The whole truth about multi-core processors. Six-core Intel Core i5 and Core i7 processors (Coffee Lake) for the "new" LGA1151 Six-core processor

INTRODUCTION Intel has long established itself as the world's fastest desktop processor. And if about what processes for computers of the middle and lower price category should be recognized as the most optimal choice for today, one can argue, in the upper price category there is not even any hint of a choice. Intel Core i7 is a family of processors that AMD cannot offer worthy alternatives. At least at the moment, when the six-core Phenom II, also known as Thuban, is still several weeks away. At the same time, we can say that the existing four-core Phenom II processors are more profitable: they are inferior in performance to the Core i7 by only a couple of tens of percent, and at the same time they are several times cheaper, but this does not change the state of affairs. The most demanding computer enthusiasts are willing to pay extra for high performance, which is why Core i7 processors are so popular.

Even in the absence of direct competition, this consumer interest in high-performance and expensive processors is pushing Intel to continue improving its high-end products, which increase clock speeds, acquire microarchitectural enhancements, and even get more cores. The main character of this article is the recently announced representative of the Core i7 family, which became the first processor for desktop computers to receive six processing cores.

However, it should be understood that the appearance of a six-core model in the Core i7 line is far from the beginning of a six-core revolution. Today Intel is ready to offer the only such processor, the Core i7-980X, belonging to the Extreme Edition series. And this means that so far a six-core CPU is a kind of demo product that will be interesting from a practical point of view only for the most affluent enthusiasts who are ready to shell out about a thousand dollars for a processor alone. Moreover, this state of affairs will last at least until autumn, when, in addition to the Core i7-980X, another, not so expensive model of such a processor may come out. However, the general situation will not change from this - the mass arrival of products with more than four cores on the market will have to wait for a very, very long time. At least when it comes to Intel processors. Of course, AMD can make certain adjustments to the situation with "public six-core", which is going to start selling processors with six cores of the mid-price category in the near future, but so far we have no opportunity to get acquainted with these products in practice, and therefore we will postpone conclusions until a more convenient occasion.

For us, acquaintance with the Core i7-980X is more interesting for another reason. This processor is based on a new Gulftown semiconductor crystal that combines six processing cores and a 12MB L3 cache. The implementation of all these nodes in a monolithic silicon crystal was made possible by using a technological process with production rates of 32 nm. The same process is partially used in the manufacture of the Clarkdale family of processors, but the Core i7-980X is the first product for which the most modern technical process is applied from start to finish. Thus, it is on the Core i7-980X that the evolution of the Nehalem microarchitecture should be fully traced. The recently announced Core i5 and Core i3 processors turned out to be a very bad example in this regard. The distribution of processor units over two semiconductor crystals, one of which is manufactured using a 45 nm process technology, has led to the emergence of additional bottlenecks that have made a negative contribution to the consumer qualities of the final products.

In other words, the Core i7-980X is what Intel engineers are currently capable of when combining cutting-edge process technology with the latest microarchitecture. And it is from this, rather theoretical point of view, that Gulftown is interesting. In practice, in the foreseeable future, such processors will be available only in the most expensive computers, and they will not get into the mass market segment this year for sure. And in 2011, no cheaper Gulftown options are planned, since Intel is going to immediately move to the implementation of the next generation of microarchitecture, Sandy Bridge.

Core i7-980X Extreme Edition in detail

Despite the fact that we described the Core i7-980X as a revolutionary product, we cannot provide any shocking details about its microarchitecture. Intel engineers simply assembled a six-core processor from their standard Nehalem designer, combining the usual elements - computing cores, L3 cache, a memory controller and a QPI bus controller. It's just that in one case, these elements became more - the number of cores increased to six, and in the other - the size of the element increased - the capacity of the L3 cache increased to 12 MB. Yet these components fit on a single chip thanks to a new 32 nm manufacturing process. As a result, despite the fact that the Gulftown crystal consists of 1,170 million transistors, which is about 1.6 times the number of transistors in the Bloomfield crystal, its area is 248 square meters. mm versus 263 sq. mm at Bloomfield.

If you look at the photo of the Gulftown crystal and the placement of various blocks on it, the conclusion suggests itself that we are faced with the result of a simple transfer of parts of the old core to production using a new technological process with minimal adjustments.

