[Gllug] Super Computing Time (fwd)

[Nix]

On Thu, 4 Oct 2001, Iain Nicholson stipulated:
> (6) Model General Relativity problems numerically
>    (i) black hole collisions
>    (ii) stellar collapse to form white dwarves, neutron stars, black holes
>    (iii) model the properties (gravitational, electromagnetic etc. ) of
> the above

(6.5) Model substantial microscopic systems from the ground up (wave
      function)
   (i) good for protein folding and just about anything else with
       sensitive dependence on initial conditions down there
   but it eats *so* much CPU time (many of the problems in this area
   are NP-complete, the rest are exponential time or worse, and there's
   an insane amount of floating-point computation) that even a
   supercomputer will have trouble making a dent in the problem.

> (9) Show the evelolution of the Universe and large scale structure (down
> to galaxy foramtion)
>   (i) Start at the Big Bang and incorporate each of the favourite teories
> at the moment and see if they fit with observations

The problem with this one is that most of the work is in building the
model. At least for the protein folding stuff we *know* the model :)

> Now get cracking, learn the physics and write the code!!!!!!!!!

Er, the problem is the middle option.

> You will have a resource too valuable to waste on parallel PovRay...

Your own mind after you've learnt all that physics...

-- 
`A truly dreadful film destined to be a cult classic for masochists.'
                                              --- Chris Priest on _A.I._.

-- 
Gllug mailing list  -  Gllug at linux.co.uk
http://list.ftech.net/mailman/listinfo/gllug