How to choose the right CPU

we will try to answer one very simple and at the same time a very difficult question: how to choose a CPU?

Whether you are upgrading your computer or planning to build a new PC, the choice of processor will be one of the most important.

For example, higher clock speeds, the number of cores can seriously affect overall performance, a faster system, more comfortable gameplay, as well as processor-dependent applications related to video editing and rendering.

How to choose the right processor

The choice of processor will affect the determination of the parameters of the motherboard since each processor will work effectively only on a specific socket of the processor and the chipset.

Choosing a processor is a serious matter, and it is necessary to take into account all the nuances that affect its future functional purpose. It is extremely useful to immediately determine what exactly the acquired CPU will be used for: solving office tasks, video editing, working with graphics, complex calculations or launching video games and gaming.

So, the main thing in choosing a processor is the task facing the future user of a computer assembled on its basis. There can be countless such tasks, for convenience we will divide all possible solutions into five large groups, inside which we will choose the most interesting processors from different price categories. We list them in order of increasing the speed of the required processor:

As you can see, the division into groups we have adopted is rather arbitrary: all categories intersect, you can find a lot of “borderline” solutions. But we had to somehow share all the offers on the market for greater clarity, and this is clearly not the worst way. At the same time, one should not forget that the general speed of a computer depends not only on the processor – most of the components contribute to it.

You might argue that you need to allocate a maximum budget for the processor, but this is not always reasonable. It may be better to save some of the money for other related components, such as RAM, cooling system, stable power supply. To choose the processor as efficiently as possible, you need to decide on the main purpose of using your computer

It makes no sense to spend the entire budget on a powerful processor, and “finish off” the rest from the rest. Accessories of a lower level simply will not allow the expensive decision to “turn around in full.” Most often, it is better to choose a processor with slightly less impressive characteristics, but then pick up the remaining components to suit it. Well, the characteristics that you need to pay attention to will be discussed below. To some readers the information on what should be guided by the choice may seem superfluous, we can advise them immediately to go to the tables in the final part of the article.

What is better AMD or Intel

For a long time on the market, there are only two manufacturers of desktop processors – AMD and Intel. The struggle between the adherents of a particular manufacturer lasts for years and most often ends with nothing. On the taste and color, as you know, all the markers are different. We will try to consider interesting solutions from each of the manufacturers.

Today in the market we have Intel latest top end processors i9, whose performance is higher than that of its closest competitor from AMD. But their cost is much lower. And the entry-level and mid-level processors at both companies are excellent, and in recent times they have not differed in price too much.

However, in conclusion, we still note that Intel is steadily occupying the lion’s share of the processor market, for which there are a number of reasons, the analysis of which is an occasion for a separate article. In the meantime, just say: if you have no specific reason to choose a processor from AMD, most likely you do not need it – feel free to choose Intel.

AMD or Intel what to choose?

For a long time, AMD was clearly inferior to Intel and was a clear outsider. However, there was a serious turnaround when the Ryzen / Threadripper-2000 series chips were released. From this moment, the company entered a fairly good competitive level in the market. In workloads involving many cores, the latest AMD Ryzen 3000 processors have even come forward, especially considering the security fixes introduced in the last year or so.
Some fans will have a biased opinion when choosing, but if you are not a fan of one or another brand, it is better to be objective in choosing a processor.

Intel still holds a small lead in games with a resolution of 1080p if you want to extract the maximum possible frames per second from your video card for display on a monitor with a high refresh rate. But AMD has significantly narrowed this gap thanks to its new Zen2 architecture and holds the direction of increasing the number of cores and threads, which can provide leadership in the field of video editing, animation, rendering.

So which to choose? What do you need to know about it? AMD or Intel?

Processor Generation:
If you are considering options for the current generation of processors (AMD Ryzen 2000 or Intel 9th ​​generation), the differences are mostly negligible. With Intel, 1080p is a little better in games and watching videos.
AMD in tasks such as video editing, rendering, will be faster.

We recommend choosing processors, of course, the latest generation, but not the ones that have just appeared on the market, because of the initially somewhat overpriced price. Choosing processors of older generations, you can doom yourself to a dead end platform without the possibility of an upgrade …

Clock frequency, number of cores:
Higher clock speeds will show higher performance in common, common tasks, such as office applications, while a larger number of cores will help you work faster with time-consuming workloads and applications.

