Emre Sevinc
2008-08-05 19:09:53 UTC
Original address: http://groups.google.com/group/comp.lang.lisp/msg/c6f416636b914d5d
---------- Forwarded message ----------
From: Viktor Winschel <***@rumms.uni-mannheim.de>
Date: Aug 5, 3:25 pm
Subject: lisp on grid
To: comp.lang.lisp
Salute Metalinguists,
I am about to start an academic project on a 1120 grid machine
(140 blades with 2 quad cores or 4 dual cores, dono, however, many :-)
The project is to implement a Bayes and decision net software targeted
at economists and econometricians. Now I have some academic
collaboration with the European, Italian and other central banks. A
spin-off is possible and likely, some hedge fonds are
interested...there
should be a market. The grid market itself is about to be designed, in
need of economic decision agents and market makers....
I guess Bayes networks could be a nice application for cells,
since the basic Bayes net algorithm, the junction tree algo, is
some kind of constraint propagation.
However, I need some advices what concurrency library could be of use
for my huge >1KCPU machine...
NetClos
cl-muproc
Or anything else? Any ideas and experiences?
Lispers to collaborate (research funds or start-up money should be
available...) are highly welcomed.
ciao ciao
Viktor Winschel
---------- Forwarded message ----------
From: Viktor Winschel <***@rumms.uni-mannheim.de>
Date: Aug 5, 3:25 pm
Subject: lisp on grid
To: comp.lang.lisp
Salute Metalinguists,
I am about to start an academic project on a 1120 grid machine
(140 blades with 2 quad cores or 4 dual cores, dono, however, many :-)
The project is to implement a Bayes and decision net software targeted
at economists and econometricians. Now I have some academic
collaboration with the European, Italian and other central banks. A
spin-off is possible and likely, some hedge fonds are
interested...there
should be a market. The grid market itself is about to be designed, in
need of economic decision agents and market makers....
I guess Bayes networks could be a nice application for cells,
since the basic Bayes net algorithm, the junction tree algo, is
some kind of constraint propagation.
However, I need some advices what concurrency library could be of use
for my huge >1KCPU machine...
NetClos
cl-muproc
Or anything else? Any ideas and experiences?
Lispers to collaborate (research funds or start-up money should be
available...) are highly welcomed.
ciao ciao
Viktor Winschel