▷ Parts of a processor outside and inside: basic concepts?

Surely we all know roughly what a CPU is, but do we really know what the parts of a processor are ? Each and every one of the main ones, which are necessary for this small square of silicon to be able to process large amounts of information, being able to transfer humanity to an era where, without having electronic systems, would be a complete debacle.

Processors are already part of our daily lives, especially of people who have been born in the last 20 years. Many have grown completely mixed with technology, not to mention the little ones who bring a Smartphone under their arms instead of a loaf… In all these devices, there is a common element called a processor, which is responsible for giving "intelligence" to the machines around us. If this element did not exist, neither would computers, mobiles, robots and assembly lines, in short, everyone would have work… but it would be impossible to get to where we have made them, there is still no world like "Matrix" but Everything will go.

Index of contents

What is a processor and why is it so important

First of all, we must be aware that not only a computer has a processor inside. All electronic devices, all, have within them an element that functions as a processor, whether it is a digital clock, a programmable automaton or a Smartphone.

But of course, we must also be aware that, depending on their capabilities and for what they are manufactured, processors can be more or less complex, from simply executing a succession of binary codes to light an LED panel, to handling huge amounts of information, including learning from them (Machine Learning and Artificial Intelligence).

The CPU or Central Processing Unit in Spanish is an electronic circuit capable of executing the tasks and instructions contained in a program. These instructions are greatly simplified, and boil down to basic arithmetic calculations (addition, subtraction, multiplication, and division), logical operations (AND, OR, NOT, NOR, NAND), and input / output (I / O) control. of the devices.

Then the processor is the element in charge of carrying out all the operations that form the instructions of a program. If we put ourselves in the machine's point of view, these operations are reduced to simple chains of zeros and ones, called bits, and that represent the current / non-current states, thus forming binary logical structures that even the human being is capable of. to understand and program in machine code, assembler or through a higher level programming language.

The transistors, the culprits of everything

Processors wouldn't exist, at least as small, if it weren't for the transistors. They are the basic unit so to speak, of any processor and integrated circuit. It is a semiconductor device that closes or opens an electrical circuit or amplifies a signal. In this way, it is how we can create ones and zeros, the binary language that the CPU understands.

These transistors started out as vacuum valves, huge light-bulb-like devices capable of performing the transistor's own commutations, but with mechanical elements in a vacuum. Computers like the ENIAC or EDVAC had vacuum valves inside them instead of transistors and they were immensely large and practically consumed the energy of a small city. These machines were the first with Von Neumann architecture.

But in the 1950s to 1960s, the first transistor CPUs began to be created - in fact, it was IBM in 1958 when it created its first semiconductor transistor-based machine with the IBM 7090. Since then the evolution was spectacular, manufacturers like Intel and later AMD began to create the first processors for desktop computers, implementing the revolutionary x86 architecture, thanks to the Intel 8086 CPU. In fact, even today, our desktop processors are based on this architecture, later we will see the parts of the x86 processor.

After this, the architecture began to become increasingly complex, with smaller chips and also with the first introduction of more cores inside, and then with cores specially dedicated to graphics processing. Even ultra-fast memory banks called cache memory and the connection bus with the main memory, RAM, were introduced inside these small chips.

The external parts of a processor

After this brief review of the history of processors until we are in our day, we will see what external elements a current processor has. We are talking about physical elements that can be touched and that are in view of the user. This will help us better understand the physical and connectivity needs of a processor.

Socket

The CPU socket or socket is an electromechanical system fixedly installed on a motherboard that is responsible for interconnecting the processor with the other elements on the board and the computer. There are several basic types of socket on the market and also with many different configurations. There are three elements in your name or denomination that will make us understand which one we are talking about:

The manufacturer can be Intel or AMD in the case of personal computers, this is something simple to understand. As for the type of connection we have three different types:

• LGA: (grid contact array), means that the contact pins are installed in the socket itself, while the CPU only has a flat contact array. PGA: (grid array of pins), it is just the opposite of the previous one, it is the processor that has the pins, and the socket the holes to insert them. BGA: (ball grid array), in this case the processor is directly soldered to the motherboard.

As for the last number, it identifies the type of distribution or number of connection pins that the CPU has with the socket. There is an enormous amount of them both in Intel and AMD.

Substratum

The substrate is basically the PCB where the silicon chip that contains the electronic circuit of the cores, called DIE, is installed. Today's processors may have more than one of these elements installed separately.

But also this small PCB contains the entire matrix of connection pins with the socket of the motherboard, almost always gold plated to improve the transfer of electricity, and with protection against overloads and current surges in the form of capacitors.

DIE

The DIE is precisely the square or chip that contains all the integrated circuit and internal components of a processor. Visually, it is seen as a small black element protruding from the substrate and making contact with the heat dissipation element.

Because the entire processing system is inside it, the DIE reaches incredibly high temperatures, so it must be protected by other elements.

