#being sent more readyoks than I send isready?

25 messages · Page 1 of 1 (latest)

north moss
#

Is it normal that stockfish send a greater number of readyok than I send it isready?

fathom ether
#

send the log

#

readyok is only ever sent as a response to isready pretty sure the issue lies within your code

north moss
#

Okay, thank you so much!

north moss
#

Does stockfish have an echo/trace mode for debugging? Essentially it echos back the commands you sent it in addition to its own responses?

#

I wish I could wrap boost::process::opstream

#

Rather, was skilled enough to do that faster than adding print statements

fathom ether
#

mh theres a log option but i dunno what it does tbh

north moss
#

Okay, so why after my first "go depth" is "readyok" printed before the depth information is?

#

THAT is apparently the case I'm not handling well

#

I would expect "go depth 20" to finish before readyok is sent to me

signal veldt
#

go isn't a blocking command. When you send isready the engine is expected to reply as soon as possible without waiting on existing searches

#

(this is what the uci spec states)

north moss
#

hmm.... how do I know when the depth search is finished then?

#

Also, hilariously, that invalidates a metric ton of numbers I crunched

#

Because I now don't know if the race condition meant I was recording the wrong eval for a position

signal veldt
#

unless it's an infinite search

north moss
#

oh, fantastic

signal veldt
#

then it won't ever end until you send stop

north moss
#

weird, there is apparently no best move here?

position startpos moves f2f3 e7e5 g2g4 d8h4
go depth 20
info string NNUE evaluation using nn-baff1ede1f90.nnue
info string NNUE evaluation using nn-b1a57edbea57.nnue
info depth 0 score mate 0
bestmove (none)

#

Oh

#

🤦‍♀️