If the human mind had CPU architecture, what would it be like and how many MHz?

[Lorn]

Suppose one's brain power could be translated into a modern processor form, how might it be represented?

Ideas?

 

[nageov3t]

I think the OP is running on a single P1 100MHz

 

[crumpet19]

Originally posted by: loki8481
I think the OP is running on a single P1 100MHz

only when its on turbo

 

[So]

Can't compare.

What MHZ does an analog integrator run at?
How about an analog sound wave? How many 'bits' does it REALLY have?

 

[SilentZero]

Originally posted by: loki8481
I think the OP is running on a single P1 100MHz

I was thinking along the 1mhz cpu in the Apple IIe

 

[vital]

do you know how fast a 100MHz cpu can calculate numbers compared to a human brain? there is no comparison.

 

[JoeKing]

a few terahertz... hyperthreaded!

 

[Lorn]

Originally posted by: vital
do you know how fast a 100MHz cpu can calculate numbers compared to a human brain? there is no comparison.

What if all of our senses held numerical values?

 

[So]

Originally posted by: vital
do you know how fast a 100MHz cpu can calculate numbers compared to a human brain? there is no comparison.

Hah. Throw a ball up in the air. Point to the spot where it will reach the apex of it's travel.

Congrats, you just did hundreds of advanced calculus equations in your head in extremely
quickly.

We aren't evolved to do the symbolic math that we invented a mere handful of centuries ago. It's like emulation, we have to run symbolic mathematics on the parts of our brain that evolved for logic and decision making, which are not efficient at mathematics as we know think of it.

 

[Koenigsegg]

Originally posted by: loki8481
I think the OP is running on a single P1 100MHz

He can do 100 million calculations in one second?

More realistically it'd be somewhere around 1hz. A calculation a second seems like an accurate estimate. The slowest CPUs ever made can still do calculations MUCH MUCh faster than the fastest human brain ever can. Why do you think computers were invented in the first place

You can also bring up the idea of memory being like a hard drive and consider access time. Since the brain can recall things relatively quickly, we'd be more comparable to hard drives with our memories than with CPUs and our processing abilities.

 

[bleeb]

Learning computer

 

[Ricemarine]

We operate at a hertz!

Like most of people around the world that don't know about computers, we might fall at half a hertz, or .01 hertz and belong to a company named Intel

Course, Einstein probably had 2 hertz.

 

[So]

quoted from New scientist, since it requires a subscription:

The Math Instinct by Keith Devlin

CAN dogs do calculus? Well no, not the way college students do it. For one thing, dogs usually get the right answer. When chasing a ball thrown into a lake, dogs run along the beach and then choose just the right point to jump into the water to reach the ball in the shortest time. To solve this problem on paper takes time and calculus, but canine brains seem to work it out swiftly and instinctively.

Keith Devlin asserts that there are two kinds of maths: natural and symbolic. Natural maths has evolved over millions of years, giving both animals and humans incredible mathematical abilities, each linked to specific survival needs, such as navigation or catching prey. Symbolic maths is unique to humans and is at most 3000 years old. Our brains have not had time to evolve specialised structures to perform symbolic maths, so we have to co-opt abilities that evolved for other purposes. He argues that this is why so many people struggle with the abstract rules they are required to learn at school.

Devlin spends the bulk of The Math Instinct describing the hard-wired abilities of animals. The Tunisian desert ant finds its way around by dead reckoning, which requires precise measurements and computations. Lobsters can navigate by sensing the Earth's magnetic field as accurately as if they were connected to the GPS network.

If animals can achieve all this, why does the average human have so much trouble with arithmetic? According to Devlin, we don't, as long as it relates to a concrete problem, such as picking the best buy at the supermarket or giving the correct change. But devoid of context, the rules for manipulating symbolic representations of numbers make little sense to most people, who begin to make mistakes they never would while shopping.

How does Devlin's research help in the classroom? I did rather wonder. We do not want our students feeling inferior to lobsters. Fortunately he ends on a positive note. If you understand how symbolic concepts connect to the natural maths you can already do, then it all boils down to practising until those abstract rules take on a more concrete reality in your mind. Back to those times tables, then.
From issue 2494 of New Scientist magazine, 09 April 2005, page 47

 

[So]

clearly, nobody is interested in an intelligent discussion.

<cartman>Screw you guys, I'm going home</cartman>

Alright. :moon: all. Hasta manana.

 

[suicidalpigeon]

shouldnt this be in the highly technical section? and if our brains/body did compare to numbers, can i oc myself? what kind of steroids do you recommend? will my heart provide enough blood to my head?

edit: oh and i will be using this for school work, partying, and basketball.

