[Beowulf] HPC in Luguano - and chess!

[Vincent Diepeveen]

Summarized if you don't want to read the PDF: As for a simple  
technical overview, there is 2 discussions important.

First one is the interview i had with many of the top researchers at  
that point.
Second is the huge discussions on rec.games.chess.computer

Not sure whether google indexes those years, otherwise you will be  
able to find it at altavista.

The thread is called something that Deep Blue should be looking ahead  
like 18-20 moves (plies) instead of the 10-12 they got
(which was similar to the top engines in 1996-1997).

There i claimed that with that huge nps and what i saw in my program  
adding modern algorithms, 18-20 moves (ply) would have been possible
to look ahead, as i saw how it was possible to get branching factor  
3.0 to get 1 move (iteration) deeper.

After months of discussions on RGCC where majority of scientists  
spoke out that it was total impossible to get a branching factor
like that (time to get another ply deeper), i got a surprising  
support email.

Regrettably it was a private e-mail yet it was written by Frans  
Morsch, at that point the reigning world champion. As usual no public  
support,
yet he reported that the branching factor of Fritz was 2.8 to get 1  
ply deeper.

So the most optimistic report and prediction of what was possible was  
by Frans Morsch. As he saw 2.8

Todays branching factor of most top chessprograms is UNDER factor 2.0  
in time to get 1 move (iteration ply depth) deeper search.

No one had predicted THAT in 90s.

If we would extrapolate that onto cryptography then some government  
officials, when realizing that, will get very sick of worry.

Vincent


On Mar 1, 2013, at 1:13 PM, Hearns, John wrote:

> Vincent,
>
> You have kept this quiet on the list! I was checking the agenda for  
> the HPC Advisory Council in Luguano.
>
> I saw a familiar name:
>
> HPC for Chess (Vincent Diepeveen, creator of Diep and Fide-Master)
>
>
>
>
>
> Dr John Hearns | CFD Hardware Specialist | McLaren Racing Limited
> McLaren Technology Centre, Chertsey Road, Woking, Surrey GU21 4YH, UK
>
>
> T:  +44 (0) 1483 262000
>
> D:  +44 (0) 1483 262352
>
> F:  +44 (0) 1483 261928
> E:  john.hearns at mclaren.com
>
> W: www.mclaren.com
>
>
>
> The contents of this email are confidential and for the exclusive  
> use of the intended recipient. If you receive this email in error  
> you should not copy it, retransmit it, use it or disclose its  
> contents but should return it to the sender immediately and delete  
> your copy.
>
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin  
> Computing
> To change your subscription (digest mode or unsubscribe) visit  
> http://www.beowulf.org/mailman/listinfo/beowulf