Oliprog's Engines and Book Tests

[Hurnavich]

HI, 1536 hash is a strange setting 512, 1024, 2048 so on normal

[kramnik]

HI, 1536 hash is a strange setting 512 1028 2056 so on normal

Theorically in this case only 1024 MBs would be used, but years ago I saw different kilonodes per second between 16364 and 18GBs.

[ChessMan]

HI, 1536 hash is a strange setting 512 1028 2056 so on normal

Theorically in this case only 1024 MBs would be used, but years ago I saw different kilonodes per second between 16364 and 18GBs.

Stockfish can use 1536 MB hash without any problems.

[Oliprog]

HI, 1536 hash is a strange setting 512, 1024, 2048 so on normal

Hello Ian! Before, I was using 2048 mb but my pc started to fail and they told me that it was possibly consuming a lot of data, so I went down to 1536 mb. I really don't understand much, so I accept your suggestions, since without a doubt you friends on the forum know much more than me about this.
Greetings to all

[Skynet]

I really don't understand much

Hello my friend! Preferable hash usage - from 16 mb to 32mb = (16x2), 64, 128, 256, 512, 1024 and so on. Engine doesn't have a problem with using 1536 mb hash, it's just not preferable for tests. Also, there is no need to keep engine in the same folder with source code, source code is only for making/building engines, it has absolutely no effect/impact on the engine.

[Peter Grayson]

HI, 1536 hash is a strange setting 512 1028 2056 so on normal

Theorically in this case only 1024 MBs would be used, but years ago I saw different kilonodes per second between 16364 and 18GBs.

That was certainly the case previously but I recall reading a while ago that Stockfish hash capability had changed and that is confirmed in Windows Task Manager when values other than 2^n are used judging by the RAM allocation to the engine.

For example, allocating 3Gb hash to Stockfish 8 and Task Manager records 2Gb allocated but for Stockfish 9 and above Task manager indicates 3 Gb.

Stockfish hash.jpg

Peter

[kramnik]

With your PC configuration (average quad core) and your settings (5 minutes) 1024 should be the best. Take a look at the usage percentage while the engine is playing: if the average is around 40% you're good to go.

Small hashtable = faster engine = stronger.

High end computers use 4GB for analysis, supercomputers like the resenschieber's Stockfish 11 (64 cores) uses 8 GB.

HI, 1536 hash is a strange setting 512, 1024, 2048 so on normal

Hello Ian! Before, I was using 2048 mb but my pc started to fail and they told me that it was possibly consuming a lot of data, so I went down to 1536 mb. I really don't understand much, so I accept your suggestions, since without a doubt you friends on the forum know much more than me about this.
Greetings to all

[Oliprog]

GM Graham Banks bin vs. Goi 6.2.2

Processor AMD A10 – 9700 R7,10 Compute Cores 4C + 6G - 3.50 GHz - RAM 8.00 GB - 64-bit operating system - x64 processor - Windows 10 - 2048 mb hash – Fritz 16 GUI – opening book: GM Graham Banks bin vs. Goi 6.2.2 - final tablebases: Syzygy 3-4-5 pieces - both engines play with 4 threads and large pages

Book Settings

GM Graham Banks bin

Book Depth 200
Best Book Line Off

Goi 6.2.2

1.Tournament ON
2.Min games 0-40
3.variety of play 50%
4.influence of learn 100%
5.Learning strength 500%

DESKTOP-TORM2PL, Blitz 5min 0


1 Raubfisch X41d3._sl GM Graham +64 +23/=96/-1 59.17% 71.0/120
2 Raubfisch X41d3._sl Goi 6.2.2 -64 +1/=96/-23 40.83% 49.0/120

Congratulations JairoSawka!

Juegos en PGN http://www.mediafire.com/file/yaqgx8b4s1cniz8/GMGraham_Banks_bin_vs._Goi_6.2.2.pgn/file

[timetraveller]

1 Raubfisch X41d3._sl GM Graham +64 +23/=96/-1 59.17% 71.0/120
2 Raubfisch X41d3._sl Goi 6.2.2 -64 +1/=96/-23 40.83% 49.0/120

wow, quite clearly

[pottboy]

sorry if i ask here, but is it possible to load the book somewhere

dear ask

[kramnik]

Lol, quite an obsolete book. I guess Oliprog is indirectly asking for 6.2.3... Well, you'll be satisfied.

