Re: [computer-go] SlugGo v.s. Many Faces

["Jim O'Flaherty, Jr." <jim_oflaherty_jr@xxxxxxxxx>]

David,

I don't think I understand.

At the beginning of a game of go, the branchiness at a ply depth of 16 is (361 * 360 * 359 * 358 *
... * 346).  Are you saying you have that many processing nodes (computers, CPUs, ?) to be able to
handle the ply depth of 16?  If not, I am not sure how you are taking the huge branchiness (just
shy of a exponential curve) and reducing it to just the exponent.

If you have the time, perhaps you could explain in greater detail what it is you are doing...i.e.
the general process, perhaps with a description of how the system starts to play a game, and in
evaluating a particular move, how the exporations engages additional computing resources.  Or a
URL to a paper or website available that describes this, it would be helpful.


Jim O'Flaherty


--- David G Doshay <ddoshay@xxxxxxxxxxxxxxxxx> wrote:

> On Aug 25, 2004, at 4:56 AM, Don Dailey wrote:
> 
> >
> >> The primary reason  the code is so  very slow is that we  have found a
> >> lookahead depth  of 16 works very well,  and it thus takes  a bit more
> >> than 16 times as long as GNU Go to settle on its move.
> >
> > Can you  explain this statement?  You  make it sound  that your search
> > takes linear time.  Would a 32 ply search take only twice as long?
> >
> 
> If by ply you are referring to the number of end leaves in a branching
> tree then the answer is that it would take no more than twice as long.
> 
> The time is linear in the depth of search. The width of the search takes
> virtually no wall clock time because different search paths are each
> done on a different cpu in the cluster. At each branch new cpus are
> brought into the calculation.
> 
> 
> Cheers,
> David
> 
> 
> _______________________________________________
> computer-go mailing list
> computer-go@xxxxxxxxxxxxxxxxx
> http://www.computer-go.org/mailman/listinfo/computer-go/
> 

_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/