[unatkot]

Energy is not power. Any writer that does not know this should not be writing about it, it really brings the validity of the entire article down.

I'm fairly certain the editor didn't use the term "power" in the physical sense, just simply as a term for the amount of energy it takes to "power" 10,000 houses for a full year.

Total average household power usage in Southern CA:

6,000 kWh electricity plus 12,000 kWh gas equivalent = 18,000 kWh per year.


Of course, the editor did leave out how long the computer needs to be running to consume that much energy. That information certainly would have helped.

[erubresen]

Of course, the editor did leave out how long the computer needs to be running to consume that much energy. That information certainly would have helped.

That little element of time that you point out there is [Doc Voice] precisely [/Doc Voice] the difference between energy and power I was driving at before.

A technically correct statement -in terms of units, not magnitude- would be:

"consumes enough electricity to power about 10,000 homes" (power to power comparison)

Or

"consumes enough electricity in some time period to power about 10,000 homes for a full year" (energy to energy comparison)

[Immonnaornach]

not only that:
1) they cant cure cancer with it
2) they cant cure aids with it,
3) they cant cure any dicease with it,
4) they can not prodict anything with it
5) they CAN provide left wing associations their right in thinkin,
6) they CAN show later on their thinking was wrong!

the only thing a supercompater is required is to kinf of the hill, max e-penis size, i am taller than you! Na-na-na-na-naah

Are you stupid, or just playing stupid?

Pure number crunching power has a lot of useful applications - not just for the use of the owning company, but for leasing to other companies, institution, etc, who can run massive problem in a short time, rather than the ages it may take on a lesser computer, the reduction in research and development time can be worth millions to billions over a year.


OP, I assume you meant "the next five top supercomputers", as this one would have been included in that title, otherwise.

[unatkot]

That little element of time that you point out there is [Doc Voice] precisely [/Doc Voice] the difference between energy and power I was driving at before.

A technically correct statement -in terms of units, not magnitude- would be:

"consumes enough electricity to power about 10,000 homes" (power to power comparison)

Or

"consumes enough electricity in some time period to power about 10,000 homes for a full year" (energy to energy comparison)

Exactly. Not disagreeing with you, just confused as to why the editor left out the obvious detail of time x energy consumption.

[adolfadsermens]

3) they cant cure any dicease with it,

You do realise that is exactly what F@H is intended to do; and F@H is essentially one big super computer, albeit one made up of millions of PCs.

4) they can not prodict anything with it Wrong again. Computers can be used to predict lots of things, the weather for example.

[erubresen]

Wrong again.

Yes you are.

Models can produce, based on our current understanding, the outcome of a system given the current (and past) inputs. We are sometimes bad at determining the current output of a system, let alone the future.

This is a non-causal world, plain and simple. The path of a kicked football is pretty predictable in an environment with no atmosphere, but as soon as air, spin, wind, surface imperfections, moisture, materials and speed are introduced, the predictability nosedives past a few hundred milliseconds into the future. Computers can be used to predict lots of things, the weather for example. That has to be the worst example possible. The randomness of weather is unfathomable. It is harder* to predict than the value of a stock, or the next digit of pi.

Imagine the size of the transfer function needed to model current temperature. Think about all the inputs that one has to consider. What about temperature one minute from now? A near infinite amount of probabilistic equations have to be considered, it becomes a combination of a lot of statistics.

Weather is a lot more than current temperature. Similar modeling has to be done for wind speed, humidity, precipitation and all the other components that comprise of definition of weather. Good luck.

*- For those of you allowed to leave your basement: Before you crucify me, this information is based on an article I read which considered the probability that a correct guess could be made. Usually limited to the number of possible outcomes and number of outputs to consider, etc.

[adolfadsermens]

Yes you are.

Models can produce, based on our current understanding, the outcome of a system given the current (and past) inputs. We are sometimes bad at determining the current output of a system, let alone the future.

This is a non-causal world, plain and simple. The path of a kicked football is pretty predictable in an environment with no atmosphere, but as soon as air, spin, wind, surface imperfections, moisture, materials and speed are introduced, the predictability nosedives past a few hundred milliseconds into the future.That has to be the worst example possible. The randomness of weather is unfathomable. It is harder* to predict than the value of a stock, or the next digit of pi.

Imagine the size of the transfer function needed to model current temperature. Think about all the inputs that one has to consider. What about temperature one minute from now? A near infinite amount of probabilistic equations have to be considered, it becomes a combination of a lot of statistics. And weather is a lot more than current temperature. Similar modeling has to be done for wind speed, humidity, precipitation and all the other components of weather. Good luck.

*- For those of you allowed to leave your basement: Before you crucify me, this information is based on an article I read which considered the probability that a correct guess could be made. Usually limited to the number of possible outcomes and number of outputs to consider, etc.

Why don't you ring what ever your local equivalent of the Met Office is and ask how they come by their forecasts?


