Re:Matrioshka Brains: computing at the limits of physics
by varjag on Saturday January 23, 2010 @12:25PM
(#1072083)
One would imagine that on a Dyson sphere sized computer, intercommunication (limited by speed of light) will become major bottleneck. At some point amassing components will stop winning you performance due to physical dimensions.
[ Parent
| on HN ]
Re:Matrioshka Brains: computing at the limits of physics
by cstross on Saturday January 23, 2010 @03:03PM
(#1072319)
Yes, lightspeed is a bottleneck. A particular problem is that the computing elements are presumably vulnerable to radiation damage, thus placing a lower limit on the internal radius of the MB's innermost shell. If MBs are migratory, that's a good reason for expecting them to be associated with cooler, dimmer red dwarf stars -- or to be concentrated some way away from their stellar energy source, leaving a classical Dyson sphere in place to harvest energy and beam it to them via laser or maser.
(This stuff got chewed over a lot on the extropians list in the early 90s, when the subject of "Jupiter Brains" -- what you get if you dismantle Jupiter and turn it into computing material -- was first raised. Googling on "neuron star" might provide some transitory amusement ...)
Re:Matrioshka Brains: computing at the limits of physics
by khafra on Saturday January 23, 2010 @08:34PM
(#1072780)
I found your exposition on the other limit--bandwidth outside the brain--even more interesting. Assuming life expands to fill the available resources, once you create a Matrioshka brain you're pretty irrevocably tied to your star, for good or for ill.
Re:Matrioshka Brains: computing at the limits of physics
by eru on Sunday January 24, 2010 @11:38AM
(#1073425)
You might still want to send out `seeds' to other stars. Even if you can't communicate with them fully.
Re:Matrioshka Brains: computing at the limits of physics
by dhume on Saturday January 23, 2010 @01:48PM
(#1072214)
If you're pooling enough people's resources to make a machine that big, you're probably going to get enough independent tasks to make good use of it too.
Re:Matrioshka Brains: computing at the limits of physics
by Confusion on Sunday January 24, 2010 @06:00AM
(#1073200)
I think that limitation would become a major bottleneck long before you reach Dyson sphere size. Internode communication is already the major bottleneck for modern supercomputers, which limits the possibly achievable computational speed for every problem except for strictly parallellizable problems that do not require such internode communication.