If we do not take into account the appearance of two additional nuclei, it is so. By themselves, the processor cores and memory controller of the Core i7-980X are completely similar to the cores and memory controller of the Core i7-900 processors, which have been produced for over a year. In fact, the difference is only in the production technology. The only innovation is the appearance of seven new AES-NI instructions aimed at speeding up the operation of cryptographic algorithms. However, these instructions are already familiar to us from Clarkdale processors.

So we just have to report the main technical characteristics of the new product, comparing them with the characteristics of the Core i7-975 - the older processor of the Bloomfield generation, which is being replaced by a new six-core flagship.

The fact that the memory controller and QPI bus controller used in Gulftown do not differ in characteristics from the corresponding blocks of Bloomfield processors means that they can be used on the same platforms. There is no PCI Express bus controller in Gulftown, and a set of logic is responsible for supporting the graphics subsystem, in the role of which is the well-known Intel X58 Express.

Based on this, it is quite logical that the Core i7-980X has an LGA1366 design and works without problems in motherboards equipped with this connector. All that is required to support the new CPU with older boards is a BIOS update.

By the way, despite the 1.5-fold increase in the number of processor cores, the Core i7-980X fits into the same thermal package as its four-core predecessors. Moreover, the transition to a more advanced technological process did not entail a decrease in the processor voltage - this is clearly seen in the CPU-Z screenshot.

However, Intel has equipped its six-core processor with a new tower cooler using four 6mm heat pipes and a two-speed fan with 100mm impeller.

But this was done not in connection with the increased heat generation, but as another step towards enthusiasts, who now, after purchasing an Extreme Edition processor, may well use a standard cooling system with good efficiency.

L3 cache and memory subsystem

In presenting Gulftown as the most powerful processor at the moment, Intel relies on two of its key features - an increased number of processing cores and an increased amount of cache memory. At the same time, it is quite obvious that at the moment there are not so many applications capable of loading six processor cores simultaneously, and most of them relate to the field of either three-dimensional modeling, or the creation and processing of digital content. Therefore, from the point of view of common applications, another property of Gulftown is much more important - L3 cache memory, the volume of which has been brought to 12 MB. It is thanks to it that in systems based on a new processor, performance gains can be noticeable in old, not optimized for multi-threaded environments, tasks. Moreover, the cache of the third level is common for all cores, which means that, depending on the nature of the load, it can be monopolized by one or several cores.

However, we remember well that even a simple increase in the amount of processor cache always entails some negative consequences. It happened this time too. Since Intel engineers did not touch the logical organization of the L3 cache, leaving it with 16-channel associativity, the increase in volume and the need for arbitration between the increased number of cores led to a 33% increase in its latency.

The second factor that can negatively affect the performance is that the Gulftown processors have reduced the frequency of the Uncore part, which includes, in addition to the L3 cache, the memory controller. Deceleration of Uncore has already been practiced by Intel engineers in Lynnfield processors, in which, due to a decrease in the frequency and voltage of the L3 cache and memory controller, they managed to significantly reduce power consumption. Similar motives were motivated by the developers this time too. The memory subsystem speed in Gulftown-based platforms has been sacrificed for two additional processing cores. Otherwise, the six-core Core i7-980X simply would not fit into the 130-watt thermal package installed for the LGA1366 processors.

As a result, when comparing the characteristics of the cache memory of the older Gulftown, Bloomfield and Lynnfield processors, a rather contradictory picture emerges.

It is quite natural that Gulftown loses to its predecessor in the speed of working with cache and memory. The size of this loss can be estimated, for example, by the results of Everest Cache & Memory Benchmark. During testing, we used DDR3-1600 SDRAM with 9-9-9-24 timings.

Core i7-980X (Gulftown)

Core i7-975 (Bloomfield)

The difference in practical cache performance is immediately apparent. Bloomfiled outperforms Gulftown by about 33% in read speed from the L3 cache and 25% in its latency. The novelty is inferior in terms of the speed of working with memory. The practical memory bandwidth and latency of a six-core processor is about 15-20% worse than that of its four-core predecessor, which has a three-channel DDR3 SDRAM controller similar at first glance.

Thus, despite the larger number of processing cores and a more spacious cache, in real applications the Core i7-980X can be inferior in performance to the Core i7-975 - there are quite objective prerequisites for that. Actually, now it becomes clear why Intel gave the new product such a small processor number. After all, the new Gulftown turns out to be better than the old Bloomfield in far from everything, and its weaknesses cannot be called insignificant.

