What is a processor? Features and characteristics. ARM is the future

A computer's processor is the main component of a computer, its "brain", let's put it this way. It performs all logical and arithmetic operations that the program specifies. In addition, it manages all computer devices.

What is a modern processor

Processors today are manufactured in the form of microprocessors. Visually, a microprocessor is a thin, rectangular-shaped crystalline silicon plate. The area of ​​the plate is a few square millimeters, it contains circuits that provide the functionality of the PC processor. As a rule, the plate is protected by a ceramic or plastic flat case, to which it is connected by means of gold wires with metal lugs. This design allows the processor to be connected to the computer's motherboard.

• address buses and data buses;
• arithmetic logic unit;
• registers;
• cache (fast memory of a small size 8-512 Kbytes);
• command counters;
• math coprocessor.

What is processor architecture?

Processor architecture is the ability of a processor to execute a set of machine codes. This is from the point of view of programmers. But the developers of computer components adhere to a different interpretation of the concept of "processor architecture". In their opinion, processor architecture is a reflection of the basic principles of the internal organization of certain types of processors. Let's say the Intel Pentium architecture is designated P5, Pentium II and Pentium III - P6, and not so long ago the popular Pentium 4 - NetBurst. When Intel closed the P5 to competing vendors, AMD designed its K7 architecture for the Athlon and Athlon XP, and the K8 for the Athlon 64.

Even processors with the same architecture can differ significantly from each other. These differences are due to the variety of processor cores that have a specific set of characteristics. The most common differences are the different system bus frequencies, as well as the size of the L2 cache and the technological characteristics of the processors. Very often, changing the core in processors from the same family also requires replacing the processor socket. And this entails problems with the compatibility of motherboards. But manufacturers are constantly improving the kernels and making constant, but not significant, changes to the kernel. Such innovations are called kernels revision and, as a rule, are denoted by alphanumeric combinations.

The system bus or processor bus (FSB - Front Side Bus) is a collection of signal lines that are combined according to their purpose (addresses, data, etc.). Each line has a specific communication protocol and electrical characteristics. That is, the system bus is the glue that connects the processor itself and all other PC devices (hard drive, video card, memory, and much more). Only the CPU is connected to the system bus itself, all other devices are connected through controllers that are located in the north bridge of the system logic set (chipset) of the motherboard. Although in some processors, the memory controller is plugged directly into the processor, which provides a more efficient memory interface to the CPU.

Cache or fast memory is an indispensable component of all modern processors. The cache is a buffer between the processor and the controller of reasonably slow system memory. The buffer stores blocks of data that are being processed at the moment, and the processor does not need to constantly access the slow system memory. Naturally, this greatly increases the overall performance of the processor itself.

In processors in use today, the cache is divided into several levels. The fastest is the first level L1, which deals with the processor core. It is usually divided into two parts - the data cache and the instruction cache. L2 interacts with L1 - the second level cache. It is much larger in size and is not divided into instruction cache and data cache. Some processors have L3 - the third level, it is even larger than the second level, but an order of magnitude slower, since the bus between the second and third levels is narrower than between the first and the second. Nevertheless, the speed of the third level is still much higher than the speed of system memory.

There are two types of cache - exclusive and non-exclusive.

An exclusive type of cache is one in which information at all levels is strictly delimited to the original.

A non-exclusive cache is a cache in which information is repeated across all levels of the cache. It is difficult to say which type of cache is better, and the former and the latter have their merits and demerits. An exclusive cache type is used in AMD processors, not an exclusive one - Intel.

The processor connector can be slotted or female. In any case, its purpose is to install the central processor. The use of the connector facilitates the replacement of the processor during upgrades and removal for the repair of the PC. The slots can be used for installing the CPU card and the processor itself. Sockets are distinguished by their purpose for certain types of processors or CPU cards.

The first place is occupied by the Intel Core i5 processor. An excellent option for a powerful gaming machine.

Second place - Intel Celeron E3200, despite the fairly decent price. The best option for an office car.

The third place is taken by intel again - this time a 4-core Core 2 Quad.

Fourth place - AMD Athlon II X2 215 2.7 GHz 1Mb Socket-AM3 OEM processor. A good choice for home and office, for those who want to save money and do not need a super powerful machine. In addition, this processor model has a lot of room for overclocking.

Fifth place - AMD Phenom II X4 945. Good price, excellent performance, large cache and 4 cores on board.

If you are ready to pay about $ 1000 for a processor, then you can purchase Intel Core 2 Extreme. But such a processor is unlikely to be suitable for a wide range of consumers. Therefore, we will consider more affordable options.

If you are a simple PC user who works with texts, watches movies, listens to music and surfs the Internet, either the Celeron E1200 or the lower-end Athlon 64 X2 will be fine for you. The latter has certain advantages over the former and will be enough for you for many years.

If you use your computer for fun, occasionally playing games, then you need to look at Core 2 Duo processors. This is the best processor for your needs.

If you are one of the users who use all the capabilities of the computer, work with audio, Internet, video, large programs and heavy games, the Core 2 Duo E8200 is most suitable for you. This processor has high performance, low heat dissipation, sufficient overclocking capability, while being affordable.

Finally, are you an uncompromising gamer and your PC needs to be a powerful foothold for gaming? You just need either a dual-core or quad-core processor, no less.

(or central processing unit, CPU) Is the main component of any computer. It is the processor that is responsible for executing the program code, so the faster the processor, the faster the tasks assigned to the computer are performed.

