Testing Stockfish 11-03-13. 480 Games.

Discussion of computer chess matches and engine tournaments.

Moderator: Ras

APassionForCriminalJustic
Posts: 417
Joined: Sat May 24, 2014 9:16 am

Re: For Tom- or Maybe Anyone Who Knows!!

Post by APassionForCriminalJustic »

geots wrote:Is there a place I can go to pick out maybe this version or a few others to download. Please don't tell me we have private beta stockfishes. I would really like to see what this one and a few more will do in comparison tests with 2.3.1 ag. same opponents on 1 core.

I have a couple 4 core i5s, as you know, and am ordering an i7 6-core system when I decide on the best one where hyperthr. can be disabled. Because they don't make them without it. And I would NEVER run one game with it enabled.

I want to get 1-core results because I know the great majority of users just don't have 6-core systems- and I think they would like to see what the versions will do for them. Plus you know I have always admired your work- but I am not sold on the idea that 2+2 is a suitable control for 6-cores.

At any rate, are these private or can I download a few?



Best,

george
The best one where hyper threading can be disabled? What are you talking about? All of Intel's CPUs can have hyper threading disabled easily within the BIOS. At any rate, hyper threading should always be enabled as it is one of Intel's awesome CPU features that will provide some benefit when generally running applications in Windows. What a lot of people probably do not know much about is that you can simply set the CPU affinity. In essence, you can get say SF or Komodo to run on ONLY the physical cores while keeping the hyper threaded cores enabled; this will help when background applications need some of the CPU resources which is basically always. Disabling hyper threading is not smart; it never was, and it never will be. Furthermore, it is a very noob-like thing to do.

Most users do not have one-core machines anymore either, so one-core tests are only good for testing engine strength without including how MP may or may not help (affect) the engine(s).

There is no such thing as what a suitable time control may be for a six-core. It does not matter. The correct question is what time control is most suitable for whichever particular engine. For instance, Komodo will always increase in ELO as the time control increases, and as the number of cores increase.
Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 9:09 pm

Re: For Tom- or Maybe Anyone Who Knows!!

Post by Tomcass »

TESTING STOCKFISH IPMAN 230515 = 480 GAMES

6 real cores
Ponder: Off.
GUI: Fritz 14
Book: Perfect 2014c – Limit 5 moves. Sedat –
No tablebases. No RTB used.
Hash 512
Relative Speed: 29.54
Knodes per second: 14.177

Time Control 4 + 0
SF 230515IPx 64 POPCNTx6 - Komodo 9.01 64-bit_X6 22.0 - 18.0 +9/=26/-5 55.00%
SF 230515IPx 64 POPCNTx6 - Houdini 4 x64_st_X6_CT0 24.5 - 15.5 +12/=25/-3 61.25%
SF 230515IPx 64 POPCNTx6 - Gull 3 x64 XP 28.0 - 12.0 +18/=20/-2 70.00%

Time Control 2 + 2
SF 230515IPx 64 POPCNTx6 - Komodo 9.01 64-bit_X6 19.5 - 20.5 +5/=29/-6 48.75%
SF 230515IPx 64 POPCNTx6 - Houdini 4 x64_st_X6_CT0 25.0 - 15.0 +12/=26/-2 62.50%
SF 230515IPx 64 POPCNTx6 - Gull 3 x64 XP 24.5 - 15.5 +13/=23/-4 61.25%

Score: 143.5 – 96.5 = 59.79%

240 Games using 6 cores:
http://www.mediafire.com/download/20pv0 ... 0Games.pgn

8 real cores
Ponder: Off
GUI: Fritz 14
Book: Perfect 2014c – Limit 5 moves. Sedat –
No tablebases. No RTB used.
Hash 512
Relative Speed: 44.92
Knodes per Second: 21.562

