The newest AMD processor. Which processor to choose for your home PC, review of the best models

Although every device installed in system unit, has its own processor; it is without a central processor that a computer cannot work. The central processing unit is the “brain” and “heart” of the computer and is responsible for:

• for controlling computer peripheral devices. This is a flash drive, built-in speakers, printer, mouse, floppy disk, scanner, webcam, microphone, speakers, monitor, keyboard, joystick, steering wheel, video card, as well as headphones, modem, hard drives and other input and output devices, storage, and also I/O;
• data processing operations;
• operation of system parts;
• execution of program object code.

Now the best processor manufacturers are Intel and ADM. For a long time Intel occupied a leading position, distinguished by high performance, efficiency and modern processor technologies. IN lately it became difficult to name the best company, buyers are often lost and don’t know which company is better to purchase a processor, since ADM has proven its devices well in all price segments of the market.

The editors of the VyborOK website have prepared for you a rating of high-quality processors from ADM, which contains only the best models selected, according to customers. Also from the article you will learn:

• about other processor names;
• what the device consists of;
• about what you should pay attention to when choosing;
• about the best models selected based on consumer reviews.

The two names mentioned above for the “brain” and “heart” of a computer are the processor and the central processing unit. In addition to these two names, the processor can also be called:

• central processing unit, the abbreviation CPU or the English version CPU is more often used;
• central processing device;
• integrated circuit;
• electronic unit;
• microprocessor;
• main part hardware computer;
• the main part of a programmable logic controller.

What does the central processing unit consist of?

The CPU is a thin square plate with an area of ​​several mm2. Inside the processor, in the silicon case (often called the rock), are logic, assembled using more than ten million transistors.

The central processing unit consists of:

• From a control device (CU), which is responsible for organizing and executing programs, and also coordinates the communication of all devices of an electronic computer during its operation;
• Arithmetic logic unit (ALU) - necessary to perform comparison and division, multiplication and subtraction, and other arithmetic and logical operations over data;
• The storage device is internal memory microprocessor, which consists of a register and cache memory. To save intermediate results and perform calculations, an intermediate fast memory in the form of registers. And to swap data and commands from RAM, as well as to speed it up, cache memory is used. The register and cache form the core of the processor;
• The clock generator produces electrical impulses that clock the operation of all nodes on the computer.

Selection criteria: how to choose the right processor and what you need to pay attention to

Below are the main characteristics of the processor, on which its performance depends (the main characteristics will also help you make the right choice):

• Bit capacity - expressed by the maximum amount of information in bits that can be simultaneously transmitted and processed. The processor capacity is characterized by the register capacity. For example, if the register width is 2 bytes, the processor width will be 16 bits (that is, 2x8). There are 16, 32 and 54-bit architectures. Accordingly, the larger it is, the higher the productivity;
• Performance is the average value of operations performed by the processor per second. Speed ​​is also called processing power;
• The clock frequency is the length of time between the current pulse of the clock generator and the start of the next pulse. Clock speed is measured in clock cycles (MHz) per second. The minimum clock frequency is 40 MHz, the maximum is 3 GHz;
• The core is the part of the processor that is responsible for executing a single stream of commands. It is necessary to select a particular number of cores based on the tasks they will perform. Dual core the processor will do for surfing the Internet, playing games with minimum requirements, as well as office work. Quad core with 4 threads – for standard tasks and games with average requirements. Quad-core with 8 threads will pull powerful games. Six and eight cores - copes with all highly demanding games and programs. A processor with more than eight cores is suitable for specialized computers;
• Simultaneous multi-threading – recently introduced by AMD SMT. SMT - Simultaneous MultiThreading, which in translation means simultaneous multithreading, necessary for simultaneous execution of different tasks. The principle of this technology is as follows: so for 1 physical core it is represented as 2 virtual or logical ones. And, accordingly, where one thread is processed, 2 are processed;
• Cache memory is a small number of memory cells that act as a buffer. The buffer is necessary to increase the speed of the computer. There are 3 memory levels - L1, L2 and L3. L1 is located on the processor die and is the smallest in volume and the fastest in speed. L2 is also included in the chip, the frequency corresponds to the frequency of the processor core. L3 – the slowest, but largest memory;
• A connector, also known as a socket, is necessary to connect the processor to the motherboard and further interact with it. Please note that the motherboard socket is suitable for a specific type of processor. Therefore, before purchasing a processor, it is imperative to find out the motherboard socket;
• Heat generation and energy consumption depend on production technology. The best energy efficiency and heat dissipation are facilitated by the increased frequency, small size of the processor, good system cooling and load control system;
• Integrated graphics processor – is a graphics core built directly into the processor. This association is used for desktop computer, as well as for laptops in the budget segment. The integrated chip allows you to reduce the cost of devices, reduce hardware power consumption and create compact hardware;
• The internal frequency multiplier factor is the result of multiplying the internal frequency factor by the reference frequency received from the motherboard;
• The technical process is the size of the transistors; the performance of the processor depends on them. The smaller size of transistors makes it possible to place a large number of them on a chip.

Review of the best ADM processors

A10 Kaveri is a budget class of processors that are ideal for office PCs. The microprocessor is represented by 4 cores with a clock frequency of 3,700 MHz and an integrated Radeon R7 Series graphics core with the highest possible frequency of 720 MHz. The processor is built on a 28 nm process technology and has 2 levels of memory: 64 KB and 4,096 KB. The manufacturer indicated the maximum possible temperature of the device - 72.4 degrees and heat dissipation - 95 W.

Advantages:

• fairly good built-in graphics core;
• low cost;
• low heating
• the presence of restrictions on the removal of thermal power;

Flaws:

• at high loads performance decreases;
• low power in single thread mode.

Ryzen 3 2300X Pinnacle Ridge

An inexpensive processor with a 12 nm process technology and 4 cores with a frequency of 3,500 MHz (maximum turbo frequency of 4,000 MHz) has quite good performance. The Ryzen 3 2300X supports DDR4-2933MHz memory with 2 channels. Cache capacities are 96, 2,048, and 8,192 KB. Typical heat dissipation is 65 W, operating temperature is 95 degrees. When running at a clock speed of 3500 MHz, the processor remains quite cool even with a regular fan. When overclocking the device to 4,000 MHz, you should be prepared for a lot of heat and, accordingly, an improved cooling system.

Ryzen 3 2300X supports Virtualization Technology, NX Bit and AMD64/EM64T.

Ryzen 3 2300X Pinnacle Ridge

Advantages:

• fairly good performance;
• 12 nanometer process technology;
• moderate cost;
• good turbo frequency.

Flaws:

• It gets very hot when accelerating. A good cooling system will help.

The Ryzen 5 2400G has four cores, clocked at 3600 and 3900 MHz, with SMT support and eight processing threads. A processor with an integrated Radeon RX Vega 11 graphics core and an unlocked multiplier is positioned as a gaming processor. Of course, the Ryzen 5 2400G is not for high-demand games, but with minimal settings the graphics chip will perform great. The maximum possible temperature is 105 degrees, and the heat dissipation is 65 W.

Ryzen 5 2400G can be supplied in BOX and OEM configurations.

Advantages:

• good overall performance, as well as the performance of the integrated graphics core;
• low power consumption;
• unlocked multiplier.

Flaws:

• High sensitivity to RAM speed.

Ryzen 5 2600 Pinnacle Ridge

The gaming processor is manufactured using a 12 nanometer technological process. The clock speed of each of Pinnacle Ridge's six cores is 3,400 MHz, and with Turbo Boost reaches 3,900 MHz. Sufficiently high clock speeds and 12 threads provide high performance and efficiency of the CPU.

If you want to get a higher clock frequency, you can purchase the Ryzen 5 2600X Pinnacle Ridge model, where the clock frequency is 3,600 MHz, when using Turbo Boost - 4,200 MHz.

The Ryzen 5 2600 has three levels of DDR4-2933 memory, which are 96 KB, 3072 KB and 16,384 KB respectively. The maximum possible number of memory channels is 2, there is a built-in memory controller. The processor supports Virtualization Technology, NX Bit and AMD64/EM64T.

The CPU supports OEM and BOX configurations. Customer reviews indicate that the cooler included in the BOX package does not do its job well. Therefore for better cooling It's better to buy a good fan yourself. So, for example, when working with Gammax 300, the maximum heating temperature is 50 degrees, when working at minimum speed - 30-40 degrees.

Ryzen 5 2600 Pinnacle Ridge

Advantages:

• reasonable price;
• good acceleration;
• high performance;
• multithreading;
• support for installing old fans;
• excellent performance;
• little heat generation.

Flaws:

• weak stock cooler.