IHS

Also called DTS or Integrated Thermal Diffuser, and its function is to capture all the temperature of the processor cores and transfer it to the heatsink that this element has installed. It is made of copper or aluminum.

This element is a sheet or encapsulation that protects the DIE from the outside, and can be in direct contact with it by means of thermal paste or directly welded. In custom gaming equipment, users remove this IHS to place heatsinks directly in contact with the DIE using thermal paste in a liquid metal compound. This process is called Delidding and its purpose is to substantially improve processor temperatures.

Heatsink

The final element that is responsible for capturing as much heat as possible and transferring it to the atmosphere. They are small or large blocks made of aluminum and a copper base, provided with fans that help cool the entire surface by means of a forced air current through the fins.

Every PC processor needs a heatsink to function and keep its temperatures under control.

Well these are the parts of a processor externally, now we are going to see the most technical part, its internal components.

Von Neumann architecture

Today's computers are based on the architecture of Von Neumann, who was the mathematician in charge of giving life to the first computers in history in 1945, you know, ENIAC and its other big friends. This architecture is basically the way in which the elements or components of a computer are distributed so that its operation is possible. It consists of four basic parts:

• Program and data memory: it is the element where the instructions to be executed in the processor are stored. It consists of storage drives or hard drives, random access RAM, and programs that contain the instructions themselves. Central Processing Unit or CPU: this is the processor, the unit that controls and processes all the information that comes from the main memory and input devices. Input and output unit: allows communication with peripherals and components that are connected to the central unit. Physically we could identify them as the slots and ports of our motherboard. Data buses: are the tracks, tracks or cables that physically connect the elements. In a CPU they are divided into the control bus, data bus and the address bus.

Multi-core processors

Before we begin listing the internal components of a processor, it is very important to know what the cores of a processor are and their function in it.

The core of a processor is the integrated circuit that is responsible for performing the necessary calculations with the information that passes through it. Each processor operates at a certain frequency, measured in MHz, which indicates the number of operations it is capable of performing. Well, current processors not only have a core, but several of them, all of them with the same internal components and capable of executing and solving instructions simultaneously in each clock cycle.

So if a core processor can execute one instruction in each cycle, if it had 6, it could execute 6 of these instructions in the same cycle. This is a dramatic performance upgrade, and it's precisely what today's processors do. But we not only have cores, but also processing threads, which are like a kind of logical cores through which the threads of a program circulate.

Visit our article on: what are the threads of a processor? Differences with the nuclei to know more on the subject.

Internal parts of a processor (x86)

There are many different microprocessor architectures and configurations, but the one that interests us is the one that is inside our computers, and this is undoubtedly the one that receives the name of x86. We could see it directly physically or schematically to make it a little clearer, know that all this is within the DIE.

We must bear in mind that the Control Unit, Arithmetic-Logic Unit, Registers and FPU will be present in each of the processor cores.

Let's first look at the main internal components:

Control unit

In English called Conrol Unit or CU, it is in charge of directing the operation of the processor. It does this by issuing commands in the form of control signals to the RAM, the arithmetic-logic unit, and the input and output devices so that they know how to manage the information and instructions that are sent to the processor. For example, they collect data, perform calculations, and store results.

This unit ensures that the rest of the components work in synchronization using clock and timing signals. Virtually all processors have this unit inside, but let's say it is outside of what is the core of processing itself. In turn, we can distinguish within it the following parts:

• Clock (CLK): it is responsible for generating a square signal that synchronizes the internal components. There are other clocks that are in charge of this synchrony between elements, for example, the multiplier, which we will see later. Program counter (CP): contains the memory address of the next instruction to be executed. Instruction Register (RI): saves the instruction that is being executed Sequencer and Decoder: interprets and executes the instructions through commands

Arithmetic-Logical Unit

You will surely know this by its acronym “ALU”. The ALU is responsible for carrying out all the arithmetic and logical calculations with integers at the bit level, this unit does work directly with the instructions (operands) and with the operation that the control unit has instructed it to do (operator).

The operands can come either from the internal registers of the processor, or directly from the RAM memory, they can even be generated in the ALU itself as a result of another operation. The output of this will be the result of the operation, being another word that will be stored in a register. These are its basic parts:

• Registers of entrance (REN): they keep in them the operands to be evaluated. Operation code: the CU sends the operator so that the operation will be carried out Accumulator or Result: the result of the operation comes out of the ALU as a binary word Status register (Flag): it stores different conditions to take into account during the operation.

Floating point unit

You will know it as FPU or Floating Point Unit. Basically it is an update carried out by the new generation processors that specializes in the calculation of floating point operations using a mathematical coprocessor. There are units that can even carry out trigonometric or exponential calculations.

Basically it is an adaptation to increase the performance of the processors in the graphics processing where the calculations to be performed are much heavier and more complex than in normal programs. In some cases, the functions of the FPU are performed by the ALU itself using an instruction microcode.

Records

Today's processors have their own storage system, so to speak, and the smallest and fastest unit is the registers. Basically it is a small warehouse where the instructions that are being processed and the results obtained from them are stored.

