I wonder if there is a list of best engines for different limitations?

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

Uri Blass
Posts: 10965
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

I wonder if there is a list of best engines for different limitations?

Post by Uri Blass »

For example Ice4 is probably the best engine with compressed source that is less than 4Kbytes
iCE 4.0 64-bit 4CPU 3160 +21 −21 45.7% +27.1 52.6% 506

What about different limitation?
1)compressed source less than X for different X than 4 kbytes.
2)source not compressed less than X.
3)memory use when it run that is less than X(it means of course also not being able to use big hash tables).
chesskobra
Posts: 356
Joined: Thu Jul 21, 2022 12:30 am
Full name: Chesskobra

Re: I wonder if there is a list of best engines for different limitations?

Post by chesskobra »

I am interested in such a list. ice seems very interesting. I wonder if it can be compiled for linux. I once searched in Jim Ablett's linux compiles for small engines.

Among very strong engines, I think Alexandria is impressive. A top 10 engine, with under 4 MB executable, so I believe very small NN.
Dann Corbit
Posts: 12803
Joined: Wed Mar 08, 2006 8:57 pm
Location: Redmond, WA USA

Re: I wonder if there is a list of best engines for different limitations?

Post by Dann Corbit »

I am more interested in chess engines for a purpose.
For instance, for proving mates in ten plies or less, Chest319 is hard to beat.
For game play, it is probably Stockfish.
For opening analysis, LC0 is an excellent second voice.
For tactical problems Rems and Sun are great.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.
Jouni
Posts: 3722
Joined: Wed Mar 08, 2006 8:15 pm
Full name: Jouni Uski

Re: I wonder if there is a list of best engines for different limitations?

Post by Jouni »

Where You got Rems and Sun?
Jouni
Gian-Carlo Pascutto
Posts: 1260
Joined: Sat Dec 13, 2008 7:00 pm

Re: I wonder if there is a list of best engines for different limitations?

Post by Gian-Carlo Pascutto »

Uri Blass wrote: Tue Nov 19, 2024 11:43 am For example Ice4 is probably the best engine with compressed source that is less than 4Kbytes
iCE 4.0 64-bit 4CPU 3160 +21 −21 45.7% +27.1 52.6% 506
Those versions are outdated, and they also did not test 4ku in the same 4 CPU configuration.

https://github.com/kz04px/4ku/releases/tag/v5.1
https://github.com/MinusKelvin/ice4/releases
Dann Corbit
Posts: 12803
Joined: Wed Mar 08, 2006 8:57 pm
Location: Redmond, WA USA

Re: I wonder if there is a list of best engines for different limitations?

Post by Dann Corbit »

Jouni wrote: Tue Nov 19, 2024 4:47 pm Where You got Rems and Sun?
I forgot those are private. But the last version of Leptir was public I think and it performs just as well tactically as those.
Interestingly, they also do very well at game play where as most tactical versions flop (such as The Huntsman) when actually playing games.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.
Uri Blass
Posts: 10965
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: I wonder if there is a list of best engines for different limitations?

Post by Uri Blass »

Dann Corbit wrote: Tue Nov 19, 2024 4:01 pm I am more interested in chess engines for a purpose.
For instance, for proving mates in ten plies or less, Chest319 is hard to beat.
For game play, it is probably Stockfish.
For opening analysis, LC0 is an excellent second voice.
For tactical problems Rems and Sun are great.
For mates I would like to have some mate competition between engines that may be in the following way:
1)Take a lot of random positions from human games.
2)give 2 engines 1 minute to search every position

a)If no engine find mate result is a draw.
b)if both engines find mate then the winner is the engine that found mate in less moves.
c)if both engines found a mate in the same number of moves then the engine that claimed mate faster is the winner.

If a claim of the engine for mate is wrong(engine claimed mate in 5 when the shortest mate is mate in 6 then the engine is not allowed to compete).
I wonder if there is an engine that provide a tree proof for mates including size of the tree when the tree may be a file with millions of lines in order to help verifying that a mate claim is correct when you can also give the engine initial small tree to help it to search only moves of white inside the tree assuming white is the side that gives the mate.
User avatar
RubiChess
Posts: 647
Joined: Fri Mar 30, 2018 7:20 am
Full name: Andreas Matthies

Re: I wonder if there is a list of best engines for different limitations?

Post by RubiChess »

Uri Blass wrote: Tue Nov 19, 2024 11:43 am For example Ice4 is probably the best engine with compressed source that is less than 4Kbytes
iCE 4.0 64-bit 4CPU 3160 +21 −21 45.7% +27.1 52.6% 506
You are mixing two engines here which have nothing in common but the almost identical name.

http://www.fam-petzke.de/cp_ice_en.shtml
127 iCE 4.0 64-bit 4CPU 3160 +21 −21 45.7% +27.1 52.6% 506

https://github.com/MinusKelvin/ice4
161‑162 ice4 v5 64-bit 3023 +23 −23 47.9% +15.1 45.5% 453
Dann Corbit
Posts: 12803
Joined: Wed Mar 08, 2006 8:57 pm
Location: Redmond, WA USA

Re: I wonder if there is a list of best engines for different limitations?

Post by Dann Corbit »

Uri Blass wrote: Fri Nov 22, 2024 6:44 am {snip}
I wonder if there is an engine that provide a tree proof for mates including size of the tree when the tree may be a file with millions of lines in order to help verifying that a mate claim is correct when you can also give the engine initial small tree to help it to search only moves of white inside the tree assuming white is the side that gives the mate.
Yes, Chest can provide the complete proof tree.
Chest319 option:
-l -L increment solution print depth / complete

Chest324 prints the refutation table:
--refutab[=B] (dft=yes) print refutation table
and:
--move-trace-all trace all moves
and:
--solution-depth-all
print complete solution

ChestUCI has:
Defender moves
FullMoveList
CompleteDuals
ShowAnalysis

If the mate is deep, you'll be sorry you asked.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.