Inexpensive microprocessor designed for motherboards with AM3+ socket. The model copes well with demanding games and programs, showing high frame rates. There are 8 Vishera cores on board, with a standard frequency of 3,500 and a maximum of 4,000 MHz. The FX-8320 Vishera has a 32nm process technology and 8 threads.

8 threads supported. DDR3-1866 memory has cache capacities of 48, 8,192 and 8,192 KB. The declared operating temperature reaches 61.1 degrees, and the heat dissipation is 125 W.

Advantages:

• fast acceleration;
• low cost with high performance;
• competent energy saving system;
• cache memory capacity;
• multitasking;

Flaws:

• A powerful cooler is needed to maintain low temperatures.

The FX-8350 Vishera is presented with 8 cores, but in fact there are 4 of them. The clock frequency is 4,000 MHz, with Turbo Boost technology – 4,200 MHz. The technological process is 32 nm, 8 threads. The model has three-level DDR3-1866 memory with capacities of 48, 8,192 and 8,192 KB. The processor copes well with multi-threaded workloads. An unlocked multiplier is supported, which will help overclock the CPU to 4,700 MHz.

Advantages:

• excellent overclocking results;
• excellent results with resource-intensive tasks.

Flaws:

• low performance in single-threaded mode;
• low energy efficiency.

The 14nm Ryzen 7 1700 with 8 Summit Ridge cores is an excellent choice for both gamers and casual users who value high performance and low power consumption. The CPU has 16 computing threads and a clock frequency of 3,000 MHz. Turbo Boost technology allows you to overclock the frequency to 3,700 MHz. The unlocked multiplier provides overclocking potential. The supported memory type is DDR4-2667 with 2 maximum possible channels. Cache capacities are 96, 4,096 and 16,384 KB.

OEM and BOX configurations with stock cooler are available.

Advantages:

• excellent multitasking;
• capacious cache;
• low heat generation;
• unlocked multiplier;
• good voltage reduction potential;

Flaws:

• high RAM requirements.

The Ryzen 7 1800X is a powerful processor that can easily handle any complex task, including all highly demanding games. The CPU runs on 8 Summit Ridge cores with 16 threads. The maximum clock frequency reaches 4,000 MHz. The gaming processor is built on a 14 nm process technology and has three-level memory with a large capacity of 96 KB, 4,096 and 16,384 KB.

The processor supports a set of commands such as AMD64/EM64T, NX Bit and Virtualization Technology.

Advantages:

• great acceleration;
• high power;
• high quality solder;
• low heat generation;
• excellent performance.

Flaws:

• The processor needs high speed RAM.

The top-end gaming processor has impressive functionality and, according to buyers, is the best in the Pinnacle Ridge series. The eight-core CPU has a 12 nm process technology, 16 threads and a 3,700 MHz operating frequency (with Turbo Boost - 4,300 MHz). SenseMI technology takes care of the highest possible performance and efficiency of the processor, based on the needs of the consumer. And Radeon FreeSync eliminates stuttering, game interruptions, and distortion. The technology will provide the user with a smooth frame change and high performance. The cache consists of three levels, where L1 is 96 KB, L2 is 4,096 KB, L3 is 16,384 KB.

The Ryzen 7 2700X performs well in multi-threaded mode. As for single-threaded mode, compared to FX this model provides much better results. The stated typical heat dissipation is 105 W, the maximum temperature is 85 degrees.

Advantages:

• high multi-threaded performance;
• good single-threaded performance;
• SenseMI and Radeon FreeSync technologies;
• automatic acceleration;
• low heating temperature;
• excellent value for money;
• high productivity and efficiency.

Flaws:

• high demands on RAM speed.

Ryzen Threadripper 1950X

Our rating ends with the most powerful and expensive processor. The Ryzen Threadripper 1950X runs on 16 cores and 32 threads. Clock frequency – from 3,400 MHz to 4,000 MHz. Quad-channel DDR4 memory has error correction. Combined cache memory capacities are 96, 8,192 and 32,768 KB. The CPU is built on a 14nm process technology and has 64 PCIe lanes.

The 16-core Ryzen Threadripper 1950X will provide:

• optimal data transfer speed and lightning-fast processing of high loads;
• quick access to large data sets and high-demand video cards;
• good optimization and load parallelization.

Ryzen Threadripper 1950X

Advantages:

• combined cache memory with error correction function;
• 64 PCIe lanes;
• good price with high power;
• high data transfer speed;
• high optimization;
• excellent results when working with parallelized loads.

The article is devoted to the best AMD processors. The “Reds” were always catching up and were lost against the background of Intel in processors and NVidia in video cards. But under the clear leadership of Lisa Su, the company achieved serious success and pretty much frayed Intel’s nerves, and the latest generation of processors and the transition to a new technical process became completely humiliating for the “blues”.

What are the characteristics of “red” processors? Each line is interesting in its own way:

• FX. The previous generation of processors made using a 32nm technical process. The market knows both the inexpensive FX4300 for simple machines and monsters like the FX9590, which can literally fry eggs.
• Athlon. Atlons have become a budget segment for office machines. It is worth noting that this series has retained its name to this day.
• Ryzen. Currently, the 3rd generation of processors has entered the market. They are not only better than Intel in price, but also more productive in multi-threaded tasks. In addition, the “Reds” prefer to “lick” drivers for their components rather than release a new generation every year. These processors are manufactured using the most advanced 7nm process technology.

We have selected for you the top best AMD processors in different price segments.

The best processors for gaming

4 AMD Ryzen 3 2200G

The most budget among gaming
Country: USA
Average price: 6219 rub.
Rating (2019): 4.5

Ryzen does a great job of showing how modern technologies exceed those that were relevant just a year or two ago. The model is only a thousand rubles more expensive than the A12, but how much higher the performance is! In synthetic tests the increase is almost one and a half times. And in real use, the model turns out to be much faster. Due to the presence of a built-in graphics core, it is perfect for photo editors and novice video editors, not to mention ordinary office work. And with a discrete graphics card of the GTX 1050ti level it will be able to provide acceptable fps at medium-high settings in modern games.

This increase is due to the use of modern Raven Ridge cores and the appearance of a 4MB L3 cache. The clock frequencies are also encouraging - 3.5 GHz in stock and up to 3.9-4.1 GHz in overclocking. Heat dissipation 65W. With a regular cooler, temperatures will remain at 40 degrees at idle, and up to 75 degrees under load. Separately, we note the use of the “old” socket – AM4. Thanks to this, when switching to a new CPU, you will not have to change the motherboard and RAM.

3 AMD Ryzen 3600

High performance
Country: USA
Average price: 14,690 rub.
Rating (2019): 4.5

Due to its newness, this processor looks very expensive. But new technologies are making themselves felt and this stone is definitely faster than its predecessor. Of course, from the start you get 3.6 GHz with the ability to overclock to 4.2 GHz. It is noteworthy that the requirements for RAM have also increased. The optimum will now be dies with a frequency of 3200 MHz. Level 3 cache has been doubled and is now 32 MB. It is in no way inferior to the Ryzen 2700X and looks much more profitable.

From the factory such a processor is overclocked to the maximum and taking the 3600X is throwing money away. Paired with the RTX 2070 it will handle any game. It can work on any motherboard on socket AM4, you just need to reflash it. However, it is not recommended to go below B450, as the processor is demanding of the motherboard's power subsystem. In many tests it is faster than 9600K and 8700K. According to reviews, even with the Gammax 300, it heats up to no more than 55 degrees.

2 AMD Ryzen 1600

Best price/performance ratio
Country: USA
Average price: 7140 rub.
Rating (2019): 5.0

You won't find anything better for this price. On AliExpress, such a stone can generally be purchased for 5,500 rubles. It is made using a 14 nm technical process, which affects the performance and speed, which are 15% lower than that of the 2600. The frequencies are also lower - the starting 3.2 GHz can be “boosted” in automatic mode to 3.6, and if you have a motherboard on X370 or The X470 chipset can be raised to 4.0 GHz, you just need to buy a good cooler. The only problem is that the latest X570 boards do not support first-generation processors. The processor supports dual-channel RAM operation with frequencies up to 2667 MHz.

Otherwise, these are the same 6 cores and 12 threads. For guys who want to assemble a full-fledged gaming computer at a reasonable price, this is just a find. It will always maintain a stable low temperature due to the solder. Old problems also remain, for example, dependence on the memory controller and its frequency.

Ryzen performance is largely tied to RAM, and the FX series generally requires complex overclocking. If you are the owner of a Ficus, then to significantly improve performance you will have to spend money on a cool motherboard, overclock the RAM, memory controller and other elements.

Ryzen performs increasingly better if higher frequency RAM is used in the system.

