Stefan Pohl Computer Chess

private website for chessengine-tests


Latest Website-News (2020/01/15): AB-testrun of Slow Chess 2.5 avx2 finished: +50 Elo to Slow Chess 2.4 - nice progress! 

NN-testrun of Lc0 0.26.3 66988 net finished. See the result and download the games in the "NN vs SF testing"- section.

Next NN-testrun: Lc0 0.26.3 J94-100

Next AB-testrun: Stockfish 210111 avx2

 

I released the new V2.00 of my Unbalanced Human Openings. All of the 400000 raw-data endpositions were re-evaluated with KomodoDragon 1.0 (instead of Komodo 14 in UHO V1.0) and all UHO openings-sets and opening-books were rebuilt.

Learn more in the "Unbalanced Human Openings"- section or download them right here

 

Stay tuned.


Stockfish testing

 

Playing conditions:

 

Hardware: Since 20/07/21 AMD Ryzen 3900 12-core (24 threads) notebook with 32GB RAM. Now, 20 games are played simultaneously (!), so from now, each testrun will have 6000 or 7000 games (instead of 5000 before) and will take only 2 days, not 6-7 days as before! From now, all engine-binaries are popcount/avx2, of course, because bmi2-compiles are extremly slow on AMD. To keep the rating-list engine-names consistent, the "bmi2"- or "pext"-extension in the engine-name is still in use for older engines - otherwise ORDO will not calculate all played games by this engine as one engine...

Speed: (singlethread, TurboBoost-mode switched off, chess starting position) Stockfish: 1.3 mn/s, Komodo: 1.1 mn/s

Hash: 256MB per engine

GUI: Cutechess-cli (GUI ends game, when a 5-piece endgame is on the board)

Tablebases: None for engines, 5 Syzygy for cutechess-cli

Openings: HERT_500 testset (by Thomas Zipproth) (download the file at the "Download & Links"-section or here)

Ponder, Large Memory Pages & learning: Off

Thinking time: 180''+1000ms (= 3'+1'') per game/engine (average game-duration: around  7.5 minutes). One 7000 games-testrun takes about 2 days.The version-numbers of the Stockfish engines are the date of the latest patch, which was included in the Stockfish sourcecode, not the release-date of the engine-file, written backwards (year,month,day))(example: 200807 = August, 7, 2020). The used SF compile is the AVX2-compile, which is the fastest on my AMD Ryzen CPU. SF binaries are taken from abrok.eu (except the official SF-release versions, which are taken form the official Stockfish website).

Download BrainFish (and the Cerebellum-Libraries)here

 

To avoid distortions in the Ordo Elo-calculation, from now, only 3x Stockfish (latest official release + the latest 2 dev-versions) and 1x Brainfish are stored in the gamebase (all older engine-versions games will be deleted, every time, when a new version was tested). Stockfish and BrainFish older Elo-results can still be seen in the Elo-diagrams below. BrainFish plays always with the latest Cerebellum-Libraries of course, because otherwise BrainFish = Stockfish.

 

Latest update: 2020/01/13: Slow Chess 2.5 (+50 Elo to Slow Chess 2.4)

 

(Ordo-calculation fixed to Stockfish 12 = 3684 Elo)

 

See the individual statistics of engine-results here

Download the current gamebase here

 

