We’re rewarding the question askers & reputations are being recalculated! Read more.
6 deleted 43 characters in body
source | link

There are many techniques to encode or decode a large amount of information into one "communication bucket/wagon" (which are called symbols). The R-2R-type Digital-to-analog converter (illustrated below) is one of the simplest yet effective methods of transmitting e.g. 16x faster than the communication clock - provided that the receiver's Analog-to-digital converter converts within a clock period (main limitation: noise).

enter image description here

Nowadays, analog communication not only use amplitude, but phase information to stuff even more bits in a communication clock cycle, as illustrated in This topic (Disclaimer: I have one answer there, but it's the only topic I can remember of which discusses this). And this is only a mere example of a plethora of techniques that become more and more complex with every year that goes by.

enter image description here

For digital communication, it's mostly down to: 1) number of parallel lines (1 for serial communication, USB is a serial link even though it has two data lines, they are always opposite each other in value to increase robustness to common mode noise by difference), and 2) compression. When you send 100MB worth of files to a 10MB-limit mailbox by compressing it, you're effectively transmitting in one clock cycle more than the physical uncompressed datarate limit. Just as there are many modulation techniques (seen above), there are many compression techniques. Note that there are dedicated compression/decompression chips available to take full advantage of compression as a means to increase transmission speed. Note that I am not saying that's what is being done for video transfers, since compression/decompression adds latency. It's however one method used for satellite links though, for example.

enter image description here

Finally, datarate limits are really limits for a given error rate. If you add to your communication protocol some encoding which detects errors and corrects them automatically (using for example state machines on both ends), you can increase the datarate past the stated limit for the same error rate.

That's how data gets in and out of a computer faster than the CPU clock, or how "one man/woman can move more or less buckets from your house to another" (amplitude and/or phase modulation, compression, encoding...).

enter image description here

Now, for the "how to get the buckets from your door to wherever you want it in your house", that's down to the width of your databus: 64bits computers took over from 32 bits computers for this reason, handle twice more data for every clock cycle. That's the number of buckets that can be grabbed at the same time by the same person in your house (assume you live on top of radioactive waste). I think processing cores could have an impact on pure transfers if the clocks of the cores are phased out to move data quicker on the data bus but generally they share not only the same bus but the same clock (the maximum clock rate is certainly set to the maximum propagation time of the data along the bus). This is obviously a simplistic view, many factors come into play - but it has the benefit of adding more reasons why it IS possible.

There are many techniques to encode or decode a large amount of information into one "communication bucket/wagon" (which are called symbols). The R-2R-type Digital-to-analog converter (illustrated below) is one of the simplest yet effective methods of transmitting e.g. 16x faster than the communication clock - provided that the receiver's Analog-to-digital converter converts within a clock period (main limitation: noise).

enter image description here

Nowadays, analog communication not only use amplitude, but phase information to stuff even more bits in a communication clock cycle, as illustrated in This topic (Disclaimer: I have one answer there, but it's the only topic I can remember of which discusses this). And this is only a mere example of a plethora of techniques that become more and more complex with every year that goes by.

enter image description here

For digital communication, it's mostly down to: 1) number of parallel lines (1 for serial communication, USB is a serial link even though it has two data lines, they are always opposite each other in value to increase robustness to common mode noise by difference), and 2) compression. When you send 100MB worth of files to a 10MB-limit mailbox by compressing it, you're effectively transmitting in one clock cycle more than the physical uncompressed datarate limit. Just as there are many modulation techniques (seen above), there are many compression techniques. Note that there are dedicated compression/decompression chips available to take full advantage of compression as a means to increase transmission speed.

enter image description here

Finally, datarate limits are really limits for a given error rate. If you add to your communication protocol some encoding which detects errors and corrects them automatically (using for example state machines on both ends), you can increase the datarate past the stated limit for the same error rate.