1 AMD Ryzen 2600

Best processor
Country: USA
Average price: 10035 rub.
Rating (2019): 5.0

The best processor today. After the release of the 3rd generation, the average price tag dropped to 10,000 rubles, but if you try, you can find a stone for less than 9,000 rubles. There is no need to look for a boxed version; just screw on the Gammax 400 for cooling and it will be enough for your eyes. The 12 nanometer process technology set the thermal package at 65 W. A starting frequency of 3.4 GHz with automatic overclocking to 3.9 is an ideal solution for gamers of any level. It will also perform well in modeling and mathematical calculations thanks to its 12 threads.

Compared to top-end FX, things with the cache are several times better. If the top 9580 had about 8 MB, then this pre-pot stone has all 16. The 2600 stably interacts with DDR4 RAM with a frequency of up to 2933 MHz on paper and up to 3200 in practice. To disperse the stone you will need motherboard on the X470 chipset, and to work in the factory range a B450 chipset is sufficient. What about the 2600X? You can bypass it, since it is essentially a 2600 overclocked to the ceiling.

The best processors for office work

5 AMD A12-9800E

Balanced office option
Country: USA
Average price: 4880 rub.
Rating (2019): 4.4

The next line is occupied by the inexpensive A12-9800E processor. Its technical characteristics make it look like a classic office model with virtually no overclocking potential and little applicability to games. The L1 and L2 cache sizes are 80 and 2048 KB, respectively, and there is no third level at all. Works with RAM up to 2400 MHz. There is also built-in graphics. Not quite fresh, the graphics core frequency is only 900 MHz.

In general, it allows you to save money and not buy a separate video card, but you can only play undemanding online projects.

4 AMD Ryzen 5 1400

Budget universal processor
Country: USA
Average price: 5920 rub.
Rating (2019): 4.5

So, the 1st generation Ryzen 1400 model has 4 cores and 8 threads, which is enough for most games, but in high-end projects there will be certain problems. The overclocking potential is extremely low - 3.2 GHz in stock and 3.4 in overclocking. Level 3 cache is also reduced to 8 MB.

According to reviews, the processor has a number of advantages and is perfectly balanced thanks to good performance per core and support for all modern protocols and interfaces. In synthetic tests it bursts and rushes to the fullest; it can easily be overclocked to 4.1 GHz. If you buy the boxed version, the branded cooler will not only delight you with a powerful air flow, but also guarantees low temperatures.

3 AMD FX-8320 Vishera

Powerful last generation processor
Country: USA
Average price: 4163 rub.
Rating (2019): 4.6

The legendary FX8320 processor has 8 cores and 8 threads, and costs only 4,100 rubles. For that kind of money, it simply has no analogues, and if you work in an office, the power of your computer is important to you and you want to play during breaks, you won’t find a better purchase. In addition, the motherboards available on the market are also relatively inexpensive, but overclocking models are almost impossible to find.

Let's be honest - it's outdated as a game. The 32nm process has a number of disadvantages, such as low memory bandwidth and the need for overclocking, which requires a board with an FX990 or AMD 970 chipset, but in the second case it will be minimal. The starting frequency of the processor is 3.5 GHz, and if you take the 8320E, you will pay 100 rubles more and get 3.3 GHz out of the box, but with a reduced thermal package. According to reviews, the processor can be overclocked to 5.0 GHz, but only with a powerful tower cooler, or even better - water cooling.

2 AMD Ryzen 3 1200

New technologies at an affordable price
Country: USA
Average price: RUB 26,080.
Rating (2019): 4.7

An unsurpassed price-performance processor in the budget segment. There is no built-in video card, but 4 full-fledged cores with 4 threads work like 8 cores of the FX 8320. Especially for office tasks, the Level 3 cache has been increased to 8 MB, which means working with a large number of files will be much easier.

As for overclocking, it is very limited. Out of the box, this inexpensive processor starts at 3.1 GHz and, with automatic boost, reaches 3.4 GHz. But there is also a surprise. Buy a good motherboard, for example, on the X370 chipset, a powerful tower cooler, and then you can reach 3.8 GHz, and with some luck, 4.2 GHz. In addition, it is faster than the i5-6400 and only slightly loses to the i5-7400. According to reviews, it heats up little under overclocking, but saving on cooling is still not recommended, even during assembly budget computer.

AMD uses solder, not thermal paste, in its processors. Simply put, a heat distribution cap is soldered to the crystal using a special compound, which eliminates the need to scalp the processor. In addition, heat transfer does not deteriorate over a long period of time.

1 AMD Athlon 200GE

Best office processor
Country: USA
Average price: 3119 rub.
Rating (2019): 5.0

A great budget processor with integrated Vega 3 graphics. In addition, this is a great temporary replacement for waiting until better times and buying a Ryzen 3600 or 2600. 2 cores and 4 threads do the job and they are cheaper than the G4620 or G5400. You can manually overclock it to 3.8 GHz and turn it into a semi-gaming one, but you shouldn’t get your hopes up; after all, the stone is not for gaming.

The processor is indirectly demanding on RAM; it is advisable to set it at 2667 MHz. Then performance and responsiveness will be maximum. The boxed cooler is plenty enough for it. Compared to full-fledged Ryzen, the level 3 cache is reduced here and is limited to 4 MB. The thermal package will also be reduced, only 35 W. There are also more powerful options on sale, for example 220GE or 240GE, but in terms of price and performance ratio this is the most optimal processor.

The best processors for simulation and complex calculations

3 AMD Ryzen 7 2700X

Best price
Country: USA
Average price: 18830 rub.
Rating (2019): 4.5

Ryzen 7, unlike the previous participant, is seen as a more mainstream product, although it can hardly be called budget. The average cost is 24,000 rubles - almost three times lower, with a difference in performance of a third. In games in conjunction with Nvidia GeForce GTX 1080 at maximum graphics settings in FullHD, all tested games (FarCry 5, Hitman, NFS Payback, etc.) produced 75-110 fps. With such performance, it’s time to move to 2K, or even 4K if you’re willing to put up with 30 frames per second. The use of 8 cores has a positive effect on multi-threaded tasks - compiling projects or rendering is noticeably faster than on similar Intel processors price category. For most users, these indicators will be enough.

The technical data numbers are encouraging. 8 cores, made using a 12 nm process technology, 16 threads. The base clock frequency under load can automatically increase from 3.7 to 4.3 GHz. The L1-L3 cache sizes are 96, 4096 and 16384 KB, respectively. DDR4-2933 RAM is supported. The heat dissipation is 105 W, which requires proper cooling. But unlike the previous participant, here you can get by with high-quality air cooling.

2 AMD Ryzen Threadripper 2990WX Colfax

Ultimate server processor
Country: USA
Average price: RUB 127,148.
Rating (2019): 4.6

This is real server power, which is not recommended for purchase by the average consumer, and for most offices it will be redundant. But for rendering, exporting, data encoding and other professional tasks, it is great. It requires a motherboard with sTR4 socket and at the moment We found only one board on the 399 chipset. It supports 16 PCI-E lanes. Enthusiasts in reviews claim that under the water block the processor can be overclocked to 4.2 GHz. At the same time, in single-threaded modes it loses to its Intel counterparts.

Standard frequencies are relatively low, 3.0 GHz will not surprise anyone. It is not recommended to install anything else besides the water block, since the thermal package of the processor is 250 W. The maximum permissible operating temperature is 68 degrees. Optimal for it will be DDR4 RAM with a frequency of 2933 MHz; in addition, 4 memory channels are supported.

1 AMD Ryzen 9 3900X

The best in terms of quality and price ratio
Country: USA
Average price: 47,770 rub.
Rating (2019): 5.0

Powerful next generation processor for simulation. The manufacturer's approach to the production process is interesting and technical specifications. The company decided not to limit the frequencies of its brainchild, which is why its starting frequency is set at 3.8 GHz with an overclock to 4.6 GHz. There are already 12 cores and 24 threads. Modern games will not be able to use all this power, and the level 3 cache has a capacity of 65 MB. With such parameters, the level of heat dissipation is also high at 105 TDP.

Out of the box, the processor is of little interest to overclockers, since there is nothing to overclock here, since the design uses the most selected crystals and components, and the stone itself is overclocked to the maximum. In multi-threaded tasks it copes with absolutely all programs. This model will require a top-end motherboard with an X570-level chipset; the minimum is an X370 with good cooling of the VRM zone.

How to choose an AMD processor?

So, you wanted to buy a processor from the “red” ones and you did the right thing. Their price-performance ratio is now ideal. Let us note those models that are most interesting to buyers:

• The Ryzen 3600 will be the ideal modern purchase, guaranteeing high performance in all games;
• Ryzen 2600 is currently the most profitable investment, which will last for several years ahead and without overpayments;
• Ryzen 1600 for those who want a modern gaming computer and want to save money;
• Athlon 200GE will be the best assistant in the office, as it has built-in graphics.

