Is the speed of electrical signals always constant within a CPU? [closed]
Microprocessors rely on assumptions about how long it takes for electrical signals to propagate through combinational logic circuits. As far as I understand from reading "Modern processor design", correct functioning seems to require that the signals neither take too long nor too short to arrive at their destination. I know that electrical signals travel roughly at the speed of light.
This made me wonder: is the signal time within a cpu constant? Or does it depend on e.g. the temperature of the cpu or some other variable? It seems to me that this would introduce necessary complications in the design.
light cpu electricity
asked Nov 29 '18 at 10:11
user56834user56834
1092
closed as too broad by Andy aka, Finbarr, RoyC, Dwayne Reid, Dmitry Grigoryev Dec 3 '18 at 11:44
Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer. Avoid asking multiple distinct questions at once. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.
add a comment |
Microprocessors rely on assumptions about how long it takes for electrical signals to propagate through combinational logic circuits. As far as I understand from reading "Modern processor design", correct functioning seems to require that the signals neither take too long nor too short to arrive at their destination. I know that electrical signals travel roughly at the speed of light.
This made me wonder: is the signal time within a cpu constant? Or does it depend on e.g. the temperature of the cpu or some other variable? It seems to me that this would introduce necessary complications in the design.
light cpu electricity
asked Nov 29 '18 at 10:11
user56834user56834
1092
closed as too broad by Andy aka, Finbarr, RoyC, Dwayne Reid, Dmitry Grigoryev Dec 3 '18 at 11:44
Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer. Avoid asking multiple distinct questions at once. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.
1
Gates propagate signals at about a one-hundredth to one-thousandth of the speed of light given the sort of dimensions they have and many can be significantly longer. I think speed of light might be a red herring in your understanding but I'm interested in what others say.
– Andy aka
Nov 29 '18 at 10:22
I've a relative working in CPU design. The comment is "the wiring became the limiting factor in speed, around 1990". This means the wiring-resistance and the CMOS gate capacitances are defining the time constants.
– analogsystemsrf
Nov 30 '18 at 3:49
add a comment |
Microprocessors rely on assumptions about how long it takes for electrical signals to propagate through combinational logic circuits. As far as I understand from reading "Modern processor design", correct functioning seems to require that the signals neither take too long nor too short to arrive at their destination. I know that electrical signals travel roughly at the speed of light.
This made me wonder: is the signal time within a cpu constant? Or does it depend on e.g. the temperature of the cpu or some other variable? It seems to me that this would introduce necessary complications in the design.
light cpu electricity
asked Nov 29 '18 at 10:11
user56834user56834
1092
Microprocessors rely on assumptions about how long it takes for electrical signals to propagate through combinational logic circuits. As far as I understand from reading "Modern processor design", correct functioning seems to require that the signals neither take too long nor too short to arrive at their destination. I know that electrical signals travel roughly at the speed of light.
This made me wonder: is the signal time within a cpu constant? Or does it depend on e.g. the temperature of the cpu or some other variable? It seems to me that this would introduce necessary complications in the design.
light cpu electricity
light cpu electricity
asked Nov 29 '18 at 10:11
user56834user56834
1092
asked Nov 29 '18 at 10:11
user56834user56834
1092
asked Nov 29 '18 at 10:11
user56834user56834
1092
asked Nov 29 '18 at 10:11
user56834user56834
1092
asked Nov 29 '18 at 10:11
user56834user56834
1092
1092
closed as too broad by Andy aka, Finbarr, RoyC, Dwayne Reid, Dmitry Grigoryev Dec 3 '18 at 11:44
Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer. Avoid asking multiple distinct questions at once. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.
closed as too broad by Andy aka, Finbarr, RoyC, Dwayne Reid, Dmitry Grigoryev Dec 3 '18 at 11:44
Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer. Avoid asking multiple distinct questions at once. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.
1
Gates propagate signals at about a one-hundredth to one-thousandth of the speed of light given the sort of dimensions they have and many can be significantly longer. I think speed of light might be a red herring in your understanding but I'm interested in what others say.