• Number of computing cores... The number of processing cores determines the number of tasks that the processor can perform in parallel. Currently, desktop computers use processors with 1, 2, 4, 6 and 8 cores. When choosing a processor, the number of cores must be selected for specific tasks that will be performed on the computer. For example, a 2-core processor is enough to work with simple applications or use the Internet, but a 4- or 6-core processor is needed to use professional graphics programs or run demanding games. In addition, some Intel processors can create 2 virtual cores for each actually existing computing core (Hyper-threading technology). Thanks to this trick, it is possible to load the kernel with work more efficiently, and therefore to increase the performance of the computational cores.
• Clock frequency... This is the number of operations performed in one second. This value is measured in megahertz (MHz). The higher the clock frequency, the higher the performance of each individual computing core.
• Bit depth. This parameter affects the ability of the processor to execute 32- or 64-bit programs. The bitness also affects the available amount of RAM, which is limited to 4 GB in 32-bit systems and 16 GB in 64-bit systems.
• Cache size... The cache memory is used by the processor to store data that it regularly uses. Thanks to the cache memory, the processor needs less frequent access to the RAM, which is significantly slower.
• Technological process... Process technology is the size of the transistors in the processor. The smaller the technical process, the lower the power consumption of the processor, as well as its heat dissipation. In addition, a decrease in the technical process allows you to fit more transistors in the same die area, which means an increase in processor performance.
• Processor heat dissipation (TDP)... This parameter shows how much heat the cooling system should remove from the processor.
• Integrated GPU... Usually integrated into a computer motherboard or CPU. With integrated graphics, you can build a computer without separate graphics cards, significantly reducing the cost and power consumption of your system.
• Socket. This is a special connector on the motherboard where the processor is installed. If the socket of the processor and the motherboard does not match, then the processor cannot be installed.

What a PC processor looks like

Major PC Processor Manufacturers

Currently, the largest CPU manufacturers are Intel and AMD.

Intel is undoubtedly the leader in this pair. Intel processors are characterized by higher performance per core and lower heat dissipation, but at the same time their cost is much higher. And this is not always justified - for example, when assembling inexpensive computers.

AMD processors have a good price-performance ratio, but some models are characterized by excessive heat generation. But, despite the fact that processors from this manufacturer lag behind in terms of technology and do not always have high quality, they are still very popular.

Today, millions of people in developed countries have one or even more personal computers and laptops in their homes, which is not surprising, since these devices are used today for a variety of purposes, from entertainment to scientific and communication. However, a significant part of PC owners do not know their devices, since they prefer to contact specialists when problems arise. This approach is quite reasonable. But nevertheless, it is worth finding out what are the main components of this iron and intellectual “friend” of man. For example, many are interested in what a processor (CPU) is, what types it is and how to choose it correctly.

Where is the PC's brains

Obviously, in order to carry out a huge number of operations smoothly and without human intervention, a certain control center is needed, which, like a brain, will transmit commands to various components of the system and peripheral devices. In a computer, this role is assigned to the processor, which performs all logical and arithmetic operations specified by a special program. In addition, it manages all other PC devices.

How does it work

To understand what a computer processor is, you need to know how it works. Unlike their counterparts of past decades, modern devices of this type are miniature in size. At first glance, a microprocessor is a rectangular thin plate made of solid crystalline silicon. Its relatively small area contains circuits that provide the functionality of the "brain" of a personal computer. The plate is enclosed in a ceramic or plastic flat case, to which it is connected by means of very thin gold wires equipped with metal tips. This design makes the processor easy and secure to connect to the PC motherboard.

Components

Those who have already learned what a processor is, want to understand what constituents it consists of. Despite its small size, this device includes many components. Among them:

• bus addresses;
• registers;
• data buses;
• arithmetic logic unit;
• cache, or fast memory with a small size of 8-512 kbytes;
• mathematical coprocessor;
• command counters.

Processor core

There are many concepts hidden under this term. If we are talking about what a processor is and what parts it consists of, then the kernel is its component, designed to execute one stream of instructions. In addition, there are multi-core options, capable of executing multiple threads of commands.

The "nuclear" characteristics include:

• command system;
• microarchitecture;
• the number of functional blocks;
• supply voltage;
• the amount of built-in cache memory;
• crystal area;
• logical and physical interface;
• maximum and typical heat dissipation;
• clock frequencies;
• production technology.

At the same time, in the physical sense, the words “processor core” mean its part containing the main functional blocks or usually an open chip of the microprocessor. In any case, it is a necessary part of the PC “brain”. Thus, the question "what is a nuclear processor" sounds somewhat incorrect, if, of course, only the CPU is considered, and not all those devices and programs that are also called processors.

System bus

Anyone who has already learned what a computer processor is, for sure, will also be interested in how it controls the rest of the PC components. It is obvious that such a task can only be done by a complex system. It is called the processor bus and is a collection of signal lines combined according to their purpose. Each of them has a specific communication protocol and electrical characteristics. Only the CPU is connected to the processor bus itself, or, as it is also called, the system bus, and all other devices are connected through the controllers of the motherboard. At the same time, there are options when the memory is connected directly to the processor, thereby ensuring its greater efficiency. It is pertinent to ask a question about what the processor capacity is, since, for example, the expression “processor capacity x 64” means that this device is equipped with a 64-bit data bus, and it processes this number of bits per cycle.

Cache

Fast memory, or cache, is a buffer between the processor and the system memory controller, which is quite slow. This component is intended to increase the overall performance of the entire device as a whole. To achieve this goal, data blocks that are being processed at the moment are transferred to the buffer and stored, and therefore the processor is not forced to constantly access the system memory.

The cache is divided into three levels:

• First level L1

It is divided into two caches - instructions and data, is the fastest and works directly with the processor core.

• Second level L2

The L2 cache interacts with L1. It is several times larger in volume and is holistic.

• Third level L3

Some modern microprocessors also have a third level, which is more than the previous two, but works an order of magnitude slower. The fact is that the bus between the 2nd and 3rd level is narrower than between the 1st and 2nd. However, the speed of the 3rd level is still significantly higher than the speed of system memory. Depending on whether the information entering the cache is repeated at different levels or not, two types of this component of the processor are distinguished: exclusive and non-exclusive. Each of them has its own advantages and disadvantages, so it is difficult to say which one is better. It can only be noted that the first type is used in AMD microprocessors, and the second in Intel.

Connector

Talking about what a computer processor is, you should pay attention to all the components, since this complex device functions only thanks to the well-coordinated work of each of them. For example, it would seem, what importance can such a primitive device as a connector have? However, its use makes it easier to replace the processor when upgrading a PC or removing it for repair.