Edit: http://www.metoffice.gov.uk/news/in-...supercomputers

[erubresen]

Why don't you ring what ever your local equivalent of the Met Office is and ask how they come by their forecasts?

Edit: http://www.metoffice.gov.uk/news/in-...supercomputers

The world is a scary place full of magic and predictions.

Never leave your mother's basement.

[bumxumer]

The world is a scary place full of magic and predictions.

Never leave your mother's basement.

I'm pretty sure he stays in the garage.

[adolfadsermens]

The world is a scary place full of magic and predictions.

Never leave your mother's basement.

In other words you don't have an intelligent answer to my previous post. As usual you're more interested in being an arse to me.

[Markdogas]

Yes you are correct in saying that weather is extremely hard to predict well, it does not mean that we don't try. I am not a meteorologist but during my Physics degree we looked at chaos theory and non-linear system. Planetary and stellar atmosphere exhibit this sort of behaviour, which means they are extremely sensitive to initial conditions and can behave radically different from one moment to the next. Due to these effects one has no choice but to simulate the behaviour to figure out the outcome and you would use super-computers to do that.

Personally I make use of the GRID to analyse data collected by the ATLAS experiment at CERN, it would be impossible to do my job without the help of it or other computer farms around the world just like it. So no, these facilities are not just about e-penis you ignorant moron!

This is pretty impressive but I really want to see quantum computing, and not just because my department is doing research into that.

[Centurnion]

I want to install F@H on it, and then run outside the building and watch the electric meters go nuts.

[erubresen]

In other words you don't have an intelligent answer to my previous post. As usual you're more interested in being an arse to me.

Lack of reading comprehension on your part does not necessitate a response from me.

Point is: Simulation is not prediction, prediction is magic.

Probabilistic models are best guesses. Dutch's point about not being able to "predict" things is silly, I agree with you. But, do not give "forecasting" more credit than it deserves. It is still a massive amount of probabilistic models and a best guess effort, that is all. The computers and the software were designed by humans (whom also do not have the ability to predict anything).

Here is an example: I predict that you will post something that is reductive and short-sighted before the end of the day. I can write a program that posts this prediction every day on these forums and be correct 100% of the time. This is not prediction, this is probability. But, in all cases, you are still a jackass.

[Immonnaornach]

Define "prediction"...
In most cases I'm familiar with, it's an assessment of what is expected to happen, often as a percentage or proportional probability.
I can predict, with confidence, that if I release a knife above my kitchen table, and there's nothing preventing it, then the knife will fall and hit the table. I cannot say it will with 100% confidence as something may occure between the release and before it can hit the table that will prevent it, such as an asteroid hitting the house or an earthquake, but

[erubresen]

It is pedantic, I admit. But it is more like this:

If I throw a ball towards Bungle's face 100 times from 50 yards in a random field of wind, what happens? Repeating this a tremendous amount of times you will see that most of the time it will miss, but some times when the wind is just right, it will knock him out cold and solve world stupidity. Apply the same concept to asteroids in space or something less useful that hurting old Bungleroo. An advantage is the repeating process, but it is still code, still written by humans and still not magic.

I am all about super-computing, but completely against woo woo.

[FsQGF1Mp]

I agree that supercomputers make alot of calculations worth doing them.
But if people are going to claim it will help diceases, predict earthquakes or weather I think you are too naive.

You may call me stupid, i don't care Gordo.

[new-nickname-zanovo]

Yes you are.

Models can produce, based on our current understanding, the outcome of a system given the current (and past) inputs. We are sometimes bad at determining the current output of a system, let alone the future.

This is a non-causal world, plain and simple. The path of a kicked football is pretty predictable in an environment with no atmosphere, but as soon as air, spin, wind, surface imperfections, moisture, materials and speed are introduced, the predictability nosedives past a few hundred milliseconds into the future.That has to be the worst example possible. The randomness of weather is unfathomable. It is harder* to predict than the value of a stock, or the next digit of pi.

I would love to see you use this argument against the sniper who has the world record for the longest distance confirmed kill.

Just sayin.

[erubresen]

I would love to see you use this argument against the sniper who has the world record for the longest distance confirmed kill.

Just sayin.

I would love to see you use context to make your point.

With the sniper, you better believe that luck* was involved, luck was essentially the factor that compensated for the chaos.

Also, the sniper was probably Russian or Finnish and I do not interact with those people.

* - By "luck" I mean, an ideal set of outputs from the probabilistic models.

[new-nickname-zanovo]

I would love to see you use context to make your point.

With the sniper, you better believe that luck* was involved, luck was essentially the factor that compensated for the chaos.

Also, the sniper was probably Russian or Finnish and I do not interact with those people.

* - By "luck" I mean, an ideal set of outputs from the probabilistic models.

I think he was British actually.

My only point is he had to overcome a lot of variables you specifically mentioned. Don't read too deep into it. It's merely an example of how you can obtain an intended result in a scenario with an unimaginable number of variables.