Testing Houdini 4.0 Pro B: 480 games.

Discussion of computer chess matches and engine tournaments.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
ouachita
Posts: 454
Joined: Tue Jan 15, 2013 3:33 pm
Location: Ritz-Carlton, NYC
Full name: Bobby Johnson

Re: Testing Houdini 4.0 Pro B: 480 games.

Post by ouachita » Fri Nov 29, 2013 2:39 pm

where can I find the Perfect 2012c book?
SIM, PhD, MBA, PE

Adam Hair
Posts: 3222
Joined: Wed May 06, 2009 8:31 pm
Location: Fuquay-Varina, North Carolina

Re: Testing Houdini 4.0 Pro B: 480 games.

Post by Adam Hair » Fri Nov 29, 2013 3:01 pm

The link to the book can be found here:
http://www.sedatcanbaz.com/chess/?page_id=209

Nelson Hernandez
Posts: 101
Joined: Sun Nov 14, 2010 8:36 pm
Location: U.S.

Re: Testing Houdini 4.0 Pro B: 480 games.

Post by Nelson Hernandez » Fri Nov 29, 2013 11:34 pm

To be honest I think that most of them are useless or obsolete.
How wrong you are, Tom! One man's trash is another man's treasure!

I and I am sure others would be happy to have your 110,000 games, and we would be perfectly willing to give you step by step help in zipping and uploading them.

Games provide data points. Games played a decade ago by computers still turn up occasionally as key games in opening theory as more expansively defined.

Drop me a line at nhernandezsr@hotmail.com and I'll help you ascend the learning curve very quickly. It's all quite easy.

Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 7:09 pm

Re: Testing Houdini 4.0 Pro B: 480 games.

Post by Tomcass » Sun Dec 01, 2013 6:32 pm

Thanks for your kind words, Nelson. I will contact you privately to try to get your help to open my games to anyone interested. By the way, I have been checking it and the number of games is substantially lower: Between 80.000 and 85.000 games, if I am not wrong.

Regards,

Tom.

Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 7:09 pm

Re: Testing Houdini 4.0 Pro B: 480 games.

Post by Tomcass » Sun Dec 01, 2013 6:35 pm

TESTING HOUDINI 4.0 PRO B = 480 GAMES _ 2nd leg.

i7 980 3.33 Ghz.
6 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012c
No tablebases. No RTB used.
Hash 512
Relative Speed: 28.66
Knodes per second: 13.759

Time Control= 4+0.

Houdini 4 Pro x64x6 - Critter 1.6a 64-bitX6_NOB_ok 27.5 - 12.5 +15/=25/-0 68.75%
Houdini 4 Pro x64x6 - Stockfish 191113 64 SSE4.2x 18.0 - 22.0 +10/=16/-14 45.00%
Houdini 4 Pro x64x6 - Deep Rybka 4.1 SSE42 x64 (x6) 27.0 - 13.0 +21/=12/-7 67.50%

Time Control= 2+2

201311Houdini4.0ProB_2+2_120games_2nd 2013

Houdini 4 Pro x64x6 - Critter 1.6a 64-bitX6_NOB_ok 29.0 - 11.0 +19/=20/-1 72.50%
Houdini 4 Pro x64x6 - Stockfish 191113 64 SSE4.2x 19.5 - 20.5 +10/=19/-11 48.75%
Houdini 4 Pro x64x6 - Deep Rybka 4.1 SSE42 x64 (x6) 29.0 - 11.0 +21/=16/-3 72.50%

Games (240) : http://www.mediafire.com/download/7mju3xqsh7l7y2z/
Score using 6 cores= 150.0 – 90.0 = 62.50%

i7 975 3.33 Ghz.
4 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012c
No tablebases. No RTB used.
Hash 256
Relative Speed: 20.62
Knodes per second: 9.899

Time Control= 4+0

Houdini 4 Pro x64x4 - Critter 1.6a 64-bitnob_4 27.5 - 12.5 +17/=21/-2 68.75%
Houdini 4 Pro x64x4 - Stockfish 191113 64 SSE4.2x 23.5 - 16.5 +14/=19/-7 58.75%
Houdini 4 Pro x64x4 - Deep Rybka 4 x64 (x4) 30.0 - 10.0 +23/=14/-3 75.00%

