Naraku Chess 1.12

[Damir]

If it was not for Jury's curiosity to find out what make Rybka so strong, would we even have known, that Rybka is not original engine its author claimed it to be, after over 5 years of lies and deception. Would we even have known, that Rybka is based on Fruit&Crafty source, which makes it a 100 % clone and illegal to use in upcoming chess events, much less make profit from it. Yes, in your opinion, I am sure we would have, without Jury's help and findings... We would still be buying Fruit&Crafty clone, which is open source with search and evaluation tuned, but hardly any original...
If Mr Perkins is guilty of the same thing, why should he have the easy ride?
What I am saying is if Mr Perkins is guilty of illegally copy&pasting source of others he should be grilled, and Jury should post the lines of the code he copied for all to see. If there is nothing, than it is simple, Jury should post nothing, if the engine is original, but I very much doubt it is.

[Christopher Conkie]

If it was not for Jury's curiosity to find out what make Rybka so strong, would we even have known, that Rybka is not original engine its author claimed it to be, after over 5 years of lies and deception. Would we even have known, that Rybka is based on Fruit&Crafty source, which makes it a 100 % clone and illegal to use in upcoming chess events, much less make profit from it. Yes, in your opinion, I am sure we would have, without Jury's help and findings... We would still be buying Fruit&Crafty clone, which is open source with search and evaluation tuned, but hardly any original...
If Mr Perkins is guilty of the same thing, why should he have the easy ride?
What I am saying is if Mr Perkins is guilty of illegally copy&pasting source of others he should be grilled, and Jury should post the lines of the code he copied for all to see. If there is nothing, than it is simple, Jury should post nothing, if the engine is original, but I very much doubt it is.

I would like to point out to you that when Strelka first came out it was declared an original work by Juri. There was no "take a look at Rybka" from him.

That is not an opinion. It is a fact.

[Damir]

Fact is after Jury released Strelka Mr Rajlich claimed Strelka as his own, because according to him, it contained and was based on closed Rybka source, which has no validity whatsoever...

Here is a little riddle for you to solve: If Strelka is based on Rybka, and Strelka is based on Fruit ( from Fabien's words) cause he has taken a look on Strelka source and has compared its similarity to match Fruits, than what is Rybka ??

[Christopher Conkie]

Fact is after Jury released Strelka Mr Rajlich claimed Strelka as his own, because according to him, it contained and was based on closed Rybka source, which has no validity whatsoever...

Here is a little riddle for you to solve: If Strelka is based on Rybka, and Strelka is based on Fruit ( from Fabien's words) cause he has taken a look on Strelka source and has compared its similarity to match Fruits, than what is Rybka ??

Strelka was made by putting parts of disassembled Rybka into a Fruit framework according to........Juri. Little wonder then that Fabian looks at it and says that he can see Fruit in it. You should also remember that Strelka was originally closed source before you decorate your hero.

It is also an entirely different thing from the work that Zach did directly on Rybka 1.0 Beta where he found certain parts of the code to be semantically the same as Fruit.

To answer your question "What is Rybka?", that is what the ICGA seek to ascertain as we speak with the help of those that can. It is not a short process and requires much work. You whining about some other engine because its author helped to dish the dirt on your precious little Ippo's is not helping. It would be better for you to wait for their conclusion than to be flapping your gums about things you know nothing about. I have put that as nicely as I can Damir.

In the meantime I suggest you burn your riddle. You are obviously confused.



Chris

[Damir]

Nope, no confusion here at all... It is true that engines coming after Ippolit source has been released, might be based or contain Ippolite source, so might every other strong engine which appears out of the blue, but that is no way to threat author of these engines in a badly manner, like Naraku author has been treated. He fully admitted it is his hobby, and he offered his engine for free, and people attacked him, and called him cloner etc...

Fact is, Thinker is more suspicious than the so called clones, who might appear out of the blue with the release of Ippolite source because people are aware... and they always shout clone clone


Why? Because it has obfuscated PV nodes, which no other engines do have.
It is also well known fact that Mr Rajlich obfuscated Rybka's nodes counts, to not show how fast Rybka really was... It showed, still do show fake kN/s node counts.
Could it be that Mr Perkins is doing the same thing ?
He should be able to fix this. Yet he does not want to. All more reason to believe he has something to hide, and does not want anyone else to know...

[bob]

Hi Tony,
A good catch I have to admit


Naraku is a free strong UCI engine that I realized in my spare time. It uses many of the most diffuse algorithms used in modern chess engines: alpha-beta and null move pruning, bitboard and so on. It is in an early stage of development but already has opening book support and ponder mode. It’s written in C++ and it is extremely fast (about 1-1.5 MNodes/s on a single CPU). In the next weeks i will add multi core support. My goal is to create a very strong free engine, over 3000 elo points, and some other engines derived from it, but a lot weaker, in the range 600-1600 ELO points.


Now what do you think guys

It looks like the statement early stage of develpment nowadays means a +2700 Elo chess engine

I like the "in the next weeks..." quote. SMP is not a "few weeks" of work. Unless one is copying something, that is...

I can tell that... I am currently working on the SMP version of my engine. I have already discarded the entire smp code of my previous attempt because it didn't scale well. I started over and hope to come up with something which works this time...
Although, if you are implementing something simpler, like shared transposition table, a few weeks is more than enough.

Yes, although it is a +lousy+ approach that produces terrible speedups for any number of cpus > 1.

BTW, I hate to bring this question into light, but is there any criterion that makes Rybka acceptable in the community and Naraku not? AFAIK both are claimed to be bitboard derivatives of Fruit. What makes Rybka/Houdini okay to test/use and Naraku not?
I am not trying to create a controversy, but we should not create double standards in this community.

Regards

About all you can say is "different strokes for different folks"...