This is very important questions considering the speed limit of the processor. The speed of the processor is depends upon the number of operations it performed per unit time. This speed is depends upon, one of the key component of the processor that is a transistor.
The most important parts in a processor are the transistors, the electronic devices that act as switches in order to construct logical gates. These logical gates are the hard-working components of our processors. Put together in different combinations, they form units capable of arithmetic and complex logical operations.
So the speed of the processor is depends upon the number of transistors on the processor chip. According to Moor's Law the integration density of transistors doubles every 18 to 24 months. So the Moor;s Law is true, since the most powerful processor has billions of transistors on the processor chip. But there is a limitation for increasing the number of transistors integrated on the processor chip.
Thermal losses occur when you are putting several billions of
transistors together on a small area and switching them on and off again
several billion times per second. The faster we switch the transistors
on and off, the more heat will be generated. Without proper cooling,
they might fail and be destroyed. One implication of this is that a
lower operating clock speed will generate less heat and ensure the
longevity of the processor. Another severe drawback is that an increase
in clock speed implies a voltage increase and there is a cubic
dependency between this and the power consumption. Power costs are an
important factor to consider when operating computing centres.
There is also the size issue of the processor chip. The smallest feature size to be found in a processor today is some 150 nano meters (this is the so called 'channel length' in a MOSFET). You can go smaller than that, but with ever increasing difficulty. Conventional transistors may still work down to about 50 nano meters channel length, giving your processor clock speeds around 20 GHz or so, but beyond that, its anybody's guess. At such small lengths, a transistor isn't really well behaved, so one can't use it as one has been.
For the above reason we cant increase the speed and performance of the processor beyond certain limit. So there is death of single core means one processor. Now a days every computers come with more than one processors embed on a chip.
But by increasing the number of cores the complexity of the system increases. The programmer has to program such that the all of the cores are utilised for solving a certain problem. So the need of parallel and concurrent programming is must. So there is concept of multi threading. A programmer can create threads and give each thread a task which can be executed parallel on the cores. To achieve this the problem must be divided in parallel parts. But again there is a limit for parallelizing the task.
This concept is specified by Amdahl's Law. It states that the performance improvement to be gained from using some faster mode of execution is limited by the fraction of the time the faster mode can be used.
The most important parts in a processor are the transistors, the electronic devices that act as switches in order to construct logical gates. These logical gates are the hard-working components of our processors. Put together in different combinations, they form units capable of arithmetic and complex logical operations.
So the speed of the processor is depends upon the number of transistors on the processor chip. According to Moor's Law the integration density of transistors doubles every 18 to 24 months. So the Moor;s Law is true, since the most powerful processor has billions of transistors on the processor chip. But there is a limitation for increasing the number of transistors integrated on the processor chip.
Moor's Law (singularitysymposium.com)
There is also the size issue of the processor chip. The smallest feature size to be found in a processor today is some 150 nano meters (this is the so called 'channel length' in a MOSFET). You can go smaller than that, but with ever increasing difficulty. Conventional transistors may still work down to about 50 nano meters channel length, giving your processor clock speeds around 20 GHz or so, but beyond that, its anybody's guess. At such small lengths, a transistor isn't really well behaved, so one can't use it as one has been.
For the above reason we cant increase the speed and performance of the processor beyond certain limit. So there is death of single core means one processor. Now a days every computers come with more than one processors embed on a chip.
But by increasing the number of cores the complexity of the system increases. The programmer has to program such that the all of the cores are utilised for solving a certain problem. So the need of parallel and concurrent programming is must. So there is concept of multi threading. A programmer can create threads and give each thread a task which can be executed parallel on the cores. To achieve this the problem must be divided in parallel parts. But again there is a limit for parallelizing the task.
This concept is specified by Amdahl's Law. It states that the performance improvement to be gained from using some faster mode of execution is limited by the fraction of the time the faster mode can be used.
No comments:
Post a Comment