Rybka 4 just around the corner it seems

[dadij]

Suppose that you have a quad and you are interested in analyzing 4 positions from different games.

You have 2 options to use the quad:

option 1:analyze every position for 6 hours with a quad.
option 2:analyze every position for 24 hours by a single core.
.
[snip]
.
Edit:Maybe you can use option 2 by running 4 processes of rybka at the same time but this seems to be less user friendly then having interface that allows it.
Uri

This is actually supported in Aquarium.

[Gian-Carlo Pascutto]

I don't see why it is bogus. Running on 4 cores gives you an effective speed-up of 3 due to SMP "losses", but 4 x 1 = 4, i.e. running 4 instances of single core, means there is no SMP loss. So I see the logic in that statement. Whether it gives a better result or not is another story.

What's the gain in running 4 unrelated instances?

If the instances are related in any way (such as being part of a bigger search tree, as in IDeA), there will be SMP loss, because you won't analyze the tree 4 times faster.

Claiming there is no loss is claiming you have found an 100% scalable tree searcher. It's easy to make such a claim, if you're using minimax. If you're using alpha-beta or any tree search algorithm worth using, claiming such a result should either get you a Nobel prize or a trip to the nuthouse.

[Dr.Wael Deeb]

Uri, you seem to understand the advantages.

He's talking about analyzing multiple unrelated positions. Which is something totally different from analyzing related positons from the same tree (=what you are talking about).

Since the speed up of Rybka running on 4 cores is approximately 3 (compared to a single core version), I need to let the single core versions in 2) run for 90 seconds to get the same analysis quality.

This math doesn't work out. It's making the mistaken assumption that IDeA analysis is 100% parallelizable, but that's just not the case.

IDeA makes different decisions about which position to analyze next depending on the result of previous analysis. Parallel analysis will hence end up analyzing positions that serial analysis does not, because it launches several searches at the same time. Hence: inefficiency.

I repeat the point I made: If IDeA would work as claimed, then "Deep Rybka" is a scam because it should use the IDeA algorithm instead. But of course IDeA also has parallel losses (maybe worse so than Rybka SMP), so the claim that it gives a free core is just marketing, because IDeA does not scale perfectly either.

I am shocked from the users blindness GCP,that they can't accept the fact that Vasik is trying to squeeze every single $ from their pockets by any means necessary

For me,the whole new marketing strategy is totaly anti-customer oriented

_The online "rent" issue is an ugly as it can be....if you are so despered to use your promised Rybka 3+ or the current version of Rybka 4,come on,buy our Aquarium 2010 for 124 $ and prepare your money for the rent

_No updates will be made,EVER ....well,we already saw that with the Rybka 3+ and we,as a close friend of mine like to say ,don't believe anymore....

Dr.D

[Dr.Wael Deeb]

I don't see why it is bogus. Running on 4 cores gives you an effective speed-up of 3 due to SMP "losses", but 4 x 1 = 4, i.e. running 4 instances of single core, means there is no SMP loss. So I see the logic in that statement. Whether it gives a better result or not is another story.

What's the gain in running 4 unrelated instances?

If the instances are related in any way (such as being part of a bigger search tree, as in IDeA), there will be SMP loss, because you won't analyze the tree 4 times faster.

Claiming there is no loss is claiming you have found an 100% scalable tree searcher. It's easy to make such a claim, if you're using minimax. If you're using alpha-beta or any tree search algorithm worth using, claiming such a result should either get you a Nobel prize or a trip to the nuthouse.

[dadij]

IDeA makes different decisions about which position to analyze next depending on the result of previous analysis. Parallel analysis will hence end up analyzing positions that serial analysis does not, because it launches several searches at the same time.

Either I don't understand you or you don't understand how IDeA works. It's true that IDeA decides about the next positions to analyze based on the results of previous analysis. However, IDeA only does that at the end of each stage when it finds the next batch of positions to analyze. The number of engines available for analysis doesn't matter when deciding which positions to analyze. Then these positions are analyzed one after the other. If there are several engines available at that time, each one will get a position to analyze, but in all cases the same set of positions will be analyzed, regardless of the number of available engines.

With this in mind, what do you mean when you say "Parallel analysis will hence end up analyzing positions that serial analysis does not"?