– Andy aka
Nov 29 '18 at 10:22
I've a relative working in CPU design. The comment is "the wiring became the limiting factor in speed, around 1990". This means the wiring-resistance and the CMOS gate capacitances are defining the time constants.
– analogsystemsrf
Nov 30 '18 at 3:49
add a comment |
1
Gates propagate signals at about a one-hundredth to one-thousandth of the speed of light given the sort of dimensions they have and many can be significantly longer. I think speed of light might be a red herring in your understanding but I'm interested in what others say.
– Andy aka
Nov 29 '18 at 10:22
I've a relative working in CPU design. The comment is "the wiring became the limiting factor in speed, around 1990". This means the wiring-resistance and the CMOS gate capacitances are defining the time constants.
– analogsystemsrf
Nov 30 '18 at 3:49
1
1
Gates propagate signals at about a one-hundredth to one-thousandth of the speed of light given the sort of dimensions they have and many can be significantly longer. I think speed of light might be a red herring in your understanding but I'm interested in what others say.
– Andy aka
Nov 29 '18 at 10:22
Gates propagate signals at about a one-hundredth to one-thousandth of the speed of light given the sort of dimensions they have and many can be significantly longer. I think speed of light might be a red herring in your understanding but I'm interested in what others say.
– Andy aka
Nov 29 '18 at 10:22
I've a relative working in CPU design. The comment is "the wiring became the limiting factor in speed, around 1990". This means the wiring-resistance and the CMOS gate capacitances are defining the time constants.
– analogsystemsrf
Nov 30 '18 at 3:49
I've a relative working in CPU design. The comment is "the wiring became the limiting factor in speed, around 1990". This means the wiring-resistance and the CMOS gate capacitances are defining the time constants.
– analogsystemsrf
Nov 30 '18 at 3:49
add a comment |
2 Answers
2
active
oldest
votes
Electromagnetic waves travel at a fraction of the speed of light, depending on the dielectric constant of the medium they are propagating through. If you had ideal components (lossless traces, circuit elements that switch instantaneously, power supplies with no internal impedance, no noise in your signal etc.) maybe you could approach these speeds. The reality is the non-idealities of the circuits will dominate the performance. Signals in a modern processor will be limited by the node capacitance, switching speeds of internal transistors/mosfets and voltage levels necessary to register high/low signals. These metrics will dictate the rise/fall times necessary to transmit information.
Yes, this introduces lots of complexity but that is why there are very smart people with PhDs and entire engineering teams necessary to design integrated circuits. Monitoring these metrics across process variation, temperature variation and voltage variation is part of a robust integrated circuit design and there is no getting away from that complexity.
answered Nov 29 '18 at 10:24
Gonzik007Gonzik007
2,7111023
My favorite fact about these things is how many modern processors must have multiple clocks on the chip spread around because the clock speeds are so fast, the clock signal wouldn't have time to reach from one side of the chip to the other in the span of a single clock cycle.
– Shufflepants
Nov 29 '18 at 16:11
@Shufflepants you don't actually need clock edges to arrive at the same cycle as they were produced. You just need them to arrive all at once. Still pretty hard though.
– John Dvorak
Nov 29 '18 at 16:18
add a comment |
Microprocessors consist of many logic circuits which are made out of standard "gates" like inverters, AND, OR NAND etc. These gates consist of (CMOS) transistors and the behavior of these transistors is indeed dependent on many parameters like temperature and variations in the manufacturing process. This influences the speed at which these gates can reliably operate. The speed is also influenced by for example how many inputs are connected to a certain output. I mean that the architecture of the circuit also plays a significant role.
It would be nearly impossible to make a reliable chip by only relying on these varying delays. You would have to slow down everything a lot to ensure proper operation, for example to make a faster circuit wait for the input of a slower circuit. That's not so efficient and does not result in a fast circuit.
That's why nearly all complex digital circuits like microprocessors are clocked, meaning there is a clock that gives a "hart beat" and every circuit works (only changes state) when this "beat" comes (on the edge of the clock).