The past year could well be declared the year of central processors. Announcements of fresh products of this kind occurred more frequently in 2017 than ever before. Moreover, we are really talking about the appearance of fundamentally new chips, and not the next cosmetic changes and a slight increase in clock frequencies. As a result, while over the past few years, final articles about the processor market for personal computers had to be literally tortured, with difficulty trying to remember at least something worthy of attention, last year gave so many hot topics that it was simply not clear what to grab onto . Stories poured out like from a cornucopia.

Last year began with the announcement of the Kaby Lake family of processors, which can hardly be called any outstanding new product. But then it exploded. AMD set the tone and began churning out fundamentally new solutions one after another, built on the Zen microarchitecture. Ryzen 7, 5, 3, and then Threadripper processors became the wind of change that was able to stir up an industry in the computer market that had long been considered stable or even stagnant.

At the same time, last year was not a benefit for AMD. Intel also presented many interesting products. The microprocessor giant has completely updated its high-performance HEDT platform and released the Skylake-X processor family, which has not only standard processors for this segment with 6-10 computing cores, but also monster chips like the Core i9-7980XE with a number of cores reaching 18 pieces. In addition, this platform has become easier to “enter”, since relatively inexpensive processors with the Kaby Lake-X design have also begun to be produced for it. The current mainstream Intel platform has also undergone major changes. With an eye on it, the company now has six-core Coffee Lake processors in its assortment, which debuted along with the updated processor socket LGA1151 (second version) and 300 series chipsets.

Thus, without much difficulty, we were able to recall at least six announcements from last year, which can be called cornerstone, but the matter is not limited to them. For comparison: in 2016, only one new CPU family came to the market - Intel Broadwell-E, and in 2015, the only notable event in the processor market was the release of Skylake processors and the LGA1151 platform (first version). In the previous couple of years, AMD was only able to distinguish itself by launching the Godawari family of hybrid processors, which not only did not bring anything fundamentally new, but also could not even offer a better level of performance compared to the FX processors of the Piledriver family from 2012.

This example perfectly shows how the disposition in the processor market has changed. And if you are only interested computer games, then the changes that have occurred can still be somehow ignored, since many older processors are still capable of providing acceptable performance in tasks of this type. But when working with intensive workloads, in particular with the creation or processing of digital content, the situation is completely different. Here, in 2017, users received opportunities at a fundamentally new level: the previously conservative processor market could become a source for noticeably accelerating work processes.

It's no exaggeration to say that these are golden times for performance enthusiasts. Intel, which claims to hold the crown for making the fastest desktop processors, can now offer Core i9-7980XE and Core i9-7960X CPUs with nearly double the number of cores of any desktop chip released before 2017. AMD, on the other hand, has in its arsenal Threadripper, a processor with similar characteristics, but sold at half the price, which is impressive in itself.

Many interesting offers have also appeared in more mundane price segments. There are countless fresh options for upgrading your PC platform. A series of turbulent events that took place in the processor market led us to the fact that today, at the same costs as a year ago, you can get at your disposal a configuration that is approximately one and a half times more productive. I think this is an excellent reason to refresh your memory of how everything happened.

It is unlikely that anyone will argue with the fact that the main processor event in the past year was the appearance of the AMD Ryzen family built on the Zen microarchitecture. As we can now say with complete confidence, these processors were able to return AMD to the list of suppliers of chips for productive computers, which revived competition and pushed progress in this area. It is generally accepted that it was thanks to AMD that we came to the point where mass processors in the upper price segment began to have more than four processor cores at their disposal. And although this is not entirely true, such a rapid development of events certainly would not have happened without AMD.

In other words, the “Ryzen effect” is a completely objective fact and great success AMD, which has not pleased us with any noteworthy performance innovations for a very long time. In the process of developing the Zen microarchitecture that underlies Ryzen processors, AMD engineers abandoned all of their past heritage. The new processors received a fundamentally new design that has nothing in common with Bulldozer - the original microarchitecture, which never managed to live up to the hopes placed on it. At Zen, engineers decided to return to classic “wide” cores, and this reflects a fundamental change in AMD’s ideas about what basis modern x86 processors should be based on.

Chips built on the Zen microarchitecture switched to the use of a modern 14-nm FinFET process technology, became noticeably more economical (up to 3.5 times) and, due to a complete redesign of the execution pipeline, gained the ability to execute a greater number of instructions per clock cycle (an advantage in IPC compared to with the latest generations of Bulldozer reaches 52 percent). Zen cores now support SMT technology, which allows two threads to be executed in parallel on one core. The cache memory subsystem was completely redesigned, and support for DDR4 SDRAM was introduced.

At the same time, AMD, which does not have such resources as Intel, included two very important things in the Zen project: high efficiency and modularity. Thanks to the first feature, the Zen microarchitecture made it possible to create relatively compact computing cores that are almost one and a half times larger in area occupied on a semiconductor chip than competitor solutions. In terms of modularity, processors with Zen microarchitecture are somewhat reminiscent of Bulldozer. However, in Zen, the basic CCX (Core Complex) building block consists of four rather than two cores, and does not have any shared parts other than the L3 cache.

It was these properties inherent in the Zen design that allowed AMD to overtake Intel in the number of computing cores contained in mainstream processors. However, you need to understand that eight-core Ryzens did not appear out of a good life. By increasing the number of cores, AMD compensates for the lower specific performance of the microarchitecture, designed to execute four instructions per clock, while modern competitor processors can process up to five x86 instructions in parallel on one core. Zen also has other weak points. For example, new AMD processors are significantly inferior to Intel ones when working with vector instructions from the AVX2 set.

In addition, the modularity of Ryzen also brought with it a fair share of negativity, since the scheme chosen by AMD engineers for combining quad-core modules is not very different high speed transactions. Communications between the cores are provided by a 256-bit proprietary two-way Infinity Fabric bus, which operates at the same frequency as the memory controller. And very often, in those tasks that involve intensive inter-core interaction, the responsiveness of such a bus is not enough. This is why there are entire classes of applications in which users have well-founded complaints about Ryzen performance. One of these classes includes processor-dependent games - in them, the performance of new AMD processors is generally worse than that of Intel's competitors.

But thanks to modularity, AMD, using the same Zeppelin semiconductor chip with two modules and eight cores, was able to prepare and bring to market an extensive family of consumer products: eight-core Ryzen 7, six-core and quad-core Ryzen 5, as well as quad-core Ryzen 3 with disabled SMT technology and reduced cache memory. Moreover, the distribution of these processors across price segments was done in such a way that each of them offers more advanced multi-core and multi-threading capabilities compared to Intel competitors.

The chosen strategy gave very good results. Besides, what about AMD's return to the market? desktop systems both experts and regular users, the company was able to gradually begin to win back market share. According to the latest published analytical data, AMD now controls approximately 13 percent of the desktop processor market, which is two and a half to three percent more than a year ago. If we translate these figures into absolute numbers, then we can say that over the past year AMD was able to increase sales compared to 2016 by about one and a half million processors. It should be emphasized that in the upper price segment, where Ryzen 7 and Ryzen 5 are fighting against processors of the Core i7 and Core i5 series, AMD’s current share is even larger and exceeds 15 percent.

The fact that older Ryzen processors are attracting the attention of enthusiasts is also evidenced by very positive sales data for AMD, which is shared by major retailers. For example, the German retailer Mindfactory.de claims that the most popular desktop CPU has been the Ryzen 5 1600 for several months in a row. And the American online store Amazon.com puts the Ryzen 5 1600 in fourth place in sales statistics, right after the overclockers Core i7 and Core i5 latest generations.

AMD Ryzen Threadripper

AMD's elegant approach to scaling its solutions, where a single semiconductor die with two quad-core Zen modules can be found in any of its desktop processors, has allowed the company to minimize rejection. Most of the blanks that AMD receives from its manufacturing partner, GlobalFoundries, are somehow used in some kind of processor. For example, partially inoperative crystals can be used in lower-end Ryzen 3, where half the cores and half the cache memory are supposed to be disabled.

A similar unification works in the other direction. Zeppelin semiconductor crystals can be used not only in CPUs with eight or fewer processing cores. They can also be combined into clusters to form processors with more than eight cores. EPYC server processors are built on this principle: four Zeppelin crystals are installed in them at once, which allows AMD to produce solutions based on the Zen microarchitecture with a number of cores reaching 32. Linking the crystals into a single whole at the logical level is carried out exactly according to the same principle as combining modules in one chip - via the Infinity Fabric bus. Physically, the crystals are assembled into a single whole within one processor package - under the processor cover in the EPYC there are four crystals at once.

