With each release of a new CPU What manufacturers usually pay attention to in presenting is such as higher performance increased ability Support for multimedia work better than ever And most importantly, the rate of energy consumption is reduced compared to the previous model. Causing new processor chips to be more powerful, but consume less power Compared to older chips in the same level In this article we will look at why this is so.
for measuring power consumption Power consumption of the processor chip In most cases, it will mainly refer to the TDP value from the manufacturer. For those who are wondering what the TDP value is, you can read it from this article. First of all, let’s see the picture first that this new CPU is stronger. But how do you consume less power?
from the picture above Compared to the light blue Intel Core i9-12900HK and the darker blue Intel Core i9-11980HK, both are similar chips. But there was a massive change in the production architecture. It can be seen that at the same level of energy consumption The percentage of performance of the Gen 12 chips will be significantly higher than the Gen 11. At the 45W point, for example, the i9-11980HK chip has a relative efficiency of 100%, while the i9-12900HK is already around 125% at that power point. equal electricity
Which the efficiency per energy consumption is better. inevitably affects the actual work as well Obviously, it makes newer notebooks stronger and work faster than older models. Plus it’s more energy efficient. This makes it easier for manufacturers to select high-end chips to put in. Allows us to see more of these gaming notebooks with a thinner body today.
For why the new processor chip is stronger but consumes less power as follows
CPU production architecture modifications
Both Intel and AMD are the two big players in the PC processor market. There is research and development to improve the chip manufacturing architecture. (architecture) always and usually introduces a new architecture to their products every 1-2 years (or maybe longer in the past), which manufacturers mainly rely on process improvements. semiconductor manufacturing (semiconductor) to make the transistor inside the CPU smaller when smaller The chip package can contain more transistors. With more transistors on a chip It will also increase the ability to process.
which in terms of marketing is familiar Each manufacturer has a numeric architecture level identification. It is in nanometers (nm), the smaller the number. It will be a newer architecture, numbers that many of you may have seen through your eyes often, such as 90nm and 65nm in the Pentium 4 era with AMD Athlon 64, including 22nm in the Intel Core i era that uses the code name Ivy. Bridge, etc. At present, we are now in the 10nm, 7nm, and 5nm eras and will soon move to the 2nm era. It will change the unit to the Angstrom level. which is a unit that is 10 times smaller than a nanometer
However, the nm identification of each manufacturer Each chip may be different. From before, it was often used to refer to the transistor size of each node. but nowadays It sometimes refers to the distance between the gate and the metallic contacts that make up the internal structure of the transistor. Therefore, now it is almost impossible to directly compare the nm numbers. to measure who’s technology is more advanced Including may need to see notes Take a look at the technical information to see what the nm numbers on each point mean.
In terms of why smaller transistors As a result, the chip consumes less power accordingly. The main reasons are as follows.
1. The smaller size of the transistor makes it possible toPlace the transistors closer together. Helps to reduce the amount of electricity that is lost.while traveling between transistors back and forth and when there is less electricity loss is equivalent to the system not having to draw more power to compensate for the loss as much as the previous model
2. When the transistor is small Resulting in a reduced electrical capacity as well (But relying on compensation by increasing the number of transistors instead), where the capacitance decreases. cause the charge to the transistor and discharging It uses less power than a large transistor. Resulting in a loss of power as well
3. A small transistor canIt is easier to change the state of the switch to turn on or off the built-in circuit.large transistor which changes status quickly Makes the processor chip work faster. Job done faster This results in lower energy consumption.
4. The use of small transistors that consume less power allows The CPU can run at lower power than the previous generation.
5. Improve the materials used in the production of transistors, such as the use of High-K Metal Gate (HKMG) that has the ability to prevent loss of current, making it possible tobring the light to its fullest potential therefore consumes less power
Summary is
Transistors are smaller, consume less power, fit more in the CPU, and therefore perform better. Use less power than before
Improving work styles and command sets
In addition to the internal hardware of the processor chip that has been developed. In terms of work processes, there are always improvements as well, such as
Modify the operation of the pipeline system. to be able to support more simultaneous tasks Distribute tasks appropriately to the capacity of each core, each thread, improving how cache data is accessed faster. high efficiency including improving predictions And how to fetch data from SSD to cache to reduce the problem of not finding the desired data in the cache (cache miss) causing the system to frequently have to fetch data from the SSD, which has to be fetched many times Wasting both time and energy as well Add new commands to better support specific tasks, such as media-related commands. security instruction set to reduce the workload of the main processor Add modules that perform specific tasks, such as neural network modules that handle AI tasks, image processing modules, etc.