Time Control 4 + 0
SF 230515IPx 64 POPCNTx8_BO - Komodo 9.01 64-bit-X8 18.5 - 21.5 +6/=25/-9 46.25%
SF 230515IPx 64 POPCNTx8_BO - Houdini 4 Pro x64_Ct0_8 24.0 - 16.0 +10/=28/-2 60.00%
SF 230515IPx 64 POPCNTx8_BO - Gull 3 x64 XP_8 29.5 - 10.5 +22/=15/-3 73.75%

Time Control 2 + 2
SF 230515IPx 64 POPCNTx8_BO - Komodo 9.01 64-bit-X8 17.5 - 22.5 +7/=21/-12 43.75%
SF 230515IPx 64 POPCNTx8_BO - Houdini 4 Pro x64_Ct0_8 25.0 - 15.0 +12/=26/-2 62.50%
SF 230515IPx 64 POPCNTx8_BO - Gull 3 x64 XP_8 27.5 - 12.5 +18/=19/-3 68.75%

Score: 142.0 – 98.0 = 59.17%

240 Games using 8 cores:
http://www.mediafire.com/download/a9f4l ... 0Games.pgn

GLOBAL SCORE AFTER 480 GAMES= 285.5 – 194.5 = 59.48%

Against : Komodo 9.01 (3219) = 48.44% Houdini 4 (3136) = 61.56%, Gull 3 XP (3103) = 68.44%

Average Estimated Elo Opponents = 3153
Estimated Elo Performance= 3219


Regards,

Tom
User avatar
Graham Banks
Posts: 44665
Joined: Sun Feb 26, 2006 10:52 am
Location: Auckland, NZ

Re: For Tom- or Maybe Anyone Who Knows!!

Post by Graham Banks »

APassionForCriminalJustic wrote:Most users do not have one-core machines anymore....
I'd imagine that even if most have quads, they'd use them for things other than chess as well, so they'd probably be using one or two cores for chess.
gbanksnz at gmail.com
Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 9:09 pm

Re: For Tom- or Maybe Anyone Who Knows!!

Post by Tomcass »

TESTING AGAIN STOCKFISH 230415= 1440 ADDITIONAL GAMES

6 real cores

Ponder: Off.
GUI: Fritz 14
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

Stockfish 230415b 64 SSE42_ - Komodo 9.01 64-bit_X6 63.5 - 56.5 +25/=77/-18 52.92%
Stockfish 230415b 64 SSE42_ - Houdini 4 x64_st_X6_CT0 79.5 - 40.5 +44/=71/-5 66.25%
Stockfish 230415b 64 SSE42_ - Gull 3 x64 XP 81.0 - 39.0 +54/=54/-12 67.50%

Time Control 2 + 2

Stockfish 230415b 64 SSE42_ - Komodo 9.01 64-bit_X6 59.0 - 61.0 +17/=84/-19 49.17%
Stockfish 230415b 64 SSE42_ - Houdini 4 x64_st_X6_CT0 72.5 - 47.5 +38/=69/-13 60.42%
Stockfish 230415b 64 SSE42_ - Gull 3 x64 XP 79.0 - 41.0 +48/=62/-10 65.83%
Score:

720 Games: 434.5 – 285.5 = 60.35%
http://www.mediafire.com/download/d3...t_720Games.pgn

8 real cores

Ponder: Off
GUI: Fritz 14
Book: Perfect 2014t – Limit 8 moves. Sedat –
No tablebases. No RTB used.
Hash 1024
Relative Speed: 44.92
Knodes per Second: 21.562

Time Control: 4+0

Stockfish 230415b 64 SSE42_ - Komodo 9.01 64-bit-X8 61.5 - 58.5 +21/=81/-18 51.25%
Stockfish 230415b 64 SSE42_ - Houdini 4 Pro x64_Ct0_8 77.0 - 43.0 +45/=64/-11 64.17%
Stockfish 230415b 64 SSE42_ - Gull 3 x64 XP_8 86.0 - 34.0 +56/=60/-4 71.67%

