[ExI] simulation

[Eugen Leitl]

On Sat, Dec 15, 2007 at 10:53:24AM -0600, Bryan Bishop wrote:

> No, I do not have any ideas. We do not have enough computing power. We'd 
> need to have some way of inserting more nodes onto the internet so that 

no, No, NO. You're not listening. Brain simulations aren't embarassingly
parallel. You'd need some few kNodes, each with <1 us latency, and 
some 10 GBytes/s throughput. On WAN, you'd get some 10-100 ms, and
few 10 kByte/s thoughput.

To be blunt, this kills it dead. You need to use very large scale
distributed resources correctly, orelse you're wasting their
strenghts.

> we could do such a project. It might be possible to have a 
> self-replicating factory that builds computational nodes for this very 

Whoa, whoa, whoa. Once we have these, then we're right in the
middle of that Singularity already. Self-rep is hard, and it's
mostly a control issue on macroscale, and a bootstrap issue on
nanoscale. We can't do either yet.

> project. What projects can we do first to assure us that making enough 
> nodes to run those 250mil sims would be useful?

-- 
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