please wait
- Forum
- General C++ Programming
- Processing Power and Electricity
Processing Power and Electricity
Sep 21, 2011 at 8:54pm
He everyone,
Maybe I didn't come to the right place to ask this question, but I was wondering if anybody knew. If electricity travels so fast and at a constant speed, why do different CPU's in different computers work at different speeds?
I would appreciate any help offered,
Muhasaresa
Maybe I didn't come to the right place to ask this question, but I was wondering if anybody knew. If electricity travels so fast and at a constant speed, why do different CPU's in different computers work at different speeds?
I would appreciate any help offered,
Muhasaresa
Sep 21, 2011 at 9:31pm
The speed of light is constant too (in a vacuum anyway) but I can make a light flash faster by flicking a switch faster. Same thing.
Sep 21, 2011 at 10:22pm
This is NOT by any means a simple question.
- The speed of light in the sense of the speed of a photon (coincidently notated as c) is constant, period. It is not influenced by the density of the medium it travells through.
- Electricity does NOT travel at the speed of light, and it is in fact dependent on the medium it travels through. On a copper wire it travells at roughly two thirds of c. When it finally gets to the components another factor called Propagation Delay comes into play. This is the time it takes for elelctronic components to "react" to the electricity that was applied to them.
- The "speed" of the components that you are refering to has nothing to do with the speed at which electricity or light travel. The measurement of "speed" in the case of computer components refers to the frequency at which the components operate. Usually this frequency is limited by the material that the component is made out of, if the specific composition of the material is unable to oscillate (change states) fast enough then the result will be loss of data. In addition to this we have another consideration because of the Nyquist–Shannon sampling theorem, in order for us to recieve all of the data on a line that is being fed to us we need to "read" that line at at least twice the rate that the data is coming in.
- Another factor that limits the speed, one that is a part of the Propagation Delay I mentioned above, at which a component can operate is it's capacitence. Because electricity does NOT travel at the speed of light, it takes a certain amount of time for a 'bit' to go from 1 to 0. In a single instance this is not a significant amount of time, the formula for the rate is <Time = Resistence * Capacitance> (notice that the amount of electricity applied, voltage or amperage, here does NOT come into play, this is CRITICAL for understanding the current industry trends) but if you are planning on doing this a few million times a second on the same data path then even this small variable comes into play. In order to reduce this amount of time manufactorers are using smaller and smaller data paths on their components, the idea here is that a physically smaller amount of material would require a smaller amount of time to switch from a high (1) state to a low (0) state or visa versa. BUT they are limited again in how small they can go by an attribute of the components called the "Trigger Voltage", that is the minimum amount of power required before the component sees a difference between a high and a low state because if the data path is too small then it burns up.
I haven't even began to scratch the surface of this topic. As much as I enjoy talking about it I can't get myself to focus on one subject long enough to adequitly explain to you how much it plays into the grand scheme of things. I hope I gave you enough to start some research on your own though as this is a very interesting topic for geeks like me.
- The speed of light in the sense of the speed of a photon (coincidently notated as c) is constant, period. It is not influenced by the density of the medium it travells through.
- Electricity does NOT travel at the speed of light, and it is in fact dependent on the medium it travels through. On a copper wire it travells at roughly two thirds of c. When it finally gets to the components another factor called Propagation Delay comes into play. This is the time it takes for elelctronic components to "react" to the electricity that was applied to them.
- The "speed" of the components that you are refering to has nothing to do with the speed at which electricity or light travel. The measurement of "speed" in the case of computer components refers to the frequency at which the components operate. Usually this frequency is limited by the material that the component is made out of, if the specific composition of the material is unable to oscillate (change states) fast enough then the result will be loss of data. In addition to this we have another consideration because of the Nyquist–Shannon sampling theorem, in order for us to recieve all of the data on a line that is being fed to us we need to "read" that line at at least twice the rate that the data is coming in.
- Another factor that limits the speed, one that is a part of the Propagation Delay I mentioned above, at which a component can operate is it's capacitence. Because electricity does NOT travel at the speed of light, it takes a certain amount of time for a 'bit' to go from 1 to 0. In a single instance this is not a significant amount of time, the formula for the rate is <Time = Resistence * Capacitance> (notice that the amount of electricity applied, voltage or amperage, here does NOT come into play, this is CRITICAL for understanding the current industry trends) but if you are planning on doing this a few million times a second on the same data path then even this small variable comes into play. In order to reduce this amount of time manufactorers are using smaller and smaller data paths on their components, the idea here is that a physically smaller amount of material would require a smaller amount of time to switch from a high (1) state to a low (0) state or visa versa. BUT they are limited again in how small they can go by an attribute of the components called the "Trigger Voltage", that is the minimum amount of power required before the component sees a difference between a high and a low state because if the data path is too small then it burns up.
I haven't even began to scratch the surface of this topic. As much as I enjoy talking about it I can't get myself to focus on one subject long enough to adequitly explain to you how much it plays into the grand scheme of things. I hope I gave you enough to start some research on your own though as this is a very interesting topic for geeks like me.