Time Control 2 + 2

Stockfish 230415b 64 SSE42_ - Komodo 9.01 64-bit-X8 60.0 - 60.0 +22/=76/-22 50.00%
Stockfish 230415b 64 SSE42_ - Houdini 4 Pro x64_Ct0_8 73.5 - 46.5 +33/=81/-6 61.25%
Stockfish 230415b 64 SSE42_ - Gull 3 x64 XP_8 82.0 - 38.0 +49/=66/-5 68.33%

Score: 440.0 – 280.0 = 61.11%
720 Games:
http://www.mediafire.com/download/1r...t_720Games.pgn

GLOBAL SCORE: 874.5 – 565.5 = 60.73%

Against : Komodo 9.01 (3219) = 50.83% Houdini 4 (3136) = 63.02%, Gull 3 XP (3103) = 68.33%

Average Estimated Elo Opponents = 3153
Estimated Elo Performance for Stockfish 230415 Bullet in this test of 1440 Games= 3228
Estimated Elo Performance for Stockfish 230415 Bullet after 2.880 Games= 3233


After such intensive testing Stockfish 230415 Bullet compile has performed around:

15 Elo points better than Stockfish6 (3218)
14 Elo points better than Komodo 9.01 (3219) and
9 Elo points better than the latest and best SF Development Version I tested (180515 - Timestamp: 1431982770 Bench: 8098149) Elo 3224

Error bars= +/- 9.

Regards,

Tom.
Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 9:09 pm

Re: For Tom- or Maybe Anyone Who Knows!!

Post by Tomcass »

TESTING STOCKFISH 050615 MZ= 1440 GAMES

6 real cores

Ponder: Off.
GUI: Fritz 14
Book: Perfect 2014c – Limit 5 moves. Sedat –
No tablebases. No RTB used.
Hash 512
Relative Speed: 29.54
Knodes per second: 14.177

Time Control 4 + 0

SF050615MZ 64 POPCNT_6 - Komodo 9.02 64-bitx6 23.0 - 17.0 +11/=24/-5 57.50%
SF050615MZ 64 POPCNT_6 - Houdini 4 x64_st_X6_CT0 29.5 - 10.5 +21/=17/-2 73.75%
SF050615MZ 64 POPCNT_6 - Gull 3 x64 XP 30.5 - 9.5 +23/=15/-2 76.25%

SF050615MZ 64 POPCNT_6 - Komodo 9.02 64-bitx6 38.0 - 42.0 +12/=52/-16 47.50%
SF050615MZ 64 POPCNT_6 - Houdini 4 x64_st_X6_CT0 51.5 - 28.5 +30/=43/-7 64.38%
SF050615MZ 64 POPCNT_6 - Gull 3 x64 XP 54.0 - 26.0 +31/=46/-3 67.50%

Time Control 2 + 2

SF050615MZ 64 POPCNT_6 - Komodo 9.02 64-bitx6 23.5 - 16.5 +11/=25/-4 58.75%
SF050615MZ 64 POPCNT_6 - Houdini 4 x64_st_X6_CT0 23.5 - 16.5 +12/=23/-5 58.75%
SF050615MZ 64 POPCNT_6 - Gull 3 x64 XP 28.0 - 12.0 +18/=20/-2 70.00%

SF050615MZ 64 POPCNT_6 - Komodo 9.02 64-bitx6 44.5 - 35.5 +19/=51/-10 55.63%
SF050615MZ 64 POPCNT_6 - Houdini 4 x64_st_X6_CT0 51.5 - 28.5 +31/=41/-8 64.38%
SF050615MZ 64 POPCNT_6 - Gull 3 x64 XP 55.5 - 24.5 +36/=39/-5 69.38%

Score using 6 cores: 453.0 – 267.0 = 62.92%
720 Games:
http://www.mediafire.com/download/xz..._720_Games.pgn