Turbo Boost and Hyper-Threading Technologies

Introduced in the very first Bloomfield processors, Turbo Boost and Hyper-Threading Technologies are now confident that they have stood the test of time and proven to be effective. And if Hyper-Threading allows you to increase the speed of the system under a multi-threaded load, then Turbo Boost technology plays the opposite role - it helps to increase the speed when loading only a part of the cores. Unsurprisingly, both of these technologies have been carried over to the new six-core Gulftown processor.

With six computing cores in the Core i7-980X, Hyper-Threading technology adds six more virtual cores to this processor, resulting in as many as twelve cores visible in the operating system at once.

Looking at this funny screenshot, a very reasonable question arises: are there such applications that are able to use all these resources to the fullest? In addition, a single memory bus is shared between all cores, so it is not excluded that computing resources will spend too much time waiting for data, since the memory bus bandwidth may not be enough for simultaneously running cores. To dispel all these doubts, we conducted a simple experiment - we checked the system performance level in a popular 3D shooter while a number of processes are running in the background in the system, using computing power and the memory bus. More specifically, we tested the speed in Far Cry 2 by running in parallel several copies of the performance test built into the WinRAR archiver (which itself also supports multithreading). During these tests, the memory worked in DDR3-1600 mode, and for comparison with Gulftown, a similar test was performed on platforms with older processors from the Bloomfield and Linnfield families.

In general, Gulftown handles multi-threaded work much better than its quad-core counterparts. The drop in performance with an increase in the background load of this processor is much slower, which means that the bandwidth provided by the three-channel memory subsystem is generally sufficient when working in multi-threaded environments.

As for the Turbo Boost technology, its implementation in the Core i7-980X is somewhat disappointing. After the Lynnfield processors for the LGA1156 platform got the opportunity to increase their frequency by 667 MHz higher than the nominal within the framework of this technology, we expected to see a similar frequency increase in Gulftown. However, Intel engineers judged differently, and in the new six-core, the Turbo Boost technology turned out to be as conservative as in Bloomfield. As a result, the frequency of the Core i7-980X with a nominal frequency of 3.33 GHz can increase by only 266 MHz - up to 3.6 GHz. Details on the frequencies of the older processors in the Gulftown, Bloomfield and Linnfield families when the turbo mode is turned on are shown in the table.

As a result, the maximum frequency of all senior processors with the Nehalem microarchitecture is the same - it is 3.6 GHz. At the same time, according to official data, the Core i7-980X is able to maintain this frequency even when loading two computing cores. But in practice, we managed to observe the operation of the Core i7-980X at a frequency of 3.6 GHz exclusively with a single-threaded load, while loading the second processor core with the work led to a decrease in the frequency to 3.46 GHz.

However, it should be remembered that the ability to overclock a processor using Turbo Boost technology is determined not only by the activity of the cores, but also by the power consumption of the processor at any given time. So the impossibility of operating the Core i7-980X at a frequency of 3.6 GHz with a dual-thread load is probably due to the fact that the power consumption of this processor in this mode goes beyond the limits set by the specification.

How we tested

There is no doubt that the Core i7-980X is one of the fastest processors. Therefore, in performance tests, to compare with it, we took a pair of the fastest quad-core Intel processors of the Core i7 series and the senior processor of the Phenom II X4 family. As a result, the test systems included the following set of components:

Processors:

AMD Phenom II X4 965 (Deneb, 3.4 GHz, 4 x 512 KB L2, 6 MB L3);
Intel Core i7-980X (Gulftown, 3.33 GHz, 6 x 256 KB L2, 12 MB L3);
Intel Core i7-975 (Bloomfield, 3.33 GHz, 4 x 256 KB L2, 8 MB L3);
Intel Core i7-870 (Lynnfield, 2.93 GHz, 4 x 256 KB L2, 8 MB L3).

Motherboards:

ASUS P7P55D Premium (LGA1156, Intel P55 Express);
Gigabyte MA790FXT-UD5P (Socket AM3, AMD 790FX + SB750, DDR3 SDRAM);
Gigabyte X58A-UD5 (LGA1366, Intel X58 Express).

Memory:

2 x 2 GB, DDR3-1600 SDRAM, 9-9-9-24 (Kingston KHX1600C8D3K2 / 4GX);
3 x 2 GB, DDR3-1600 SDRAM, 9-9-9-24 (Crucial BL3KIT25664TG1608);

Graphics Card: ATI Radeon HD 5870.
Hard Drive: Western Digital VelociRaptor WD3000HLFS.
Power supply: Tagan TG880-U33II (880 W).
Operating system: Microsoft Windows 7 Ultimate x64.
Drivers:

Intel Chipset Driver 9.1.1.1025;
ATI Catalyst 10.3 Display Driver.