[Spock]

I am shocked from the users blindness GCP,that they can't accept the fact that Vasik is trying to squeeze every single $ from their pockets by any means necessary

That's totally ridiculous. Rybka is his full time job, he needs to make a living.

Care to give up your job and try and make a living say from selling opening books ? I'm getting really sick of the Vas-bashing that is going on here.

[Dr.Wael Deeb]

I am shocked from the users blindness GCP,that they can't accept the fact that Vasik is trying to squeeze every single $ from their pockets by any means necessary

That's totally ridiculous. Rybka is his full time job, he needs to make a living.

Care to give up your job and try and make a living say from selling opening books ? I'm getting really sick of the Vas-bashing that is going on here.

I am only expressing my opinion related to the topic,is this against the charter or something

I am not happy with the new prices of Aquarium 2010,I am not happy that the promised Rybka 3+ has never been released and I am not happy with this online rent system of the chess engine that will be affordable for a certain circle of people,what's wrong to express my position

No one is bashing Vasik....

Don't get mad at me when I say that it's becoming a religion,not a simple fanboy club

[Henrik Dinesen]

The cost of Rybka+Aquarium was, if I remember correctly, about if not the same last. It's possible to check, but I'm too lazy.

However, since it's the price for the package, not the singular product, I can't imagine that Rybka 4 UCI will be priced much different than Rybka 3, meaning around 100$. The standalone Quarium 3 costs 19$ now, which I believe is reduced from 25$. This means there's a little discount buying the package Deep Rybka Aquarium 2010 which is set to 119$.

For the thickheaded among us; the point is that the preorder prices we see, is the price for Rybka and Aquarium, not just Rybka 4.

And for the online Rybka: Just another item on the shelf. If you don't want it, don't buy it!

And yes, Vas has been dragged through alot here recently, mostly bad, despite he surely wasn't the one initiating all that.
And it's pity that people can't have fun with an engine without being called fanboys or part of some religion! The "religion" here as far as I'm concerned, is computerchess.

[Dr.Wael Deeb]

The cost of Rybka+Aquarium was, if I remember correctly, about if not the same last. It's possible to check, but I'm too lazy.

However, since it's the price for the package, not the singular product, I can't imagine that Rybka 4 UCI will be priced much different than Rybka 3, meaning around 100$. The standalone Quarium 3 costs 19$ now, which I believe is reduced from 25$. This means there's a little discount buying the package Deep Rybka Aquarium 2010 which is set to 119$.

For the thickheaded among us; the point is that the preorder prices we see, is the price for Rybka and Aquarium, not just Rybka 4.

And for the online Rybka: Just another item on the shelf. If you don't want it, don't buy it!

And yes, Vas has been dragged through alot here recently, mostly bad, despite he surely wasn't the one initiating all that.
And it's pity that people can't have fun with an engine without being called fanboys or part of some religion! The "religion" here as far as I'm concerned, is computerchess.

Hi,
After reading your post,I've decided to call it an end of my comments regarding this issue....It's only getting worse and I am afraid that I am about to lose friends whose friendship has been gained through many years just because our opinions and visions related to this topic are different....
Chilling out regards,
Dr.D

[oreopoulos]

What's the gain in running 4 unrelated instances?

If the instances are related in any way (such as being part of a bigger search tree, as in IDeA), there will be SMP loss, because you won't analyze the tree 4 times faster.

Wait you got it wrong. IDEA is not a chess search engine. Nothing close to that. It is just a chess crawler. It will use the engine as a compass to build a tree of positions.

Now the gain is simple (compare to previous iteration). Previously you had a queue of positions to be analysed by a Nxcore engine. Now you will have the SAME queue to be analyzed by N single core engines. If for example you choose to analyse to fixed depth, it obvious that the N single cores will finish the queue much faster.

Just that.

A chess engine is nothing like that. It is obvious that you cannot split a tree and claim that you doubled speed. No.

But the approach IDEA goes for, is helped by many single core instances. Its an asynchronous process, with a lot of redundant steps, compared to a chess engine. Different story though

It would be nice to see IDEA-enabled engines, that would build a tree, instead of a hash table and support all the interactivity. Not hard i suppose, with the right API