What is cpu frequency

To compare any technical devices and in order to give an idea of ​​their capabilities, certain characteristics are used that have a numerical expression. For processors, the main one is the clock frequency. Moreover, this concept has cardinal differences when it comes to single-core and multi-core options. So what is the clock speed of a processor if it is only capable of executing one stream of instructions? It turns out that this parameter shows how many calculations per unit of time a particular single-core device can perform. Accordingly, the higher the clock speed, the more the processor can perform operations per unit of time. Most often it is 1.0-4 GHz and is determined by multiplying the external frequency by a certain constant. It is quite another matter if you need to find out what the clock speed of the processor is. In this case, some would-be experts recommend calculating this parameter for the entire device, multiplying the data for the single-core version by the number of components. However, this is fundamentally wrong, since the clock speed of the entire device does not change from the number of cores, and the positive effect concerns only the processor's performance. To top it off, it should be noted that when choosing a processor, the frequency should not be a decisive factor, but the values ​​of all its characteristics should be considered as a whole.

What is a GPU

As you know, modern PCs provide excellent "picture". This is achieved through the GPU, a special device that performs graphics rendering. In addition, they are intended to be used as a 3D graphics accelerator. Due to the pipelined architecture, such devices are much more efficient at processing images and other graphics than the CPU, which was described above.

What is a word processor?

There is a certain amount of confusion at the moment in matters related to PC architecture, since the same terms are often used to refer to completely different things. In particular, the term processor also refers to programs for formatting text, changing fonts, paragraphs, spell checking, and more. The most famous examples are OpenOffice.org, Writer and the super popular Microsoft Word. Therefore, from the name you can safely cite when you need to answer the question of what a word processor is.

A few words about the most common personal computer processors

The first place in popularity is occupied by the Intel Core i5 processor. It is considered a great option when you need a powerful gaming machine. It is followed by a model from Intel - Celeron E3200, which is not cheap, but is the best choice for serious office work. Another processor from Intel, the four-core Core 2 Quad, also has a lot of fans among specialists. If you are not striving to become the owner of a super-powerful machine and want to save money, then pay attention to the AMD Athlon II X2215 or AMD Phenom II X4945.

Now you know what a processor is, what types it can be and what characteristics it has.

Perhaps the key advantage of the personal computer as a platform is its impressive flexibility and customization capabilities, which today, thanks to the emergence of new standards and types of components, seem almost limitless. If ten years ago, pronouncing the abbreviation "PC", one could confidently imagine a white iron box, entangled with wires and buzzing somewhere under the table, today there are no such unambiguous associations and cannot be.

Today's PC can be a powerful workstation focused on computing performance, or a designer's workstation geared toward 2D quality and fast data handling. It could be a top-end gaming machine or a humble multimedia system that lives under the TV ...

In other words, each PC today has its own tasks, which correspond to a particular set of hardware. But how do you choose the right one?

You should start with the central processor. The graphics card will determine the performance of the system in games (and a number of work applications that use GPU computing). Motherboard - the format of the system, its functionality "out of the box" and the ability to connect components and peripherals. However, it is the processor that will determine the capabilities of the system in everyday household tasks and work.

Let's take a look at what is important when choosing a processor and what is not.

What you should NEVER pay attention to

Processor manufacturer

As is the case with video cards (and, by the way, with many other devices), our compatriots are always happy to turn an ordinary consumer product into something that can be raised to the standard and go to war with the supporters of the opposite camp. Can you imagine a situation in which lovers of pickled cucumbers and canned tomatoes have divided the store with a barricade, cover each other with the last words and often resort to assault? Agree, it sounds like complete nonsense ... but in the field of computer components this happens all the time!

Moreover, like any sectarian, brand fans see the world as exclusively divided into black and white. Everything, absolutely all products with their favorite logo are the absolute ideal and perfection itself, and the solutions opposing them are the very embodiment of evil, the container of all possible shortcomings.

The fact that each of the two manufacturers of central processing units - respectively, Intel and AMD, - there are fully formed product lines, consisting of devices of completely different characteristics with completely different prices, the sectarians prefer to remain silent. As, in fact, and that in different price segments the real leader can change.

Recommendation # 1: When planning to build a new PC or upgrade an old one, decide on the budget first. Calculate the amount that you have on hand, add to it a certain reserve, which you, if necessary, are ready to add, and then see which models of central processing units fit into this budget.

Clearly realize that you are choosing these particular models, and it is their characteristics that are important to you. What is happening and who is leading in segments above or below your budget is none of your business. All that matters to you is how much productivity you get now for the money you have.

"Game" or "non-game" processor

The processor does not have a feature or feature that would or would not allow it to run games (although the parents of some buyers would gladly pay for it). It has a performance that may or may not be sufficient for a comfortable game. The division into game and non-game models is nothing more than artificial marketing. Moreover, the division is very strange and often does not correspond to the real capabilities of the CPU.

Recommendation # 2: Whatever goals you set for your future PC - whether it will be a gaming system, a workstation or the main element of a home multimedia system - the simplest parameter is guided by the simplest parameter: how much processor performance is sufficient for these tasks.

Openers

The crisis year 2016, in which incomes of the population fell, and, consequently, sales of everything and everyone, including central processors, gave us another myth, which will now sit on the Internet for a long time. And even longer in the minds of ordinary buyers.

The essence of the phenomenon is simple: "old processors cannot work with new video cards, keep running to buy new ones!" Recommendations are especially delivered here to replace the quite suitable and current Core i5 processors of the old generations with the Core i3 processors of the new generations, which are worse in all respects. Well, and, of course, advice to spend 40 thousand on upgrading the platform for the sake of games with a video card for 20 thousand.

Recommendation # 3: Actually, and. The task of any opener is not to help you choose a suitable processor, but to "insert" a newer and more expensive device, preferably complete with a motherboard and memory. If you see the opening, step aside and do not listen. Otherwise it will be more expensive for itself.

What can SOMETIMES be important

OEM and BOX-complete set, it is also "cooling system included"

Central processing units can be supplied in two versions: "boxed" and OEM configuration... The difference is very simple: a "box" is, in fact, a box in which, in addition to the processor itself, there are a warranty card and a standard cooling system (although in rare cases, like the FX 9000 series processors, it may be absent). OEM is just a processor, absolutely nothing. No box, no cooler, no warranty card.

This is due to the fact that the OEM-complete set, as intended by the processor manufacturer, is intended for companies that assemble and sell ready-made PCs. Processors in this case are purchased in large quantities and are supplied in pallets containing more than 20 pieces. Again, according to the manufacturer's logic, from these pallets they should go straight to computers.

But in our country, an OEM processor can be freely purchased at retail (see angry reviews on the topic "They took out the processor in a bag"). Such a configuration is cheaper than a boxed one, and sometimes it is very significant.

