Dorky 4.0 released

[mmcknight]

For the first time in over 7 years, there is a new version of Dorky. Now including Nalimov egtb and parallel search. Please download here: http://mmcknight.home.insightbb.com/

Enjoy!

Matt

[Dr.Wael Deeb]

For the first time in over 7 years, there is a new version of Dorky. Now including Nalimov egtb and parallel search. Please download here: http://mmcknight.home.insightbb.com/

Enjoy!

Matt

Thanks Matt,will be tested at once
Dr.D

[Graham Banks]

For the first time in over 7 years, there is a new version of Dorky. Now including Nalimov egtb and parallel search. Please download here: http://mmcknight.home.insightbb.com/

Enjoy!

Matt

Thanks Matt. Roughly how strong is this latest version?

Cheers,
Graham.

[SzG]

For the first time in over 7 years, there is a new version of Dorky. Now including Nalimov egtb and parallel search. Please download here: http://mmcknight.home.insightbb.com/

Enjoy!

Matt

Great news, Matt. Added to my testing queue.

Cheers,

[mmcknight]

Hi Graham,

Thanks for asking. The newest version tests at 150 ELO higher than version 3.3 which is the strongest of all the releases. It turns out my testing for 3.45 and 3.48 was not good. This value is with many games at short time controls 60+1 and 180+0 using only one CPU per engine. The new version is now aware of the conventional time control (like 40/5) whereas previous versions were not, so at that format the strength is even +10 ELO more. Also, since the newest version supports SMP, strength is even higher when using multiple cores.

Matt

[Dr.Wael Deeb]

Thanks Matt for the notification,all good now
Dr.D

[mmcknight]

I updated the binaries after the release because I had used a compiler flag that left some debug code in. The code wasn't being called but made for a slightly larger binary. Please redownload if you are going to be testing. I didn't want to change the version number because there are no code changes, just some code being removed by re-compile.

Thanks,
Matt

[Graham Banks]

Hi Graham,

Thanks for asking. The newest version tests at 150 ELO higher than version 3.3 which is the strongest of all the releases. It turns out my testing for 3.45 and 3.48 was not good. This value is with many games at short time controls 60+1 and 180+0 using only one CPU per engine. The new version is now aware of the conventional time control (like 40/5) whereas previous versions were not, so at that format the strength is even +10 ELO more. Also, since the newest version supports SMP, strength is even higher when using multiple cores.

Matt

Thanks Matt. Here's the logo I'll use.

[SzG]

Matt, I noticed that Dorky run on 2 cores always uses a core even if it's the opponent's turn. I hope it is not pondering as that would make test results unusable for the ranking list.

Cheers,

[mmcknight]

Gabor,

It's not pondering if you have given the 'easy' command, but the child thread is spin-waiting. I am going to fix this and release a new version, because I don't like burning CPU unless i'm searching chess trees. Thanks for the heads-up.

Matt