[Computer-go] Zombie processes

Nick Wedd nick at maproom.co.uk
Sun Oct 26 03:30:10 PDT 2014

Yesterday F J Dickhut won his game in the "codecentric go challenge
2014" five-game match against Crazy Stone, bringing the score to 3-1,
and winning the match. My congratulations to Franz Dickhut!  I hope
such an event will be held again next year.

I have just read a posting by 'leichtloeslich' at
which reads

 > Crazystone had several disconnections during the game.
 > After the first disconnect (and subsequent reconnect), appearently
 > a second instance of Crazystone was running in the background (I
 > think all the way till the second disconnect, you can read about this
 > in the KGS kibitz from the end of the game), so that for about 1/3 of
 > the game Crazystone was running on merely half the processing
 > power/memory it should have been able to use.

I have come across a similar problem before. In the 2010 European Go
Congress, I was sent a copy of Leela to operate. It had a few
disconnects (after which I simply restarted it), but did badly, losing
all its games.  At the end of the event, when I shut down the machine it 
had been using, I found multiple copies of Leela running, which
would all have been competing for resources.

I wonder how practicable it would be for a bot, on startup, to search
for other copies of itself running, and issue a warning if it finds any.

Nick Wedd
nick at maproom.co.uk

More information about the Computer-go mailing list