Recommendation # 4: The box configuration is always a compromise. The stock cooler is not the most efficient, not the quietest, and certainly not the most profitable in terms of price. Someone may be bribed by a longer warranty period from the "boxing" against the OEM, but the processor is an extremely tenacious device, and it is not easy to break it (except purposefully and mechanically). If he lived with you for the first day, with a 95% probability he will live for the next 10 years. Alternative coolers, again, may turn out to be both cheaper and more efficient than the standard one.

On the other hand, everything depends on the price. If the cost of "boxing" is only slightly higher than the OEM - take the box, it won't get any worse.

Free multiplier and processor frequency

Not every user of even the most ordinary gaming PC is interested in overclocking, not to mention platforms on which such overclocking is generally not needed or contraindicated. However, in some cases, this parameter can be useful.

The frequency of modern processors consists of two parameters: the base frequency, set by the system bus, and the multiplier, which varies from model to model. Accordingly, by changing one of the two parameters or both at once, we can change the final clock frequency of the processor and its performance. Nevertheless, not all modern platforms allow overclocking the processor via the bus (and even fewer platforms allow you to do it officially). So, if you plan to overclock in advance - choose models CPU Unlocked Multiplier, this will greatly facilitate your task.

As for the processor clock speed (how basic and in turbo mode) is a very specific parameter. All other things being equal, yes, processor performance is determined by frequency. For example, if we compare two processors from the line Core i5 belonging to the same generation and based on the same core, the faster will be the one with the higher frequency.

But if you compare the Core i5 with the Core i3 of the same generation or with the Core i5 of the previous generation, the frequency will not be the determining factor at all! In the first case, the number of execution units will be important, in the second - architectural differences and support for individual technologies and instructions.

Recommendation # 5: The free factor is a useful parameter, but not for everyone. Whether you need it or not, depends on the situation, and you cannot give unequivocal recommendations here. As for the frequency, use this parameter with caution. It is only important if all other parameters are the same.

Integrated graphics core

Most modern processors, with rare exceptions, are equipped with integrated graphics... Some buyers are dissatisfied with this - they say, why am I overpaying for something that I will not use? However, in reality, the integrated graphics core does not take away, but SAVES your money.

How so? It's simple. You bought a computer with a powerful processor, an overclocking motherboard and a large amount of memory, and postponed the purchase of a gaming video card for later. Just 8-10 years ago, in such a situation, you would have had to look for a "plug" for a slot at flea markets - an outdated or weak video card, on which you could sit out until a more powerful modern device was purchased. Simply because otherwise the computer would not work - processors did not know how to output video then, and top-end motherboards and integrated video were incompatible things.

Today, you simply connect a monitor to the outputs on the motherboard and use a PC without wasting time and money. Moreover, the performance of modern integrated graphics is such that undemanding users and those who need a computer not for games do not need a video card at all!

Stand alone here AMD APU... Their key advantage is precisely the powerful integrated graphics, which makes these processors an excellent option for HTPC and multimedia systems, but at the same time, their use with discrete video becomes meaningless. To be fair - the top models of modern Intel processors are equipped with a video core no worse, but they cost much more than APUs, and the performance of their processor part for an HTPC is extremely redundant.

Who lives today without integrated graphics? These are the top Intel processors for the platform LGA 2011-3- according to their status, they are supposed to work either with the most powerful gaming video cards, or with professional computing accelerators. Also, AMD processors are deprived of graphics for the outgoing already AM3 + platform... And processor family Athlon II- the same APUs, only with the graphics part disabled: extremely cheap and just as productive for their price tag.

In addition, some (but not all) processors do without integrated graphics. Intel Xeon designed for the mainstream LGA 115x platforms. These processors are worth a special mention. Despite the "server" name, they are actually analogous to the desktop Core i5 / i7. Significant differences are the ability to install in motherboards that support multiprocessor configurations and support for Error Correcting RAM (ECC).

Recommendation # 6: You shouldn't be afraid of integrated graphics - this is a great bonus, which will also soon become the standard for all platforms with the exception of LGA 2011-3 and possibly its descendants. The built-in kernel can be very useful in some cases, or even save you the trouble of buying a discrete graphics card. But it's not worth chasing after it either: processors without integrated graphics can also have many advantages.

What is REALLY important for you to know

Socket

A socket is a connector where a processor is installed on the motherboard. Like any other connector, it has certain physical dimensions, design, number of contacts, and so on. Accordingly, with rare exceptions, only one processor family can be installed in one socket. For example, a processor for socket AM4 in a motherboard with an FM2 + or LGA 1151 socket cannot be physically installed (or rather, it is possible once, but after that you will need a new processor and a new motherboard).

Accordingly, the choice of socket determines which processors will be available to you at the time of purchase, and which you will be able to install in the future (if at all). It affects the performance of the system, the capabilities and price of the future upgrade, and often the number of peripheral devices that can be installed in a PC.

Recommendation No. 7: Decide what you want to get from the PC. Yes, some modern platforms are absolutely universal (and some future platforms promise to be so) and can be flexibly configured for any task if you have the right amount of money, but this does not mean that they have no analogues. Some of your tasks can be solved with much less expenses, and some - much more efficiently for the same expenses.

If you choose a processor for an existing motherboard, do not be lazy to spend a few minutes to go to the manufacturer's official website and see a list of CPU models compatible with it. It is free, not difficult at all, and does not require any special knowledge, but in some cases it will help you save time and money.

It so happens that the processor matches the socket, but is not supported by the motherboard at all, or requires a BIOS update to start. The second can be done in advance before buying a new CPU, and the first is better to find out right away than to return a serviceable product to the store later, for the incompatibility of which with your hardware is not your fault, nor the store employees are to blame.

There are also cases when the processor is nominally supported, but in fact cannot work in a specific motherboard - for example, when the power subsystem of the motherboard is too weak, and the processor, on the contrary, is too power-hungry and demanding on power. It is also better to learn about this in advance than to deal with the consequences later.

If you choose a processor for a completely new system, you should pay attention to the current sockets:

AM1 is an AMD platform designed for nettops, embedded systems and entry-level multimedia PCs. Like all APUs, it has relatively powerful integrated graphics, which is its main advantage.

