Notes for the Lifebox, the Seashell, and the Soul - Rudy Rucker

Notes for The Lifebox, the Seashell, and the Soul, by Rudy Rucker
almost bizarre that the conclusions I ended up reaching had never been reached before.” 5
And then he gives some thought to why complexity has so long been ignored.
One reason seems to be that in normal technology we figure out what we want a
system to do, and then reverse engineer it, and to be sure it will work, we stick to very simple
and predictable systems. But nature is under no such constraint. It’s in fact common for
natural systems to achieve their effects by the unpredictable accumulation and interaction of
a myriad of small computational steps. A second reason why many of Wolfram’s ideas are
historically new is that it’s only the coming of computers that made his investigations
possible. Historically, people simply never in history looked at things like Rule 30. The
most complex patterns one tends to see in earlier art and architecture are fairly regular
fractals three or four levels deep, such as, for instance, the curlicues of Celtic illuminations or
iron-work vines.
Parallel
But in point of fact, word-processing, emailing, and Web browsing all work just fine
on single-processor machines. Electronic parallel computers are being built, but only for
very intense supercomputing tasks like weather prediction or, depressingly, for nuclear
weapons simulation.
***
And if there’s more than one process, are they all copies of the same process, or are
they different from each other?
Programming parallel machines is in fact much easier if you synchronize the
processors and only let them write to nearby memory. If you let the processes run at their
own speeds and read and write wherever they like, the situation can get exceedingly hairy.
***
I saw my first parallel hardware machine like this in 1984 at a Naval Research Center
in Beltsville, Maryland. I’d been invited there to give a talk on the fourth dimension ⎯
maybe the scientists had some faint hope I might have an idea for a faster-than-light
hyperdrive! At the time, I was just getting interested in cellular automata such as the game of
Life; cellular automata are a particularly simple kind of parallel computation. In fact I was
on my way back from interviewing Wolfram and the cellular automata wizards Normal
Margolus and Tommaso Toffoli for an article that ended up appearing in, disappointingly for
me, a science fiction magazine.
In Beltsville, one of my hosts let me look through a glass window at an entire room
full of computer chips, constituting a “Massively Parallel Processor” which was used to
process terabyte arrays of satellite photo data, for instance removing stray “turd bits” by
averaging pixels with their neighbors. I asked if he’d ever run the Game of Life on it, but he
said no.
In today’s paper, I see that the Lawrence Livermore labs are building a
supercomputer called Thunder which will use several thousand Intel microprocessors in a
parallel architecture. The machine is expected to run some ten thousand times as fast as an
ordinary PC. Thunder won’t quite match the power of today’s fastest supercomputer, a
Japanese machine called the Earth Simulator, which is used for simulations of climate
change, such as global warming. America’s Thunder machine, on the other hand, will be
5 ibid, p. 22.
p. 69