     Program                      Elo    +    -   Games   Score   Av.Op.  Draws

   1 CFish 12 3xCerebellum      : 3725    9    9  7000    86.1 %   3388   27.3 %
   2 Stockfish 201214 avx2      : 3723    8    8  7000    78.4 %   3471   41.2 %
   3 Stockfish 201231 avx2      : 3721    8    8  7000    78.1 %   3473   42.5 %
   4 Stockfish 201225 avx2      : 3721    7    7  7000    78.2 %   3471   41.7 %
   5 CFish 12 avx2              : 3703    9    9  7000    84.6 %   3388   29.1 %
   6 Stockfish 12 200902        : 3684    4    4 25000    77.3 %   3442   40.8 %
   7 KomodoDragon 1.0 avx2      : 3648    5    5 14000    71.0 %   3464   44.8 %
   8 SF 200910 miniNNue avx2    : 3615    7    7  7000    72.1 %   3436   43.2 %
   9 Stockfish 200731 popc      : 3600    8    8  7000    80.5 %   3344   36.2 %
  10 Stockfish 11 200118        : 3563    4    4 17000    69.5 %   3402   41.6 %
  11 Stockfish 10 181129        : 3524    5    5 15000    78.5 %   3287   37.7 %
  12 KomodoDragon 1.0 MCTS      : 3480    6    6  7000    57.6 %   3426   57.1 %
  13 Stockfish 9 180201         : 3474    9    9  5000    74.9 %   3272   41.7 %
  14 Komodo 14.1 x64            : 3453    6    6  8000    56.3 %   3410   55.6 %
  15 Komodo 14 bmi2             : 3444    4    4 19000    51.3 %   3438   51.6 %
  16 Houdini 6 pext             : 3440    2    2 51000    56.4 %   3393   46.8 %
  17 Nemorino 6.00 avx2         : 3438    4    4 21000    47.7 %   3463   50.5 %
  18 Komodo 13.3 bmi2           : 3437    6    6  8000    62.8 %   3341   49.9 %
  19 Komodo 13.1 bmi2           : 3424    5    5 11000    62.0 %   3333   48.8 %
  20 Slow Chess 2.5 avx2        : 3422    6    6  7000    45.0 %   3467   51.0 %
  21 Komodo 12.3 bmi2           : 3411    7    7  7000    62.7 %   3313   49.4 %
  22 Ethereal 12.75 avx2        : 3398    4    4 20000    42.1 %   3468   49.9 %
  23 Ethereal 12.62 avx2        : 3389    6    6  8000    49.1 %   3401   54.6 %
  24 Slow Chess 2.4 popc        : 3372    5    5 15000    37.2 %   3486   46.3 %
  25 Ethereal 12.50 popc        : 3356    6    6  7000    45.8 %   3396   55.2 %
  26 Slow Chess 2.3 popc        : 3343    4    4 14000    42.3 %   3407   52.2 %
  27 Komodo 14 MCTS             : 3339    8    8  5000    44.4 %   3384   53.4 %
  28 Ethereal 12.25 pext        : 3337    5    5 12000    35.2 %   3470   46.4 %
  29 Pedone 3 avx2              : 3330    6    6  8000    33.4 %   3466   46.7 %
  30 Slow Chess 2.2 popc        : 3328    5    5 11000    32.9 %   3482   42.7 %
  31 Igel 2.9.0 popavx2         : 3328    6    6 10000    30.4 %   3493   45.1 %
  32 RubiChess 1.9dev nnue      : 3319    6    6 10000    27.2 %   3521   40.1 %
  33 Ethereal 12.00 pext        : 3316    6    6  9000    43.1 %   3370   50.8 %
  34 Ethereal 11.75 pext        : 3308    5    5  9000    39.3 %   3391   53.2 %
  35 Xiphos 0.6 bmi2            : 3303    3    3 31000    36.2 %   3421   48.9 %
  36 Fire 7.1 popc              : 3300    3    3 41000    41.9 %   3372   50.7 %
  37 Xiphos 0.5.6 bmi2          : 3288    7    7  7000    41.2 %   3355   54.6 %
  38 Minic 2.51 nasc_nutr       : 3282    6    6  7000    31.2 %   3436   45.0 %
  39 Ethereal 11.53 pext        : 3280    6    6  7000    42.2 %   3342   53.4 %
  40 Komodo 12.3 MCTS           : 3275    7    7  7000    42.7 %   3333   46.3 %
  41 Ethereal 11.25 pext        : 3271    7    7  6000    38.4 %   3361   51.0 %
  42 rofChade 2.3 bmi2          : 3257    5    5 11000    33.8 %   3387   47.5 %
  43 Booot 6.4 popc             : 3244    7    7  6000    31.1 %   3393   46.5 %
  44 Schooner 2.2 popc          : 3241    7    7  6000    31.3 %   3390   50.3 %
  45 Laser 1.7 bmi2             : 3218    7    7  6000    30.8 %   3370   45.8 %
  46 Fizbo 2 bmi2               : 3212    8    8  5000    36.0 %   3324   39.0 %
  47 Fritz 17                   : 3212    7    7  6000    29.4 %   3376   44.2 %
  48 Shredder 13 x64            : 3209    8    8  6000    31.9 %   3358   42.6 %
  49 RubiChess 1.8 popc         : 3207    6    6  7000    32.0 %   3344   46.1 %
  50 Defenchess 2.2 popc        : 3205    8    8  5000    26.6 %   3394   41.8 %
  51 Booot 6.3.1 popc           : 3199    8    8  5000    34.0 %   3327   44.1 %
  52 Andscacs 0.95 popc         : 3167    8    8  5000    23.1 %   3390   35.4 %

The version-numbers (180622 for example) of the engines are the date of the latest patch, which was included in the Stockfish sourcecode, not the release-date of the engine-file. Especially the asmFish-engines are often released much later!!

Below you find a diagram of the progress of Stockfish in my tests since August 2020

And below that diagram, the older diagrams.

 

You can save the diagrams (as a JPG-picture (in originial size)) on your PC with mouseclick (right button) and then choose "save image"...

The Elo-ratings of older Stockfish dev-versions in the Ordo-calculation can be a little different to the Elo-"dots" in the diagram, because the results/games of new Stockfish dev-versions - when getting part of the Ordo-calculation - can change the Elo-ratings of the opponent engines and that can change the Elo-ratings of older Stockfish dev-versions (in the Ordo-calculation / ratinglist, but not in the diagram, where all Elo-"dots" are the rating of one Stockfish dev-version at the moment, when the testrun of that Stockfish dev-version was finished).

 

 

 

 

 

 


Sie sind Besucher Nr.