Using a clock allows for a much easier and more automated design flow. Still, at the end of the design when the real delays can be determined by specialist tools, simulations can prove if there are still timing issues or not and also suggest what to change to solve these.
answered Nov 29 '18 at 13:18
BimpelrekkieBimpelrekkie
47.1k240104
add a comment |
2 Answers
2
active
oldest
votes
2 Answers
2
active
oldest
votes
active
oldest
votes
active
oldest
votes
Electromagnetic waves travel at a fraction of the speed of light, depending on the dielectric constant of the medium they are propagating through. If you had ideal components (lossless traces, circuit elements that switch instantaneously, power supplies with no internal impedance, no noise in your signal etc.) maybe you could approach these speeds. The reality is the non-idealities of the circuits will dominate the performance. Signals in a modern processor will be limited by the node capacitance, switching speeds of internal transistors/mosfets and voltage levels necessary to register high/low signals. These metrics will dictate the rise/fall times necessary to transmit information.
Yes, this introduces lots of complexity but that is why there are very smart people with PhDs and entire engineering teams necessary to design integrated circuits. Monitoring these metrics across process variation, temperature variation and voltage variation is part of a robust integrated circuit design and there is no getting away from that complexity.
answered Nov 29 '18 at 10:24
Gonzik007Gonzik007
2,7111023
My favorite fact about these things is how many modern processors must have multiple clocks on the chip spread around because the clock speeds are so fast, the clock signal wouldn't have time to reach from one side of the chip to the other in the span of a single clock cycle.
– Shufflepants
Nov 29 '18 at 16:11
@Shufflepants you don't actually need clock edges to arrive at the same cycle as they were produced. You just need them to arrive all at once. Still pretty hard though.
– John Dvorak
Nov 29 '18 at 16:18
add a comment |
Electromagnetic waves travel at a fraction of the speed of light, depending on the dielectric constant of the medium they are propagating through. If you had ideal components (lossless traces, circuit elements that switch instantaneously, power supplies with no internal impedance, no noise in your signal etc.) maybe you could approach these speeds. The reality is the non-idealities of the circuits will dominate the performance. Signals in a modern processor will be limited by the node capacitance, switching speeds of internal transistors/mosfets and voltage levels necessary to register high/low signals. These metrics will dictate the rise/fall times necessary to transmit information.
Yes, this introduces lots of complexity but that is why there are very smart people with PhDs and entire engineering teams necessary to design integrated circuits. Monitoring these metrics across process variation, temperature variation and voltage variation is part of a robust integrated circuit design and there is no getting away from that complexity.
answered Nov 29 '18 at 10:24
Gonzik007Gonzik007
2,7111023
My favorite fact about these things is how many modern processors must have multiple clocks on the chip spread around because the clock speeds are so fast, the clock signal wouldn't have time to reach from one side of the chip to the other in the span of a single clock cycle.
– Shufflepants
Nov 29 '18 at 16:11
@Shufflepants you don't actually need clock edges to arrive at the same cycle as they were produced. You just need them to arrive all at once. Still pretty hard though.
– John Dvorak
Nov 29 '18 at 16:18
add a comment |
Electromagnetic waves travel at a fraction of the speed of light, depending on the dielectric constant of the medium they are propagating through. If you had ideal components (lossless traces, circuit elements that switch instantaneously, power supplies with no internal impedance, no noise in your signal etc.) maybe you could approach these speeds. The reality is the non-idealities of the circuits will dominate the performance. Signals in a modern processor will be limited by the node capacitance, switching speeds of internal transistors/mosfets and voltage levels necessary to register high/low signals. These metrics will dictate the rise/fall times necessary to transmit information.
Yes, this introduces lots of complexity but that is why there are very smart people with PhDs and entire engineering teams necessary to design integrated circuits. Monitoring these metrics across process variation, temperature variation and voltage variation is part of a robust integrated circuit design and there is no getting away from that complexity.
answered Nov 29 '18 at 10:24
Gonzik007Gonzik007
2,7111023
Electromagnetic waves travel at a fraction of the speed of light, depending on the dielectric constant of the medium they are propagating through. If you had ideal components (lossless traces, circuit elements that switch instantaneously, power supplies with no internal impedance, no noise in your signal etc.) maybe you could approach these speeds. The reality is the non-idealities of the circuits will dominate the performance. Signals in a modern processor will be limited by the node capacitance, switching speeds of internal transistors/mosfets and voltage levels necessary to register high/low signals. These metrics will dictate the rise/fall times necessary to transmit information.
