CCRL Discussion Board

To simulate any number of pieces, try limiting hash table size and see how engines handle a small number of pieces. The ones it gets wrong when searching for a few minutes, try searching a few hours and see how often that fixes the problem. I think it won't work for all positions, but seeing how oft...

Here is my plan to make the problem easier so someone with my ability might solve it: Initial limitation: Only compute white wins as it simplifies the code a little. Also don't worry about pawns. Make a naïve implementation including code working for array. Save result. Write query code for files. I...

Sounds like there might be some fun optimizations there, esp. with the 2 duck positions idea.

I hope you meant "no trouble" or "little trouble". An ominous indication of the complexity of your scheme. No, I mean to make the problem easier, so someone like myself, who has trouble juggling all the details, is able to do it. One part of it is applying the good advice I got ...

After 18 months of failing I have decided I am not as good at programming as I thought. Next step is writing a plan for solving the problem for someone who has trouble juggling all the details required for the solution I have in mind. First part of it will be a very naïve way of doing the computatio...

Or, if the L is allowed is allowed to rotate, 4 times as many, as its orientation also becomes part of the game state. Or yet another factor 2 more if it is also allowed to flip over to a mirro image. Actually, the rotations and mirrorings would be cancelled out exactly by being able to rotate and ...

Treating it as one piece seems reasonalble. Would it have to stay on the board in its entirety? Would it be allowed to capture even if no capture of itself was allowed? Would it be allowed to rotate, and if so, when and around what axis? Have you tried playing with it on a chess board to see what ca...

Try defining the rules for your new piece. Then play with it and you should be able to see if you want a tablebase. Fantasy pieces are generally not hard to implement a tablebase for. If you have a queen which can lay eggs, the number of eggs need to be limited in order for a tablebase to be practic...

Is the idea that interleaving bits from white and black pieces in the index improves data access locality, reducing memory requirements? Or is the indexing geared towards applying 8-way symmetry through simple bitwise operators, thereby improving calculating speed? (Or both.) Yes, the idea is to im...

I got some good advice here which I didn't take, but now I am trying to do better. To do so I am starting with writing a design document: https://github.com/Koistinen/Endgame/blob/master/algo.txt The goal is to avoid optimizing much except what is needed to fit in ram. In the beginning extra disk i/...

I think Google, Facebook, et.c. have shown us that centrally controlled computing easily turns becomes evil. That is why I prefer decentralized computing, not just distributed. That way I don't get to lord it over people. Besides computer chess I have also been interested in crypto techniques for qu...

I've been thinking of how to do 8-man and beyond and a decentralized way of computing it seems interesting as storage is one of the biggest problems. For 8-man each pawnless endgame table need only take 2-3 months (with 50 move rule, proportional with dtc and worse for dtm) given each computing part...

How about this answer: When computers have 64 times as much ram, i.e. 64*16 GB = 1 TB of ram. So, perhaps in 12 years assuming doubling every 2 years?

Nice, C and GNU Licence!

Have downloaded and will study it.

Unless I'm missing something, it should be 16 (4x4), because with this piece placement both white and black can have 4 different possibilities of castling rights. Ok, that is very reasonable, because it could get complicated if you had to consider if those rights might ever be used when counting. 4...

I like it! Metrics have many benefits. To clarify could you look at the position and explain?

rnb1kbnr/pppp1ppp/8/4p3/5PPq/8/PPPPP2P/RNBQKBNR w


How many ULPs have the pieces on the squares as above and why?

After changing line 802 to: for(tb=p;4095&(long)tb;) ++tb; /* align with cash line */ and including string.h it appears to compile without complaints from GCC on my Debian system. Running with argument "KQ.K" results in output: 2602010 2603000 # N: 18,3 14,3 -14,3 -18,3 33,3 31,3 -31,3...

petter wrote:How long do you think it may take to calculate one of the 7 chessmen tablebases in a i-950 home computer with
6 gigabyte of ram?

My estimation is: at least 4000 times the time it takes to calculate a 5-man or 60 times the time it takes to calculate a 6-man.

Has anyone got performance data for this generator?
Say, how long does it take to generate all 5-man tablebases? On what computer? What metric?

I'd be willing to put this speed into bounty specs (relaxed a bit for safety). Assuming there is no other choice. If there is other choice, then I'd try to aim for a faster-performing generator. What is the performance of the currently available 6-man generators on current hardware when the table f...

I think it would be nice if the demand for speed is lax as it is much easier to implement something when you are allowed to be half as fast as optimum, and more laxity gives more ease. In my scheme, there is one 46.4 GiBy table that needs to be read twice and written once every iteration. With dupli...

Distributed probing should be fun. One way for a server to prove to a client that an honest answer with hash function F constructed using cryptographic hash function H is this: F(T)-> if small_enough(T)->H(T); else H([H(first_half(T)),H(second_half(T))]). Client must know F(Table). Server has a tabl...

A comment on 7-men EGTB Bounty Reborn - Summary . While requiring that probing code be available in C is sensible, having programming language requirements for the generator might be harmful. Allow any language as long as there are no dependencies on nonfree code. Readability requirements are fine, ...

With an the clean code and OSI approved license requirements it seems likely that those who want a different metric would be able to adapt the code to their needs. I'd like to see any of the metrics WDL50-DTM be allowed. (That is: any one of the metrics mentioned in this thread if computed should co...

Maybe specify a benchmark computer exactly would be good, what processor, motherboard, memory and disks are connected. Because 4 GB(gigabytes) memory modules now are reasonably priced, allowing a requirement of 24 GB at 10-20 GB/s for generating computer should be ok. (I would not need more than 8GB...