R1p/P7/5PPK/5r2/8 b - -
info time 1002 nps 1063599 fen 8/2Q2p2/5qpk/4p3/8/5P1P/6K1/8 w - -
info time 1005 nps 929624 fen 5rk1/1pr2bbp/pNp1n1p1/2P1p3/1PBpP3/1N5P/1P3PP1/3RR1K1 b - -
info time 1004 nps 1053890 fen 8/1b3pk1/rp5p/6p1/6B1/1R4P1/4P2P/4K3 w - -
info time 1003 nps 692620 fen r1bqk2r/pp2ppbp/3p1np1/8/2PQP3/2N5/PP2BPPP/R1B1K2R w KQkq -
info time 1004 nps 906040 fen 8/1b4pk/7q/p3pP2/1p2P3/1Pp5/2P1Q3/5BK1 b - -
info time 1005 nps 801413 fen r2qkb1r/2p2ppp/p1p5/3pP3/4n1b1/3Q1N2/PPP2PPP/RNB2RK1 w kq -
info time 1003 nps 829217 fen 4r1k1/pp2pp1n/3p2p1/6Bp/2rN1P1P/q1P5/2PQ2P1/1K1RR3 b - -
info time 1002 nps 798970 fen rnbq1rk1/pp3pbp/2pp1np1/4p3/2PPP3/2N2N2/PP2BPPP/R1BQ1RK1 w - -
info time 1001 nps 1041555 fen 8/8/p3k1p1/Bb5p/1P5P/5P2/1r4P1/3R2K1 b - -
info time 1000 nps 948843 fen 4k3/q4p2/2n1p1p1/p1Np4/3P1QP1/PP3P2/6K1/8 b - -
info time 1000 nps 1241384 fen 3bn3/4k3/2p3p1/ppN2P2/P5P1/1P1KB3/8/8 b - -
info time 1000 nps 780187 fen r1b1r1k1/5pbp/pp1p1np1/2qPn3/P1p1P3/2N1BPN1/1P1QB1PP/R1R4K b - -
info time 1000 nps 885228 fen r1q1rnk1/p1p2pp1/bp2p2p/n2pP3/Q1PP1N1B/P1PB1P2/6PP/R4RK1 w - d6
info time 1001 nps 871202 fen r3r1k1/1bpnqppp/1p1p4/p2Pp3/1PP1P3/P5P1/1BQ2PBP/1R1R2K1 b - -
info time 1000 nps 1128302 fen 1r6/4Pk2/1p1p1P2/p7/8/1P1K2p1/P7/5R2 b - -
info time 1001 nps 821699 fen 1rbqk2r/2pp1ppp/p1n2n2/1pb1p3/P3P3/1B3N2/1PPP1PPP/RNBQ1RK1 w k -
info time 1005 nps 1040345 fen r4bk1/6pp/1B2n3/4P3/pp6/1P2N1P1/P5KP/3R4 w - -
info time 1005 nps 733834 fen r1bqkb1r/pp1n1ppp/2n5/2ppP3/3P1P2/5N2/PP4PP/RNBQKB1R w KQkq -
info time 1004 nps 809897 fen r1qr2k1/2p2pPp/p3n3/1p6/8/4BQ2/PP3PPP/R2R2K1 w - -
info time 1004 nps 1068439 fen 4r3/1p4k1/p1p1P1pp/3p4/1P1N4/P6B/7P/4K3 w - -
info time 1005 nps 813504 fen 8/R4pk1/4pbq1/8/2P4p/5Q2/8/5K2 b - -
info time 1005 nps 1148896 fen 4k3/R4pp1/P7/4P2p/5P2/3p2P1/r7/4K3 w - -
info time 1005 nps 752562 fen 2r3k1/1p4p1/p3bp2/8/N3Pb1p/2R4P/1PP1Q1P1/2q3BK b - -
info time 1001 nps 742962 fen rnbqkb1r/ppp2ppp/4pn2/3p4/2PP4/2N5/PP2PPPP/R1BQKBNR w KQkq -
info time 1000 nps 1612399 fen 8/6N1/8/1k4K1/3n4/8/5P2/8 b - -
info time 1000 nps 815187 fen rnbqkbnr/ppp1pppp/3p4/8/4P3/8/PPPP1PPP/RNBQKBNR w KQkq -
info time 1000 nps 1172503 fen 8/PRn5/3k4/r4p2/5PB1/8/5P2/6K1 w - -
info time 1000 nps 1302123 fen r5k1/4Rpp1/3p3p/2pP4/p1P3P1/1P6/P4PP1/5K2 w - -
info time 1005 nps 1264079 fen 5n2/2k1KB2/7P/8/3n1P2/8/p1N5/8 w - -
info time 1002 nps 807123 fen rnbqkbnr/pp1ppppp/8/2p5/4P3/5N2/PPPP1PPP/RNBQKB1R b KQkq -
result nps 933163
I would to take a shot and do a mac compile for you. I have sort of become an expert compiling for my machines.(iMac and Mac Pro)