Cache

The next level of storage is cache memory, which is also extremely fast memory, much more than RAM memory that is responsible for storing the instructions that will be imminently used by the processor. Or at least you will try to store the instructions that you think will be used, as sometimes there is no choice but to request them directly from RAM.

The cache of current processors is integrated into the same DIE of the processor, and is divided into a total of three levels, L1, L2 and L3:

• Level 1 Cache (L1): It is the smallest after logs, and the fastest of the three. Each processing core has its own L1 cache, which in turn is divided into two, the L1 Data which is responsible for storing the data, and the L1 Instruction, which stores the instructions to carry out. It is usually 32KB each. Level 2 Cache (L2) - This memory is slower than L2, but also larger. Typically, each core has its own L2, which may be about 256 KB, but in this case it is not directly integrated into the core circuit. Level 3 Cache (L3): It is the slowest of the three, although much faster than RAM. It is also located outside the nuclei and is distributed among several nuclei. It ranges between 8 MB and 16 MB, although in very powerful CPUs it reaches up to 30 MB.

Inbound and outbound buses

The bus is the communication channel between the different elements that make up a computer. They are the physical lines through which the data circulates in the form of electricity, the instructions and all the elements necessary to process. These buses can be placed directly inside the processor or outside it, on the motherboard. There are three types of buses on a computer:

• Data bus: surely the easiest to understand, because it is the bus through which the data sent and received by the different components circulates, to or from the processor. This means that it is a bidirectional bus and through it will circulate words with a length of 64 bits, the length that the processor is capable of handling. An example of a data bus are LANES or PCI Express Lines, which communicate the CPU with the PCI slots, for example, for a graphics card. Address bus: the address bus does not circulate data, but memory addresses to locate where the data stored in memory is. RAM is like a large data store divided into cells, and each of these cells has its own address. It will be the processor that asks the memory for the data by sending a memory address, this address must be as large as cells have the RAM memory. Currently a processor can address memory addresses of up to 64 bits, that is, we could handle memories of up to 2 ⁶⁴ cells. Control bus: the control bus is in charge of managing the two previous buses, using control and timing signals to make synchronized and efficient use of all the information that circulates to or from the processor. It would be like the air traffic control tower of an airport.

BSB, input / output unit and multiplier

It is important to know that the current processors do not have the traditional FSB or Front Bus, which served to communicate the CPU with the rest of the elements of the motherboard, for example, chipset and peripherals through the north bridge and the south bridge. This is because the bus itself has been inserted into the CPU as an input and output (I / O) data management unit that directly communicates the RAM with the processor as if it were the old north bridge. Technologies like AMD's HyperTransport or Intel's HyperThreading are responsible for managing the exchange of information on high-performance processors.

The BSB or Back Side Bus is the bus that is in charge of connecting the microprocessor with its own cache memory, normally that of L2. In this way the Front Bus can be freed from quite a load, and thus bring the speed of the caches even closer to the speed of the core.

And finally we have the multipliers, which are a series of elements located inside or outside the processor that are responsible for measuring the relationship between the CPU clock and the clock of the external buses. At this point we know that the CPU is connected to elements such as RAM, the chipset and other peripherals through buses. Thanks to these multipliers, it is possible that the CPU frequency is much faster than the external buses, in order to be able to process more data.

A multiplier of x10 for example, will allow a system that works at 200 MHz, to work on the CPU at 2000 MHz. In current processors, we can find units with the multiplier unlocked, this means that we can increase its frequency and thus its processing speed. We call this overclocking.

IGP or internal graphics card

To finish with the parts of a processor we can not forget the integrated graphics unit that some of them carry. Before we have seen what a FPU is, and in this case we are facing something similar, but with much more power, since basically they are a series of cores capable of independently processing the graphics of our team, which for mathematical purposes, are a massive amount of floating point calculations and graphics rendering that would be very processor intensive.

The IGP does the same function as an external graphics card, the one we installed over the PCI-Express slot, only at a smaller scale or power. It is called the Integrated Graphics Processor because it is an integrated circuit installed in the same processor that relieves the central unit of this series of complicated processes. It will be useful when we do not have a graphics card, but for now, it does not have a performance comparable to these.

Both AMD and Intel have units that integrate IGP in the CPU, thus being called APU (Accelerated Processing Unit). An example of this is almost all the Intel Core of the i family, along with the AMD Athlon and some Ryzen.

Conclusion on the parts of a processor

Well, we come to the end of this long article where we see in a more or less basic way what the parts of a processor are, both from an external and internal point of view. The truth is that it is a very interesting topic but damn complex and long to explain, the details of which are beyond the understanding of almost all of us who are not immersed in the assembly lines and manufacturers of this type of device.

Now we leave you with a few tutorials that may be interesting to you.

If you have any questions or want to clarify any issue in the article, we invite you to write it in the comment box. It is always good to have the opinion and wisdom of others.