Fat Titz 1.0 released

[Sopel]

Fat Titz 1.0

This is a FREE UCI-compatibile chess engine. It is a fork of cfish.

Thanks to Norman Schidt and Albert Sliver for inspiration on creating my own fork of Stockfish. Also big thanks to Robert Houdart for great optimization ideas, overall making the engine almost 20% faster!

This engine uses a HalfKAv2-2048x2-64-64-1 evaluation network, which contains whopping 4 times the knowledge of Stockfish 14. It was trained partially on Lc0 data, which gives it a unique positional style, while still preserving the tactical sharpness of Stockfish. The network was trained using a modification of the state-of-the-art NNUE trainer, utilizing publicly available datasets 1, 2, 3

Due to the large size the release is available only on google drive here. It includes Windows and Linux binaries for all supported architectures, along with the network. This is the only place where the network can be found.

Additional features

• Polyglot support
• Anarchy mode

  • Code: Select all

    setoption name Anarchy value true

  • makes en-passant forced
• 64-bit hash key
  • reduces the amount of hash collisions and allows for more robust long analysis
  • resizing the transposition table preserves the contents as much as possible
• Persistent transposition table

  • Code: Select all

    setoption name PersistentTTMinDepth value 4

    (min 0, max 255). The minimum entry depth to store/load.

  • Code: Select all

    setoption name PersistentTTFileName value filename.ptt

    . The file which contains the persisted TT. Doesn't do anything on itself.

  • Code: Select all

    setoption name PersistentTTSerialize

    . Serializes the current transposition table according to the options above. The file is overwritten.

  • Code: Select all

    setoption name PersistentTTDeserialize

    . Deserializes the current transposition table according to the options above. Only worse entries are replaced.

[Eduard]

Thanks!

Wow, it's the biggest network I've seen so far. On Ryzen 3900X (22 threads):



Greeting position - Testposition opening:
[d]r1b2rk1/pp5p/1n2pp2/3p4/2P2B2/b1P2NPq/P1Q1BP1P/1K1R3R w - - 0 1

19. g4!

Analysis by Fat Titz 260821 64 AVX2 NUMA:

19.Ng1 Qf5 20.Qxf5 exf5 21.cxd5 Na4 22.Kc2 Bd7 23.Bf3 Nb6 24.Kb1 Ba4 25.Rd2 Nc4 26.Rd4 b5 27.Ne2 Rad8 28.Bc7 Rd7 29.d6 Re8 30.Bd5+ Kf8 31.Bxc4 bxc4 32.Rd2 Bc6 33.Rg1 Be4+ 34.Ka1 Kf7 35.Nd4
White is better: +/- (0.83 ++) Depth: 25/37 00:00:06 63757kN
19.g4
White is better: +/- (0.91 ++) Depth: 25/39 00:00:07 74680kN
19.g4
White is better: +/- (1.03 ++) Depth: 25/39 00:00:07 78404kN
19.g4
White is better: +/- (1.21 ++) Depth: 25/39 00:00:07 81491kN
19.g4
White is clearly better: +/- (1.46 ++) Depth: 25/39 00:00:08 83790kN
19.g4
White is clearly better: +- (1.78 ++) Depth: 25/39 00:00:08 87695kN
19.g4 e5 20.Be3 Bxg4 21.Rhg1 Kh8 22.Ng5 fxg5 23.Bxg4 Qh6 24.Be2 Qc6 25.Rxg5 Nxc4 26.Bxc4 Qxc4 27.Rxe5 Bc5 28.Rexd5 Bxe3 29.fxe3 Rae8 30.Rd7 Rf7 31.Rxf7 Qxf7 32.Qd2 Qf5+ 33.Kc1
White is clearly better: +- (1.63) Depth: 25/39 00:00:08 87832kN
19.g4 e5 20.Be3 Bxg4 21.Rhg1 Kh8 22.Ng5 fxg5 23.Bxg4 Qh6 24.Be2 Qc6 25.Rxg5 Nxc4 26.Bxc4 Qxc4 27.Rxe5 Bc5 28.Rexd5 Bxe3 29.fxe3 Rae8 30.Rd7 Rf7 31.Rxf7 Qxf7 32.Qd2 Qf5+ 33.Kc1
White is clearly better: +- (1.71 ++) Depth: 26/36 00:00:09 104MN
19.g4 e5 20.Be3 Bxg4 21.Rhg1 Kh8 22.Ng5 fxg5 23.Bxg4 Qh6 24.Be2 Qc6 25.Rxg5 Nxc4 26.Bxc4 Qxc4 27.Rxe5 Bc5 28.Rexd5 Bxe3 29.fxe3 Rae8 30.Rd7 Rf7 31.Rxf7 Qxf7 32.Qd2 Qf5+ 33.Kc1
White is clearly better: +- (1.79 ++) Depth: 26/42 00:00:10 116MN, tb=6
19.g4 e5 20.Be3 Bxg4 21.Rhg1 Kh8 22.Ng5 fxg5 23.Bxg4 Qh6 24.Be2 Qc6 25.Rxg5 Nxc4 26.Bxc4 Qxc4 27.Rxe5 Bc5 28.Rexd5 Bxe3 29.fxe3 Rae8 30.Rd7 Rf7 31.Rxf7 Qxf7 32.Qd2 Qf5+ 33.Kc1
White is clearly better: +- (1.92 ++) Depth: 26/42 00:00:10 123MN, tb=52