Time Control= 2+2

Houdini 4 Pro x64x4 - Critter 1.6a 64-bitnob_4 28.0 - 12.0 +19/=18/-3 70.00%
Houdini 4 Pro x64x4 - Stockfish 191113 64 SSE4.2x 19.0 - 21.0 +10/=18/-12 47.50%
Houdini 4 Pro x64x4 - Deep Rybka 4 x64 (x4) 28.5 - 11.5 +22/=13/-5 71.25%


Games (240)= http://www.mediafire.com/view/1271nr...0games_2nd.pgn
Score using 4 cores= 156.5 – 83.5 = 65.21%

Fixed = 130.5 – 109.5 = 54.37%
Incremental = 134.0 -106.0 = 55.83%

Global Score: 306.5 – 173.5 = 63.85%
Against Critter 1.6 (3093): % Stockfisk 191113 (3204): 60.00% Deep Rybka 4 (3016) 47.81%

Average ELO of opponents= 3.104
Estimated ELO Performance= 3.201


-------------------------------

Only as a reminder, the score in the first 480 games leg was:

Global Score: 264.5-215.5 = 55.10%
Against Houdini 3.0 Pro (3172): 57.50% Komodo 6 (3154): 60.00% Stockfish 111113SL Ipman (3224) 47.81%

Average ELO of opponents= 3.183
Estimated ELO Performance= 3.219

Total Global Score after 960 Games: 571.0 – 389.0 = 59.48%
Average Elo of opponents = 3.143
Estimated Elo Performance for Houdini 4.0 Pro B = 3.210


Even after 960 games I can only give an approach to the Houdini 4.0Pro B strength. Under my testing conditions and using 4 and 6 cores computers:

- The improvement of Houdini 4.0 Pro over Houdini 3.0 Pro is substantial: Around 38 Estimated Elo Points (EEP).
- Houdini 4.0 Pro is about 56 EEP ahead of Komodo 6
- Houdini 4.0 Pro is about 14 EEP below the SF 11113SL compile by Ipman.
- Houdini 4.0 Pro is about 91 EEP ahead of Stockfish 4.
- Houdini 4.0 Pro is about 6 EEP ahead of the latest Stockfish development I have tested: 191113. (Within error bars).

Regards from Barcelona,

Tom.

fauzi
Posts: 61
Joined: Wed Nov 20, 2013 9:42 am

Re: Testing Houdini 4.0 Pro B: 480 games.

Post by fauzi » Sun Dec 01, 2013 11:05 pm

the second link contains dots in between and its not complete, so it doesn't work can You please fix it Tom.

User avatar
Ajedrecista
Posts: 1473
Joined: Wed Jul 13, 2011 7:04 pm
Location: Madrid, Spain.
Contact:

Re: Testing Houdini 4.0 Pro B: 480 games.

Post by Ajedrecista » Mon Dec 02, 2013 8:04 am

Hello:
fauzi wrote:the second link contains dots in between and its not complete, so it doesn't work can You please fix it Tom.
The solution was in post #75 of this ImmortalChess Forum thread:

http://immortalchess.net/forum/showthread.php?p=436645

Here is the full link:

http://www.mediafire.com/view/1271nrjw5 ... es_2nd.pgn

But surely you want to download the PGN instead of view it online:

http://www.mediafire.com/download/1271n ... es_2nd.pgn

The size of this PGN file is circa 1.64 MB.

Thank you very much for your tests, Tom. Great job as usual.

Regards from Spain.

Ajedrecista.

Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 7:09 pm

Re: Testing Houdini 4.0 Pro B: 480 games.

Post by Tomcass » Mon Dec 02, 2013 8:41 am

Ajedrecista wrote:Hello:
fauzi wrote:the second link contains dots in between and its not complete, so it doesn't work can You please fix it Tom.
The solution was in post #75 of this ImmortalChess Forum thread:

http://immortalchess.net/forum/showthread.php?p=436645

Here is the full link:

http://www.mediafire.com/view/1271nrjw5 ... es_2nd.pgn

But surely you want to download the PGN instead of view it online:

http://www.mediafire.com/download/1271n ... es_2nd.pgn

The size of this PGN file is circa 1.64 MB.

