Understanding CPU, understanding the various parameters of CPU.

2024-03-16

The CPU is the core of a computer, responsible for all the computations and controls of the computer. Choosing a CPU is the most important part of a DIY computer, so before we choose a CPU, we must first understand some of the most basic knowledge about the CPU. So let me introduce some information about the CPU.

01 Understanding the CPU

I. The meaning and function of the CPU

CPU is the abbreviation of central processing unit in English, which means the central processor and is the core of the computer. It is usually a super-large scale integrated circuit and is the computing center and control center of the computer.

Its main function is to interpret computer instructions and process data in computer software.

The central processor (CPU) mainly includes two parts, namely the controller and the calculator, which also includes high-speed buffer memory and several numerical control, control buses that realize the connection between them.

II. The manufacturing process of the CPU

The CPU is made of semiconductor silicon and some metals and chemical raw materials. The manufacturing of the CPU is an extremely precise and complex process, and only a few manufacturers have the ability to research and produce CPUs today. So how complex is it, and what processes need to go through to be completed?

Advertisement

1. Silicon purificationThe material used to produce chips such as CPUs is a semiconductor, and the main raw material at this stage is silicon (Si), which is a non-metallic element. From a chemical perspective, because it is located at the boundary between metals and non-metals in the periodic table, it has semiconductor properties, making it suitable for manufacturing various tiny transistors. It is one of the most suitable materials for manufacturing modern large-scale integrated circuits.

During the production process of CPUs, the purity of silicon is required to be very high, almost without any impurities. On average, there can be no more than one impurity atom in every 100 million silicon atoms. In the process of purifying silicon, the raw material silicon is melted and placed into a huge quartz furnace. At this time, a crystal seed is placed into the furnace so that the silicon crystal can "grow" around this crystal until a nearly perfect single-crystal silicon ingot is formed, and the diameter of this silicon ingot is mostly 300 millimeters.

Single-crystal silicon ingot

2. Cutting wafers

After purification, the next step is to cut the single-crystal silicon ingot into slices. Because it is cut horizontally from a cylindrical body, the section is circular, so it is called a wafer.

Wafers are the real materials for making CPUs.

Cutting wafers is to use a machine to cut a silicon wafer of a predetermined specification from a single-crystal silicon rod and divide it into multiple small areas. Each area will become a CPU core. Generally speaking, the thinner the wafer is cut, the more CPU products can be made from the same amount of silicon material.

Next, the cut wafers are polished and checked for deformation or other issues. Quality inspection directly determines the final yield of CPUs. The polished wafers are almost flawless, and the surface can even be used as a mirror. So just the process of cutting wafers is so complicated, let alone other processes. Even the famous Intel company does not produce such wafers, but purchases finished products from third parties and then uses its own technology for other processes.

3. Photolithography

After polishing is completed, a photosensitive resist material called photoresist or photoresist is applied on the silicon oxide layer obtained through thermal treatment.4. Etching

This step is the most crucial in the CPU manufacturing process and is a key technology in the CPU industry. Etching technology pushes the application of light to the extreme. The etching process uses ultraviolet light with a very short wavelength, combined with a large lens. The short-wavelength light passes through the holes in these quartz masks and is exposed to the photoresist.

To avoid exposing areas that should not be exposed to light, masks must be made to shield these areas. The chemical reaction that occurs is similar to the change in film after an old-fashioned camera is pressed. The area exposed to ultraviolet light dissolves the light-sensitive material. Next, stop the light exposure and remove the mask, and use a specific chemical solution to wash away the exposed photosensitive anti-etching film and a layer of silicon closely attached to the anti-etching film underneath. This is a rather complex process, and the complexity of each mask needs to be described with 10GB of data.

Then, the exposed silicon will be bombarded by atoms, causing the exposed silicon substrate to be doped locally, thus changing the conductivity of these areas to create N-wells or P-wells. Combined with the substrate made above, the CPU's gate circuit is completed.

5. Repeat and Layer

To process a new layer of circuit, grow silicon oxide again, then deposit a layer of polysilicon, apply a photosensitive substance, repeat the printing and etching process to obtain a groove structure containing polysilicon and silicon oxide. Repeat this process multiple times to form a 3D structure, which is the final CPU core. Fill in metal as a conductor between every few layers. The number of layers depends on the layout of the CPU during the design and the size of the current passing through.

6. Wafer Testing and Slicing

After the wafer is made, it needs to be tested. This step will test the electrical performance of the wafer to check for any errors and which step the errors occurred in. Next, the wafer is cut into blocks, each of which is a processor core. The cores with defects found during the testing process are discarded, and the good ones are prepared for the next step. Each CPU core on the wafer will be tested separately. The key characteristics of each processable processor can be identified, such as the highest frequency, power consumption, heat generation, etc., and the processor's grade is determined. If the performance is good and stable, it is used as a high-end processor core; otherwise, it is defined and locked according to the stable frequency of the core, packaged, and sold as a mid-range processor. In this way, the grading of the entire series is completed.