Although server processors are not the direct topic of this material, we had to talk about them because AMD decided to adopt Intel’s approach to producing high-performance desktop solutions and, based on the EPYC design, prepare a separate Ryzen Threadripper HEDT platform. However, unlike Intel HEDT processors, which are usually close analogues of products for servers and workstations, Ryzen Threadripper is a somewhat separate product. It has its own infrastructure (Socket TR4 platform), and in design such processors are significantly different from their server counterparts. While EPYC is a combination of four Zeppelin dies, Ryzen Threadripper is logically formed from just a couple of these building blocks. As a result, the maximum number of cores in the Ryzen Threadripper is sixteen, but it is also an excellent work tool for those users who are busy creating digital content.

The use of two Zeppelin crystals for Ryzen Threadripper doubles the capabilities of this processor not only in terms of support for multi-threaded computing, but also in terms of expandability. As a result, a platform based on such processors acquires all the features necessary for productive workstations. Systems with Ryzen Threadripper can use quad-channel DDR4 SDRAM, and the number of PCI Express 3.0 processor lines increases to 60, which allows you to create powerful multi-GPU configurations or disk arrays, composed of NVMe drives. From this position, Ryzen Threadripper turns out to be even better than Intel's offerings in the HEDT segment.

But the most important advantage of Ryzen Threadripper lies in the price. AMD has made its HEDT platform much more affordable than similar Intel options and offers a 16-core processor 40 percent cheaper than a competitor's 16-core processor. Thanks to favorable price Ryzen Threadripper is in good demand among professionals. As sales statistics show, the older 16-core Ryzen Treadripper 1950X is now sold in noticeably larger quantities than the older multi-core processors of the Core i9 family.

It is curious that, despite all that has been said, AMD did not initially intend to release Ryzen Threadripper. Such processors appeared thanks to the private initiative of individual company employees who were developing a multi-core consumer solution based on the Zen microarchitecture in their free time from their main jobs. However, its creation ultimately did not require too much effort. In terms of hardware design, Ryzen Threadripper is as unified as possible with EPYC server processors. For example, the Ryzen Threadripper processor package is similar to EPYC: in reality, it contains four semiconductor crystals inside, a couple of which are locked at the manufacturing stage.

AMD Raven Ridge (Ryzen Mobile)

Despite the fact that processors for mobile computers are not the topic of this material (you can read about them in other “results”), we could not ignore Ryzen Mobile. The fact is that this is a fundamentally different product from existing desktop Ryzen products, which has noticeable architectural features. In addition, over the coming months, Ryzen Mobile will also have desktop twins - APUs, codenamed Raven Ridge.

At the moment, AMD has introduced two Ryzen processors classified as mobile. They have similar desktop model names Ryzen 7 2700U and Ryzen 5 2500U, but at the same time they are seriously different from the usual representatives of the Ryzen 7 and Ryzen 5 series. There are two reasons for such differences. Firstly, Ryzen Mobile is an APU equipped with an integrated graphics core, which is based on the most advanced Vega architecture. Secondly, mobile media based on Zen microarchitecture fit into narrow 15-watt frames. When AMD was just about to release a new generation of its processors, its representatives claimed that the Zen microarchitecture had enviable versatility and could be used in processors of different classes without any problems. Ryzen Mobile clearly illustrates this: as you can see, there were no problems with the introduction of Zen into energy-efficient processors for laptops.

However, you need to keep in mind that Ryzen mobiles have only four cores with Hyper-Threading support, that is, they are based on a single CCX module. Moreover, CCX in in this case somewhat different than in Zeppelin. The third level cache memory in it is cut in half - to 4 MB. However, due to the fact that in mobile processors built into the crystal GPU Vega with 11 computing units (some of them in the Ryzen 7 2700U and Ryzen 5 2500U are deactivated), its area approximately corresponds to the area of ​​the eight-core Zeppelin. This means that in terms of cost, quad-core mobile APUs are no more profitable than desktop processors.

Despite the low power consumption and heat dissipation, Ryzen Mobile can boast good clock speeds. The base is set at 2.0-2.2 GHz, but the turbo mode is very aggressive and can raise the frequency to 3.8 GHz, typical for desktops. The graphics core operates at a frequency of 1.1-1.3 GHz.

The described characteristics allow us to roughly imagine what capabilities desktop APUs based on the Zen microarchitecture will receive. Obviously, their thermal package will be expanded to 35-65 W, which will allow increasing the base frequency of such processors to the level of 3 GHz and higher, but as for the size of the cache memory and parameters of the graphics core, changes are hardly possible here. This means that in terms of processor performance, Raven Ridge desktops will be something between Ryzen 3 and Ryzen 5. However, in order for the new generation APUs to desktop computers Compared to existing solutions, we were able to increase productivity several times, and this is quite enough.

AMD Bristol Ridge

Raven Ridge processors for the desktop segment have not yet been officially announced, and this event should only be expected in the coming spring. However, this does not mean at all that the Socket AM4 platform released this year has not yet received its APUs. You just need to keep in mind that the Bristol Ridge processors currently offered for this role with an integrated GPU, although released quite recently, are based on an outdated design: their computing cores have an Excavator microarchitecture, and their graphics belong to the GCN 1.3 generation (that is, Fury ).

Thus, the first generation of APUs, which appeared in the fresh Socket AM4 desktop platform, is a re-release of the Carrizo hybrid processors, which AMD has offered as mobile solutions since mid-2015. The only significant difference between the new hybrid processors and their prototype from the past is related to the implementation of compatibility with a more modern platform. To achieve this, Bristol Ridge has added a new memory controller that supports DDR4 SDRAM.

It turns out that hybrid processors with the Bristol Ridge design are unlikely to attract enthusiasts. It is enough just to mention that such APUs are the last refuge of the Bulldozer architecture. And although certain minor optimizations have been made in the Excavator cores, in terms of performance they are very close to previous versions of this microarchitecture. To this we must add that the number of computing cores in Bristol Ridge is limited to four, plus such processors do not have a third-level cache at all. To visualize how dismal performance this configuration can produce, it’s worth recalling that when AMD spoke of Zen’s 52 percent advantage over previous microarchitectures, it was precisely the comparison with Excavator that was in mind.

The external interfaces implemented in Bristol Ridge processors are also not encouraging. The DDR4 memory controller they use is very slow, seriously inferior in latency to the controller of Ryzen processors and cannot work with any high-speed DDR4 SDRAM modules. The bus for connecting discrete graphics accelerators in Bristol Ridge is presented only in a stripped down form to PCI Express 3.0 x8.

The only bright spot in the Bristol Ridge design is the integrated graphics, which in older versions of these processors has 512 stream processors with GCN 1.3 architecture and operates at frequencies exceeding 1 GHz. Thanks to this, Bristol Ridge is superior to Kaveri (Godavari) design carriers and can boast of the title of desktop APU with the most powerful graphics core at the moment. However, it is obvious that this title will be taken away from them as soon as Raven Ridge processors in the Socket AM4 version come to the market.

Apparently, AMD initially had no intention of launching Bristol Ridge in open sale due to the obvious backwardness of their architecture. Such processors were to be distributed on special orders among OEM partners. But later, the company decided to present to the public a limited range of desktop Bristol Ridge models in order to somehow justify the presence of video outputs on most Socket AM4 boards. Nevertheless, they did not attract any noticeable attention to themselves, which, however, is completely natural.

Intel Kaby Lake

Last year, Intel began with the release of a family of mass-produced Kaby Lake processors, but this event did not attract too much attention. The thing is that Kaby Lake should have been called Skylake Refresh, because the new family turned out to be not so new, but only a cosmetic update of the previous design, aimed at common desktop systems. However, we cannot ignore this step in the final article. Despite the fact that there are only a few technical improvements in Kaby Lake, from an organizational point of view these processors mean a lot.

Firstly, the advent of Kaby Lake marked the untimely death of Intel’s “tick-tock” principle, according to which the transition to new manufacturing technologies alternated with microarchitecture updates. Neither one nor the other happened in Kaby Lake, and Intel announced the transition to a new sequence of stages in the development of its CPUs: “process - architecture - optimization.” However, as it became clear later, the company is not able to maintain such a rhythm, and by now everything has slipped into iterative optimization without introducing new technical processes and introducing new microarchitectures, the end of which is not yet visible. But we'll talk about this later.