8 real cores

Ponder: Off
GUI: Fritz 14
Book: Perfect 2014c – Limit 5 moves. Sedat –
No tablebases. No RTB used.
Hash 1024
Relative Speed: 44.92
Knodes per Second: 21.562

Time Control: 4+0

SF050615MZ 64 AVX_8 - Komodo 9.02 64-bitx8 20.5 - 19.5 +4/=33/-3 51.25%
SF050615MZ 64 AVX_8 - Houdini 4 Pro x64_Ct0_8 28.5 - 11.5 +20/=17/-3 71.25%
SF050615MZ 64 AVX_8 - Gull 3 x64 XPx8 28.0 - 12.0 +17/=22/-1 70.00%

SF050615MZ 64 AVX_8 - Komodo 9.02 64-bitx8 38.5 - 41.5 +8/=61/-11 48.13%
SF050615MZ 64 AVX_8 - Houdini 4 Pro x64_Ct0_8 53.5 - 26.5 +30/=47/-3 66.88%
SF050615MZ 64 AVX_8 - Gull 3 x64 XPx8 51.5 - 28.5 +26/=51/-3 64.38%

Time Control 2 + 2
SF50615MZ 64 AVX_8 - Komodo 9.02 64-bitx8 23.0 - 17.0 +9/=28/-3 57.50%
SF050615MZ 64 AVX_8 - Houdini 4 Pro x64_Ct0_8 27.5 - 12.5 +16/=23/-1 68.75%
SF050615MZ 64 AVX_8 - Gull 3 x64 XPx8 29.0 - 11.0 +19/=20/-1 72.50%

SF050615MZ 64 AVX_8 - Komodo 9.02 64-bitx8 39.0 - 41.0 +13/=52/-15 48.75%
SF050615MZ 64 AVX_8 - Houdini 4 Pro x64_Ct0_8 54.5 - 25.5 +31/=47/-2 68.13%
SF050615MZ 64 AVX_8 - Gull 3 x64 XPx8 54.0 - 26.0 +30/=48/-2 67.50%

Score: 447.5 – 272.5 = 62.15%
240 Games: http://www.mediafire.com/download/sz..._240_Games.pgn
480 Games: http://www.mediafire.com/download/mp..._480_Games.pgn

GLOBAL SCORE: 900.5 – 539.5 = 62.53%
Against : Komodo 9.02 (3211) = 52.09% Houdini 4 (3136) = 66.67%, Gull 3 XP (3103) = 68.86%

Average Estimated Elo Opponents = 3150
Estimated Elo performance: 3238

Error Bars: +/- 13

This is a new best score ever in my tests, 5 elo points better than the previous leader:
Estimated Elo Performance for Stockfish 230415 Bullet after 2.880 Games= 3233

Congratulations to Marco Zerbinati and to the Stockfish Team!.

Regards from Barcelona.

Tom.
Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 9:09 pm

Re: For Tom- or Maybe Anyone Who Knows!!

Post by Tomcass »

TESTING STOCKFISH DEVELOPMENT 270615 = 480 GAMES

Timestamp: 1435394759 Bench: 8646407

6 real cores

Ponder: Off.
GUI: Fritz 14
Book: Perfect 2014c – Limit 5 moves. Sedat –
No tablebases. No RTB used.
Hash 512
Relative Speed: 29.54
Knodes per second: 14.177

Time Control 4 + 0

Stockfish 270615 64 POPCNTx6 - Komodo 9.1 64-bit_x6 20.5 - 19.5 +8/=25/-7 51.25%
Stockfish 270615 64 POPCNTx6 - Houdini 4 x64_st_X6_CT0 22.5 - 17.5 +12/=21/-7 56.25%
Stockfish 270615 64 POPCNTx6 - Gull 3 x64 XP 23.0 - 17.0 +9/=28/-3 57.50%

