NUCLEUS

[Laskos]

Whatever time the human component needs, to make a decision when playing as Centaur, runs in parallel to the one used by the engines. The TD won't give you a bonus for selecting that playing mode. When the clock runs dry, the human will eventually stop "wasting" time and start following blindly one engine's output. At that point, play will be worse than the engine running on its own, because you have to deduct the human lag from the time per move available. That's one of the many quirks to Centaur play, your configuration doesn't represent that mode of play at all.

I don't understand what you are saying. I have no experience in that, but there is possibility of a "Centaur" running as I describe. Human runs 2 engines simultaneously (say on half the threads each) and makes the decision in the latter stages of analysis (thus 50% for LC0 and not 50%+50%=100%, but keep in my my LC0 is weak on CPU) what move is better. The result is encouraging insofar, which might show that a "Centaur" using SF (the strongest Elo-wise) and a bit weaker Komodo against a pure SF is beneficial, if the human part of the "Centaur" is as skillful as LC0. You seem to say that a "Centaur" will always be weaker than the pure strongest engine, which is probably wrong to say.

If you don't understand what I'm saying, how can you say anything about it? There's no possibility of a "Centaur" running as you describe, because it would go against the rules of any Freestyle tournament ever held. The rest of the post stems from a misinterpretation, so I'll just refer you to my part of the quoted message.

Ok, so this is not a "Centaur" according to some "rules of Freestyle tournaments", thanks for letting me know. So, the Freestyle tournaments must be a pretty silly thing, if this is not allowed, didn't know that. Therefore I must reformulate: squeezing out the best out of 3 entities at fixed hardware and time: skillful human (LC0), strongest engine SF, second strongest engine Komodo, one seems to be able to be better than the single strongest engine SF on the identical hardware and time. If this holds, it surely would help in correspondence chess, and generally, in analysis.

[Ozymandias]

Ok, so this is not a "Centaur" according to some "rules of Freestyle tournaments", thanks for letting me know. So, the Freestyle tournaments must be a pretty silly thing, if this is not allowed, didn't know that.

When have you seen a chess tournament of any kind, where some players are allowed 50% more time? Freestyle tournaments have several problems and I've talked at length about the most prominent ones, but such a time surplus, wouldn't be a proposal I'd put forward.

[Laskos]

Ok, so this is not a "Centaur" according to some "rules of Freestyle tournaments", thanks for letting me know. So, the Freestyle tournaments must be a pretty silly thing, if this is not allowed, didn't know that.

When have you seen a chess tournament of any kind, where some players are allowed 50% more time? Freestyle tournaments have several problems and I've talked at length about the most prominent ones, but such a time surplus, wouldn't be a proposal I'd put forward.

No, I meant that human (LC0 say on strong GPU) analyzes shortly the replies to the 2 engines' PVs, starting shortly before the time is up. Time parity is implied, I just cannot reproduce it with LC0 without a strong GPU, that's why I gave it some reasonable time on my weak CPU. Human is completely independent of CPU and GPU, so the time parity is even easier to insure effectively.

[Ozymandias]

Ok, so this is not a "Centaur" according to some "rules of Freestyle tournaments", thanks for letting me know. So, the Freestyle tournaments must be a pretty silly thing, if this is not allowed, didn't know that.

When have you seen a chess tournament of any kind, where some players are allowed 50% more time? Freestyle tournaments have several problems and I've talked at length about the most prominent ones, but such a time surplus, wouldn't be a proposal I'd put forward.

No, I meant that human (LC0 say on strong GPU) analyzes shortly the replies to the 2 engines' PVs, starting shortly before the time is up. Time parity is implied, I just cannot reproduce it with LC0 without a strong GPU, that's why I gave it some reasonable time on my weak CPU. Human is completely independent of CPU and GPU, so the time parity is even easier to insure effectively.

That changes things, in such a setup, you'd only need to take a little bit off the slaves' time, to compensate for the excess used by the master. How much, it would depend on how much time a good GPU would need to perform the task you're assigning your (relatively) weak CPU.

[peter]

I get the message too if I try to install Nucleus as a WB engine. But Nucleus is UCI.

For me it's the same with UCI (which I tried first of course as said in readme).
Does your installed Nucleus have any config.- file stored?
Maye I could copy it and change it then?
Simply starting the .exe itselft didn't help neither.
Any command- line instruction maybe?

[Laskos]

When have you seen a chess tournament of any kind, where some players are allowed 50% more time? Freestyle tournaments have several problems and I've talked at length about the most prominent ones, but such a time surplus, wouldn't be a proposal I'd put forward.

No, I meant that human (LC0 say on strong GPU) analyzes shortly the replies to the 2 engines' PVs, starting shortly before the time is up. Time parity is implied, I just cannot reproduce it with LC0 without a strong GPU, that's why I gave it some reasonable time on my weak CPU. Human is completely independent of CPU and GPU, so the time parity is even easier to insure effectively.

That changes things, in such a setup, you'd only need to take a little bit off the slaves' time, to compensate for the excess used by the master. How much, it would depend on how much time a good GPU would need to perform the task you're assigning your (relatively) weak CPU.

Yes, and a strong GPU with CUDA LC0 is literally 30 or so times faster than LC0 on my CPU, so the time to slaves would decrease by less than 2% each.

[Ozymandias]

No, I meant that human (LC0 say on strong GPU) analyzes shortly the replies to the 2 engines' PVs, starting shortly before the time is up. Time parity is implied, I just cannot reproduce it with LC0 without a strong GPU, that's why I gave it some reasonable time on my weak CPU. Human is completely independent of CPU and GPU, so the time parity is even easier to insure effectively.

That changes things, in such a setup, you'd only need to take a little bit off the slaves' time, to compensate for the excess used by the master. How much, it would depend on how much time a good GPU would need to perform the task you're assigning your (relatively) weak CPU.

Yes, and a strong GPU with CUDA LC0 is literally 30 or so times faster than LC0 on my CPU, so the time to slaves would decrease by less than 2% each.

Ok, in that case there's no much difference between 50-50-50 and 50-49-49. We can call such a player a "Barebones Centaur".

[yanquis1972]

i think the 150% example is the one that seems most representative, though. the whole point is that a sufficiently strong, GPU-bound NN engine unlocks considerable resources on a fairly common consumer machine, & i don't see why an unassisted hybrid shouldn't be stronger than a dedicated stockfish on the same machine.

it'd also just be interesting, at the present time, to see what leela with a tactical search and/or 'watchdog' is capable of.

i'll have to read up on how the 3 engines are implemented with nucleus, but in kai's (limited) simulation, the "CPU" is split between 2 engines, & even with a sick adolescent NN guiding, it was still a match for full-strength SF. that strikes me as extremely promising

[Rebel]

I get the message too if I try to install Nucleus as a WB engine. But Nucleus is UCI.

For me it's the same with UCI (which I tried first of course as said in readme).
Does your installed Nucleus have any config.- file stored?
Maye I could copy it and change it then?
Simply starting the .exe itselft didn't help neither.
Any command- line instruction maybe?

I am sorry, but I don't have an explanation either. Maybe Ferdinand when he reads this.

[peter]

I am sorry, but I don't have an explanation either. Maybe Ferdinand when he reads this.

I found out the reason for not succeeding by myself:
I renamed the engines master.exe and slave1(2).exe, should have let the.exe- ending away, Windows does add it automatically when renaming an already existing .exe- file, so obviously the files were internally named .exe.exe.


Thanks again!