As for optimization in Kaby Lake, to produce these Intel processors launched an improved manufacturing process with a resolution of 14+ nm, which, due to changes in the semiconductor structure of transistors, slightly pushed back the frequency potential of the chips. As a result, representatives of the Kaby Lake family were able to obtain approximately 200 MHz higher clock speeds and proportionally increased performance compared to Skylake. The achievements of overclocking models have grown within approximately the same limits, although, as before, Intel added headaches to enthusiasts by continuing to use its proprietary polymer thermal interface under the processor cover.

The second fundamental point is that when introducing the Kaby Lake lineup to the market, Intel placed a serious emphasis on further improving the characteristics of low-cost processors. In the Core i3 series, for the first time in the history of the microprocessor giant, an overclocking model with a free multiplier appeared. And the Pentium series unexpectedly gained support Hyper-Threading technology, which in the eyes of the vast majority of users put it on the same level as the Core i3. Indeed, the difference between these variations of chips was almost completely erased: Core i3, compared to the updated Pentiums of the Kaby Lake generation, could only offer slightly higher frequencies and support for AVX instructions, which is not present in Pentium. As a result, the new dual-core, four-thread Pentiums immediately became extremely successful processors for low-cost gaming systems.

In fairness, it is necessary to mention one addition at the architectural level, which nevertheless found a place in Kaby Lake. True, we are only talking about the graphics core of these processors. In them, Intel was able to expand multimedia capabilities: the new GPU, classified by the developers as generation 9.5, received full support for hardware acceleration of encoding and decoding of 4K video in HEVC format with the Main10 profile. IN Skylake processors such support was partially implemented by the driver and used the computing cores, but now everything began to work without any load on processor resources.

Intel Coffee Lake

During 2017, Intel managed to carry out the second “act of optimization” of the Skylake processor design. Just nine months after the release of Kaby Lake, the next generation of mass-produced chips, Coffee Lake, was introduced to the market. As in the previous step, no microarchitectural improvements were made, and from the point of view of the structure of the Coffee Lake computing cores they continue to remain complete analogues Skylake. Only the technical process has changed again. It still uses 14 nm resolution, but another “lift” of the transistor structure has given another increase in their operating efficiency. As a result, Intel has the right to talk about a re-improved 14++ nm process technology, which opened the way to another extensive step in increasing the performance of the proposed mass solutions.

But this time the developers did not focus on increasing clock speeds. Development took a different path - along the path of increasing parallelism and multi-threading, for which additional computing cores were added to Coffee Lake. Intel has completely reviewed its entire lineup and decided that, starting in the second half of 2017, the main striking force that will be thrown at the mass market segment should be processors with six computing cores, previously offered by the company only for the HEDT platform.

As a result, the Core i7 series, which until now included processors with four cores and support for up to eight threads simultaneously, became six-core and, thanks to the preservation of Hyper-Threading technology, twelve-threaded. The Core i5 series, which previously included quad-core processors without Hyper-Threading, now includes simple six-core processors without multi-threading support. And the name Core i3 with the introduction of the Coffee Lake design was given to quad-core processors without Hyper-Threading, completely similar to the representatives of the Core i5 series of the Kaby Lake generation. It is important that the actual clock frequencies of the updated processors have hardly decreased, which was greatly facilitated by both the new 14++ nm process technology and the aggressive Turbo Boost 2.0 technology.

Ultimately, Coffee Lake was able to provide a very noticeable jump in performance. While usually, when changing generations of mainstream Intel processors, the performance increase was at the level of 5-10 percent, Coffee Lake was able to offer approximately 40 percent superiority over the predecessors of the Kaby Lake family. It is believed that AMD forced the microprocessor giant to take such an unprecedented step. Indeed, the Ryzen processors that it began offering in the mainstream segment have four to eight cores, so some kind of response from Intel should have followed. However, in reality, the development of six-core Coffee Lake began long before Ryzen entered the market, and, most likely, Intel made the decision on the urgent need to add cores to mass processors independently. But what the “Ryzen effect” absolutely influenced was the time frame in which Coffee Lake was brought to market.

The announcement of Coffee Lake was pushed forward by several months, and this created many problems for the new products. The most important of them is that Intel was unable to organize deliveries of Coffee Lake in the required volumes. New processors still remain in short supply: far from the full model range is available for sale, and the chips that reach the shelves are sold at inflated prices. This significantly restrains the growth of popularity and distribution of Coffee Lake, which, despite all its attractiveness, is still losing in sales to AMD’s offerings.

There is one more problem. Simultaneously with the announcement of Coffee Lake, Intel was going to update the entire platform, for which it was planned to release a new range of chipsets with improved capabilities such as support for USB 3.1 Gen 2 and the presence of a built-in WiFi 802.11ac link-level controller. However, the development of these chipsets was delayed, and Intel had to urgently construct a “makeshift” - a new Z370 chipset made from the Z270. This set of system logic is so far the only option compatible with Coffee Lake. For this reason, the range of motherboards for new processors is very limited, and all of them are quite expensive. This factor, along with the CPU shortage, seriously hinders the spread of inexpensive Coffee Lake modifications: motherboards for budget processors of other families can be purchased for almost half the price.

Along the way, serious questions arose regarding the reasons for the incompatibility of Coffee Lake with earlier platforms. Intel claims that the need for new boards is caused by the increased demands of processors with an increased number of cores on power supply circuits. However, motherboard manufacturers deny this and claim that the incompatibility is completely artificial. Moreover, there are even examples of successful launch of new processors on old motherboards. Because of this, Coffee Lake's reputation was somewhat tarnished from the very beginning.

However, over time the situation should change. In theory, Coffee Lake really looks very tempting, and as soon as the problems with the availability of processors and motherboards are resolved, buyers will certainly appreciate such an interesting offer in terms of price and performance. Apparently, this will happen sometime in March. At least, it was for this period that the release of Coffee Lake and the accompanying platform was originally planned.

Intel Skylake-X

In addition to the intensive development of the basic desktop platform, which, in the absence of obvious progress in the implementation of new technological processes and without introducing any microarchitectural improvements, was updated twice last year, Intel also brought to the market a new premium platform LGA2066, which belongs to the HEDT class. As before, the processors for it were designed according to server patterns, but in the new Skylake-X design, when transferring server developments to the desktop platform, the company went much further than usual.

Until now, for HEDT processors it was customary to use only the simplest version of the Xeon semiconductor crystal, which had the smallest number of computing cores. Therefore, older processors for high-performance desktop systems until 2017 could offer their owners no more than 8 or 10 computing cores. Last year everything changed. Now Intel has considered it possible to allow desktop users not only to the low-end (LCC), but also to the middle (HCC) version of the Xeon semiconductor crystal. This immediately put Intel’s new HEDT platform in a special position, because the Skylake-X processor lineup included not only eight- and ten-core processors, but also more serious chips with up to 18 cores.

To commemorate this event, Intel even launched a new Core i9 processor line, which included offerings with more than ten cores. But the appearance of a new name is not the only surprise that Intel marketers decided to coincide with the release of multi-core processors. The second surprise was in the pricing policy. With the release of Skylake-X, the maximum cost of a desktop CPU has moved up to two thousand dollars. Considering the significant increase in the number of cores, such a change can hardly be called unfounded, but we have never seen such high numbers on the price tags of desktop processors.

The name Skylake-X clearly indicates which microarchitecture such chips are based on. Indeed, the HEDT platform, which traditionally lags behind mass solutions in terms of development rates, has just received a design that consumer segment appeared in 2015 with Intel's initial introduction of the 14nm process technology. However, this did not stop engineers from introducing a number of unique innovations into Skylake-X at the level of the overall processor structure. However, it should be borne in mind that these innovations were made to meet the needs of server clients, and their effect in the desktop segment is far from ambiguous.

An example is a change in the scheme for combining cores into one. If previously the cores were connected to an intraprocessor ring bus (Ring Bus), of which two could coexist at once in multi-core processors, in Skylake-X a mesh network structure superimposed on an array of cores began to be used for inter-core interaction. In theory, this approach makes it possible to simplify data transfer routes during inter-core interactions in multi-core chips. However, in practice, the benefit of the new connection scheme is manifested only in processors on the HCC chip, and in ten-core or even simpler CPUs, on the contrary, it increases latency.

The second important change concerns the cache memory subsystem. The size of the individual L2 cache for each core in Skylake-X has been increased from the usual 256 KB to 1 MB, but in return, the L3 cache common to the entire processor has been approximately halved - now its volume is calculated at 1.375 MB for each core. At the same time, the algorithm for operating the third-level cache also changed: it became non-inclusive and victimized, which should ultimately increase the efficiency of the caching system without increasing the size of the semiconductor crystal.

