You could what they do in bridge tournaments, and provide two sets of results 
from the same tournament.

Hardware would be unrestricted for everyone

The Open result would include all participants, exactly as today.

A "single machine" result would only include participants that ran on a single 
node, perhaps with no more than 4 or 6 cores, or perhaps a single CPU 
(typically 4 to 6 cores).  The idea would be to compare programs on hardware 
that is generally available to consumers.

So a single machine participant would get two results, perhaps something like 
4th in open, 2nd in single machine.  The KGS tournament and pairings would 
include everyone, so there is no overhead for the organizer other than making a 
second results table that only includes single machine results.

You would have to decide whether to allow two entries for the cluster programs, 
one using the cluster, and one single-machine.  In that case, the open result 
would only report the better of the two for that program.

I don’t think it is a good idea to make a general restriction, and it's 
impractical to require identical hardware for all participants.  If you want to 
do a special tournament with identical hardware, you should consider requiring 
an AWS instance, since they are identical, and inexpensive.  It does take a few 
days to get AWS figured out and tested, so the prep is not completely trivial.  
The New DNN technology works best with a GPU, and other programs get no benefit 
from GPU, so including or not including a GPU in your standard machines will 
highly favor one type of program over the other.

Conflict of interest disclosure.  Many Faces is designed for single machine, 
shared memory, because that's what my customers run it on.  I had multinode MPI 
code for a Microsoft cluster in 2008 when I won, but I ripped it out long ago.  
I'm interested in seeing how programs compare on consumer hardware.

David

> -----Original Message-----
> From: Computer-go [mailto:[email protected]] On Behalf
> Of Gonçalo Mendes Ferreira
> Sent: Saturday, October 10, 2015 10:40 AM
> To: [email protected]
> Subject: *****SPAM***** Re: [Computer-go] KGS bot tournaments - what are
> your opinions?
> 
> 
> 
> On 10/10/2015 18:30, David Doshay wrote:
> > I agree completely that there is no way to enforce computational limits
> over the internet.
> >
> > I am against  identical hardware  tournaments because people have worked
> to get their programs working on the hardware they have, and some people
> will be on the other side of any hardware decision, Mac v.s. PC being the
> most obvious.
> There is no "Mac hardware".
> >
> > I am left wondering what the point is for such a tournament. Is it to
> show who is the most efficient programmer? Is it to show how these
> programs might run on somebody s home computer? These things are not
> important for research code that is not intended for resale.
> I'm also against identical hardware restrictions, but divisions can be
> very flexible. Not everyone cares for research and you wouldn't be using
> open tournaments for research results either way.
> 
> Gon alo F.
> _______________________________________________
> Computer-go mailing list
> [email protected]
> http://computer-go.org/mailman/listinfo/computer-go

_______________________________________________
Computer-go mailing list
[email protected]
http://computer-go.org/mailman/listinfo/computer-go

Reply via email to