Complete system kit:
You need balance and consistency (for example, it is very important to choose a video card for the processor) Do not bundle a strong CPU with weak memory, a video card and a low-budget motherboard.

What generation of CPU do you need?

Every year or so, Intel and AMD update their processor lines with a new architecture. The current generation for Intel is the “9th Generation Core Series,” such as the Intel Core i7-9700K and the more powerful Intel Core i9-9900K. The latest AMD chips are the Ryzen 3000 lineup, such as the AMD Ryzen 9 3900X, Ryzen 7 3800X and Ryzen 7 3700X. When looking at the model number, you can see the generation as the first digit of four numbers (for example: 8 on the Core i7-8400 or 3 on the Ryzen 7 3700X).

Note that both companies tend to lag behind when their extreme chips come out of their current architectures. Thus, the latest and best Intel X-Series processors you can get are still the 7th generation, and the second-generation AMD Threadripper chips came out a few months after the Ryzen 2000 processors were launched.

How to decipher model names and numbers?

The many Latin characters and numbers that make up the name of a CPU product can be confusing. Intel and AMD divide most of their chips into three categories: “good, better, best”, good – Core i3 / Ryzen 3, better – Core i5 / Ryzen 5 and best (at least for AMD) Core i7 / Ryzen 7. Intel has a Core i9-9900K on top of its core product suite, as well as an extreme/premium level like the Core i9-9980XE, priced at around $ 2,000. But for the vast majority of users, these chips are not needed and are outside the price range of most buyers.

For low-budget users, Intel introduces Pentium and Celeron chips (Celeron is a stripped-down version of Pentium), AMD has low-end Athlon processors. At the highest level are AMD Threadripper and Intel Core X series, also Core X / i9 and Xeon W.

What do the model numbers that come after 3, 5 or 7 mean?

The first digit stands for generation. The Intel Core i7-8700 is the 8th generation Core processor, and the AMD Ryzen 5 2600 is the 2nd generation Ryzen processor. Other numbers indicate less significant numbers in the lineup, with higher ones usually better. (more cores and/or more clock frequencies). Also, the letter “K” at the end of Intel models means that it is unlocked for overclocking. Quite a small part of Intel chips with “K”, but almost all AMD Ryzen processors are unlocked for overclocking (“K” is not required). At the end of the AMD model, Letter X means higher clock speed (than without X).

Do I need to overclock?

For some, it makes sense to spend extra money on buying higher-end chips (overclocked from the factory), but overclocking Ryzen is a common practice.

The practice of overclocking the processor to the limit due to the fact that it works at clock speeds above the specified ones is an art that many enthusiasts like to practice. But, if you are not interested in seeing how fast you can get your chip to work without failures, overclocking is often not worth the money for the average user.

In order to overclock the processor to higher frequencies than the original ones, you are likely to spend additional funds on an improved cooling system and a motherboard with good cooling of the chipset and other elements, favorable for overclocking. With regard to AMD chips, overclocking is almost always allowed, but if you want to overclock an Intel chip, you will have to buy a processor with the letter “K”. By the time you take all these additional costs into account.

It is important to remember that even having collected all the necessary equipment, it may turn out that the chip does not accelerate well – or worse, if you do not know what you are doing, then damage your processor or shorten its service life by increasing its voltage too much.

What are the main Processor Specifications?

If you look at processor specifications, you will see many numbers. What do you need to pay attention to?

CPU specs

• Clock speeds
• Cores
• Threads
• TDP: The Thermal Design Profile/Power (TDP) is the maximum amount of heat that a chip generates
• Cache
• IPC

Clock frequency

One of the few universal parameters for all processors from any manufacturers. In this case, clock is called one simple operation, and clock frequency is the ability of the processor to process a certain number of such operations over a certain period of time. It is measured in gigahertz – generally accepted units of the frequency of periodic processes. For example, a processor clock speed of 2.9 GHz means that it can perform 2 billion 900 million operations per second. And the higher this indicator, the more productive the processor is. But this characteristic is not able to take the place of the main and only reference point for choosing, as we have already said, today the processor speed depends on a whole series of parameters. For example, the number of cores.