That's how data gets in and out of a computer faster than the CPU clock, or how "one man/woman can move more or less buckets from your house to another" (amplitude and/or phase modulation, compression, encoding...).

enter image description here

Now, for the "how to get the buckets from your door to wherever you want it in your house", that's down to the width of your databus: 64bits computers took over from 32 bits computers for this reason, handle twice more data for every clock cycle. That's the number of buckets that can be grabbed at the same time by the same person in your house (assume you live on top of radioactive waste). I think processing cores could have an impact on pure transfers if the clocks of the cores are phased out to move data quicker on the data bus but generally they share not only the same bus but the same clock (the maximum clock rate is certainly set to the maximum propagation time of the data along the bus). This is obviously a simplistic view, many factors come into play - but it has the benefit of adding more reasons why it IS possible.

There are many techniques to encode or decode a large amount of information into one "communication bucket/wagon" (which are called symbols). The R-2R-type Digital-to-analog converter (illustrated below) is one of the simplest yet effective methods of transmitting e.g. 16x faster than the communication clock - provided that the receiver's Analog-to-digital converter converts within a clock period (main limitation: noise).

enter image description here

Nowadays, analog communication not only use amplitude, but phase information to stuff even more bits in a communication clock cycle, as illustrated in This topic (Disclaimer: I have one answer there, but it's the only topic I can remember of which discusses this). And this is only a mere example of a plethora of techniques that become more and more complex with every year that goes by.

enter image description here

For digital communication, it's mostly down to: 1) number of parallel lines (1 for serial communication, USB is a serial link even though it has two data lines, they are always opposite each other in value to increase robustness to common mode noise by difference), and 2) compression. When you send 100MB worth of files to a 10MB-limit mailbox by compressing it, you're effectively transmitting in one clock cycle more than the physical uncompressed datarate limit. Just as there are many modulation techniques (seen above), there are many compression techniques. Note that there are dedicated compression/decompression chips available to take full advantage of compression as a means to increase transmission speed. Note that I am not saying that's what is being done for video transfers, since compression/decompression adds latency. It's however one method used for satellite links though, for example.

Finally, datarate limits are really limits for a given error rate. If you add to your communication protocol some encoding which detects errors and corrects them automatically (using for example state machines on both ends), you can increase the datarate past the stated limit for the same error rate.

That's how data gets in and out of a computer faster than the CPU clock, or how "one man/woman can move more or less buckets from your house to another" (amplitude and/or phase modulation, compression, encoding...).

enter image description here

Now, for the "how to get the buckets from your door to wherever you want it in your house", that's down to the width of your databus: 64bits computers took over from 32 bits computers for this reason, handle twice more data for every clock cycle. That's the number of buckets that can be grabbed at the same time by the same person in your house (assume you live on top of radioactive waste). I think processing cores could have an impact on pure transfers if the clocks of the cores are phased out to move data quicker on the data bus but generally they share not only the same bus but the same clock (the maximum clock rate is certainly set to the maximum propagation time of the data along the bus). This is obviously a simplistic view, many factors come into play - but it has the benefit of adding more reasons why it IS possible.

5 added 89 characters in body
source | link

There are many techniques to encode or decode a large amount of information into one "communication bucket/wagon" (which are called symbols). The R-2R-type Digital-to-analog converter (illustrated below) is one of the simplest yet effective methods of transmitting e.g. 16x faster than the communication clock - provided that the receiver's Analog-to-digital converter converts within a clock period (main limitation: noise).

enter image description here

Nowadays, analog communication not only use amplitude, but phase information to stuff even more bits in a communication clock cycle, as illustrated in This topic (Disclaimer: I have one answer there, but it's the only topic I can remember of which discusses this). And this is only a mere example of a plethora of techniques that become more and more complex with every year that goes by.

enter image description here

For digital communication, it's mostly down to: 1) number of parallel lines (1 for serial communication, USB is a serial link even though it has two data lines, they are always opposite each other in value to increase robustness to common mode noise by difference), and 2) compression. When you send 100MB worth of files to a 10MB-limit mailbox by compressing it, you're effectively transmitting in one clock cycle more than the physical uncompressed datarate limit. Just as there are many modulation techniques (seen above), there are many compression techniques. Note that there are dedicated compression/decompression chips available to take full advantage of compression as a means to increase transmission speed.

enter image description here

Finally, datarate limits are really limits for a given error rate. If you add to your communication protocol some encoding which detects errors and corrects them automatically (using for example state machines on both ends), you can increase the datarate past the stated limit for the same error rate.

