Re: [Bug-gnubg] Problems using Stop rollout when one move appears to be best option

[Christian Anthon]

Could sound like a racing condition or similar. I'll take a look and
see what I can figure out. May be that my computer isn't fast enough
to trigger the bug.

Christian.

On Sun, Aug 23, 2009 at 10:44 PM, Michael Petch<address@hidden> wrote:
>
> I started the same rollout (its not finished yet) of all the possible moves
> with one thread instead of 16. After the minimum required trials, all moves
> except the top 6 stopped rolling out at 217 including 13, 21 and 24. It
> would appear that number of threads *may* be related to the rollouts not
> properly stopping when the JSD limit has been reached.
>
> I can understand with 16 threads that it won't stop on 217 exactly (+/-10),
> but going the full trial when above the JSD limit from nearly the start is
> peculiar.
>
> Michael
>
> On 23/08/09 2:08 PM, "Michael Petch" <address@hidden> wrote:
>
>> *********
>> One other anomaly I saw (I was using 16 threads) was the status bar was
>> moving along and a few occasions the progress bar jumped backwards
>> considerably and then caught back up within 20 seconds. This happened a few
>> times, and seems to be reproducible.
>> *********
>>
>>
>
>
>