Number of Cores

Initially, all processors were single-core, and developers achieved their productivity by increasing the clock frequency. But this could not go on forever – having reached a certain limit caused by technology, manufacturers were forced to look for other ways for development. One of them was the increase in the number of cores, which in a simplified form can be represented as “processor in the processor.” This allowed us to simultaneously set two (four, six, and so on) tasks for the processor.

Let’s try to see this with a simple example. While working on this article, we are simultaneously using a text editor and an audio player. A single-core processor would easily allow us to do this, but the parallelism of the tasks we set would be only visual – in fact, the processor would switch between the threads with a high frequency, processing each of them alternately. Since we are talking about milliseconds, we would not notice anything until we wanted to create more such flows – here the notorious “brakes”, freezes and other troubles would begin.

A multi-core processor allows you to physically distinguish between multiple threads of tasks, one core is responsible for one, the other for another. In an ideal situation, this eliminates the need for switching from stream to stream and improves overall performance. Moreover, a number of resource-demanding programs can “scatter” their tasks into several cores, uniformly loading the processor, which ultimately leads to an increase in its efficiency. Unfortunately, in reality, everything is not so rosy – there are always a number of unpleasant nuances. But in general, the situation is just that. At the moment, single-core processors are almost a thing of the past, even low-cost solutions have at least two cores – you can still find a single-core processor, but the meaning of this venture tends to zero.

A bit of alphabet

Often in the name of the processor, you can find a letter designation, for example, Intel Core i5-2550K. In this way, “specialized” solutions are labeled – for example, the letter “K” means an open multiplier. Dealing with what this means is a lot of enthusiasts seeking to “overclock” their processor. But for this, we need quite extensive knowledge in the matter.

As for processors with the letters “S” and “P”, which mean lower power consumption and heat dissipation … Honestly, the heat dissipation of modern processors is already at an acceptable level.

Processor threads

They determine how many independent processes the chip can process at a time. In theory, this is as much as the number of cores. But now many processors have the possibility of a lot of threading, which allows one core to process two threads at once. Intel has Hyper-Threading, AMD has SMT (simultaneous multithreading). A larger number of threads increase multitasking and better performance in applications that require the processor to process a large number of tasks at once, for example, video editors, editing programs, processing mathematical arrays and transcoders.

Processor TDP

What does tdp stand for processor?
This is the maximum amount of heat that the chip emits during operation (at standard speeds), measured in watts (W). Given this parameter, for example, the Intel Core i7-8700K has a TDP of 95 W, you should understand that a processor cooler must be able to cope with such an amount of heat, and also that your power supply can provide a sufficient amount of current (usually than higher TDP the more power is needed PSU). It is also important to consider that during acceleration much more heat is generated.

Without knowledge of TDP, you will not be able to select the necessary cooling device (air or water) for the entire system to work stably, not to mention that overheating can cause serious consequences for the entire system. Often, higher performance is usually combined with higher TDP.

CPU Cache

What is CPU cache? A relatively small amount of processor memory (buffer) is located on the processor board and is used to speed up access to data and instructions between the processor and RAM.

There are three types of cache:

• L1Cache is the fastest, but cramped,
• L2 Cache is more spacious, but slower,
• L3 Cache is spacious, but relatively sluggish.

When the data needed for the CPU is not available in any of the cache areas, it accesses the RAM, which is much slower, partly because it is physically farther than the CPU cache. Most likely, it’s good to pay attention to the size of the cache,but keep in mind it is difficult to equate to real performance, and there are more important factors to consider.

CPU IPC

What is CPU IPC?
A little-known indicator, but very important. The number of instructions per cycle. Even if you have two CPUs with the same frequency, number of threads, cache cores, they can have completely different IPCs. Sometimes even the processors of the new generation with a low frequency are more productive than the old with the best other indicators. IPC is highly dependent on the processor architecture, which is why chips for example, the 9th generation Core i7 will be better than the 8th generation Core i7.

IPC is usually not specified and you can get this information only by looking at performance tests

What will be more important to you: CPU threads vs cores vs clock speed

To answer this question, you need to understand what tasks are usually set for your PC. High clock speeds will provide faster response and shorter application loading time (RAM and reading/writing speed of hard or SSD drives are important at the same time). With a higher frequency, single-threaded tasks (audio editors, various old programs) will be faster. Many popular games are still single-threaded.