That's how data gets in and out of a computer faster than the CPU clock, or how "one man/woman can move more or less buckets from your house to another" (amplitude and/or phase modulation, compression, encoding...).

enter image description here

Now, for the "how to get the buckets from your door to wherever you want it in your house", that's down to the width of your databus: 64bits computers took over from 32 bits computers for this reason, handle twice more data for every clock cycle. That's the number of buckets that can be grabbed at the same time by the same person in your house (assume you live on top of radioactive waste). I tend to say the number ofthink processing cores has nocould have an impact on pure transfers because I am pretty sureif the clocks of the cores are phased out to move data quicker on the data bus but generally they share not only the same bus but the same clock (the maximum clock rate is certainly set to the maximum propagation time of the data along the bus, otherwise the clocks of the cores could be different and phased out to move data quicker on the data bus). This is obviously a simplistic view, many factors come into play - but it has the benefit of adding more reasons why it IS possible.

There are many techniques to encode or decode a large amount of information into one "communication bucket/wagon" (which are called symbols). The R-2R-type Digital-to-analog converter (illustrated below) is one of the simplest yet effective methods of transmitting e.g. 16x faster than the communication clock - provided that the receiver's Analog-to-digital converter converts within a clock period (main limitation: noise).

enter image description here

Nowadays, analog communication not only use amplitude, but phase information to stuff even more bits in a communication clock cycle, as illustrated in This topic (Disclaimer: I have one answer there, but it's the only topic I can remember of which discusses this). And this is only a mere example of a plethora of techniques that become more and more complex with every year that goes by.

enter image description here

For digital communication, it's mostly down to: 1) number of parallel lines (1 for serial communication, USB is a serial link even though it has two data lines, they are always opposite each other in value to increase robustness to common mode noise by difference), and 2) compression. When you send 100MB worth of files to a 10MB-limit mailbox by compressing it, you're effectively transmitting in one clock cycle more than the physical uncompressed datarate limit. Just as there are many modulation techniques (seen above), there are many compression techniques. Note that there are dedicated compression/decompression chips available to take full advantage of compression as a means to increase transmission speed.

Finally, datarate limits are really limits for a given error rate. If you add to your communication protocol some encoding which detects errors and corrects them automatically (using for example state machines on both ends), you can increase the datarate past the stated limit for the same error rate.

That's how data gets in and out of a computer faster than the CPU clock, or how "one man/woman can move more or less buckets from your house to another" (amplitude and/or phase modulation, compression, encoding...).

Now, for the "how to get the buckets from your door to wherever you want it in your house", that's down to the width of your databus: 64bits computers took over from 32 bits computers for this reason, handle twice more data for every clock cycle. That's the number of buckets that can be grabbed at the same time by the same person in your house (assume you live on top of radioactive waste). I tend to say the number of cores has no impact on pure transfers because I am pretty sure they share not only the same bus but the same clock (the maximum clock rate is certainly set to the maximum propagation time of the data along the bus, otherwise the clocks of the cores could be different and phased out to move data quicker on the data bus). This is obviously a simplistic view, many factors come into play - but it has the benefit of adding more reasons why it IS possible.

There are many techniques to encode or decode a large amount of information into one "communication bucket/wagon" (which are called symbols). The R-2R-type Digital-to-analog converter (illustrated below) is one of the simplest yet effective methods of transmitting e.g. 16x faster than the communication clock - provided that the receiver's Analog-to-digital converter converts within a clock period (main limitation: noise).

enter image description here

Nowadays, analog communication not only use amplitude, but phase information to stuff even more bits in a communication clock cycle, as illustrated in This topic (Disclaimer: I have one answer there, but it's the only topic I can remember of which discusses this). And this is only a mere example of a plethora of techniques that become more and more complex with every year that goes by.