Performance

Overall performance

The SYSmark 2007 test, which shows the performance of systems when executing typical scenarios in real applications, immediately highlights the shortcomings of Gulftown, which we talked about above. In the event that the applications used do not have high-quality optimization for multi-core processor architectures, the Core i7-980X can easily lag behind its predecessor, the four-core Core i7-975. This is exactly the picture observed in the E-Learning and Productivity scenarios - in them, the higher result is shown not by the processor with more cores, but by the one with the faster L3 cache and memory controller. Scenarios that simulate the creation and processing of digital content put Gulftown in the first place, which is not surprising, since the applications used for this type of activity are usually good at distributing the load across several computational cores. But as a result, the overall SYSmark 2007 score of the new Core i7-980X is practically the same as the Core i7-975 score.

Gaming performance

Many modern games can already effectively use the resources of dual-core processors. Some of them are also capable of loading quad-core CPUs. To fully load the six-core Gulftown with work, and besides, it has support for Hyper-Threading technology, modern games are clearly not able to do it. Therefore, the differences in the results between Core i7-980X and Core i7-975 are not so striking. Much more important for gaming applications is another factor - increased to 12 MB L3 cache. It is thanks to him that the new Intel CPU can become a useful acquisition for gamers.

3DMark Vantage

The popular 3DMark Vantage benchmark is able to efficiently load any number of processor cores. That is why the result of the Core i7-980X looks very impressive in it. So new world records in this test will now be set mainly by systems based on this processor.

Application performance

Adobe Photoshop is an application optimized for multi-core architectures. But not all operations and filters performed in it use the maximum number of cores. Therefore, the advantage of the six-core processor turned out to be not so significant, and it is partly explained not so much by the number of Gulftown cores as by its increased L3 cache.

Video transcoding is a perfectly parallelizable task. Therefore, here the new Core i7-980X with six cores naturally demonstrates more than 40 percent superiority over the Core i7-975, which has only four processing cores.

A similar picture is observed with nonlinear editing of high-definition video in Premiere Pro.

WinRAR can also use several processor cores, but with an increase in their number more than three, the performance gain becomes almost imperceptible. Therefore, the Core i7-980X and Core i7-975 show similar speeds. And by the way, the 12 MB L3 cache of a six-core processor does not give a visible effect either: its large volume, unfortunately, is neutralized by high latency.

Arithmetic calculations in Excel 2007 can be effectively parallelized. As a result, our test task is calculated much faster on a new processor with a large number of cores.

Sonar 8 Producer audio software also performs slightly faster in the final mixdown on a six-core processor system. The advantage of the Core i7-980X over the Core i7-975 is about 5%.

Final rendering refers to those types of workload that always respond positively to an increase in the number of cores in the system. So at least a 20% superiority of the Core i7-980X over its competitors is quite a natural result.

Single threaded performance

To see how the processors cope with a single-threaded load, we included two additional tests in the study: the MaxxPi computational test and the Fritz chess program, in which the number of processor cores involved was manually set to one. This test is interesting because the older processors of the Core i7 family have Turbo Boost technology, due to which their clock frequency when loaded with a single processor core is equalized at around 3.6 GHz.

As you can see, in these tests, the Core i7-980X and Core i7-975 show relatively similar results with a slight advantage of the older processor, which has a more efficient cache memory in terms of speed. Moreover, the Core i7-870 is catching up with them, the slight lag in which in this case is mainly due to the lower bandwidth of the memory subsystem.

power usage

Formally, the increase in the number of cores in the new Core i7-980X processor did not lead to a change in the calculated heat dissipation. Its TDP compatibility with the LGA1366 platform is ensured both by the more modern technical process used in the production of Gulftown semiconductor crystals, and by the lower frequency and voltage of the Uncore. As a result, the estimated typical heat dissipation of the Core i7-980X, as well as the Core i7-975, is 130 watts.

However, to get a more detailed picture, we also conducted practical energy consumption tests. The following graphs show the total system consumption (without monitor) measured "after" the power supply, which is the sum of the power consumption of all components involved in the system. The efficiency of the power supply itself is not taken into account in this case. During the measurements, the load on the processors was created by the 64-bit version of the LinX 0.6.3 utility. In addition, to correctly estimate idle power consumption, we have activated all available energy-saving technologies: C1E, AMD Cool "n" Quiet and Enhanced Intel SpeedStep.

