FGRL - 16 Cores vs RTX 2070 rating list - 46. Ethereal 12.50, 49. SlowChess Blitz Classic 2.3

[fastgm]

New: 46. Ethereal 12.50, 49. SlowChess Blitz Classic 2.3

Code: Select all

   # Engine                            :    Elo  Error   Games    Points    (%)      W      D      L     D(%)   CFS
 -------------------------------------------------------------------------------------------------------------------
   1 Stockfish 12                      :   3507      8    2500    1477.0   59.08    532   1890     78   75.60    69
   2 Stockfish 200813 82215d0fd0df     :   3504      8    2500    1477.5   59.10    541   1873     86   74.92    96
   3 Stockfish NNUE SV20200725-1313    :   3493      9    2500    1475.0   59.00    520   1910     70   76.40   100
   4 Lc0 0.25.1 LS 15                  :   3478      8    4750    3464.0   72.93   2392   2144    214   45.14    89
   5 Lc0 0.24.1 LS 14.3                :   3471      7    4750    3433.0   72.27   2347   2172    231   45.73    68
   6 Lc0 0.24.1 LS 14.2                :   3469      9    3250    2472.5   76.08   1799   1347    104   41.45    89
   7 Lc0 0.24.0 LS 14                  :   3460      9    3500    2677.0   76.49   1964   1426    110   40.74    52
   8 Lc0 0.26.1 384x30-J92-115         :   3460      9    3250    2395.5   73.71   1652   1487    111   45.75    92
   9 Lc0k 0.23.2 256x20-T40-1541       :   3452      6    6500    4580.0   70.46   3041   3078    381   47.35    77
  10 Lc0 0.26.1 384x30-T60-4585        :   3448      9    3250    2295.5   70.63   1509   1573    168   48.40    54
  11 Lc0 0.24.1 LS 13                  :   3447      9    3000    2187.5   72.92   1508   1359    133   45.30    66
  12 Stockfish NNUE GK200627           :   3445      8    2500    1305.5   52.22    321   1969    210   78.76    74
  13 Stockfish 200502 C0               :   3441      8    2500    1301.5   52.06    383   1837    280   73.48    72
  14 Lc0 0.26.1 384x30-T60-4619        :   3437      9    3000    2218.0   73.93   1544   1348    108   44.93    57
  15 Lc0 0.26.1 64498                  :   3436      8    3750    2627.5   70.07   1749   1757    244   46.85    84
  16 Lc0 0.24.1 384x30-T60-3010        :   3431      7    5000    3477.5   69.55   2256   2443    301   48.86    59
  17 Fat Fritz 1.1                     :   3429      8    4000    2921.0   73.03   2066   1710    224   42.75    54
  18 Lc0 0.25.1 63681                  :   3429      7    4750    3222.0   67.83   2027   2390    333   50.32    59
  19 Lc0 0.25.1 3972_20k_TCEC          :   3427      9    2750    1964.0   71.42   1305   1318    127   47.93    62
  20 Stockfish 11 C0                   :   3426      5    6750    3390.0   50.22    851   5078    821   75.23    55
  21 Lc0 0.22.0 42850                  :   3425      7    4500    3173.5   70.52   2142   2063    295   45.84    51
  22 Lc0 0.22.0 T40B.4-160             :   3425      7    4500    3173.0   70.51   2131   2084    285   46.31    80
  23 Stockfish 11                      :   3421      7    4000    2157.5   53.94    777   2761    462   69.03    71
  24 Lc0 0.25.1 63305                  :   3419      7    4750    3176.5   66.87   1926   2501    323   52.65    74
  25 Lc0 0.25.1 63305 (Kayra4)         :   3415      9    3250    2285.5   70.32   1477   1617    156   49.75    83
  26 Lc0k 0.23.2 62172                 :   3409      7    4750    3160.0   66.53   1924   2472    354   52.04    80
  27 Lc0 LS-LS 12.2                    :   3405      8    3750    2680.5   71.48   1809   1743    198   46.48    58
  28 Lc0 0.22.0 J13B.2-188             :   3404      7    4500    3068.0   68.18   1955   2226    319   49.47    76
  29 Lc0 0.25.1 703042                 :   3400      7    4750    3073.5   64.71   1917   2313    520   48.69    64
  30 Lc0 LS-LS 12.1                    :   3398      9    3000    2056.0   68.53   1299   1514    187   50.47    53
  31 Lc0k 0.23.2 384x30-T40-2036       :   3398      8    4250    2905.0   68.35   1923   1964    363   46.21    99
  32 Fat Fritz 1.0                     :   3385      8    3000    2012.0   67.07   1311   1402    287   46.73    86
  33 Lc0 0.22.0 32930                  :   3379      7    4500    2942.0   65.38   1821   2242    437   49.82    65
  34 Lc0 0.22.0 LS 12 preview          :   3377      9    2750    1876.0   68.22   1258   1236    256   44.95    58
  35 Stockfish 10                      :   3376      8    2750    1337.5   48.64    428   1819    503   66.15    72
  36 Lc0 0.23.2 61950                  :   3373      8    3750    2482.0   66.19   1550   1864    336   49.71    74
  37 Lc0 0.23.2 58613+                 :   3369      8    3500    2268.5   64.81   1426   1685    389   48.14    63
  38 Lc0 0.22.0 384x30-T40-1573        :   3367      9    2750    1846.0   67.13   1220   1252    278   45.53    96
  39 Lc0 0.25.1 3972_20k_TCEC+         :   3355      9    2500    1646.5   65.86    944   1405    151   56.20    96
  40 Komodo 14                         :   3346      6    5000    1928.0   38.56    368   3120   1512   62.40    89
  41 Lc0 0.22.0 11260                  :   3339      9    2750    1755.5   63.84   1044   1423    283   51.75   100
  42 Houdini 6.03                      :   3326      5    8250    3166.0   38.38    726   4880   2644   59.15    66
  43 Komodo 13.3                       :   3324      7    3750    1379.0   36.77    294   2170   1286   57.87   100
  44 Komodo 13.2.5                     :   3311      6    4500    1749.5   38.88    503   2493   1504   55.40   100
  45 Scorpio 3.07 NN-ens-net-20x256    :   3280      9    2500    1466.5   58.66    855   1223    422   48.92    98
  46 Ethereal 12.50                    :   3266     10    2500     692.5   27.70     77   1231   1192   49.24    96
  47 Ethereal 12.25                    :   3255      8    3500     954.5   27.27    145   1619   1736   46.26   100
  48 Ethereal 12.00                    :   3222      7    5250    1300.0   24.76    213   2174   2863   41.41    65
  49 SlowChess Blitz Classic 2.3       :   3220     10    2500     571.5   22.86     64   1015   1421   40.60    85
  50 Xiphos 0.6                        :   3214      5    8250    2075.0   25.15    294   3562   4394   43.18    88
  51 Ethereal 11.75                    :   3208      7    4500    1196.0   26.58    210   1972   2318   43.82    59
  52 SlowChess Blitz Classic 2.2       :   3207      9    3500     780.5   22.30    102   1357   2041   38.77    90
  53 Fire 7.1                          :   3200      6    8250    1959.5   23.75    353   3213   4684   38.95   100
  54 RofChade 2.3                      :   3165      7    5750    1087.0   18.90    128   1918   3704   33.36   100
  55 Booot 6.4                         :   3140      8    6250    1059.0   16.94    115   1888   4247   30.21    81
  56 Laser 1.7                         :   3135      7    7750    1401.0   18.08    201   2400   5149   30.97    86
  57 Schooner 2.2                      :   3128     12    3000     451.0   15.03     45    812   2143   27.07    65
  58 Andscacs 0.95                     :   3125      7    7250    1262.0   17.41    205   2114   4931   29.16   100
  59 Deep Shredder 13                  :   3103      9    5000     818.0   16.36    114   1408   3478   28.16    83
  60 Booot 6.3.1                       :   3097      9    4750     767.0   16.15    118   1298   3334   27.33   ---