User avatar
MikeB
Posts: 2522
Joined: Thu Mar 09, 2006 5:34 am
Location: Pen Argyl, Pennsylvania

Re: Floyd 0.5 released

Post by MikeB » Sun Oct 11, 2015 6:58 pm

Image

hmmm - the Windows version under VM is faster than the native Mac version ? - we can make that better, let me know if I can help...

User avatar
Sylwy
Posts: 3358
Joined: Fri Apr 21, 2006 2:19 pm
Location: IASI (Romania) - the historical capital of MOLDOVA

Re: Nice or not ?

Post by Sylwy » Sun Oct 11, 2015 7:08 pm

THANK YOU !

For Floyd 0.5:


Image

Image

Image

SilvianR :wink:

NR: the engine is already installed on my test PC !

mvk
Posts: 589
Joined: Tue Jun 04, 2013 8:15 pm

Re: Floyd 0.5 released

Post by mvk » Sun Oct 11, 2015 8:35 pm

I will replace the executable with a zip. I downloaded the file and the contents are ok. The problem you experienced is possibly the usual unix problem where there is metadata outside the file (permission flags) which don't get transmitted. Zip captures that correctly. Besides, some firewalls block downloading of binaries.

One difference between my compiles and Joost's is that mine have debugging checks in them for faster diagnostics in case of crashes.
[Account deleted]

mvk
Posts: 589
Joined: Tue Jun 04, 2013 8:15 pm

Re: Nice or not ?

Post by mvk » Sun Oct 11, 2015 9:05 pm

Sylwy wrote:THANK YOU !

For Floyd 0.5:


Image

Image

Image

SilvianR :wink:

NR: the engine is already installed on my test PC !
Nice logos, I like the middle one best :-) Please let me know how it goes, especially if it hangs, crashes or beats Stockfish. All of those are not supposed to happen!
[Account deleted]

User avatar
Sylwy
Posts: 3358
Joined: Fri Apr 21, 2006 2:19 pm
Location: IASI (Romania) - the historical capital of MOLDOVA

Re: The hash is a problem !

Post by Sylwy » Mon Oct 12, 2015 5:53 am

mvk wrote: Please let me know how it goes, especially if it hangs, crashes or beats Stockfish. All of those are not supposed to happen!
The default hash is 128 MB. OK, I started the match (with Fridolin 2.00) with 128 MB for each engine, but the Floyd hash progressively increased (much).
That's for sure a problem !
The style is very particular, something like an eccentric master. I like it. An original engine !

SilvianR :wink:

Image

Image

User avatar
Sylwy
Posts: 3358
Joined: Fri Apr 21, 2006 2:19 pm
Location: IASI (Romania) - the historical capital of MOLDOVA

Re: The hash is a problem !

Post by Sylwy » Mon Oct 12, 2015 12:48 pm

Some infos regarding my above post :

GUI: Arena 3.5
OS: Widows XP SP3

SilvianR :wink:

Post Reply