AM4 is AMD's universal platform for the mainstream segment. Combines desktop APUs with powerful Ryzen CPUs to build PCs for virtually any budget and user need.

TR4 is AMD's flagship platform for Threadripper processors. This is a product for professionals and enthusiasts: 16 physical cores, 32 threads of computing, a four-channel memory controller and other impressive numbers that give a serious increase in productivity in work tasks, but are practically not in demand in the home segment.

LGA 1151_v2- a socket, which in no case should be confused with a regular LGA 1151 (!!!). It is the current generation of Intel's mainstream platform, and is finally bringing processors with six physical cores to the consumer segment - and this is what is valuable. However, be sure to keep in mind that Coffee Lake processors cannot be installed in motherboards with 200 and 100 series chipsets, and older Skylake and Kaby Lake processors cannot be installed in motherboards with 300 series chipsets.

LGA 2066- the latest generation of the Intel platform, designed for professionals. It can also be interesting as a platform for a gradual upgrade. Younger processors Core i3 and Core i5 practically do not differ from their counterparts under the LGA 1151 of the first version and are relatively affordable, but later they can be replaced with Core i7 and Core i9.

Number of Cores

This parameter requires a lot of caveats, and it should be used with caution, however, it is this parameter that allows you to more or less logically build and differentiate central processors.

Models with two computing cores and also with two physical cores and four virtual threads Regardless of the clock frequency, the degree of dynamic overclocking, architectural advantages and fan mantras, today they are firmly established in the segment of office PCs, and even there - not in the most critical places. We don't have to seriously talk about using such CPUs in gaming machines, and even more so - in workstations.

Processors with four computing cores look a little more relevant, and can satisfy the needs of both office workers and not the most demanding home users. It is quite possible to assemble a budget gaming PC on them, although in modern titles the performance will be limited, and the simultaneous execution of several operations - for example, recording a game video - will be impossible or will lead to a noticeable drop in FPS.

The best option for home - six-core processors... They are able to provide high performance in games, do not faint when performing several resource-intensive tasks at the same time, allow you to use a PC as a home workstation, and with all this, they keep quite an affordable cost.

Eight-core processors- the choice of those who are busy with more serious tasks than games. While they can handle entertainment without any problems, their benefits are most noticeable in work applications. If you are engaged in video processing and editing, draw complex layouts for printing, design houses or other complex structures, then these CPUs should be chosen. You will not notice the excess performance, but the fast processing and the absence of freezes at the most crucial moment will definitely delight you.

Processors with 10 and 16 cores- this is already a server segment and very specific workstations, which differ from the previous version approximately like the work of a special effects designer for a big movie from the work of a video editor on youtube (in fact, they are used there approximately). It is difficult to recommend unequivocally, or vice versa, to discourage them from buying. If you really need such performance, you already know how and where you will use it.

Recommendation # 8: The number of cores is not the clearest parameter, and it does not always allow us to classify processors with similar characteristics to the same group. Nevertheless, when choosing a processor, you should be guided by this parameter.

Performance

The final and most important parameter, which, alas, cannot be found in any store catalog. Nevertheless, in the end, it is he who determines whether this or that processor is right for you, and how much the operation of a PC based on it will meet your initial expectations.

Before heading to the store for a processor that seems to suit you, do not be too lazy to study its detailed tests. Moreover, "detailed" are not vidos on YouTube, showing you what you should see as intended by their author. Detailed tests are a large-scale comparison of a processor in synthetic benchmarks, professional software and games, carried out according to a clear methodology with the participation of all or most of the competing solutions.

As is the case with video cards, reading and analyzing such materials will help you determine whether a particular processor is worth its money, and what, if possible, it can be replaced.

Recommendation # 9: Spending a couple of evenings reading and comparing information from different sources (it is important that they are authoritative, and highly desirable - foreign), you will make a reasoned choice and save yourself a lot of problems in the future. Trust me, it's more than worth it.

Criteria and options for selection:

According to the above criteria, the CPUs from the DNS directory can be allocated as follows:

Processors AMD Sempron and Athlon under socket AM1 suitable for assembling budget multimedia PCs, embedded systems and similar tasks. For example, if you want to install a full-fledged PC with a desktop operating system in your car or assemble a small nettop that will secretly live in the bowels of a country house or garage, you should pay attention to this platform.

For office PCs dual-core processors will do Intel Celeron, Pentium and Core i3... Their advantage in this case will be the presence of a built-in graphics core. The performance of the latter is sufficient to display the necessary information and speed up browsers, but completely insufficient for games, which should not be in the workplace anyway.

For home multimedia pc the best choice would be AMD's APUs for the current AM4 socket. Representatives of the A8, A10 and A12 lines combine a quad-core processor and quite good graphics under one cover, which can confidently compete with budget video cards. A PC based on this platform can be made very compact, but its performance is sufficient to play any content, as well as a number of work tasks and a considerable list of games.

For budget gaming PC quad-core processors will do AMD Ryzen 3 and quad-core Core i3 for socket LGA 1151_v2 ( do not confuse with dual-core Core i3 for the LGA 1151 socket !!!). The performance of these processors is sufficient for any home tasks and most games, but it is still not worth loading them with serious work or trying to perform several resource-intensive tasks at the same time.

For budget workstation compromise option could be AMD Ryzen 5 Quad-Core Processors... In addition to physical cores, they also offer virtual compute threads, which ultimately allow operations to be performed in eight threads. Of course, this is not as efficient as physical cores, but the probability of seeing 100% CPU load and a drop in FPS below playable when recording or streaming gameplay is much lower here than in the previous two options. And the subsequent editing of this video will be faster.

The best choice for home gaming pc- six-core processors AMD Ryzen 5 and Intel Core i5 for the LGA 1151_v2 socket (not to be confused with their four-core predecessors !!!). The cost of these CPUs is quite humane, they can even be called relatively affordable, unlike the top lines of Ryzen 7 and Core i7. But the performance is quite enough to play any games interesting to the user and work from home. And even at the same time, if there is such a desire.

For top-end gaming PCs or workstations processors will do without pretensions to being chosen and elitist AMD Ryzen 7 and Intel Core i7 having, respectively, 8 cores / 16 threads and 6 cores / 12 threads. As a mainstream platform, these processors are still relatively affordable and do not require expensive motherboards, power supplies or coolers. However, their performance is sufficient for almost all tasks that an ordinary user can put before a PC.

