Christopher Conkie - Strelka...

[Chris Taylor]

Hi

How is the Strelka engine? Have you found anything worth sharing yet?



Best


Chris

[Uri Blass]

Hi

How is the Strelka engine? Have you found anything worth sharing yet?



Best


Chris

My opinion is that Strelka is too fat.
I changed strelka to have a smaller source that produce the same output.

I deleted some thousands lines from part of the tables that strelka use and wrote few hundreds of lines to calculate the tables in order to make strelka significantly smaller.

Uri

[Uri Blass]

Note that one array of strelka seem to be the same as a function in movei.

strelka has a big array of 4096 chars Distance[64][64] to calculate distance between squares.

I have no array for this and

I simply have the following definition in movei to calculate the same thing

#define Max(a,b) (((a) > (b)) ? (a) : (b))
#define FileDistance(a,b) abs(((a)&7) - ((b)&7))
#define RankDistance(a,b) abs(((a)>>3) - ((b)>>3))
#define Distance(a,b) Max(FileDistance(a,b),RankDistance(a,b))

I added this definition to strelka and found that Distance(i,j) is always equal to Distance[j] in strelka for 0<=i,j<=63

Strelka is using 512 lines of code in order to get what I got in few lines and without array by a simple definition.

What is your opinion about it?

Uri

[Alessandro Scotti]

Maybe the array method was faster at the beginning and this is no longer the case due to technology changes (maybe it's still faster, I don't really know).

[Dann Corbit]

Hi

How is the Strelka engine? Have you found anything worth sharing yet?



Best


Chris

Several people have weighed in with opinions. Of course, this must be considered inconclusive.
There are only two people who matter here:
The authors of Fruit and Rybka.
In my opinion, if either of those authors say that the Strelka author has borrowed too much, then the Strelka author has borrowed too much. If neither has any objections, then there is nothing wrong with Strelka.

Personally, because of the use of Fruit at some point in the geneology, I think Strelka should be GPL as well, but I am not expert enough to know if that notion has any weight.

[Dann Corbit]

Maybe the array method was faster at the beginning and this is no longer the case due to technology changes (maybe it's still faster, I don't really know).

It probably varies from machine to machine. Especially, I would expect differences on machines with huge cache like Xeons verses machines with a lesser cache.

[Christopher Conkie]

Hi

How is the Strelka engine? Have you found anything worth sharing yet?



Best


Chris

Several people have weighed in with opinions. Of course, this must be considered inconclusive.
There are only two people who matter here:
The authors of Fruit and Rybka.
In my opinion, if either of those authors say that the Strelka author has borrowed too much, then the Strelka author has borrowed too much. If neither has any objections, then there is nothing wrong with Strelka.

Personally, because of the use of Fruit at some point in the geneology, I think Strelka should be GPL as well, but I am not expert enough to know if that notion has any weight.

Well i know one of those people would like to see all the source and its not Fabien.

If it's ok with him....then its ok. I'd encourage them to send that to him if I were you.

I would not stick to my guns about this if I did not know that. Perhaps you can arrange for that to happen?



Christopher

[Dann Corbit]

Has Vas said he wants to see the code? I have never seen him say that.
Can you provide a link where he has made that request?

Fabian does have the code and he has spent some time going over it. I am not sure what the hold-up is on his offering his opinion. I guess that it is not simple or perhaps just that he is very busy right now.

[Christopher Conkie]

Has Vas said he wants to see the code? I have never seen him say that.
Can you provide a link where he has made that request?

Do you think he would such a thing in public? Please just trust me on this. I would not lie to you about this. It is a very good solution however.

Fabian does have the code and he has spent some time going over it. I am not sure what the hold-up is on his offering his opinion. I guess that it is not simple or perhaps just that he is very busy right now.

Probably busy (but then again....) I'm sure Vas has the time. It would solve the real issue here.

If it were found that it did contain Rybka things, GPL would be out of the question unless he wanted it that way. He has expressed such a view in his own forum. It would depend although on how much Rybka and the significance of that "how much".

Of course everything could be ok in which case no need for gpl unless the Strelka author wanted that. Of course it depends on how much Fruit is in it as well.....

I think your best bet is Vas.



Christopher

[Dann Corbit]

Has Vas said he wants to see the code? I have never seen him say that.
Can you provide a link where he has made that request?

Do you think he would such a thing in public? Please just trust me on this. I would not lie to you about this. It is a very good solution however.

Fabian does have the code and he has spent some time going over it. I am not sure what the hold-up is on his offering his opinion. I guess that it is not simple or perhaps just that he is very busy right now.

Probably busy (but then again....) I'm sure Vas has the time. It would solve the real issue here.

If it were found that it did contain Rybka things, GPL would be out of the question unless he wanted it that way. He has expressed such a view in his own forum. It would depend although on how much Rybka and the significance of that "how much".

Of course everything could be ok in which case no need for gpl unless the Strelka author wanted that. Of course it depends on how much Fruit is in it as well.....

I think your best bet is Vas.



Christopher

If Vas wants to see the code, why wouldn't he ask? I can see no reason for it whatsoever.