Thank you very much for your tests, Tom. Great job as usual.

Regards from Spain.

Ajedrecista.
Thank you very much Jesús for clarifying it and for your kind words. :wink:

Un abrazo!.

Tom.

Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 7:09 pm

Re: Testing Houdini 4.0 Pro B: 480 games.

Post by Tomcass » Fri Dec 13, 2013 11:49 am

TESTING Houdini 4.0 StandardA SD14 Ct0= 480 GAMES

A friend of mine suggested that it might be useful to test Houdini 4.0 with contempt 0 and SD 14. Here you have the results:

i7 980 3.33 Ghz.
6 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012c
No tablebases. No RTB used.
Hash 512
Relative Speed: 28.66
Knodes per second: 13.759

Time Control= 4+0.

Houdini 4 x64_StA_SD14_CT0x6 - Stockwood 061213 18.5 - 21.5 +4/=29/-7 46.25%
Houdini 4 x64_StA_SD14_CT0x6 - Stockfish 071213 64 20.0 - 20.0 +6/=28/-6 50.00%
Houdini 4 x64_StA_SD14_CT0x6 - Stockfish 091213SL 20.0 - 20.0 +8/=24/-8 50.00%

Time Control= 2+2

Houdini 4 x64_StA_SD14_CT0x6 - Stockwood 061213 19.5 - 20.5 +5/=29/-6 48.75%
Houdini 4 x64_StA_SD14_CT0x6 - Stockfish 091213SL 19.5 - 20.5 +6/=27/-7 48.75%
Houdini 4 x64_StA_SD14_CT0x6 - Stockfish 071213 6 17.5 - 22.5 +6/=23/-11 43.75%

Score using 6 cores= 115.0 – 125.0 = 47.92%
240 Games= http://www.mediafire.com/view/o9j4hf403 ... 0games.pgn


i7 975 3.33 Ghz.
4 real cores
Ponder: Off.
GUI: Fritz 12
Book: Perfect 2012c
No tablebases. No RTB used.
Hash 256
Relative Speed: 20.62
Knodes per second: 9.899

Time Control= 4+0

Houdini 4 x64_StA_SD14_Ct0x4 - Stockwood 061213 22.5 - 17.5 +10/=25/-5 56.25%
Houdini 4 x64_StA_SD14_Ct0x4 - Stockfish 071213 64 21.0 - 19.0 +13/=16/-11 52.50%
Houdini 4 x64_StA_SD14_Ct0x4 - Stockfish 091213SL 20.0 - 20.0 +9/=22/-9 50.00%

Time Control= 2+2

Houdini 4 x64_StA_SD14_Ct0x4 - Stockwood 061213 15.5 - 24.5 +5/=21/-14 38.75%
Houdini 4 x64_StA_SD14_Ct0x4 - Stockfish 071213 x 21.0 - 19.0 +9/=24/-7 52.50%
Houdini 4 x64_StA_SD14_Ct0x4 - Stockfish 091213SL 20.0 - 20.0 +2/=36/-2 50.00%

Score using 4 cores = 120.0 – 120.0
240 Games= http://www.mediafire.com/download/oclhopjti54so7w/

Fixed = 122.0 – 118.0 = 50.83%
Incremental = 113.0 – 127.0 = 47.08%

Global Score: 235.0-245.0 = 48.96%

In my computers Houdini 4.0 Standard Contempt 0 SD 10 is the best version I have tested till now. The same version with SD 14, tested here, is not so strong.

Regards,

Tom.

Tomcass
Posts: 786
Joined: Sun Apr 16, 2006 7:09 pm

Re: Testing Houdini 4.0 Pro B: 480 games.