Yes, this introduces lots of complexity but that is why there are very smart people with PhDs and entire engineering teams necessary to design integrated circuits. Monitoring these metrics across process variation, temperature variation and voltage variation is part of a robust integrated circuit design and there is no getting away from that complexity.
answered Nov 29 '18 at 10:24
Gonzik007Gonzik007
2,7111023
edited Nov 29 '18 at 10:42
answered Nov 29 '18 at 10:24
Gonzik007Gonzik007
2,7111023
answered Nov 29 '18 at 10:24
Gonzik007Gonzik007
2,7111023
answered Nov 29 '18 at 10:24
Gonzik007Gonzik007
2,7111023
2,7111023
My favorite fact about these things is how many modern processors must have multiple clocks on the chip spread around because the clock speeds are so fast, the clock signal wouldn't have time to reach from one side of the chip to the other in the span of a single clock cycle.
– Shufflepants
Nov 29 '18 at 16:11
@Shufflepants you don't actually need clock edges to arrive at the same cycle as they were produced. You just need them to arrive all at once. Still pretty hard though.
– John Dvorak
Nov 29 '18 at 16:18
add a comment |
My favorite fact about these things is how many modern processors must have multiple clocks on the chip spread around because the clock speeds are so fast, the clock signal wouldn't have time to reach from one side of the chip to the other in the span of a single clock cycle.
– Shufflepants
Nov 29 '18 at 16:11
@Shufflepants you don't actually need clock edges to arrive at the same cycle as they were produced. You just need them to arrive all at once. Still pretty hard though.
– John Dvorak
Nov 29 '18 at 16:18
My favorite fact about these things is how many modern processors must have multiple clocks on the chip spread around because the clock speeds are so fast, the clock signal wouldn't have time to reach from one side of the chip to the other in the span of a single clock cycle.
– Shufflepants
Nov 29 '18 at 16:11
My favorite fact about these things is how many modern processors must have multiple clocks on the chip spread around because the clock speeds are so fast, the clock signal wouldn't have time to reach from one side of the chip to the other in the span of a single clock cycle.
– Shufflepants
Nov 29 '18 at 16:11
@Shufflepants you don't actually need clock edges to arrive at the same cycle as they were produced. You just need them to arrive all at once. Still pretty hard though.
– John Dvorak
Nov 29 '18 at 16:18
@Shufflepants you don't actually need clock edges to arrive at the same cycle as they were produced. You just need them to arrive all at once. Still pretty hard though.
– John Dvorak
Nov 29 '18 at 16:18
add a comment |
Microprocessors consist of many logic circuits which are made out of standard "gates" like inverters, AND, OR NAND etc. These gates consist of (CMOS) transistors and the behavior of these transistors is indeed dependent on many parameters like temperature and variations in the manufacturing process. This influences the speed at which these gates can reliably operate. The speed is also influenced by for example how many inputs are connected to a certain output. I mean that the architecture of the circuit also plays a significant role.
It would be nearly impossible to make a reliable chip by only relying on these varying delays. You would have to slow down everything a lot to ensure proper operation, for example to make a faster circuit wait for the input of a slower circuit. That's not so efficient and does not result in a fast circuit.
That's why nearly all complex digital circuits like microprocessors are clocked, meaning there is a clock that gives a "hart beat" and every circuit works (only changes state) when this "beat" comes (on the edge of the clock).
Using a clock allows for a much easier and more automated design flow. Still, at the end of the design when the real delays can be determined by specialist tools, simulations can prove if there are still timing issues or not and also suggest what to change to solve these.
answered Nov 29 '18 at 13:18
BimpelrekkieBimpelrekkie
47.1k240104
add a comment |
Microprocessors consist of many logic circuits which are made out of standard "gates" like inverters, AND, OR NAND etc. These gates consist of (CMOS) transistors and the behavior of these transistors is indeed dependent on many parameters like temperature and variations in the manufacturing process. This influences the speed at which these gates can reliably operate. The speed is also influenced by for example how many inputs are connected to a certain output. I mean that the architecture of the circuit also plays a significant role.