Time Control 2 + 2

Stockfish 270615 64 POPCNTx6 - Komodo 9.1 64-bit_x6 19.5 - 20.5 +4/=31/-5 48.75%
Stockfish 270615 64 POPCNTx6 - Houdini 4 x64_st_X6_CT0 28.5 - 11.5 +20/=17/-3 71.25%
Stockfish 270615 64 POPCNTx6 - Gull 3 x64 XP 27.5 - 12.5 +17/=21/-2 68.75%

Score using 6 cores: 141.5 – 98.5 = 58.96%

240 Games:
http://www.mediafire.com/download/z3r3g ... mes_X6.pgn

8 real cores

Ponder: Off
GUI: Fritz 14
Book: Perfect 2014c – Limit 5 moves. Sedat –
No tablebases. No RTB used.
Hash 1024
Relative Speed: 44.92
Knodes per Second: 21.562

Time Control: 4+0

Stockfish 270615 64 BMI2x8 - Komodo 9.1 64-bit_x8 22.0 - 18.0 +9/=26/-5 55.00%
Stockfish 270615 64 BMI2x8 - Houdini 4 Pro x64_Ct0_8 28.5 - 11.5 +18/=21/-1 71.25%
Stockfish 270615 64 BMI2x8 - Gull 3 x64 XPx8 31.0 - 9.0 +24/=14/-2 77.50%

Time Control 2 + 2

Stockfish 270615 64 BMI2x8 - Komodo 9.1 64-bit_x8 18.5 - 21.5 +4/=29/-7 46.25%
Stockfish 270615 64 BMI2x8 - Houdini 4 Pro x64_Ct0_8 27.5 - 12.5 +16/=23/-1 68.75%
Stockfish 270615 64 BMI2x8 - Gull 3 x64 XPx8 27.0 - 13.0 +16/=22/-2 67.50%

Score using 8 cores: 154.5 – 85.5 = 64.37%

240 Games:
http://www.mediafire.com/download/skkgu ... _Games.pgn

GLOBAL SCORE: 296.0 – 184.0= 61.67%

Against : Komodo 9.1 (3247) = 50.31% Houdini 4 (3136) = 66.88%, Gull 3 XP (3103) = 67.81 %

Average Elo of Oponents= 3.162
Estimated Elo Performance for Stockfish Development 270615 after 480 games= 3.244


Error bars: +/-23 Elo points.

This is a great score for SF Development 270615. Till now the best SF Development version in my tests was 180515, 3219 Elo. Therefore I have found an improvement of 25 Elo points in just 40 days (!!). I will continue this test until 1440 games to confirm -or not- this amazing performance, only 3 Elo points below the current leader Komodo 9.1.

Regards,

Tom.
Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 9:09 pm

Re: For Tom- or Maybe Anyone Who Knows!!

Post by Tomcass »

TESTING STOCKFISH DEVELOPMENT 270615 = 1440 GAMES. THE TEST IS FINISHED!

Timestamp: 1435394759 Bench: 8646407

6 real cores

Ponder: Off.
GUI: Fritz 14
Book: Perfect 2014c – Limit 5 moves. Sedat –
No tablebases. No RTB used.
Hash 512
Relative Speed: 29.54
Knodes per second: 14.177

Time Control 4 + 0

Stockfish 270615 64 POPCNTx6 - Komodo 9.1 64-bit_x6 20.5 - 19.5 +8/=25/-7 51.25%
Stockfish 270615 64 POPCNTx6 - Houdini 4 x64_st_X6_CT0 22.5 - 17.5 +12/=21/-7 56.25%
Stockfish 270615 64 POPCNTx6 - Gull 3 x64 XP 23.0 - 17.0 +9/=28/-3 57.50%

Stockfish 270615 64 POPCNTx6 - Komodo 9.1 64-bit_x6 42.5 - 37.5 +16/=53/-11 53.13%
Stockfish 270615 64 POPCNTx6 - Houdini 4 x64_st_X6_CT0 55.5 - 24.5 +37/=37/-6 69.38%
Stockfish 270615 64 POPCNTx6 - Gull 3 x64 XP 58.5 - 21.5 +40/=37/-3 73.13%