7. Packaging

At this time, the CPU is a piece of wafer, which cannot be used directly by users. It must be sealed in a ceramic or plastic casing. This allows it to be easily mounted on a circuit board. The packaging structure varies, but the higher the grade of the CPU, the more complex the packaging. New packaging often brings improvements in the electrical performance and stability of the chip, and indirectly provides a solid and reliable foundation for the increase in the main frequency.CPU Parameters

When selecting a CPU, the most important reference standards are the various parameters of the CPU.

1. CPU Frequency

(1) Clock Speed

Clock speed, also known as the clock rate, is measured in megahertz (MHz) or gigahertz (GHz), and it represents the speed at which the CPU performs calculations and processes data. Generally, the higher the clock speed, the faster the CPU processes data.

CPU Clock Speed = Base Frequency * Multiplier. There is a certain relationship between the clock speed and the actual computational speed, but it is not a simple linear relationship. Therefore, the CPU clock speed is not directly related to the actual computational capability of the CPU; it also depends on various performance indicators such as the CPU pipeline and bus.

(2) Base Frequency

Base frequency is the fundamental frequency of the CPU, measured in megahertz (MHz). The base frequency of the CPU determines the operating speed of the entire motherboard. In layman's terms, overclocking in desktop computers is essentially increasing the CPU's base frequency. The CPU determines the operating speed of the motherboard, and more accurately, it directly affects the operating frequency of the memory.

(3) Multiplier

The multiplier refers to the relative proportion between the CPU clock speed and the base frequency. Under the same base frequency, the higher the multiplier, the higher the CPU frequency. However, in reality, under the premise of the same base frequency, a high multiplier CPU itself does not have much significance. This is because the speed of data transmission between the CPU and the system is limited. If one blindly pursues a high clock speed and obtains a high multiplier CPU, a significant bottleneck effect will occur—the maximum speed at which the CPU can obtain data from the system cannot meet the CPU's computational speed. Generally, apart from engineering samples of Intel CPUs, other CPUs have locked multipliers.Translate the following article into English:

(4) Front Side Bus Frequency

The front side bus is the bus that connects the CPU to the northbridge chip. When selecting a motherboard and CPU, it is important to consider the compatibility between the two. Generally speaking, the front side bus is determined by the CPU. If the motherboard does not support the front side bus required by the CPU, the system will not work. In other words, both the motherboard and CPU need to support a specific front side bus for the system to function. However, the default front side bus of a CPU is unique, so the front side bus of a system is mainly determined by the CPU. This is one of the reasons why I said to choose the CPU before selecting the motherboard. The front side bus is the data channel between the processor and the motherboard's northbridge chip or memory controller hub, and its frequency directly affects the speed at which the CPU accesses memory.

Due to different technologies used by Intel and AMD, the relationship between their FSB frequencies and external frequencies is also different. The current relationship for Intel processors is FSB frequency = external frequency * 4; while for AMD, it is FSB frequency = external frequency * 2.

The difference between the external frequency and the front side bus FSB frequency is that the front side bus speed refers to the speed of data transmission, while the external frequency is the synchronous operation speed between the CPU and the motherboard. In other words, a 100MHz external frequency specifically indicates that the digital pulse signal oscillates once per second; while a 100MHz front side bus means that the amount of data the CPU can accept per second is 100MHz * 64 bits / 8 bits / Byte = 800MB/s.

2. CPU Cache

Cache refers to the area where high-speed data exchange can take place. The size of the cache is also one of the important indicators of the CPU, and the structure and size of the cache have a significant impact on the CPU speed. The smaller the cache capacity, the higher the operating frequency, which generally operates at the same frequency as the processor, and the work efficiency is much greater than that of the system memory and hard disk.

In actual operation, when the CPU needs to read data, it first searches in the high-speed cache. If found, it is used directly; otherwise, it searches and uses data from the memory, and then puts it into the cache. Because the high-speed cache is extremely fast, it directly improves the CPU's processing and computing capabilities.

Level 1 cache is the first layer of high-speed cache in the CPU, divided into data cache and instruction cache. The capacity and structure of the level 1 cache have a greater impact on the CPU's performance. However, high-speed buffer memory is composed of static RAM, with a more complex structure. Under the condition that the CPU chip area cannot be too large, the capacity of the level 1 cache cannot be too large.

Level 2 cache is the second layer of high-speed cache in the CPU, divided into internal and external chip types. The internal chip's level 2 cache operates at the same speed as the main frequency, while the external level 2 cache only has half of the main frequency. The capacity of the level 2 cache also affects the CPU's performance, and theoretically, the larger the better.

Level 3 cache is divided into two types, the early external form, and now integrated into the CPU. The level 3 cache is not as fast as the first and second level caches, but it is much larger in capacity. The current mainstream CPU level 3 cache is around 8M.

Comment