White advantage = 58.87 +/- 0.69
Draw rate (equal opponents) = 69.05 % +/- 0.20

Games        : 125.500 (finished)

White Wins   :  40.846 (32.5 %)
Black Wins   :  24.873 (19.8 %)
Draws        :  59.781 (47.6 %)

White Score  : 56.4 %
Black Score  : 43.6 %

http://www.fastgm.de

[Jouni]

Great, that somebody test with 16 cores! But is 512 MB hash really enough?

[Wolfgang]

Look at the time control and you can answer the question yourself...

[AndrewGrant]

Great, that somebody test with 16 cores! But is 512 MB hash really enough?

Probably not. I would expect to see 4GB, if you scale other lists.

[Alayan]

How did you come up with 4GB ? 64MB at 60s+0.6s scales to 1GB with 16 cores, and 64MB is what Ethereal/Stockfish use on their respective testing framework. 128MB is used in FastGM's bullet list, but that would scale to 2GB not 4.

[AndrewGrant]

How did you come up with 4GB ? 64MB at 60s+0.6s scales to 1GB with 16 cores, and 64MB is what Ethereal/Stockfish use on their respective testing framework. 128MB is used in FastGM's bullet list, but that would scale to 2GB not 4.

I was attempting to scale up Ethereal's and Stockfish's testing defaults, based on core counts and time.

[Raphexon]

Those Xeon cores are fairly slow.
I'd say 512 mb is probably a little tight, but over 1gb would get liberal very quickly.

[Jouni]

OK probably 512 MB is enough for SF, specially when NN slowed nps about 30%.