Time Control 2 + 2

Stockfish 270615 64 POPCNTx6 - Komodo 9.1 64-bit_x6 19.5 - 20.5 +4/=31/-5 48.75%
Stockfish 270615 64 POPCNTx6 - Houdini 4 x64_st_X6_CT0 28.5 - 11.5 +20/=17/-3 71.25%
Stockfish 270615 64 POPCNTx6 - Gull 3 x64 XP 27.5 - 12.5 +17/=21/-2 68.75%

Stockfish 270615 64 POPCNTx6 - Komodo 9.1 64-bit_x6 45.0 - 35.0 +17/=56/-7 56.25%
Stockfish 270615 64 POPCNTx6 - Houdini 4 x64_st_X6_CT0 54.5 - 25.5 +32/=45/-3 68.13%
Stockfish 270615 64 POPCNTx6 - Gull 3 x64 XP 48.0 - 32.0 +21/=54/-5 60.00%

Score using 6 cores: 445.5 – 274.5 = 61.87%

240 Games:
http://www.mediafire.com/download/z3r3g ... mes_X6.pgn

480 Games:
http://www.mediafire.com/download/g96w6 ... Leg_X6.pgn


8 real cores

Ponder: Off
GUI: Fritz 14
Book: Perfect 2014c – Limit 5 moves. Sedat –
No tablebases. No RTB used.
Hash 1024
Relative Speed: 44.92
Knodes per Second: 21.562

Time Control: 4+0
Stockfish 270615 64 BMI2x8 - Komodo 9.1 64-bit_x8 22.0 - 18.0 +9/=26/-5 55.00%
Stockfish 270615 64 BMI2x8 - Houdini 4 Pro x64_Ct0_8 28.5 - 11.5 +18/=21/-1 71.25%
Stockfish 270615 64 BMI2x8 - Gull 3 x64 XPx8 31.0 - 9.0 +24/=14/-2 77.50%

Stockfish 270615 64 BMI2x8 - Komodo 9.1 64-bit_x8 41.0 - 39.0 +13/=56/-11 51.25%
Stockfish 270615 64 BMI2x8 - Houdini 4 Pro x64_Ct0_8 52.0 - 28.0 +28/=48/-4 65.00%
Stockfish 270615 64 BMI2x8 - Gull 3 x64 XPx8 53.5 - 26.5 +30/=47/-3 66.88%

Time Control 2 + 2

Stockfish 270615 64 BMI2x8 - Komodo 9.1 64-bit_x8 18.5 - 21.5 +4/=29/-7 46.25%
Stockfish 270615 64 BMI2x8 - Houdini 4 Pro x64_Ct0_8 27.5 - 12.5 +16/=23/-1 68.75%
Stockfish 270615 64 BMI2x8 - Gull 3 x64 XPx8 27.0 - 13.0 +16/=22/-2 67.50%

Stockfish 270615 64 BMI2x8 - Komodo 9.1 64-bit_x8 39.0 - 41.0 +9/=60/-11 48.75%
Stockfish 270615 64 BMI2x8 - Houdini 4 Pro x64_Ct0_8 52.5 - 27.5 +29/=47/-4 65.63%
Stockfish 270615 64 BMI2x8 - Gull 3 x64 XPx8 48.5 - 31.5 +22/=53/-5 60.63%

Score using 8 cores: 441.0 – 279.0 = 61.25%

240 Games:
http://www.mediafire.com/download/skkgu ... _Games.pgn

480 Games:
http://www.mediafire.com/download/kn9gq ... &3Legs.pgn

GLOBAL SCORE: 886.5 – 553.5= 61.56%

Against : Komodo 9.1 (3247) = 51.67% Houdini 4 (3136) = 66.98%, Gull 3 XP (3103) = 66.04 %

Average Elo of Oponents= 3.162

Estimated Elo Performance for Stockfish Development 270615 after 480 games= 3.243


Error bars: +/-13 Elo points.

