[ExI] simulation

[Eugen Leitl]

On Sat, Dec 15, 2007 at 01:05:54PM -0600, Bryan Bishop wrote:

> Huh? What about a warehouse where all of the cpu nodes are being placed 
> in a neat, orderly fashion, and then having simple bots (or people) 
> come around and connect the new fiber from node to router/distributor?

But this sets us back to mice and pumpkins, aka a Beowulf cluster
worth a couple megabucks, and taking a megabuck/year to run it.

Much easier is it to become a computational neuroscientist, and to
rent time on a national facility. Now for mapping parameter spaces 
as e.g. Nano at home a BOINC or a grid client is about optimal.
If you sell it right millions contributors might join (keeping
them entertained is even more difficult, though), and you
no longer have to budget for energy, climate and sysadmins.
 
> Ah, maybe a warehouse is not enough. So the best place to do this might 
> be the desert, and then we don't have to worry about cables running 
> over the surface, it's not like we're going to be disturbed by curious 
> bystanders.

Your best bet is to tale a few megabucks, and blow them upon
a hall venting outside where power is cheap and climate is chilly.
You might have noticed Google's and Microsoft's efforts in that
direction -- only they can't afford to be topologically to far
removed from their customers.
 
> I was embarrased to order the book a few weeks ago, only to find within 
> an hour an online copy. Heh.

It is a good idea to still order the book, or at least pay for a book
or two from the Ventus series (I won't, they're not that good).

-- 
Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org
______________________________________________________________
ICBM: 48.07100, 11.36820 http://www.ativel.com http://postbiota.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE