Onno 1.0 is now available

[Gerd Isenberg]

A note on Onno's credits site. Tord Romstad has not invented magic bitboards or either magic multiplication. The inventor was Lasse Hansen, while Pradu Kannan came up with individual sized tables for each square. For further references see Magic Bitboards from cpw.

Thanks,
Gerd

[Onno Garms]

Tord Romstad has not invented magic bitboards or either magic multiplication. The inventor was Lasse Hansen, while Pradu Kannan came up with individual sized tables for each square.

Corrected. Thank you for pointing out.

I am not planning to scan all internet fora for comments on Onno, so better email me if you find a problem.

[Gerd Isenberg]

Tord Romstad has not invented magic bitboards or either magic multiplication. The inventor was Lasse Hansen, while Pradu Kannan came up with individual sized tables for each square.

Corrected. Thank you for pointing out.

I am not planning to scan all internet fora for comments on Onno, so better email me if you find a problem.

Thanks. Since your site is public, I thought a public statement here was appropriate. Still, credits for magic bitboards exclusively to Lasse and Pradu. I was somehow involved in the development with separate lines, but never thought mapping up to twelve occupied bits would work due to huge memory, until the both guys proved me wrong - specially with Java like arrays. Tord later did his own implementation with respect to find own magics by trying randoms.

Curious about the upcoming event from the Tilburg centre for Creative Computing, the Promotion of Dhr. F.M.H. Reul with the thesis: "New Architectures in Computer Chess", supervised by Prof.dr. H.J. van den Herik and Dr.ir. J.H.W.M. Uiterwijk, which apparently covers magic bitboards.

Cheers,
Gerd

[Mike S.]

http://tinyurl.com/mmm2ux

[mclane]

mike, you give a link that related onno with toga, right ?

do you want to say that onno is a toga clone ?

[Dr.Wael Deeb]

mike, you give a link that related onno with toga, right ?

do you want to say that onno is a toga clone ?

Again

[Eelco de Groot]

mike, you give a link that related onno with toga, right ?

do you want to say that onno is a toga clone ?

Again

Yes, well, exactly, not again .

Just a few points to the defence of Onno, I'm sure that there are better points:

♣ Main point, I think these kind of evaluation difference lists are of very little value to detect cloning. Because it is much too easy to manipulate the evaluation output for values not close to zero, if your intention is to obfuscate things. So any engine that seems to have tuned itself to another strong engine and shows similar outputs to this engine, I would rather trust not to have a manipulated output than suspect it is a clone.

♣ Onno is a bitboard engine so that is already a big difference compared to Toga.

♣ It has been under development for several years.

♣ It is however not multi-CPU capable, unlike Toga

♣ Onno 1.0 does according to Onno Garms not compile to faster .exes with the Intel compiler compared with MVSC, much unlike Toga

This does not scream Toga clone warnings to me...

Regards, Eelco

[Dr.Wael Deeb]

mike, you give a link that related onno with toga, right ?

do you want to say that onno is a toga clone ?

Again

Yes, well, exactly, not again .

Just a few points to the defence of Onno, I'm sure that there are better points:

♣ Main point, I think these kind of evaluation difference lists are of very little value to detect cloning. Because it is much too easy to manipulate the evaluation output for values not close to zero, if your intention is to obfuscate things. So any engine that seems to have tuned itself to another strong engine and shows similar outputs to this engine, I would rather trust not to have a manipulated output than suspect it is a clone.

♣ Onno is a bitboard engine so that is already a big difference compared to Toga.

♣ It has been under development for several years.

♣ It is however not multi-CPU capable, unlike Toga

♣ Onno 1.0 does according to Onno Garms not compile to faster .exes with the Intel compiler compared with MVSC, much unlike Toga

This does not scream Toga clone warnings to me...

Regards, Eelco

Thanks Eelco for your thoughts....much appreciated
Dr.D

[Ryan Benitez]

Yes, well, exactly, not again .

I am confidant that Onno is not a clone.

Just a few points to the defence of Onno, I'm sure that there are better points:

♣ Main point, I think these kind of evaluation difference lists are of very little value to detect cloning. Because it is much too easy to manipulate the evaluation output for values not close to zero, if your intention is to obfuscate things. So any engine that seems to have tuned itself to another strong engine and shows similar outputs to this engine, I would rather trust not to have a manipulated output than suspect it is a clone.

I agree

♣ Onno is a bitboard engine so that is already a big difference compared to Toga.

Trivial conversion that takes a matter of hours.

♣ It has been under development for several years.

Yes he has posted on well known forums for some time now.

♣ It is however not multi-CPU capable, unlike Toga

Someone would not sell an engine and call it SMP with the SMP Toga has. It is ok for Toga because people have the freedom to see what the code is and choose to accept it for what it is. It would not be a commercially acceptable solution though.

♣ Onno 1.0 does according to Onno Garms not compile to faster .exes with the Intel compiler compared with MVSC, much unlike Toga

Fruit also is not faster with Intel compiler. I don't know much about windows compilers though so it does not mean much to me.

This does not scream Toga clone warnings to me...

I agree, if it was a clone people would have come of with much more convincing evidence by now.
I see it as a promising engine in an otherwise stale field that could make some great gains over the years. Computer chess needs engines like Onno to come along.