After only 7 Seconds on my Ryzen 3900X, this is the fastest time I've seen here so far. Nice!

[OfekShochat]

high level trolling sopel, but why is it free? this couldve been sold at 100000000000000 <insert german currency here>

[Raphexon]

Thanks Sopel for letting me beta-test it.

Preliminary testing it seems to scale great:

Code: Select all

stockfish_14_x64_bmi2.exe vs fat_titz_windows_bmi2.exe: 2 - 6 - 83  [0.478]
-15 elo (calculated with elostat)

Book used HERT 250
TC= 600 + 6
1 thread each
12 concurrent games ​

Hope we'll see it soon on SP-CC, FastGM, CCRL and FRGL lists to see how well its ELO trends upwards.

[xr_a_y]

I can't wait for the promotional videos on youtube !

[Guenther]

Fat Titz 1.0

This is a FREE UCI-compatibile chess engine. It is a fork of cfish.

Thomas, you forgot to add that you have already proven in tests, that it will be at least 20 rating points
better than the Slipper and Shit engines ;-)

[Chessqueen]

Fat Titz 1.0

This is a FREE UCI-compatibile chess engine. It is a fork of cfish.

Thomas, you forgot to add that you have already proven in tests, that it will be at least 20 rating points
better than the Slipper and Shit engines

Are you trolling again? You are no different than the ChessIdiot or ChessQuack that I know?

[Damir]

Thomas I think '' Fat Tits'' are much better than Titz...

[OfekShochat]

Fat Titz 1.0

This is a FREE UCI-compatibile chess engine. It is a fork of cfish.

Thomas, you forgot to add that you have already proven in tests, that it will be at least 20 rating points
better than the Slipper and Shit engines

Are you trolling again? You are no different than the ChessIdiot or ChessQuack that I know?

chess queen isnt a better name imo

[Sopel]

Thanks Sopel for letting me beta-test it.

Preliminary testing it seems to scale great:

Code: Select all

stockfish_14_x64_bmi2.exe vs fat_titz_windows_bmi2.exe: 2 - 6 - 83  [0.478]
-15 elo (calculated with elostat)

Book used HERT 250
TC= 600 + 6
1 thread each
12 concurrent games ​

Hope we'll see it soon on SP-CC, FastGM, CCRL and FRGL lists to see how well its ELO trends upwards.

That's a very good result! I only tested it at 60+0.6, and it was about equal to Stockfish 14.