Post by Tomcass » Mon Dec 05, 2016 6:34 pm

TESTING HOUDINI 5 PRO: 1.200 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

Fixed Time Control= 4+0
Houdini 5 Pro x64-popc_X6 - SF8 ASMPD_X6 26.5 - 33.5 +5/=43/-12 44.17%
Houdini 5 Pro x64-popc_X6 - Komodo 10.2 64-bit_X6_NOB 37.0 - 23.0 +18/=38/-4 61.67%
Houdini 5 Pro x64-popc_X6 - SugaR 2.6 64 POPCNT_X6 32.5 - 27.5 +10/=45/-5 54.17%
Houdini 5 Pro x64-popc_X6 - Houdini 4 Pro x64_St_Ct0_X6 47.5 - 12.5 +37/=21/-2 79.17%
Houdini 5 Pro x64-popc_X6 - Gull 3 x64 XP_X6 45.5 - 14.5 +34/=23/-3 75.83%

Incremental Time Control= 2+2
Houdini 5 Pro x64-popc_X6 - SF8 ASMPD_X6 29.5 - 30.5 +7/=45/-8 49.17%
Houdini 5 Pro x64-popc_X6 - Komodo 10.2 64-bit_X6_NOB 35.0 - 25.0 +14/=42/-4 58.33%
Houdini 5 Pro x64-popc_X6 - SugaR 2.6 64 POPCNT_X6 31.0 - 29.0 +6/=50/-4 51.67%
Houdini 5 Pro x64-popc_X6 - Houdini 4 Pro x64_St_Ct0_X6 45.0 - 15.0 +31/=28/-1 75.00%
Houdini 5 Pro x64-popc_X6 - Gull 3 x64 XP_X6 46.0 - 14.0 +32/=28/-0 76.67%

600 Games= http://www.mediafire.com/file/hx4ae3g3q0h37yd/
Score using 6 cores= 375.5 – 224.5 = 62.58%

8 real cores
Ponder: Off
GUI: Fritz 14
Book: Perfect 2014c – Limit 5 moves. Sedat –
No tablebases. No RTB used.
Hash 1024
Relative Speed: 53.22
Knodes per Second: 25.543

Fixed Time Control= 4+0
Houdini 5 Pro x64-pext_X8 - Komodo 10.2 64-bit_X8_NOB 35.0 - 25.0 +17/=36/-7 58.33%
Houdini 5 Pro x64-pext_X8 - SF8 ASMPD_X8 28.5 - 31.5 +7/=43/-10 47.50%
Houdini 5 Pro x64-pext_X8 - SugaR 2.6 64 BMI2_X8 31.0 - 29.0 +9/=44/-7 51.67%
Houdini 5 Pro x64-pext_X8 - Houdini 4 Pro x64_Ct0_8 46.0 - 14.0 +32/=28/-0 76.67%
Houdini 5 Pro x64-pext_X8 - Gull 3 x64 BMI2_X8 47.0 - 13.0 +36/=22/-2 78.33%

Incremental Time Control= 2+2
Houdini 5 Pro x64-pext_X8 - SF8 ASMPD_X8 28.5 - 31.5 +6/=45/-9 47.50%
Houdini 5 Pro x64-pext_X8 - Komodo 10.2 64-bit_X8_NOB 32.5 - 27.5 +9/=47/-4 54.17%
Houdini 5 Pro x64-pext_X8 - SugaR 2.6 64 BMI2_X8 32.0 - 28.0 +11/=42/-7 53.33%
Houdini 5 Pro x64-pext_X8 - Houdini 4 Pro x64_Ct0_8 49.0 - 11.0 +38/=22/-0 81.67%
Houdini 5 Pro x64-pext_X8 - Gull 3 x64 BMI2_X8 44.5 - 15.5 +30/=29/-1 74.17%

http://www.mediafire.com/file/663v3rppale6ws0/
Score using 8 cores= 374.0 – 226.0 = 62.33%
Global score after 1.200 games = 749.5 – 450.5 = 62.46%

Houdini 5 Pro. Against

SF8 ASMPD (3338) 49.79%
Komodo 10.2 (3277) 55.42%
Sugar2.6 (3291) 52.71%
Houdini 4 Pro (3136) 78.13%
Gull 3 (3103) 76.25%
----------
Average ELO of rivals= 3.229

Estimated ELO for Houdini 5 Pro after 1.200 games = 3.316

Only 22 ELO points below SF8 ASMPD. 180 ELO better than Houdini 4 Pro. What an incredible improvement! Congratulations to Robert Houdart. :D

Regards,

Tom.

Post Reply