If it is still not enough - for high-performance workstations processors intended AMD Ryzen Threadripper designed for installation in the TR4 socket, and top models of Intel processors for the LGA 2066 socket - Core i7 and Core i9 with 8, 10, 12 or more physical nuclei. In addition, the processors offer a four-channel memory controller, which is important for a number of professional tasks, and up to 44 PCI-express lines, allowing you to connect a lot of peripherals without losing data exchange speed. It is impossible to recommend these CPUs for home use both due to their price and due to their "sharpening" for multithreading and professional tasks. But in operation, processors for top platforms can literally outrun their desktop counterparts by several times.

The central processing unit is a key component of any personal computer. In this article, we will talk about the main characteristics of modern processors, their technological features and basic functionality.

Introduction

Any computer device, be it a laptop, desktop PC or tablet, consists of several important components that are responsible for its functionality and overall performance. But perhaps the most important of these is the central processing unit (CPU, CPU, or CPU) - the device responsible for all basic calculations and executing machine instructions (program code). No wonder, it is the processor that is considered the brain of the computer and the main part of its hardware.

As a rule, when choosing a computer for ourselves, we first of all pay attention to what kind of processor it is based on, since the capabilities and functionality of your future PC will directly depend on its performance. That is why a person who has information about modern processor manufacturers and development trends of this market will be able to correctly determine not only the capabilities of a particular computer device, but also assess the prospects of a future purchase of a new PC or updating an old one.

It is quite obvious that processors installed in all kinds of computer and electronic devices differ among themselves not only in their performance, but also in design features, as well as in the principles of operation. As part of this cycle, we will get acquainted with processors based on architecturex86 that underlie most modern desktops, laptops and netbooks, as well as some tablets.

Surely, many readers, especially those who are just starting to get acquainted with a computer, have a certain prejudice that it is the lot of experienced users to understand all these "processor wisdom", because it is very difficult. But is everything really so problematic?

On the one hand, of course, a processor is a very complex device and it is really not easy to thoroughly study all its technical characteristics. The situation is further aggravated by the fact that the number of CPU models that you can now find on the modern market is very large, since several generations of chips are on sale at the same time. But on the other hand, processors have only a few key characteristics, having understood which, an ordinary user will be able to independently assess the capabilities of a particular processor model and make the right choice without getting confused in all the model variety.

Main characteristics of processors

The x86 architecture was first implemented in Intel's own processors in the late 1970s and was based on complex instruction set computing (CISC). This architecture got its name from the last two digits that ended the code names of the models of early Intel products - experienced users probably still remember the 286th (80286), 386th (80386) and 486th (80486) personal computers, which were the dream of any computer scientist in the late 80s, early 90s.

To date, the x86 architecture has also been implemented in processors from AMD, VIA, SiS, Cyrix and many others.

The main characteristics of processors, according to which they are usually divided in the modern market, are:

• company manufacturer
• series
• number of computing cores
• type of installation connector (socket)
• clock frequency.

Manufacturer (brand) ... Today, all desktop and laptop CPUs are divided into two big camps under the Intel and AMD brands, which together account for about 92% of the total global microprocessor market. Despite the fact that Intel's share of them is about 80%, these two companies have been competing with each other with varying success for many years, trying to lure buyers under their banners.

Series - is one of the key characteristics of the central processor. As a rule, both manufacturers divide their products into several groups according to their performance, targeting different categories of users and different market segments. Each of these groups constitutes a family or series with its own distinctive name, by which one can understand not only the price niche of the product, but also, in general, its functionality.

Today, Intel products are based on five main families - Pentium (Dual-Core), Celeron (Dual-Core), Core i3, Core i5 and Core i7... The first three are aimed at budget home and office solutions, the last two are at the heart of productivity systems.

CPUIntel Core i7

The line of chips keeps a little apart from the main families Atom, which differs from the rest in low power consumption and low cost. These processors are designed for installation in budget systems where high performance is not required but low power consumption is required. These include netbooks, nettops, tablet PCs and communicators.

It is impossible not to mention another family of processors from the Santa Clara company - Core 2... Despite the fact that it is no longer produced, and you can find it on sale only at various flea markets, until now, this family enjoys well-deserved popularity among users, and many current home computers are equipped with processors of this particular series.

AMD, to fans of its products, offers processors of the series Athlon II, Phenom II, A-Series and FX-Series... The path of the first two families is coming to a logical conclusion, while the last two are just gaining momentum. Somewhere else you can find the most budgetary processors on sale Sempron though their days are practically numbered.

CPUAMD FX-Series

Like Intel, AMD also has its own "mobile" series called E-series is a low-power microprocessor designed for low-cost desktop and laptop PCs.

Number of computing cores ... Even in the last decade, there was no separation of processors by the number of cores at all, since they were all single-core. But times are changing, and today, single-core CPUs can be called an anachronism, and they have been replaced by multi-core cousins. The most common of these are dual and quad core chips. Processors with three, six and eight processing cores are somewhat less common.

The presence of several cores in a processor at once is intended to increase its performance, and as you understand, the more there are, the higher it is. However, when working with old software that is not optimized for multicore computing, this rule may not work.

Connector type ... Any processor is installed in a motherboard, on which there is a special connector (socket) for this or, in other words, a socket (Socket). Processors from different manufacturers, series and generations are installed in different types of sockets. Now, for desktop PCs, there are seven of them - four for Intel chips and three for AMD.

The main and most widespread socket for Intel central processors is considered to be LGA 1155. The most productive and advanced solutions of this company are installed in the LGA 2011 socket. The remaining two types of sockets - LGA 775 and LGA 1156 are in their last days, since the release of processors for these types of sockets is practically terminated.

Among AMD products, Socket AM3 is the most widely used type of connector today. As a rule, most of the budget and most popular products of the company are installed in it. True, this situation is likely to change in the near future, since all the latest processors and high-performance solutions have Socket AM3 + and Socket FM1 sockets.

By the way, Intel and AMD processors can be very easily distinguished by one characteristic feature, which you may have already noticed when looking at the photos. AMD products have a plurality of pins on the back, with which they are connected to the motherboard (inserted into the connector). Intel uses a fundamentally different solution, since the contact pins are not located on the processor itself, in the inside of the motherboard connector.

