TESTING STOCKFISH 221214MZ : 480 Games.i7 980 3.33 Ghz.
6 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2014t – Limit 8 moves. Sedat –
No tablebases. No RTB used.
Hash 512
Relative Speed: 29.54
Knodes per second: 14.177
[You must be registered and logged in to see this link.]Time Control = 4+0
SF221214MZ x64 POPCNTx6 - Houdini 4 x64_st_X6_CT0 26.0 - 14.0 +14/=24/-2 65.00%
SF221214MZ x64 POPCNTx6 - Komodo 8 64-bit_6_NOB 23.5 - 16.5 +11/=25/-4 58.75%
SF221214MZ x64 POPCNTx6 - Gull 3 x64 XP 25.5 - 14.5 +14/=23/-3 63.75%
Time Control= 2+2
SF221214MZ x64 POPCNTx6 - Houdini 4 x64_st_X6_CT0 23.5 - 16.5 +13/=21/-6 58.75%
SF221214MZ x64 POPCNTx6 - Komodo 8 64-bit_6_NOB 24.0 - 16.0 +10/=28/-2 60.00%
SF221214MZ x64 POPCNTx6 - Gull 3 x64 XP 29.0 – 11.0 +19/=20/-1 72.50%
Score using 6 cores: 151.5 – 81.5= 63.12%
240 Games =
i7 975 3.33 Ghz.
4 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012t –Limit 8 moves. Sedat-
No tablebases. No RTB used.
Hash 256
Relative Speed: 20.62
Knodes per second: 9.899
Time Control: 4+0
SF221214MZ x64 POPCNTx4 - Houdini 4 Pro x64-A_OK 21.0 - 19.0 +5/=32/-3 52.50%
SF221214MZ x64 POPCNTx4 - Komodo 8 64-bit_4_NOB 22.0 - 18.0 +10/=24/-6 55.00%
SF221214MZ x64 POPCNTx4 - Gull 3 x64 XP 25.5 - 14.5 +16/=19/-5 63.75%
Time Control= 2+2
SF221214MZ x64 POPCNTx4 - Houdini 4 Pro x64-A_OK 24.5 - 15.5 +15/=19/-6 61.25%
SF221214MZ x64 POPCNTx4 - Komodo 8 64-bit_4_NOB 22.5 - 17.5 +10/=25/-5 56.25%
SF221214MZ x64 POPCNTx4 - Gull 3 x64 XP 28.0 - 12.0 +16/=24/-0 70.00%
Score using 4 cores: 143.5 – 96.5= 59.79%
240 Games:
[You must be registered and logged in to see this link.]Segmenting by Time Control:
Fixed TC = 143.5 –94.5 = 59.79%
Incremental TC = 151.5 – 88.5 = 63.12%
GLOBAL SCORE: 295.0 – 185.0= 61.46%
Against : Houdini 4.0 St. Ct0 (3227) = 59.37% ; Komodo 8 (3266) = 57.50%, Gull 3 XP (3199) = 67.50%
Average Estimated Elo Opponents = 3231
Estimated Elo Performance= 3311Error bars= +/- 23 EEP
This compile has clearly broken all the records in my tests. Well done, Marco Z!
Before trumpeting that this is the new number one SF compile under my testing conditions, lets repeat this test until 1.400 games, where the score will be much more reliable, with error bars of only +/- 13.
Regards,
Tom CasanovasTESTING STOCKFISH 221214MZ : Test 2 and 3. Games 481 to 1440.i7 980 3.33 Ghz.
6 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2014t – Limit 8 moves. Sedat –
No tablebases. No RTB used.