It is worth adding that, along with multi-core Skylake-X processors, a pair of quad-core processors with the Kaby Lake-X design were also released for the LGA 2066 platform. In terms of characteristics, they are close analogues of conventional Kaby Lake overclockers for mass-market systems and therefore do not have the opportunity to take advantage of the HEDT platform such as quad-channel memory and an increased number of PCI Express lanes. Therefore, after the appearance of Coffee Lake, Kaby Lake-X design carriers lost all attractiveness and were never able to gain any noticeable popularity.

The past year was unusually rich in processor innovations. And not at all formal and not about “plus five percent”: in 2017, chips for desktop computers took a very noticeable step forward both in terms of architecture and in terms of speed and capabilities. It is quite natural that processor developers will hardly be able to extend such a high pace and such a high concentration of innovations for another year. Therefore, we can hardly expect the same intense cycle of events from 2018. In the coming twelve months, everything will most likely be limited to evolutionary changes.

In the first half of the coming year, AMD should finally move Raven Ridge processors to the desktop segment and enable the Socket AM4 platform to become truly universal. The APUs that AMD currently offers for this platform do not stand up to criticism. The promising Raven Ridge is based on the modern Zen microarchitecture, and this should allow them to become a worthy alternative Intel Core i3 and Pentium in applications where a powerful video accelerator is not required. Thanks to this AMD platform should regain the right to use in office and multimedia computers, which is quite capable of becoming another impetus for this manufacturer to increase its market share.

In addition, around March we expect to see updated versions of Ryzen processors, which are currently in the news under the code name Pinnacle Ridge. The main circumstance determining their appearance will be the transfer of the Zen microarchitecture to the 12-nm process technology, which was launched last spring in the 12LP (Leading-Performance) version by AMD's production partner, GlobalFoundries. It is generally accepted that along with this, Ryzen processors will be able to receive higher clock speeds and some additional optimizations, such as an improved memory controller.

As for Intel, in the first half of the year it should solve all the problems inherent in Coffee Lake. The model range should be expanded, regular deliveries should be stabilized, and prices should be brought into line with the official price list. In addition, during the same period, the announcement of a full set of system logic sets of the three hundredth series is expected, which should give the green light to the emergence of inexpensive motherboards with Coffee Lake support.

There is no clarity yet regarding the further update of mass-produced Intel processors for the desktop segment. The current versions of Intel's plans do not promise any replacement for the Coffee Lake design throughout 2018, but at the same time it is known that in the second half of the year in mobile computers Whiskey Lake processors will have to arrive, which will be the third optimization of Skylake, produced using the 14 +++ nm process technology. At the same time, the new 10-nm technology, which was initially planned by the microprocessor giant for implementation at the end of 2015, together with Ice Lake processors, will be able to appear in desktop chips no earlier than 2019.

Consequently, the only new Intel desktop product in 2018 may be Cascade Lake-X processors for the high-performance HEDT platform. Intel plans to release them in the fourth quarter. However, they should not bring any special changes with them, because the design of Cascade Lake-X is simple optimization Skylake, produced by switching to a 14+ nm process technology.

However, we will certainly find something to debate about during 2018. For example, in the first quarter, the microprocessor giant promises to introduce incredible Core H processors, in which four cores with a Kaby Lake design will be adjacent to an integrated AMD Vega graphics accelerator and HBM2 memory. So far, not too much is known about these unexpected solutions, and the most we know for sure is that they are aimed at gaming laptops and compact NUC class systems. However, the example of such cooperation between AMD and Intel in itself looks very inspiring and shows that for the sake of progress, technology companies can form the most unexpected alliances. In general, it definitely won’t be boring.

The processor is one of the main components of your computer. If you want to get a high-performance gaming computer, you need to choose the best processor. First of all, you need to pay attention to the new Ryzen processors. However, if you don't want to change the motherboard or you don't have enough money, you can get a top-end processor on the AM3+ socket. There are a lot of great options available for socket AM3+, so choosing the best AM3+ processor from them can seem like quite a daunting task.

The AM3+ socket replaces AM3 and supports Vishera (Piledriver) and Zambezi (Bulldozer) processors. So, many motherboards support both AM3+ and AM3 processors, so you have a lot of choice. Now let's move on to the review of the processors themselves.

The performance of your system will depend on the processor. When choosing a processor, pay attention to the following parameters:

• Single core performance. You may be wondering why you need to consider single core performance when you have at least six cores. This setting appears in programs that are not optimized to effectively use all six or eight cores. And such programs and games are still very common.
• Multi-core performance. Almost all AM3+ processors have excellent multi-threaded performance. A large number of cores means that programs for tasks such as video editing, rendering, 3D modeling, and virtualization will run much faster.
• Core frequency. This parameter means the number of operations that the processor can perform per second. Measured in gigahertz (GHz). A higher frequency means that the processor can perform more operations in a given period of time. In multi-core processors, tasks are distributed among the cores to increase processing speed;
• Turbo Core frequency. When the system needs more computing resources, the processor frequency is automatically increased by 400-600 MHz to obtain higher performance. The more the Turbo Core frequency increases, the higher its overclocking potential;
• Technical process. This parameter determines the dimensions of one structural element of the processor. Modern processors are manufactured using a process technology less than 15 nanometers. In most cases, the smaller the process, the lower the energy consumption and heat dissipation. Processors on the AM3+ socket were manufactured using the 32 nm process technology;
• Heat dissipation. Processor performance affects its heat dissipation. The higher the clock frequency, the greater the power consumption and heat dissipation. Top AM3+ processors run on high frequencies and support overclocking, so they consume more power, which means they generate more heat. Therefore, you will need to buy a very good cooling system.

The best processors on socket AM3+

Here best processors on AM3+, which we will look at in this article:

1.AMD FX-9590

This processor is the most powerful in the list of the best AM3+ processors. It was the first processor to support frequency increases up to 5 GHz in Turbo Core mode. With a base clock speed of 4.7 GHz, you can get very fast gaming and multimedia performance with this processor.

This processor is best installed only on motherboards with a good power supply, given its 220 W thermal package. It also needs a good cooling system, since the processor gets very hot. Among other things, the processor has an unlocked multiplier, which allows you to overclock it. If you have good cooling, you will get very high performance with this processor. But for games you will also need a good discrete video card, since the processor does not have integrated video.

Advantages:

• Overclocking support using AMD Core 3.0 technology;
• Frequency increase to 5 GHz in Turbo Core;
• Eight cores developed on Bulldozer architecture;
• There is support for M.2 and USB 3.1 drives;

Flaws:

• High heat generation and power consumption, so a special motherboard and cooling system are required.

2.AMD FX-8370

The AMD FX-8370 processor is designed for those who want multi-core processor, which will work on a budget motherboard and a not very expensive cooling system. Its heat output is only 125 watts. Compared to the FX-9590, this processor can be considered a winner because its heat dissipation is much lower, but you still get the same eight cores.

If we talk about overclocking, then the AMD FX-8370 also has an unlocked multiplier and you can use software AMD Overdrive to overclock it. By default, the processor runs at 4 GHz, but it can be overclocked to 4.3 GHz without any problems. It can be used without any problems both for games and for working with multimedia and video.

Advantages:

• High clock speed: 4 GHz and 4.3 GHz in Turbo Core;
• Overclocking support using AMD overdrive;

Flaws:

• Not suitable for building ultra-powerful gaming PCs.

3.AMD FX-8320

This is the second generation processor in the FX processor line. It also has 8 computing cores with a base frequency of 3.5 GHz and a maximum frequency in Turbo Core mode of 4 GHz. If you're looking for the best socket AM3+ processor for gaming, this is a great choice. This best solution in terms of FPS. In modern games, this processor is capable of providing smooth gameplay, delivering 60 or more frames per second.

Since this processor has eight cores, it goes without saying that it has great overclocking capabilities. The base processor frequency is 3.5 GHz, and in Turbo Core mode it can increase to 3.7 GHz. It is perfect for working with multimedia and video, as well as for games.

Advantages:

• Provides high performance due to large quantity cores;
• Supports DDR3-1866 RAM;
• Supports customization of energy consumption;

4. AMD FX-6300

AMD FX-6300 is a budget processor. It appeared around the same time as the FX-8350 and was positioned as a replacement for the AMD FX-6200. It also has good overclocking headroom at a very low price.

The processor has a base clock speed of 3.5 GHz, which can be increased to 3.8 GHz. With good air or even water cooling, some users were able to achieve a frequency of 4.8 GHz. However, the processor does not have an integrated graphics chip.

The processor is perfect for building a budget computer. In normal use, it has a heat output of 95 watts. But during overclocking, the energy consumption of the processor, and therefore the heat dissipation, will increase.

Advantages:

• The best budget AM3+ processor for games;
• Unlocked multiplier;
• Well suited for video processing.

