Re: OT: CPUs

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/running_fast.html

or

http://tinyurl.com/5hrmwy

Chris

.

Relevant Pages

  • Re: OT: CPUs
    ... strategy but rather yields oscillating trends: (complexity in hardware ... There certainly are numerous examples of coevolution in predator-prey ...
    (talk.origins)
  • Re: OT: CPUs
    ... <snip stuff about CPUs> ... strategy but rather yields oscillating trends: (complexity in hardware ...
    (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: OT: CPUs
    ... <snip stuff about CPUs> ... their cpus to obtain the best performances with current software. ... strategy but rather yields oscillating trends: (complexity in hardware ...
    (talk.origins)
  • Re: WSJ article on software liability
    ... >> It is easy enough to compare complexity between software and hardware ... >> more reliable tan software regardless of the complexity. ... >hardware design vs. software design. ... My CPUs have not failed on me ...
    (comp.programming)