More and more modern programs are written to use many cores and threads. With multitasking (high-resolution video editor, other complex, time-consuming tasks loading the processor), priority tends to more cores. But for most gamers and average users, a frequency of 3-4 GHz with four to eight cores is enough.

What kind of socket does my motherboard need for this processor?

Different processors are installed only in the corresponding type of socket. If you already have a motherboard and do not want to replace it, then you need to purchase a processor that matches its socket. Accordingly, on the contrary, the motherboard you buy on the socket must be compatible with your new processor. AMD has implemented one socket – AM4 for Ryzen and Athlon of the current generation (except Threadripper) and plans to use it without changing until 2020. This means that new generations of processors can be used on an old motherboard with an AM4 socket simply by updating its BIOS.

Intel, on the other hand, has in recent years tended not to support the compatibility of its new chips with old motherboards, even if the socket is practically the same. For example, Intel LGA 1150 and 1151 sockets differ in only one conclusion, but they are already different. Socket 1551-v.2 for 8th generation chips, physically the same as for previous 6th and 7th generation processors. But motherboards with 1151 sockets no longer work with newer processors on 1151-v.2 sockets, as Intel says, newer chips (which have more cores and are obviously more interesting) have great power requirements.

This injustice often upsets those users who planned to upgrade to newer processors from Intell, and this means that they will have to buy a newer and more expensive motherboard for the current generation chip, even if the previous generation board was completely fine.

List of current sockets and corresponding chipsets for reference:

• Intel processor aging slots are labeled: LGA 775
• Modern Intel processor slots are labeled: LGA 1150, LGA 1155, LGA 2011
• Obsolete AMD processor slots are marked: AM2, AM3, FM1, FM2
• Modern AMD processor slots are marked: AM3 +, FM2 +

All you need to know about the slots, friends, is that they must be the same on the processor and motherboard, otherwise the processor simply will not install.

When choosing a processor, first ask yourself what tasks it will have to perform, evaluate the budget considering how much you will spend on other components, such as RAM, a hard drive or SSD, a cooling device, a power supply, a video card, etc.

The processor of your platform will be the main link, but it would be a mistake to combine a high-speed chip with poor graphics card (if you are not a gamer) or with a slow mechanical hard disk (HDD).

Other characteristics

The main parameters that are worth paying attention to are discussed. But there were still a number of details, which we will not talk about in detail, fearing that our readers will fall asleep long before the end of the article. But for a couple of proposals, we still give them. We must say right away that for most modern processors these parameters are at a level sufficient to not think about them too much in the process of assembling a home or office PC.

A good example is the supported type of RAM. Actual today is DDR3 memory with a frequency of 1333 MHz. It is important to remember that most modern processors have a memory controller integrated into the kernel and simply specify in the description of the acquired solution which memory suits it. We will definitely talk more about choosing memory cards in the material devoted to this issue.

Some processors may have a built-in graphics controller that allows you to refuse to buy a video card. For budget “office” computers, this is a very interesting solution, but we would not recommend assembling something more serious without an external video card, despite the presence of very powerful hybrid processors from AMD. At least, due to the impossibility of a separate upgrade of the processor and the video card in the future.

The frequency of the system bus (FSB) determines the transmission speed of incoming and outgoing information from the processor. The larger it is, the faster the exchange of information, and the processor’s work more efficiently. Also, in the descriptions of the processors, you can find the size of the cache – the “fast” built-in memory designed for temporary storage of program code and data. Its presence allows the processor to constantly store data inside itself, and not wait for its receipt from the RAM. Accordingly, the larger the cache, the better. But this is in theory, but in practice the vast majority of processors have exactly the amount of cache that is needed – you should not give it too much importance, it is better to pay attention to the key parameters described above.

The same can be said about the support of various technologies, for example – SSE4, which is a set of fifty-four teams whose task is to increase the performance of processors in working with media, gaming applications and three-dimensional modeling. If you purchase a more or less modern processor, it probably has the whole set of technologies that software developers have managed to use in their products.

With this we will finally finish the theoretical part and move on to the practical – direct selection of processors for specific “types” of computers.