the stockfish bug on time management has caused much ink to be spilled.
stockfish blames everyone except themselves, even though there's a lot of internal blame to be had.
the easy move code, broken.
the lazy smp, dubious at first, and breaking easy move (see above)
so then it's windows fault (os time slice at 10ms)
or maybe a hardware problem.
just admit "lazy smp" is dubious as a concept, and flawed in implementing.
see https://groups.google.com/d/msg/fishcoo ... VUqnQ8CgAJ
The rules clearly state that upgrading the binary between a Stage is a risk.
If you are willing to accept the risk you can reap the benefits but if it backfires then it creates a situation like we see here now where blame is thrown around like shrapnel from a grenade
STAGE 3 BROUHAHA
Moderator: Psyck
ANOTHER STOCKFISH BUG
the misery continues, round 4 and up big against protector, and...
stockfish lose on time (88ms)
almost saved by the"tcec win rule"
which makes you wonder why they have that rule there in the first place
could be lots of bugs in mayny programs, but covered up by their methods
gull is in first, drawing to houdini and komodo and winning the other two
stockfish lose on time (88ms)
almost saved by the"tcec win rule"
which makes you wonder why they have that rule there in the first place
could be lots of bugs in mayny programs, but covered up by their methods
gull is in first, drawing to houdini and komodo and winning the other two