All of these have the main purpose of reducing the load on the CPU core by delegating some of the work to more specialized modules. Work faster than handle instead. and shorten the overall processing time. which the processing chip takes less time to work It also helps to reduce energy consumption as well. That’s why the new generation of processors work faster. Multitasking Plus it consumes less power as well.
The RAM chip is also packaged in the same package as the CPU, like in Apple’s M family of chips, and when combined with other modules that are packaged together, they are called “RAM chips”. System-on-Chip (SoC) to allow the processor to access data in RAM as quickly as possible. which is the same The shorter the time it takes to access the information. The system will only run faster. resulting in high efficiency and less power consumption as well
Summary is
Add a dedicated work area to share the CPU load, so the work can be completed faster, using less power.
Improved sleep and idle states.
CPUs are not usually running at 100% full potential all the time, but they are scaled to suit the needs. For example, if it’s turning on the device to standby, leave it. The processor may not work more than 10% depending on the background process running. Or if it’s a new processor, such as Intel’s 12th and 13th Gen, that have cores categorized as performance cores (P-Cores) and mainstream cores. Emphasis on high energy efficiency (efficiency core or E-Core). Tasks that do not require high computing power will mainly use the E-Core, which already has sufficient power. For heavier tasks such as playing games, rendering, and complex calculations. It will send the job to the P-Core to do instead. Once done, it will adjust the status to idle that reduces the overall power consumption of the system. But still get the performance that is suitable for working at that time
In addition to enabling the necessary number and type of cores. The processor chip can also adjust the speed appropriately. Which will also be related to the voltage level, which on the Intel side uses the technology name SpeedStep, on the AMD side is Cool’n’Quiet.
The working state in relation to CPU power consumption is referred to as the C-state according to the ACPI standard. On Gen 13 Intel chips, it supports up to six core state levels: C0, C1, C1E, C6, C8 and C10 at each level have their work patterns. and different energy consumptions, summarized as follows:
C0 – Running normally C1 – AutoHalt the core execution C1E – The core is still in the C1 state, but speeding and lowering the voltage to the lowest allowable level C6 , C8 and C10 – Move data from L1 and L2 level caches to L3 shared by all cores. Then save the state to SRAM and reduce the voltage of that core. It may drop to 0V to disable that core altogether. depending on the depth of the state
In addition, there is also a division of the state of various packages in the processor chip (Package C-state) and the state of the system. This allows the system to manage power consumption in more detail. as in the table below
The G-state is the state of the entire computer system, while the S-state is the sleep state, to put it simply, the level of sleep. This, when matched with the processor’s C-state, gives the system the option to adjust performance levels. and power consumption levels are more varied than before that may not have been divided as thoroughly as at present thus making it possible to adjust the speed and power consumption in detail and meets the needs more To read more details can click intoDownload the datasheet file from Intel. ok
Summary is
Various improvements to the state of the chip and the system. Support speed adjustment and reduce the use of electricity as needed than ever
operating system and software that supports the CPU better.
In addition to the chip itself that has always been developed. The operating system side is also designed to work with new technologies in the processor chip as well. The most obvious is the transition from a single-core chip with the help of Hyper-Threading to a full-core, multi-core era that was not yet supported by the OS and early software. full, especially games that sometimes put heavy work on one core The rest has almost no work to process at all. We see that the performance of the system is not as high as it should be.
But after that, OS manufacturers, especially major ones like Microsoft, designed their OS to support more multi-core operations. Especially the system to distribute the work properly. Obviously, since the Windows Vista era has been fine-tuned. Optimized systems to handle multi-core chips better than Windows XP, but with a significant leap in resource usage. Before starting to fit in the Windows 7 era
which when the OS and the processor can help each other to manage the division of work Properly distribute the load between cores. It makes it possible to complete the work faster. Consume less overall energy as well, so for general use that does not require a specific program It’s a good idea to use the latest OS that is updated as regularly as possible. in order for the hardware to work at its full potential. and increase the security of the system at the same time
Summary is
Newer OSes are designed to better support the multi-core performance of newer chips. Resulting in high efficiency, the work is completed faster, the overall power consumption is reduced.
which when all assembled As a result, new processor chips are more efficient, stronger than before, along with lower power consumption. which uses lower power This also results in reduced heat released during operation. Nowadays, we have seen thin and light laptops. Comes with a high-performance chip There are more cores. Plus, the battery lasts for hours. on the desktop side The obvious thing is the heat reduction. until it can be put into a small case It also supports better overclocking. And it’s easier than ever.