5.AMD FX-8350

AMD FX-8350 is a top-end processor with a single core base frequency of 4 GHz. It is developed on the Bulldozer architecture and has good overclocking reserves. The manufacturer officially states that in Turbo Core mode the processor can operate at frequencies up to 4.2 GHz. And this is a very good indicator.

The processor runs on socket AM3+ and supports DDR3-1866 RAM. It is great for both gaming and other tasks, such as 3D modeling or HD video editing.

Advantages:

• Support Turbo technology Core;
• Technological process - 32 nm;
• Bulldozer architecture with good overclocking potential;

Flaws:

• The cooling fan is a little loud.

Conclusions

This article contains the best socket AM3+ processors that you can use in your system builds. The AMD FX-8350 is the best AM3+ processor when considering the heat dissipation to performance ratio. It is perfect for both overclocking and gaming. The processor can operate at frequencies up to 5 GHz, and therefore requires very good cooling.

AMD is one of the largest manufacturers of processors for computers and laptops. Today, all personal computers are equipped with processors either from this company or from Intel. For a number of reasons, many users choose one of the two manufacturers because each has its own advantages. Now we will look at the best AMD processors so that adherents of this company can choose the right model for themselves. The developers have an impressive range, so any user can decide which processor is best to choose for their needs.

No. 10 – AMD A10 Kaveri

Price: 6900 rubles

AMD A10 Kaveri is a budget processor with four cores that is perfect if you want to build a multimedia or office computer. The model looks great both in a work office computer and in a gaming PC.

For games, the processor will not be the most powerful, but still a worthy solution. It has quite impressive characteristics, but they are not enough to play modern demanding projects at maximum settings.

A separate advantage is the automatic frequency increase function. The processor does not need to be overclocked to get maximum power.

When the need arises, the frequency automatically increases to the maximum limit, which is especially noticeable in games. The A10 has a built-in graphics core.

Of course, its power is not enough to replace a full-fledged video card, but it is enough for everyday tasks and most current online games.

No. 9 – AMD FX-6350 Vishera

Price: 5940 rubles

The AMD FX-6350 Vishera is a six-core, six-thread processor that's great for gaming and other standard tasks. The processor clock speed is 3.9 GHz, but in turbo mode it reaches 4.2 GHz.

In addition, it is possible to overclock the processor to increase its performance.

The FX-6350 is created using a 32 nm technical processor and has an open multiplier. The downside of this solution is the lack of an integrated graphics core, so a video card will be required in any case.

Also, if the device is overclocked, powerful cooling will be required to maintain operation. The processor is really quite hot. But if you have a good cooling system in your case, you don't have to worry about this.

Overall, the FX-6350 is one of the most powerful options in the budget segment.

AMD FX-6350 Vishera

No. 8 – AMD FX-8320 Vishera

Price: 5630 rubles

If you need more performance for home computer, you can pay attention to the excellent eight-core AMD processor FX-8320 Vishera with a frequency of 3.5 GHz is one of the most powerful processors in the entry-level AMD segment.

This is already a fairly fast processor on its own, but in certain situations it is capable of achieving even higher speeds. In turbo boost, all eight cores are accelerated to 5 GHz clock speed.

Vector extensions used by AMD also improve performance, increasing speed for particularly heavy applications. For this purpose floating point calculations are used.

The 8MB L3 cache is another improvement over most other processors, providing a great user experience. Your personal computer will last a long time with this fast, cool and reliable processor and will give you the necessary power reserve to perform various tasks.

AMD FX-8320 Vishera

No. 7 – AMD Ryzen 3 1200

Price: 5640 rubles

Our top processor continues with the AMD Ryzen 3 1200. Having passed all the basic tests, this chipset showed very good results, as did the entire AMD Ryzen 3 line in general.

This model will especially appeal to those who like to overclock hardware. With a standard 3.2 GHz, it can be overclocked to 4 GHz, which makes the 1200 one of the most powerful options in the budget segment.

Ryzen 3 1200 has received good reviews from customers. This is not surprising, since the model offers high base power and great overclocking potential in a very short time. low price.

This is also a great option if such indicators as operating temperature and power consumption are important to you. Everyone is accustomed to the fact that this is precisely what AMD has the biggest problems with.

But the AMD Ryzen 3 1200 consumes quite little power and even under heavy loads does not heat up to temperatures above 64 degrees.

AMD Ryzen 3 1200

No. 6 – AMD Ryzen 3 2200G

Price: 6623 rubles

Another good quad-core processor with a clock frequency of 3.5 GHz. This model is from the same line as the previous one. But in this case we are considering a more advanced device.

It uses a 14 nm technical process, which today is considered a very good indicator. There is also an integrated graphics core, which means that many will be able to build a computer without a video card with such a processor. All this allowed the AMD Ryzen 3 2200G to join the list of powerful and cheap devices.

The Ryzen 3 2200G can be overclocked to a clock speed of 4.1 GHz. With such indicators, the chipset can already run most games at maximum settings. Unless of course you play in 4K resolution.

The processor has a fairly good built-in video core of 1650 Mhz at 1.3V. So, if you don’t play the most demanding games, you can do without a video card entirely. In a word, this is a logical continuation of a successful line.

AMD Ryzen 3 2200G

No. 5 – AMD Ryzen 5 2400G

Price: 9570 rubles

It's time to consider a more expensive AMD model Ryzen 5 2400G. It is immediately necessary to clarify that previous processors cannot be compared with it. There are many reasons for this.

The Ryzen 5 2400G is capable of providing the highest possible computer performance for a non-flagship processor.

Main feature This model is equipped with an advanced built-in video card. This is one of the most powerful types graphics hardware, making the AMD Ryzen 5 2400G truly a gaming processor.

This chip can replace a separate video card, unless of course we are talking about top models. The combination of four cores with 8 threads and a powerful built-in video card is what can provide you with sufficient performance reserves for any task.

And if you also use a program to overclock processors, you can really achieve high results, which will leave behind more expensive models.

AMD Ryzen 5 2400G

No. 4 – AMD Ryzen 5 2600X

Price: 13969 rubles

AMD Ryzen 5 2600X is an improved version of the previous model. Here, unlike the first, there are already six cores, with a clock frequency of 3.6 GHz.

The processor is created using 12 nm process technology, which is an advanced technology for our time. The 2600X has good overclocking capabilities, which is what attracts many users to this chipset. The model in question can be overclocked to 4 GHz without any problems.

The performance of the Ryzen 5 2600X is more than decent. If you combine this processor with a good video card, you can not only play modern games, but also work with graphics; the processor performs well in video rendering.

AMD Ryzen 5 2600X

No. 3 – AMD Ryzen 7 2700X

Price: 21971 rubles

The AMD Ryzen 5 line is good, but we won’t dwell on it for long and move on to the seven.

The Ryzen 7 2700X is an excellent eight-core processor with a base clock speed of 3.7 GHz and 16 MB of cache. In turbo mode, the processor can accelerate to 4.3 GHz.

Of course, there is also built-in graphics, but it is unlikely that anyone will purchase such a processor for a computer without a video card. This is a top solution for a home PC, which with this chipset will effortlessly cope with any tasks.

AMD Ryzen 7 2700X

No. 2 – AMD Ryzen Threadripper 1950X

Price: 58950 rubles

In the previous paragraphs the most popular processors from AMD, but the table of models from this manufacturer also contains premium solutions that are superior in power to most modern options on the market.

Among them is the sixteen-core AMD Ryzen Threadripper 1950X, which has a clock speed of 3.4 GHz and 32 MB of cache.

The Threadripper 1950X processor can be overclocked to 4 GHz, but even at the standard clock speed it produces very high performance. At the same time, the chipset is quite cold, its temperature usually does not exceed 60 degrees.

It makes sense to buy a Threadripper to perform professional tasks; its capabilities are excessive in a regular gaming PC.

AMD Ryzen Threadripper 1950X

#1 – AMD Ryzen Threadripper 2990WX

Price: 148,710 rubles

Well, the title itself powerful processor from AMD receives the AMD Ryzen Threadripper 2990WX.

This is a powerful unit with 32 cores, which is sold at a very high price. For the same money you can build a fairly powerful personal computer from scratch.

Threadripper 2990WX has a clock speed of 3 GHz, but in turbo mode it can accelerate to 4.2 GHz. As for the cache, there is as much as 64 MB of it.

Technical process – 12 nm. With such a chipset, you can achieve the best possible performance on your PC.

https://youtu.be/3u6dY2stOBA

AMD Ryzen Threadripper 2990WX

So, we have reviewed the best offers from AMD. Among them, any gamer or other advanced user will be able to find the ideal processor for themselves.

Fortunately, AMD does not specialize in any particular segment, but produces models for a wide variety of consumers. This is probably why they achieved such popularity.