Re: OT: CPUs

On May 17, 1:42 pm, r norman <r_s_norman@xxxxxxxxxxxx> wrote:
On Sat, 17 May 2008 10:27:34 -0700 (PDT), chris thompson



<chris.linthomp...@xxxxxxxxx> wrote:
On May 17, 12:03 pm, Vend <ven...@xxxxxxxxxxx> wrote:
On 17 Mag, 16:59, r norman <r_s_norman@xxxxxxxxxxxx> wrote:

On Sat, 17 May 2008 07:37:20 -0700 (PDT), Vend <ven...@xxxxxxxxxxx>
wrote:

<snip stuff about CPUs>

When I was studing for my Computer Architecture exam, I realized that
there is a nice coevolutionary relationship between hardware and
software:
Software developers try to optimize their software to obtain the best
performances on the current cpus, while cpu designers try to optimize
their cpus to obtain the best performances with current software.

Interestingly, this doesn't seem to converge to a stable design
strategy but rather yields oscillating trends: (complexity in hardware
vs. complexity in software, serial execution vs. parallel execution,
small memory footprint vs large memory footprint, etc.)

Do phenomena like this happen in biological coevolution too?

Perhaps you mean something like the difference between what appears to
be the rather genetically constrained (instinctive) behavior of
arthropods = complexity in hardware vs. a predominance of learned
behavior of mammals, especially primates = complexity in software. I
don't see that biology can easily switch back and forth with the
oscillations that you suggest; once you get set in a particular
pattern of development, you stick with it until you go extinct and
allow a different branch of evolution to try the alternative.

Incidentally, the contrast I suggest above is very imperfect: many
arthropods are good at various forms of learning and we mammals
including humans have a great deal of "prewired" aspects to our
behavior.

I was referring to oscillation between the traits of two or more
species which interact with each other.

The classic example is coevolution between predator and prey species.
Deer have an extremely useful spring-like mechanism in their ankles
that gives them a boost when running. Reducing the number of digits
also helps, as does lengthening the legs. Predators like big cats
(all cats, really) have what's called a floating scapula that isn't
anchored, and serves to add some power to the stride. The collarbone
is much reduced or even absent in some species, thus removing a
potential obstacle to movement of the scapula. Canids also
lengthened their limbs, mainly by lengthening the feet, though, more
than the legs themselves.

Now for all this to be valid as coevolution, it would have to have
happened in some kind of sequence, and that I cannot be sure happened.
I'd bet someone here could tell us though.

Here's a nice site discussing adaptations for speed:

http://animaldiversity.ummz.umich.edu/site/topics/mammal_anatomy/runn...

or

http://tinyurl.com/5hrmwy

There certainly are numerous examples of coevolution in predator-prey
or herbivore-plant interactions. But the best examples are in
cooperative ventures like the coevolution of flowers and pollinating
animals. However these are usually directional in nature leading to a
highly specialized endpoint, not oscillatory.

I know this kind of runs contrary to the usual thinking, but calling
pollination 'cooperative' has always seemed kind of Panglossian to me.
If those animals could get what they wanted without the added
responsibility of being an unwitting menage-a-trois member I bet they
would. And sure, repopulating the plant populations is in their best
interest, but that's another turn onto Teleology Street.

In any case, I'm not entirely unsure that contradicts Vend's intent.
I think, on rereading that post, that Vend was referring to
alternating strategies- predator gets faster, prey gets
faster...predator develops attack from the rear, prey's eyes move to
side of head (OK that one's shaky but you get the message). I don't
think Vend meant an actual back-and-forth gain/loss of traits. We
don't, after all, see software packages getting smaller in response to
more powerful CPUs, which is what the oscillation would entail in the
that sense.

Chris

.

Relevant Pages

  • Re: OT: CPUs
    ... their cpus to obtain the best performances with current software. ... strategy but rather yields oscillating trends: (complexity in hardware ... The classic example is coevolution between predator and prey species. ...
    (talk.origins)
  • Re: OT: CPUs
    ... strategy but rather yields oscillating trends: (complexity in hardware ... The classic example is coevolution between predator and prey species. ...
    (talk.origins)
  • Re: WSJ article on software liability
    ... >) hardware is simpler than that of software. ... >) hardware chips interacting in highly complex ways and yet they have ... of behavioral complexity, almost all hardware systems remain logically ...
    (comp.programming)
  • Re: WSJ article on software liability
    ... >)>size and complexity of a software emulator of a chip. ... Does that make hardware any less complex? ... A software emulation of hardware is a lot more stable ... Hardware failure is so much more expensive than software failure. ...
    (comp.programming)
  • Re: WSJ article on software liability
    ... >>'' My point is that an algorithmic software system of a given complexity ... >> is much more unstable than an equally complex hardware system. ... >>are either using the assumption that they are of equal complexity, ... hardware design vs. software design. ...
    (comp.programming)
Loading