We will not consider connectors here for mobile solutions, since this has no practical sense. After all, the type of socket is important for the user only if you plan to independently replace (upgrade) the processor in your computer. It is rather difficult to do this in portable devices, and it is almost impossible to buy mobile versions of processors themselves in retail.

Clock frequency - a characteristic that determines the performance of a processor, measured in megahertz (MHz) or gigahertz (GHz) and showing the number of operations that it can perform per second. True, it is fundamentally wrong to compare the performance of different processor models only in terms of their clock frequency.

The fact is that to perform one operation, different chips may require a different number of clock cycles. In addition, modern systems use pipelined and parallel processing for calculations, and can perform several operations at once in one cycle. All this leads to the fact that different processor models with the same clock frequency can show completely different performance.

Summary Table of Desktop Processor Families

Technological process(production technology)

In the production of microcircuits and, in particular, microprocessor crystals in an industrial environment, photolithography is used - a method by which, using lithographic equipment, conductors, insulators and semiconductors are applied to a thin silicon substrate, which form the processor core. In turn, the used lithographic equipment has a certain resolution, which determines the name of the applied technological process.

Intel

Why is the technological process by which processors are made so important? Continuous improvement of technologies allows to proportionally reduce the size of semiconductor structures, which helps to reduce the size of processor cores and their power consumption, as well as reduce their cost. Reducing power consumption, in turn, reduces the heat dissipation of the processor, which allows increasing their clock frequency, and hence the processing power. Also, a small heat dissipation allows you to use more productive solutions in mobile computers (laptops, netbooks, tablets).

Silicon wafer with processor chipsAMD

Intel's first x86 processor, still the mainstay of all modern CPUs, was manufactured in the late 1970s using a 3 micron (micrometer) process technology. By the early 2000s, almost all the leading chip manufacturers, including AMD and Intel, had mastered the 0.13 micron or 130 nm technological process. Most modern processors are manufactured using 32 nm - a technical process, and since mid-2012 and 22 nm technology.

The transition to a finer process technology is always a significant event for microprocessor manufacturers. After all, this, as noted earlier, leads to a decrease in the cost of manufacturing chips and an improvement in their key characteristics, which means that it makes the developer's products more competitive in the market.

Power consumption and heat dissipation

In the early stages of their development, microprocessors consumed very little power. But with the increase in clock frequencies and the number of transistors in the core of chips, this figure began to grow rapidly. The power consumption factor, practically not taken into account at first, has a colossal impact on the evolution of processors today.

The higher the power consumption of the processor, the more it generates heat, which can lead to overheating and failure of both the processor itself and the surrounding microcircuits. To remove heat, special cooling systems are used, the size of which directly depends on the amount of heat generated by the processor.

In the early 2000s, the heat dissipation of some processors rose above 150 W, and they had to use massive and noisy fans to cool them. Moreover, the average power of power supplies at that time was 300 W, which means that more than half of it should have been spent on servicing the "gluttonous" processor.

It was then that it became clear that further increase in the processing power of processors is impossible without reducing their power consumption. The developers were forced to radically revise processor architectures and begin to actively implement technologies that help reduce heat generation.

Processors operating at ultra-high clock speeds have to be cooled with such gigantic cooling systems.

To assess the heat dissipation of processors, a value was introduced that characterizes the requirements for the performance of cooling systems and is called TDP... TDP shows how much heat a particular cooling system should be designed to remove when used with a particular processor model. For example, the TDP of processors for mobile PCs should be less than 45W, since large and heavy cooling systems cannot be used in laptops or netbooks.

Today, in the era of the heyday of portable devices (laptops, nettops, tablets), developers have managed to achieve tremendous results in the field of reducing power consumption. This was facilitated by: the transition to a thinner technological process in the production of crystals, the introduction of new materials to reduce leakage currents, a change in the layout of processors, the use of all kinds of sensors and intelligent systems that monitor temperature and voltage, as well as the introduction of other energy saving technologies. All these measures allow developers to continue to increase the processing power of processors and use more powerful solutions in compact devices.

In practice, it is worth taking into account the thermal characteristics of the processor when buying if you want to assemble a silent compact system, or, for example, if you want your future laptop to run on battery power as long as possible.

Processor architecture and codenames

At the heart of each processor is the so-called processor architecture - a set of qualities and properties inherent in a whole family of microchips. The architecture directly determines the internal design and organization of processors.

Traditionally, Intel and AMD have codenamed their various processor architectures. This makes it possible to systematize modern processor solutions more accurately. For example, processors of the same family with the same clock speed and number of cores can be manufactured using a different technological process, which means they have different architecture and performance. Also, the use of sonorous names in the names of architectures allows manufacturers to more effectively present their new developments to us users.

Intel's developments are geographic names of places (mountains, rivers, cities, etc.) that are close to the locations of its industrial structures responsible for the development of the corresponding architecture. For example, the first Core 2 Duo processors were built on the Conroe architecture, which was named after a city located in the US state of Texas.

AMD has no clear tendency in the formation of names for its developments. The thematic focus may change from generation to generation. For example, the company's new processors are codenamed Liano and Trinity.

Tiered cache

In the process of performing calculations, the microprocessor must constantly access memory to read or write data. In modern computers, the function of the main data storage and interaction with the processor is performed by the RAM.

Despite the high speed of data exchange between these two components, the processor often has to stand idle, waiting for the information requested from the memory. In turn, this leads to a decrease in the speed of calculations and the overall performance of the system.

To improve this situation, all modern processors have a cache - a small intermediate memory buffer with very fast access, used to store the most frequently requested data. When the processor needs some data, it first looks for copies of them in the cache, since the necessary information will be retrieved from there much faster than from the main memory.

Most microprocessors for modern computers have a tiered cache, consisting of two or three independent memory buffers, each of which is responsible for speeding up certain processes. For example, the cache of the first level (L1) can be responsible for accelerating the loading of machine instructions, the second (L2) - accelerating data writing and reading, and the third (L3) - accelerating the translation of virtual addresses into physical ones.

One of the biggest challenges facing developers is finding the optimal cache sizes. On the one hand, a large cache can contain more data, which means that the percentage that the processor finds the right ones among them is higher. On the other hand, the larger the cache size, the greater the latency when fetching data from it.