Hash 512
Relative Speed: 29.54
Knodes per second: 14.177
Time Control = 4+0
SF221214MZ x64 POPCNTx6 - Houdini 4 x64_st_X6_CT0 48.5 - 31.5 +25/=47/-8 60.63%
SF221214MZ x64 POPCNTx6 - Komodo 8 64-bit_6_NOB 46.5 - 33.5 +27/=39/-14 58.13%
SF221214MZ x64 POPCNTx6 - Gull 3 x64 XP 48.5 - 31.5 +26/=45/-9 60.63%
Time Control= 2+2
SF221214MZ x64 POPCNTx6 - Houdini 4 x64_st_X6_CT0 48.5 - 31.5 +27/=43/-10 60.63%
SF221214MZ x64 POPCNTx6 - Komodo 8 64-bit_6_NOB 47.5 - 32.5 +24/=47/-9 59.38%
SF221214MZ x64 POPCNTx6 - Gull 3 x64 XP 50.0 - 30.0 +28/=44/-8 62.50%
Score using 6 cores: 289.5 – 190.5 = 60.31%
480 Games =
[You must be registered and logged in to see this link.]i7 975 3.33 Ghz.
4 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012t –Limit 8 moves. Sedat-
No tablebases. No RTB used.
Hash 256
Relative Speed: 20.62
Knodes per second: 9.899
Time Control: 4+0
SF221214MZ x64 POPCNTx4 - Houdini 4 Pro x64-A_OK 46.0 - 34.0 +20/=52/-8 57.50%
SF221214MZ x64 POPCNTx4 - Komodo 8 64-bit_4_NOB 50.5 - 29.5 +29/=43/-8 63.13%
SF221214MZ x64 POPCNTx4 - Gull 3 x64 XP 52.0 - 28.0 +31/=42/-7 65.00%
Time Control= 2+2
SF221214MZ x64 POPCNTx4 - Houdini 4 Pro x64-A_OK 49.5 - 30.5 +26/=47/-7 61.88%
SF221214MZ x64 POPCNTx4 - Komodo 8 64-bit_4_NOB 45.5 - 34.5 +18/=55/-7 56.88%
SF221214MZ x64 POPCNTx4 - Gull 3 x64 XP 51.5 - 28.5 +27/=49/-4 64.38%
Score using 4 cores: 295.0 – 185.0 = 61.46%
480 Games:
[You must be registered and logged in to see this link.]Segmenting by Time Control:
Fixed TC = 292.0 – 188.0 = 60.83%
Incremental TC = 292.5 – 187.5= 60.94%
GLOBAL SCORE: 584.5 – 375.5= 60.88%
Against : Houdini 4.0 St. Ct0 (3227) = 60.16% ; Komodo 8 (3266) = 59.38%, Gull 3 XP (3199) =
63.13%
Average Estimated Elo Opponents = 3231
Estimated Elo Performance= 3307Tests 2 and 3, with 960 additional games, have confirmed the exceptional strength of this compile.
------------------
SUMMARY OF TESTS 1, 2 AND 3 FOR STOCKFISH 221214 MZ AFTER 1440 GAMES.GLOBAL SCORE: 879.5 – 560.5 = 61.08%
Against : Houdini 4.0 St. Ct0 (3227) = 59.90% ; Komodo 8 (3266) = 58.75%, Gull 3 XP (3199) =
64.59%
Average Estimated Elo Opponents = 3231
Estimated Elo Performance= 3309 Error Bars +/- 13
After 1440 games –around 200 computer hours non stop work- the tests show that this compile is stronger than any other compile or engine, under my testing environment. Only as a reference:
- 9 ELO points stronger than Stock-Fish SF6, my second best.
- 15 ELO points stronger than SF 221214 Development
- 43 ELO points stronger than Komodo 8
- 82 ELO points stronger than Houdini 4 and
- 110 ELO points stronger than Gull 3 XP.
Congratulations to the Stockfish Team
for their great work and to Marco Z.
for this amazing compile.
Regards from Barcelona.
Estimated Elo Performance= 3309~~~~~~~~~~~~~~~~~~~
Ignore corrupt TCEC contest
SF is THE strongest chess engine in the world . and it is FREE
Thank you Tom Casanovas The # 1 (male)
Engine tester