GM Graham Banks bin vs. Goi 6.2.2

Processor AMD A10 – 9700 R7,10 Compute Cores 4C + 6G - 3.50 GHz - RAM 8.00 GB - 64-bit operating system - x64 processor - Windows 10 - 2048 mb hash – Fritz 16 GUI – opening book: GM Graham Banks bin vs. Goi 6.2.2 - final tablebases: Syzygy 3-4-5 pieces - both engines play with 4 threads and large pages

Book Settings

GM Graham Banks bin

Book Depth 200
Best Book Line Off

Goi 6.2.2

1.Tournament ON
2.Min games 0-40
3.variety of play 50%
4.influence of learn 100%
5.Learning strength 500%

DESKTOP-TORM2PL, Blitz 5min 0


1 Raubfisch X41d3._sl GM Graham +64 +23/=96/-1 59.17% 71.0/120
2 Raubfisch X41d3._sl Goi 6.2.2 -64 +1/=96/-23 40.83% 49.0/120

Congratulations JairoSawka!

Juegos en PGN http://www.mediafire.com/file/yaqgx8b4s1cniz8/GMGraham_Banks_bin_vs._Goi_6.2.2.pgn/file

[Hurnavich]

Lol, quite an obsolete book. I guess Oliprog is indirectly asking for 6.2.3... Well, you'll be satisfied.

GM Graham Banks bin vs. Goi 6.2.2

Processor AMD A10 – 9700 R7,10 Compute Cores 4C + 6G - 3.50 GHz - RAM 8.00 GB - 64-bit operating system - x64 processor - Windows 10 - 2048 mb hash – Fritz 16 GUI – opening book: GM Graham Banks bin vs. Goi 6.2.2 - final tablebases: Syzygy 3-4-5 pieces - both engines play with 4 threads and large pages

Book Settings

GM Graham Banks bin

Book Depth 200
Best Book Line Off

Goi 6.2.2

1.Tournament ON
2.Min games 0-40
3.variety of play 50%
4.influence of learn 100%
5.Learning strength 500%

DESKTOP-TORM2PL, Blitz 5min 0


1 Raubfisch X41d3._sl GM Graham +64 +23/=96/-1 59.17% 71.0/120
2 Raubfisch X41d3._sl Goi 6.2.2 -64 +1/=96/-23 40.83% 49.0/120

Congratulations JairoSawka!

Juegos en PGN http://www.mediafire.com/file/yaqgx8b4s1cniz8/GMGraham_Banks_bin_vs._Goi_6.2.2.pgn/file

Not that bad if you disable 1...e6 i have had some good games with it when you kick the french opening to the curb

[kramnik]

Not that bad if you disable 1...e6 i have had some good games with it when you kick the french opening to the curb

6.2.3 answers ...e5. I have noticed some time ago that it was a Stockfish flaw giving a lower score for ...e6 and in general not giving a correct value for the first move. You know, workarounds. Besides I had Don to take inspiration from.

[JairoSawka]

Not that bad if you disable 1...e6 i have had some good games with it when you kick the french opening to the curb

6.2.3 answers ...e5. I have noticed some time ago that it was a Stockfish flaw giving a lower score for ...e6 and in general not giving a correct value for the first move. You know, workarounds. Besides I had Don to take inspiration from.

It depends on how it is tuned. in a couple of days i am going to make a match of 30 games between goblin and french defense E4 E6 , D4 - D6


30 games all of them will be a draw goblin won't be able to win any game.. in the case of goi, I would have tuuned that defense in a different way.

perhaps french defense is not used nowadays but it was used by nakamura and caruana.

e4 d5 Scandinavian defense now is being used again why? magnus Carlsen started using again. thats it.

when Magnus starts using french defense all the people will use french defense.

[Hurnavich]

Not that bad if you disable 1...e6 i have had some good games with it when you kick the french opening to the curb

6.2.3 answers ...e5. I have noticed some time ago that it was a Stockfish flaw giving a lower score for ...e6 and in general not giving a correct value for the first move. You know, workarounds. Besides I had Don to take inspiration from.

It depends on how it is tuned. in a couple of days i am going to make a match of 30 games between goblin and french defense E4 E6 , D4 - D6


30 games all of them will be a draw goblin won't be able to win any game.. in the case of goi, I would have tuuned that defense in a different way.

perhaps french defense is not used nowadays but it was used by nakamura and caruana.

e4 d5 Scandinavian defense now is being used again why? magnus Carlsen started using again. thats it.

when Magnus starts using french defense all the people will use french defense.

What goblin are you going to use an old one? i would prefer a longer test by Oliprog