Therefore, caches of different levels have different sizes, while the first level cache is the smallest, but also the fastest, and the third is the largest, but also the slowest. The search for data in them occurs according to the principle from the smallest to the largest. That is, the processor first tries to find the information it needs in the L1 cache, then in L2 and then in L3 (if available). If the required data is absent in all buffers, the RAM is accessed.

In general, the efficiency of the cache, especially of the 3rd level, depends on the nature of the program's access to memory and the processor architecture. For example, in some applications, the presence of an L3 cache can bring a 20% increase in performance, and in some it will not affect it at all. Therefore, in practice, it is hardly worth being guided by the characteristics of a tiered cache when choosing a processor for your computer.

Integrated graphics

With the development of production technologies and, as a consequence, a decrease in the size of chips, manufacturers have the opportunity to place additional microcircuits inside the processor. The first of these was the graphics core, which is responsible for displaying the image on the monitor.

This solution allows you to reduce the total cost of the computer, since in this case there is no need to use a hotel video card. It is obvious that hybrid processors are focused on use in budget systems and the corporate sector, where the performance of the graphics component is secondary.

The first example of integrating a video processor into a "normal" CPU was demonstrated by Intel in early 2010. Of course, this did not bring any revolution, since up to this point the graphics had been successfully integrated into the chipsets of motherboards for a long time.

Once upon a time, the difference in functionality between integrated and discrete graphics was fundamental. Today, we can only talk about the different performance of these solutions, since the built-in video chips are capable of displaying images on multiple monitors in any available resolutions, performing 3D acceleration and hardware video encoding. In fact, integrated solutions in terms of their performance and capabilities can be compared with lower-end graphics cards.

Intel is integrating a graphics core into its processors under the unpretentious name IntelHDGraphics of its own design. At the same time, Core 2, Celeron processors and older Core i7 models do not have integrated graphics cores.

AMD, having carried out a merger in 2006 with the giant for the production of video cards, the Canadian company ATI, is embedding video chips of the Radeon HD family in its solutions. Moreover, some of the company's new processors are combining x86 processor cores and Radeon graphics on a single die. A single unit created by merging a central (CPU) and a graphics (GPU) processor is called an APU, Accelerated Processor Unit. This is exactly what (APU) is now called the A and E-series processors.

In general, integrated graphics from AMD are more powerful than Intel HD and appear to be preferred for gaming applications.

ModeTurbo

Many modern processors are equipped with technology that allows them in some cases to automatically increase the clock speed above the nominal, which leads to increased application performance. In fact, this technology is a "self-overclocking" of the processor. Turbo run time will vary based on operating conditions, workload, and platform design.

Intel uses its own intelligent overclocking technology called Turbo Boost in its processors. It is used in the productive Core i5 and Core i7 families.

By monitoring the parameters related to the load on the CPU (voltage and amperage, temperature, power), the built-in control system increases the clock frequency of the cores in the case when the maximum thermal package (TDP) of the processor has not yet been reached. In the presence of unloaded kernels, they are disabled and free up their potential for those used by applications. The fewer cores are involved in calculations, the higher the clock speed of the chips involved in the calculations rises. For single-threaded applications, the acceleration can be up to 667 MHz.

AMD also has its own technology for dynamic overclocking of the most loaded cores and uses it only in its 6 and 8-core chips, which include the Phenom II X6 and FX series. It is called Turbo Core and can only work if the number of loaded cores in the process of computing is less than half of their total number. That is, in the case of 6-core processors, the number of inactive cores should be at least three, and 8-core processors - four. Unlike Intel Turbo Boost, this technology does not affect the frequency increase by the number of free cores and it is always the same. Its value depends on the processor model and ranges from 300 to 600 MHz.

Conclusion

In conclusion, let's try to apply the practically acquired knowledge with benefit. For example, a popular computer electronics store sells two Intel Core i5 processors at the same 2.8 GHz clock speed. Let's take a look at their descriptions taken from the store's website and try to understand their differences.

If you look closely at the screenshots, despite the fact that both processors belong to the same family, they have not so much in common: the clock frequency and the number of cores. The rest of the characteristics differ, but the first thing you should pay attention to is the types of connectors in which both processors are installed.

Intel Core i5 760 has Socket 1156, which means it belongs to the outdated generation of processors. Buying it will be justified only if you already have a motherboard with such a socket in your computer, and you do not want to change it.

The newer Core i5 2300 is already manufactured using a thinner process technology (32 nm versus 45 nm), which means it has a more perfect architecture. Despite the slightly smaller L3 cache and "self-overclocking", this processor will certainly not yield to its predecessor in performance, and the presence of integrated graphics will make it possible to do without purchasing a separate video card.

Despite the fact that both processors have the same heat dissipation (95 W), the Core i5 2300 in equal conditions will be cooler than its predecessor, since we already know that a more modern technological process provides lower power consumption. In turn, this increases its overclocking potential, which is good news for computer enthusiasts.

Now let's look at an example based on AMD processors. Here we have specially selected processors from two different families - Athlon II X4 and Phenom II X4. In theory, the Phenom line is more productive than the Athlon, but let's look at their characteristics and decide if everything is so simple.

It can be seen from the characteristics that both processors have the same clock frequency and number of processing cores, almost identical heat dissipation, and also both lack an integrated graphics core.

The first difference that immediately catches your eye is that the processors are installed in different sockets. Despite the fact that both of them (connectors) are currently actively supported by motherboard manufacturers, Socket FM1 looks somewhat preferable from this pair from the point of view of future upgrades, since new A-series processors (APUs) can be installed there.

Another advantage of Athlon II X4 651 is the thinner and more modern technological process by which it was produced. Phenom II responds with Turbo mode and L3 cache.

As a result, the situation is ambiguous, and here the key factor may be the retail price, which for a processor from the Athlon II line is 20-25% lower than that of the Phenom II. Taking into account the more promising platform (Socket FM1), the purchase of Athlon II X4 651 looks more attractive.

Of course, in order to speak more unambiguously about the advantages of certain processor models, you need to know what architecture they are based on, as well as their real performance in various applications, measured in practice. In the next article, we will take a closer look at the modern lineup of Intel and AMD microprocessors for desktop PCs, get acquainted with the characteristics of various CPU families, and also give comparative results of their performance.