It would be nearly impossible to make a reliable chip by only relying on these varying delays. You would have to slow down everything a lot to ensure proper operation, for example to make a faster circuit wait for the input of a slower circuit. That's not so efficient and does not result in a fast circuit.
That's why nearly all complex digital circuits like microprocessors are clocked, meaning there is a clock that gives a "hart beat" and every circuit works (only changes state) when this "beat" comes (on the edge of the clock).
Using a clock allows for a much easier and more automated design flow. Still, at the end of the design when the real delays can be determined by specialist tools, simulations can prove if there are still timing issues or not and also suggest what to change to solve these.
answered Nov 29 '18 at 13:18
BimpelrekkieBimpelrekkie
47.1k240104
add a comment |
Microprocessors consist of many logic circuits which are made out of standard "gates" like inverters, AND, OR NAND etc. These gates consist of (CMOS) transistors and the behavior of these transistors is indeed dependent on many parameters like temperature and variations in the manufacturing process. This influences the speed at which these gates can reliably operate. The speed is also influenced by for example how many inputs are connected to a certain output. I mean that the architecture of the circuit also plays a significant role.
It would be nearly impossible to make a reliable chip by only relying on these varying delays. You would have to slow down everything a lot to ensure proper operation, for example to make a faster circuit wait for the input of a slower circuit. That's not so efficient and does not result in a fast circuit.
That's why nearly all complex digital circuits like microprocessors are clocked, meaning there is a clock that gives a "hart beat" and every circuit works (only changes state) when this "beat" comes (on the edge of the clock).
Using a clock allows for a much easier and more automated design flow. Still, at the end of the design when the real delays can be determined by specialist tools, simulations can prove if there are still timing issues or not and also suggest what to change to solve these.
answered Nov 29 '18 at 13:18
BimpelrekkieBimpelrekkie
47.1k240104
Microprocessors consist of many logic circuits which are made out of standard "gates" like inverters, AND, OR NAND etc. These gates consist of (CMOS) transistors and the behavior of these transistors is indeed dependent on many parameters like temperature and variations in the manufacturing process. This influences the speed at which these gates can reliably operate. The speed is also influenced by for example how many inputs are connected to a certain output. I mean that the architecture of the circuit also plays a significant role.
It would be nearly impossible to make a reliable chip by only relying on these varying delays. You would have to slow down everything a lot to ensure proper operation, for example to make a faster circuit wait for the input of a slower circuit. That's not so efficient and does not result in a fast circuit.
That's why nearly all complex digital circuits like microprocessors are clocked, meaning there is a clock that gives a "hart beat" and every circuit works (only changes state) when this "beat" comes (on the edge of the clock).
Using a clock allows for a much easier and more automated design flow. Still, at the end of the design when the real delays can be determined by specialist tools, simulations can prove if there are still timing issues or not and also suggest what to change to solve these.
answered Nov 29 '18 at 13:18
BimpelrekkieBimpelrekkie
47.1k240104
answered Nov 29 '18 at 13:18
BimpelrekkieBimpelrekkie
47.1k240104
answered Nov 29 '18 at 13:18
BimpelrekkieBimpelrekkie
47.1k240104
answered Nov 29 '18 at 13:18
BimpelrekkieBimpelrekkie
47.1k240104
47.1k240104
add a comment |
add a comment |
1
Gates propagate signals at about a one-hundredth to one-thousandth of the speed of light given the sort of dimensions they have and many can be significantly longer. I think speed of light might be a red herring in your understanding but I'm interested in what others say.
– Andy aka
Nov 29 '18 at 10:22
I've a relative working in CPU design. The comment is "the wiring became the limiting factor in speed, around 1990". This means the wiring-resistance and the CMOS gate capacitances are defining the time constants.
– analogsystemsrf
Nov 30 '18 at 3:49