After 1440 and with just +/-13 Elo error bars, Stockfish Development 270615 has fully confirmed its substantial improvement over the best SF Development scorer so far: SF Dev. 180515, 3219 Elo.

Something important has happened between both versions. 24 Elo points in just 40 days is a big improvement.

Stockfish Development 270615 has scored:

- Just 4 Elo points below the robust leader Komodo 9.1 (Although SF has won the direct match against the leader with 51.67%)
- 24 Elo better than Stockfish 6
- 34 Elo better than Komodo 9
- 90 Elo better than Komodo 8
- 107 Elo better than Houdini 4
- 140 Elo better than Gull 3 XP.

Congratulations to the Stockfish Team!

Regards,

Tom.
JJJ
Posts: 1346
Joined: Sat Apr 19, 2014 1:47 pm

Re: For Tom- or Maybe Anyone Who Knows!!

Post by JJJ »

Stockfish improve his score a lot without increment. Stockfish didn't win vs Komodo 9.1 with time increment. The average of both is meaningless for me.
Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 9:09 pm

Re: For Tom- or Maybe Anyone Who Knows!!

Post by Tomcass »

TESTING STOCKFISH DEVELOPMENT 160715 = 1440 GAMES

Timestamp: 1437027460 Bench: 6943812

FIRST LEG: 720 GAMES AT FIXED TIME CONTROL 4+0.

6 real cores

Ponder: Off.
GUI: Fritz 14
Book: Perfect 2014c – Limit 5 moves. Sedat –
No tablebases. No RTB used.
Hash 512
Relative Speed: 29.54
Knodes per second: 14.177

Time Control 4 + 0

Stockfish 160715 64 POPCNTx6 - Komodo 9.1 64-bit_x6 58.0 - 62.0 +17/=82/-21 48.33%
Stockfish 160715 64 POPCNTx6 - Houdini 4 x64_st_X6_CT0 77.5 - 42.5 +41/=73/-6 64.58%
Stockfish 160715 64 POPCNTx6 - Gull 3 x64 XP 83.0 - 37.0 +51/=64/-5 69.17%
360 Games:
http://www.mediafire.com/download/f65xksfaxmv6d16


8 real cores

Ponder: Off
GUI: Fritz 14
Book: Perfect 2014c – Limit 5 moves. Sedat –
No tablebases. No RTB used.
Hash 1024
Relative Speed: 44.92
Knodes per Second: 21.562

Time Control: 4+0

Stockfish 160715 64 BMI2x8 - Komodo 9.1 64-bit_x8 58.0 - 62.0 +18/=80/-22 48.33%
Stockfish 160715 64 BMI2x8 - Houdini 4 Pro x64_Ct0_8 77.0 - 43.0 +45/=64/-11 64.17%
Stockfish 160715 64 BMI2x8 - Gull 3 x64 XPx8 78.0 - 42.0 +44/=68/-8 65.00%

360 games:
http://www.mediafire.com/download/fs4081ex89s3k8m


GLOBAL SCORE AFTER 720 GAMES AT FIXED TIME CONTROL: 431.5 - 288.5= 59.93%

Average Elo of Oponents= 3.162

Estimated Elo Performance for Stockfish Development 160715 after 720 games= 3.232


As a reference, the score of Stockfish Development 270615 (Timestamp: 1435394759 Bench: 8646407) after 720 games at Fixed Time Control 4+0 was 450.5 – 269.5 = 62.57% . And its estimated Elo Score was: 3.250. It seems to be some regression in this version of SF. Development, beyond the error bars. Has any other tester noticed this?.

Let’s see what happen at incremental Time Control 2+2 after 720 games. Just started.

Regards,

Tom.
voyagerOne
Posts: 154
Joined: Tue May 17, 2011 8:12 pm

Re: For Tom- or Maybe Anyone Who Knows!!

Post by voyagerOne »