enter image description here

For digital communication, it's mostly down to: 1) number of parallel lines (1 for serial communication, USB is a serial link even though it has two data lines, they are always opposite each other in value to increase robustness to common mode noise by difference), and 2) compression. When you send 100MB worth of files to a 10MB-limit mailbox by compressing it, you're effectively transmitting in one clock cycle more than the physical uncompressed datarate limit. Just as there are many modulation techniques (seen above), there are many compression techniques. Note that there are dedicated compression/decompression chips available to take full advantage of compression as a means to increase transmission speed.

enter image description here

Finally, datarate limits are really limits for a given error rate. If you add to your communication protocol some encoding which detects errors and corrects them automatically (using for example state machines on both ends), you can increase the datarate past the stated limit for the same error rate.

That's how data gets in and out of a computer faster than the CPU clock, or how "one man/woman can move more or less buckets from your house to another" (amplitude and/or phase modulation, compression, encoding...).

enter image description here

Now, for the "how to get the buckets from your door to wherever you want it in your house", that's down to the width of your databus: 64bits computers took over from 32 bits computers for this reason, handle twice more data for every clock cycle. That's the number of buckets that can be grabbed at the same time by the same person in your house (assume you live on top of radioactive waste). I think processing cores could have an impact on pure transfers if the clocks of the cores are phased out to move data quicker on the data bus but generally they share not only the same bus but the same clock (the maximum clock rate is certainly set to the maximum propagation time of the data along the bus). This is obviously a simplistic view, many factors come into play - but it has the benefit of adding more reasons why it IS possible.

4 added 89 characters in body
source | link

There are many techniques to encode or decode a large amount of information into one "communication bucket/wagon" (which are called symbols). The R-2R-type Digital-to-analog converter (illustrated below) is one of the simplest yet effective methods of transmitting e.g. 16x faster than the communication clock - provided that the receiver's Analog-to-digital converter converts within a clock period (main limitation: noise).

enter image description here

Nowadays, analog communication not only use amplitude, but phase information to stuff even more bits in a communication clock cycle, as illustrated in This topic (Disclaimer: I have one answer there, but it's the only topic I can remember of which discusses this). And this is only a mere example of a plethora of techniques that become more and more complex with every year that goes by.

enter image description here

For digital communication, it's mostly down to: 1) number of parallel lines (1 for serial communication, USB is a serial link even though it has two data lines, they are always opposite each other in value to increase robustness to common mode noise by difference), and 2) compression. When you send 100MB worth of files to a 10MB-limit mailbox by compressing it, you're effectively transmitting in one clock cycle more than the physical uncompressed datarate limit. Just as there are many modulation techniques (seen above), there are many compression techniques. Note that there are dedicated compression/decompression chips available to take full advantage of compression as a means to increase transmission speed.

Finally, datarate limits are really limits for a given error rate. If you add to your communication protocol some encoding which detects errors and corrects them automatically (using for example state machines on both ends), you can increase the datarate past the stated limit for the same error rate.

That's how data gets in and out of a computer faster than the CPU clock, or how "one man/woman can move more or less buckets from your house to another" (amplitude and/or phase modulation, compression, encoding...).

Now, for the "how to get the buckets from your door to wherever you want it in your house", that's down to the width of your databus: 64bits computers took over from 32 bits computers for this reason, handle twice more data for every clock cycle. That's the number of buckets that can be grabbed at the same time by the same person in your house (assume you live on top of radioactive waste). I tend to say the number of cores has no impact on pure transfers because I am pretty sure they share not only the same bus but the same clock (the maximum clock rate is certainly set to the maximum propagation time of the data along the bus, otherwise the clocks of the cores could be different and phased out to move data quicker on the data bus). This is obviously a simplistic view, many factors come into play - but it has the benefit of adding more reasons why it IS possible.

There are many techniques to encode or decode a large amount of information into one "communication bucket/wagon" (which are called symbols). The R-2R-type Digital-to-analog converter (illustrated below) is one of the simplest yet effective methods of transmitting e.g. 16x faster than the communication clock - provided that the receiver's Analog-to-digital converter converts within a clock period (main limitation: noise).

enter image description here

Nowadays, analog communication not only use amplitude, but phase information to stuff even more bits in a communication clock cycle, as illustrated in This topic (Disclaimer: I have one answer there, but it's the only topic I can remember of which discusses this). And this is only a mere example of a plethora of techniques that become more and more complex with every year that goes by.