Last edited on Sep 21, 2011 at 10:25pm
Sep 22, 2011 at 6:55am
Wow! Thank you so much for your help! I really appreciate it :D
Muhasaresa
Muhasaresa
Sep 22, 2011 at 5:13pm
Perhaps I was drastically simplifying things to shorten my answer but just to be clear: The speed of light is NOT constant. It changes passing from medium to dissimilar medium. That's why different frequencies of light bend into the rainbow passing from air to glass in a prism. Modern scientists have slowed light to 38 MPH by passing a LASER beam though a sodium gas cloud. 38 MPH!! Incredible stuff. There's also a school of thought that maintains that the speed of light has been steadily slowing down since the Big Bang. Think about what THAT does to some of our scientific 'facts'!
Although it is true that electrons travel at about 2/3c (in copper), since that material is full of free electrons, one pops out on one end as soon as you push one in on the other end making the speed of electricity essentially c for most intents and purposes.
Man, you can go on and on with this kind of stuff and it just keeps getting more and more interesting. What will we be discussing by the time molecular switch circuitry comes along? Too bad I have a job...
Bottom line to the OP: Computer speed is measured in switching frequency, the rate of information is what counts, not the rate of physical motion of any of its components. Technically that rate would be zero. Unless you fling your PC across the room...
Although it is true that electrons travel at about 2/3c (in copper), since that material is full of free electrons, one pops out on one end as soon as you push one in on the other end making the speed of electricity essentially c for most intents and purposes.
Man, you can go on and on with this kind of stuff and it just keeps getting more and more interesting. What will we be discussing by the time molecular switch circuitry comes along? Too bad I have a job...
Bottom line to the OP: Computer speed is measured in switching frequency, the rate of information is what counts, not the rate of physical motion of any of its components. Technically that rate would be zero. Unless you fling your PC across the room...
Last edited on Sep 22, 2011 at 5:14pm
Sep 22, 2011 at 8:13pm| Although it is true that electrons travel at about 2/3c (in copper), since that material is full of free electrons, one pops out on one end as soon as you push one in on the other end making the speed of electricity essentially c for most intents and purposes. |
I'll conceed that I may have been wrong about the consistency of the speed of light, but I don't like it because it messes with the Newtonian side of physics too much.
@ OP: I would like to leave you with this almost universal truth though. The more you study about his topic, the more you'll realize how much guess work goes into EVERYTHING WE DO.
Sep 23, 2011 at 2:14pm| The more you study about his topic, the more you'll realize how much guess work goes into EVERYTHING WE DO. |
HaHa, you got that right! It's amazing how quickly so many "discussions" on the subject quickly degenerate into silly semantic arguments. Whenever you get really small, really large, really fast, etc., our simplified "macro-view" of the universe loses its consistency. Math and physics we take for granted start breaking down. Even our language becomes insufficient to discuss the subjects. Ahh, Good Times!
Sep 23, 2011 at 3:03pm| Although it is true that electrons travel at about 2/3c (in copper), since that material is full of free electrons, one pops out on one end as soon as you push one in on the other end making the speed of electricity essentially c for most intents and purposes. |
Nope, it is still 2/3c. You are confusing steady state with a instantaneous change. What is true: The rate of electrons coming out of one end is the same as the rate of electrons coming in from the other end in steady state. Since I don't know how this 2/3c speed was measured, it could be that:
A. It takes 3/(2c) seconds for the electron that came in in one end to be spat out in the other 1 meter apart, or
B. It takes 3/(2c) seconds for the electron that was already 1 meter apart to "fall out" after the other electron came in from the other end.
Since I am unaware that you can track a single electron in a copper wire (or any other medium), I would have to assume that it is B.
In any case, physics is not my forte.
Sep 23, 2011 at 5:43pm
There's our friend semantics coming into play again. That's why I changed "electrons" to "electricity" in my explanation. Take a look at this: http://www.jimloy.com/physics/electric.htm
Sep 23, 2011 at 6:45pm
Electrons travel surprisingly slow through a medium such as copper wire. The reason why light switches seem to turn on the light immediately is because they travel all at the same time.
Anyway, my understanding is that temperature is a huge factor in limiting the speed of todays integrated circuits. I believe the amperage* has been getting lower and lower (which generates less and less heat) which has led to higher and higher speed processors.
* Correct me if I'm wrong.
Anyway, my understanding is that temperature is a huge factor in limiting the speed of todays integrated circuits. I believe the amperage* has been getting lower and lower (which generates less and less heat) which has led to higher and higher speed processors.
* Correct me if I'm wrong.
Sep 23, 2011 at 7:44pm
Right. Temperature is a measure of the kinetic energy of the particles of matter. The higher the temperature, the more chaos inside the conductive material, the more difficult it is to transmit a signal across it. This is mainly the reason why superconductors have only been achieved at very low temperatures.
Sep 23, 2011 at 7:49pm| * Correct me if I'm wrong. |
I believe you're right. When we finally get room temperature superconductors (and semi-conductors) to work, CPUs will have the potential to run at ludicrous speed, but still keep their cool.
| Since I am unaware that you can track a single electron in a copper wire (or any other medium), I would have to assume that it is B. |
Only statistically speaking of course but if you could it would take about 12 hours for the little bugger to go 1 meter in a 2mm diameter wire. At least if you believe this math: http://www.jensign.com/JavaScience/www/cuwire/cuwire.html
Topic archived. No new replies allowed.