 

[BillyBobJoel71]

we can not calculate as fast as a pentium i386, by far. pentium 4 and a64 chips can calculate 32 million digits of pi in less than an hour, using just a single algorithm. we can't even do 10.

 

[SketchMaster]

Originally posted by: So
quoted from New scientist, since it requires a subscription:

The Math Instinct by Keith Devlin

CAN dogs do calculus? Well no, not the way college students do it. For one thing, dogs usually get the right answer. When chasing a ball thrown into a lake, dogs run along the beach and then choose just the right point to jump into the water to reach the ball in the shortest time. To solve this problem on paper takes time and calculus, but canine brains seem to work it out swiftly and instinctively.

Keith Devlin asserts that there are two kinds of maths: natural and symbolic. Natural maths has evolved over millions of years, giving both animals and humans incredible mathematical abilities, each linked to specific survival needs, such as navigation or catching prey. Symbolic maths is unique to humans and is at most 3000 years old. Our brains have not had time to evolve specialised structures to perform symbolic maths, so we have to co-opt abilities that evolved for other purposes. He argues that this is why so many people struggle with the abstract rules they are required to learn at school.

Devlin spends the bulk of The Math Instinct describing the hard-wired abilities of animals. The Tunisian desert ant finds its way around by dead reckoning, which requires precise measurements and computations. Lobsters can navigate by sensing the Earth's magnetic field as accurately as if they were connected to the GPS network.

If animals can achieve all this, why does the average human have so much trouble with arithmetic? According to Devlin, we don't, as long as it relates to a concrete problem, such as picking the best buy at the supermarket or giving the correct change. But devoid of context, the rules for manipulating symbolic representations of numbers make little sense to most people, who begin to make mistakes they never would while shopping.

How does Devlin's research help in the classroom? I did rather wonder. We do not want our students feeling inferior to lobsters. Fortunately he ends on a positive note. If you understand how symbolic concepts connect to the natural maths you can already do, then it all boils down to practising until those abstract rules take on a more concrete reality in your mind. Back to those times tables, then.
From issue 2494 of New Scientist magazine, 09 April 2005, page 47

That made me feel really smart!

 

[bleeb]

The complexity of the human mind cannot be matched with computer technology... yet. I suggest a neural net processor learning at a geometric rate.

 

[Ns1]

20 P4 3.2ghz underclocked to p133

 

[BillyBobJoel71]

our brains can perceive emotion through sensors, and they can calculate many things a cpu can never do (how pretty a girl looks, for instance). however, both run on electric currents.

 

[vshah]

now the real question is: if i put a "64 bit" case badge on my forehead, does that add 10mhz?

 

[JoeKing]

I was looking for a link to an article discussing this, but came across this great post by Jothaxe a loooooong time ago.

In some senses it is very helpful to compare the brain to a CPU and think of it like a processor.

But examined in detail this is not always helpful, because the algorithms the brain uses for "processing" information are not performed like those of a CPU.

Specifically the brain uses highly parallel methods for nearly all lower level thinking. The closest thing that a brain has to a "cycle" is the time it takes for a synapse to fire, but since this doesnt happen in series, it isnt very cyclical.

An interesting sidenote: it has been learned that the human mind is able to perform visual recognition tasks in times as short as 1/20 of a second. Based on synapse firing times, it has been determined that this leaves time for no more than 100 steps in a neural-net sequence. This has been labeled the "100 step rule."

Compare this to the way we would teach a computer to perform a visual recognition task: scan the entire image, performing complex edge detection routines. Using this information, determine those feature which are typically part of larger scale stuctures. Then perform a statistical analysis of these features to determing if they appear in a likely "constellation." This entire algorithm may contain many millions of sequential steps.

The brain < 100 steps
A CPU > 1000000 steps (maybe closer to 1000000000 for a really accurate routine...)

This just gives you an idea of how differently the brain operates from a CPU. The amazing thing is that we are starting to build CPU's that can compete on the same time scales as the brain, doing things sequentially.

Hopefully neuroscience will advance in the next few decades to the point that we really understand what is going on in the brain to the point that we can implement it in machines. I want to be alive to see it...

 

[sundev]

Our brains suck at "real" calculations (e.g. 8435623 x 32149234^1534), but are good at other things as people have already mentioned - e.g. if you're playing outfield in a baseball game and a pop-fly comes in your direction.. you have to do tons of calculations to know where to go, how fast, where the ball will land and when, etc. etc. All the while you're mindful of things like how fast you can move, your reach, gravity, blah blah blah. Real computers would probably take forever to figure that out.

 

[EyeMNathan]

The human brain is the perfect example of an extremely complex analog computer.

 

[HDTVMan]

Im sitting next to a Pentium 60 with math bug.