For digital communication, it's mostly down to: 1) number of parallel lines (1 for serial communication, USB is a serial link even though it has two data lines, they are always opposite each other in value to increase robustness to common mode noise by difference), and 2) compression. When you send 100MB worth of files to a 10MB-limit mailbox by compressing it, you're effectively transmitting in one clock cycle more than the physical uncompressed datarate limit. Just as there are many modulation techniques (seen above), there are many compression techniques. Note that there are dedicated compression/decompression chips available to take full advantage of compression as a means to increase transmission speed.

Finally, datarate limits are really limits for a given error rate. If you add to your communication protocol some encoding which detects errors and corrects them automatically (using for example state machines on both ends), you can increase the datarate past the stated limit for the same error rate.

That's how data gets in and out of a computer faster than the CPU clock, or how "one man/woman can move more or less buckets from your house to another" (amplitude and/or phase modulation, compression, encoding...).

Now, for the "how to get the buckets from your door to wherever you want it in your house", that's down to the width of your databus: 64bits computers took over from 32 bits computers for this reason, handle twice more data for every clock cycle. That's the number of buckets that can be grabbed at the same time by the same person in your house (assume you live on top of radioactive waste). I tend to say the number of cores has no impact on pure transfers because I am pretty sure they share not only the same bus but the same clock (the maximum clock rate is certainly set to the maximum propagation time of the data along the bus, otherwise the clocks of the cores could be different and phased out to move data quicker on the data bus). This is obviously a simplistic view, many factors come into play - but it has the benefit of adding more reasons why it IS possible.

There are many techniques to encode or decode a large amount of information into one "communication bucket/wagon" (which are called symbols). The R-2R-type Digital-to-analog converter (illustrated below) is one of the simplest yet effective methods of transmitting e.g. 16x faster than the communication clock - provided that the receiver's Analog-to-digital converter converts within a clock period (main limitation: noise).

enter image description here

Nowadays, analog communication not only use amplitude, but phase information to stuff even more bits in a communication clock cycle, as illustrated in This topic (Disclaimer: I have one answer there, but it's the only topic I can remember of which discusses this). And this is only a mere example of a plethora of techniques that become more and more complex with every year that goes by.

enter image description here

For digital communication, it's mostly down to: 1) number of parallel lines (1 for serial communication, USB is a serial link even though it has two data lines, they are always opposite each other in value to increase robustness to common mode noise by difference), and 2) compression. When you send 100MB worth of files to a 10MB-limit mailbox by compressing it, you're effectively transmitting in one clock cycle more than the physical uncompressed datarate limit. Just as there are many modulation techniques (seen above), there are many compression techniques. Note that there are dedicated compression/decompression chips available to take full advantage of compression as a means to increase transmission speed.

Finally, datarate limits are really limits for a given error rate. If you add to your communication protocol some encoding which detects errors and corrects them automatically (using for example state machines on both ends), you can increase the datarate past the stated limit for the same error rate.

That's how data gets in and out of a computer faster than the CPU clock, or how "one man/woman can move more or less buckets from your house to another" (amplitude and/or phase modulation, compression, encoding...).

Now, for the "how to get the buckets from your door to wherever you want it in your house", that's down to the width of your databus: 64bits computers took over from 32 bits computers for this reason, handle twice more data for every clock cycle. That's the number of buckets that can be grabbed at the same time by the same person in your house (assume you live on top of radioactive waste). I tend to say the number of cores has no impact on pure transfers because I am pretty sure they share not only the same bus but the same clock (the maximum clock rate is certainly set to the maximum propagation time of the data along the bus, otherwise the clocks of the cores could be different and phased out to move data quicker on the data bus). This is obviously a simplistic view, many factors come into play - but it has the benefit of adding more reasons why it IS possible.

3 added 145 characters in body
source | link
2 added 1892 characters in body
source | link
1
source | link