|
heroic!
modified 20-Oct-19 21:02pm.
|
|
|
|
|
somewhere on the Internet I read that the high-low voltage change is not immediate but happens slowly instead.
for example a voltage graph of 010 would be:
<br />
___________<br />
/ \<br />
/ \ <br />
_______/ \_________ <br />
and not<br />
<br />
____________<br />
<br />
__________ ___________<br />
<br />
<br />
I've also read that the "/" and "\" parts are a factor that slows processors' speed
why does this happen,( can you recomend me a book or something to understand this better ? )
and does it mean that 11110000 is processed faster than 10101010 for example ?
thanks a lot !!
edit: I tried putting my pretty 'graph' on inside 'code' so it would look ok... it didn't work 
|
|
|
|
|
I'm not sure of which operation would perform quicker but to understand why the voltage is not instant would be to look at electronic circuits and transistors books i think.
A value of 0 or 1 in binary is determined when the voltage goes below or above a certain threshold. For example a binary value of 0 might be when the voltage is less than 0.25v and a binary value of 1 could be above 0.75v. Because of this, coupled with the fact that the voltage does not change instantly, there is a slight delay for when the binary value changes.
This delay is also taken into consideration for setting clock speeds since you have many transistors chained together (inside a CPU for example), each of which has a slight delay when its binary value (voltage) changes so the clock speed needs to ensure that all transistors have had enough time to change their value before executing the command.
Transistors are unable to change value (voltage) instantly due to the physical properties of the material its made of and possibly requiring an infinite amount of energy to perform a change in 0 seconds... will need to check the actual reason as i can't recall it atm.
I don't think there will be any difference in performance between those two values listed since it would depend on the clock speed.
Note: this is what i can remember from university electronics class..
|
|
|
|
|
Everything in digital circuits are sync-ed to a clock that is constant.
If a bit doesn't settle to the proper value by the time of the clock pulse as expected then the whole shee-bang will fail and the clock must be slowed all the time. So due to the constant clock and do or die nature of digital circuits, it is impossible for a particular bit pattern to be processed faster than another. Current draw will fluctuate some however.
Microprocessors are all built on a sheet like cookies and the ones that can't reliably truck bits around inside with a desired clock rate are stamped as slower so they can reliably process data.
Hence the different speed grades of Pentiums for example. They don't set out to make slower ones the yield just dictates this.
|
|
|
|
|
What do you mean by "processed faster"? You are presenting numbers and not state transitions I assume. Operations on numbers are determined by processor clock speeds which are set to allow for the latency in transitions. Thus a value is set an operation is performed. These happen in fixed intervals. The CPU doesn't care what the number is - the speed of processing is the same.
Peter Wasser
"The whole problem with the world is that fools and fanatics are always so certain of themselves, and wiser people so full of doubts." - Bertrand Russell
|
|
|
|
|
No, it's not processed faster. If you're talking about a serial stream, the sample is taken at some point in the middle of the defined "bit" time. The rising and falling edges are not instantaneous due, I believe, to natural capacitance of the circuit and other factors. You can't just gate something on or off instantaneously.
Keep in mind that everything in a computer is synchronized to a clock cycle or some multiple of a clock cycle (though modern CPU's have several different clocks doing things) the point is, any bits, whether processed in serial, sampled at a particular frequency, or determined by rising/falling edges, or processed in parallel as is the case with a memory or instruction bus, everything marches to the beat of the same drummer, so just because you have non-instantaneous rising/falling edges, this doesn't affect the processing speed.
Marc
|
|
|
|
|
Rise and fall times for most technologies are nearly equal, so no, it doesn't affect the 'processing' time. As Marc mentioned, the device doing the processing is clocked at a rate consistent with the technology, and individual bits are sampled after an appropriate settling time, nearer the middle of the bit time than the transition. What is significantly affected by transition time is the power consumption of the device. Power is defined as the product of the voltage across and the current through a path. For semiconductor devices, a one is defined as a high voltage with little or no current, while a zero is a relatively high current with little or no voltage; both states consume little power. But the time spent between states can consume large amounts of power because the voltage and current are both non-zero. The quicker we move from one state to the other, the less power is consumed, and the less heat we have to get rid of.
Will Rogers never met me.
|
|
|
|
|
Apart from what the other responses mention, the processor doesn't read bits serially, it reads them over a bus, so 11110000 would be processed at the same speed as 10101010
Would 11111111 be processed at the same speed as 00000000 would be a better question!
PooperPig - Coming Soon
|
|
|
|
|
The properties of circuits that affect changing voltages and currents are capacitance and inductance. Transistors are a switch that are controlled by the voltage on the gate, depending on the type of transistor. As I understand it, it is the capacitance on that gate that determines the switching speed of the transistor and the overall clock speed that can be used for the chip.
http://en.wikipedia.org/wiki/Transistor[^]
|
|
|
|
|
11010110 00110010 000111010 11101000
|
|
|
|
|
Juts a thought. Not a very nice one, but effective.
|
|
|
|
|
It's transferred by bodily fluids, so perhaps blending them up first and spraying them with the mist?
|
|
|
|
|
I think hitting the deck from 20000 ft is going to blend them up pretty well.
|
|
|
|
|
Yeah, I had considered that, but it relies on random splatter patterns... 
|
|
|
|
|
its no different to warfare in the middle/dark ages 
'g'
|
|
|
|
|
That's a darn good idea. 
|
|
|
|
|
We could drop a few hundred neutron bombs all over Africa, it would be faster and thorougher! And just as discriminatory!
|
|
|
|
|
Africa, I thought we were working on IS not Africa, man you a glutton for punishment!
So why not neutron bomb both the middle east and Africa, and while you are at it reduce the world population and nuke India and China. Oh and I hate paedophiles so nuke any country with lots of them, include the Catholic church in that lot.
Just to be fair randomly drop a few hundred across any location not included in the above.
Never underestimate the power of human stupidity
RAH
|
|
|
|
|
Well said, my thoughts exactly!
Let's not forget to bomb those damn commies too (Puttin & pal), South America (let's not forget where the drug comes from!) and.... I think I forgot somewhere....
|
|
|
|
|
Super Lloyd wrote: I think I forgot somewhere
Australia?
M.D.V. 
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
|
|
|
|
|
Nelek wrote: Australia?
Nah the random drops, paedo and catholic drops will get Oz.
Never underestimate the power of human stupidity
RAH
|
|
|
|
|
No thanks... we're good!!
|
|
|
|
|
Funnily enough I don't have a problem with Russia nor the drug lords.
Never underestimate the power of human stupidity
RAH
|
|
|
|
|
Indeed, that is ... suspicious ... 
|
|
|
|
|
What as Africa got to do with ISIS?
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
