Why don't we nuke Rita?

amit wrote:
Kurgan Gringioni wrote:

It's proven.

In nonlinear systems, small perturbations may yield large scale
fluctuations.

dumbass,

that's a big generalization. it's not true for every nonlinear system
or every perturbation to a chaotic system.

Did you overlook the word _may_?

Kurgan Gringioni wrote:
Our group concluded that such a machine would be infeasible for the
fluids problem. It was impossible to design a simple cheap chip that
would handle the cascading Fourier transforms. The Fouriers were just
too complex compared to the factoring or the gravitational attraction
calculations made in the aformentioned one-off specialty machines. I'm
sure there have been many other groups studying that method that have
come to the same conclusion because if someone were ever able to make a
cheap, superfast fluids dedicated computer, they'd rake in the cash.

Clown****er,

You guys would have been right to bail, even if your approach had
shown promise. The history of supercomputing is that no company
has been able to show consistent profitability in that segment
since the late 1980s. The ones I'm familiar with are having
problems making payroll right now.

But mostly I wanted to call someone a clown****er. I don't respond
to Kunich any more so that made you the next best candidate.

Thanks,

Bob Schwartz

Bob Schwartz wrote:
Kurgan Gringioni wrote:
Our group concluded that such a machine would be infeasible for the
fluids problem. It was impossible to design a simple cheap chip that
would handle the cascading Fourier transforms. The Fouriers were just
too complex compared to the factoring or the gravitational attraction
calculations made in the aformentioned one-off specialty machines. I'm
sure there have been many other groups studying that method that have
come to the same conclusion because if someone were ever able to make a
cheap, superfast fluids dedicated computer, they'd rake in the cash.

Clown****er,

You guys would have been right to bail, even if your approach had
shown promise. The history of supercomputing is that no company
has been able to show consistent profitability in that segment
since the late 1980s. The ones I'm familiar with are having
problems making payroll right now.




Dumbass -

It wouldn't have been a "supercomputer". It would've only been able to
solve one very specific type of problem.


thanks,

K. Gringioni.

Kurgan Gringioni says...

If computers can do that, then why don't they use computers to model
airflow over cars and automobiles rather than windtunnels?

You can admit it or keep sticking your head in the sand: complex problem
solution depends upon the evolution not only of the hardware, such as computers,
but also the software and, probably more importantly, the experimental design
and methodology. You seem to believe that the only way to pound a nail faster
is to get a larger hammer.

The answer is: even today's fastest supercomputers aren't fast enough.
It's *because* of the "Butterfly Effect". Small scale perturbations can
and will lead to large scale fluctuations.

Yes, yes, I understand the theory. You're so blinded by it, you can't see that
REAL improvements in weather forecasting are happening in the here and now. I'm
not saying that weather forecasting is anywhere near perfect, far from it. But
just look at Katrina - Louisiana residents were warned 4 days in advance to
evacuate - many didn't and look at the results. Would a pretty butterfly
wingflap simulation have been more convincing to those folks that chose not to
evacuate? I doubt it.

BTW, I may be a loudmouth newsgroup spraybag, but I was also part of a
group that did a feasibility study on making a machine that was
specifically dedicated to doing those fluids simulations.

If this is where you think I'll bow down in awe of your background, think again.
I work with a group that is actually doing research (and suceeding, even if
slower than desired) to develop new software and hardware for weather
forecasting. And they aren't searching for butterflies to explain the recent
weather events in the gulf.

In article
. com,
"amit" wrote:

Kurgan Gringioni wrote:

It's proven.

In nonlinear systems, small perturbations may yield large scale
fluctuations.

dumbass,

that's a big generalization. it's not true for every nonlinear system
or every perturbation to a chaotic system.

I'll jump in here. Some systems of differential equations
in some portions of their solution sets exhibit
sensitivity to initial conditions. Two integrations of
the differential equations with slightly different initial
values quickly diverge. The rate of the divergence is
proportional to the size of the difference in the initial
conditions. This dependence of rate on size of the
changing quantity is the hallmark of exponential growth.

The question is: why do we run numerical simulations on
systems we know to manifest sensitivity to initial
conditions?

Hint: two words, initials S L.

The interested reader may peruse

John Guckenheimer, Philip Holmes - Nonlinear Oscillations,
Dynamical Systems, and Bifurcations of Vector Fields
(Applied Mathematical Sciences Vol. 42)
Springer-Verlag (ISBN: 0387908196)

--
Michael Press

Mad Dog wrote:
If this is where you think I'll bow down in awe of your background, think again.
I work with a group that is actually doing research (and suceeding, even if
slower than desired) to develop new software and hardware for weather
forecasting. And they aren't searching for butterflies to explain the recent
weather events in the gulf.

Cool.

You guys got enough money to keep the hardware part in business?
They look to be in some pretty deep ****.

Bob Schwartz

Bob Schwartz says...

You guys got enough money to keep the hardware part in business?

It's easier to buy toys than it is to hire people to work with them. Some day,
there will be billions of dollars worth of instrumentation gathering dust in
labs across the country. Look for me as a Walmart greeter - flipping burgers
makes me nauseous.

Mad Dog wrote:

If this is where you think I'll bow down in awe of your background, think again.
I work with a group that is actually doing research (and suceeding, even if
slower than desired) to develop new software and hardware for weather
forecasting. And they aren't searching for butterflies to explain the recent
weather events in the gulf.



Dumbass -

All you can offer is educated guesses and that's all weather
forecasting will ever be able to offer. The guesses will just become
more educated, but come on. The weatherman is still wrong often and
understandably so.

thanks,

K. Gringioni.

Kurgan Gringioni says...

All you can offer is educated guesses and that's all weather
forecasting will ever be able to offer. The guesses will just become
more educated, but come on. The weatherman is still wrong often and
understandably so.

Turdgun Butterflyholio:

The 4-day advance notice given to the residents of New Orleans was a pretty
functional educated guess. And it turned out to be pretty accurate. So go
ahead and bitch about not having enough computer power to crunch your FTs and
flap your little butterfly wings till the cows come home. Some of the rest of
us will try to do something productive.

Mad Dog wrote:
Kurgan Gringioni says...

All you can offer is educated guesses and that's all weather
forecasting will ever be able to offer. The guesses will just become
more educated, but come on. The weatherman is still wrong often and
understandably so.

Turdgun Butterflyholio:

The 4-day advance notice given to the residents of New Orleans was a pretty
functional educated guess. And it turned out to be pretty accurate. So go
ahead and bitch about not having enough computer power to crunch your FTs and
flap your little butterfly wings till the cows come home. Some of the rest of
us will try to do something productive.



Dumbass -

No one ever said that weathermen didn't do anything useful, but the
fact remains that they make educated guesses.

BTW, you never did address the fact of why
expensive-to-use-and-maintain windtunnels are used to measure airflow
over small objects like automobiles rather than modeling it on a
computer.

You know the answer, but it doesn't sink in, does it?


thanks,

K. Gringioni.