Without load, the consumption of the LGA1366 platform exceeds the consumption of other platforms, regardless of which processor is used in it. This is explained by the fact that the Intel X58 Express chipset has a very "gluttonous" disposition. The share of the consumption of the processors themselves in idle is no more than a few watts.

The situation looks much more interesting under load. The new six-core processor turns out to be even more economical than its four-core sibling, the Core i7-975. However, the 32nm process technology does not work any special miracles, and the Core i7-980X remains a very power-hungry device: its consumption significantly exceeds the consumption of older CPUs for the LGA1156 and Socket AM3 platforms. On the other hand, given the fact that Gulftown has a one and a half times more computing capacity, power efficiency (the ratio of performance to power consumption) is also reaching a new level.

Overclocking

Transferring the production of processors to a new technological process usually entails an increase in the frequency potential. The Core i7-980X is the first processor exclusively manufactured using a 32nm process technology. That is why the results of its overclocking are of particular interest.

The only Gulftown currently available is the Extreme Edition. This means that Intel does not fix its multiplier, giving the user an easy way to overclock. It was this opportunity that we used when conducting our experiments. To remove heat from the processor during the tests, we used a Thermalright Ultra-120 eXtreme air cooler.

First of all, we tried to set the overclocking limit of the Core i7-980X, which can be achieved without increasing its supply voltage above the standard 1.2 V. recent material, just such overclocking is the most energy efficient and does not lead to a catastrophic increase in power consumption and heat release.

Practical tests have shown that stability without raising the processor voltage is not lost at a maximum frequency of only 3.6 GHz.

Unfortunately, this frequency is very close to the nominal one and can hardly satisfy enthusiasts. Therefore, the second series of experiments was carried out with increasing the voltage on the CPU to 1.35 V. Moreover, as we know from the example of Clarkdale, processors manufactured using 32-nm technology should respond very well to voltage growth.

By increasing the voltage, we were able to achieve stable processor performance at a much higher frequency of 4.13 GHz.

But frankly, this is not the result that we hoped to see when overclocking the new Core i7-980X. It turns out that, despite the fact that this processor was released according to the most modern technological process, it does not overclock better than a year-old CPU built on 45 nm semiconductor crystals. In other words, when overclocked without the use of special cooling means, the frequency potential of Gulftown approximately corresponds to the potential of Bloomfield processors, the overclocking limit of which is in the region of 4.0-4.2 GHz.

By the way, I would like to note two features that we noticed when overclocking the Core i7-980X. First, Gulftown maintains a relatively low temperature even when its frequency increases with increasing supply voltage. 60 degrees at maximum load is very low compared to the temperatures at which Bloomfield Core i7 processors overclocked with increasing supply voltage usually operate. Secondly, a successful Gulftown overclocking requires a rather careful selection of voltage, and too much voltage increase leads to a deterioration in the overclocking results. For example, our processor unit ran at 4.13 GHz when its voltage was raised above the nominal by 0.15 V, but when the voltage was increased by 0.2 V, it could not pass stability tests even at 4.0 GHz.

conclusions

Despite the fact that Gulftown is not only the first six-core processor for desktop computers, but also the first CPU in the production of which exclusively 32nm process technology is used, we would not classify it as a new generation product. In fact, Intel offered us everything the same that we have already seen in Bloomfield processors, only this time for the presentation of the next model in the Core i7 family, not an increase in the clock frequency, but the addition of computational cores was chosen. That, taking into account the block structure of processors with the Nehalem microarchitecture, is not such an innovation.

As a result, the new Core i7-980X theoretically has one and a half times higher performance, which formally makes it the fastest processor for desktop computers. In practice, it all depends on application optimization. As tests have shown, there are not so many tasks that receive a commensurate gain in performance when working on a six-core processor, and they relate exclusively to the creation and processing of digital content. It turns out that the Core i7-980X is a great option for use in a workstation base, and not in a home computer.

It is not surprising that when Intel released the six-core Gulftown to the market, it was limited to offering a single model that costs $ 999. Under normal conditions, using a processor with six computational cores does not make much sense, and Gulftown, moreover, under certain circumstances, may be slower than its four-core predecessors due to the increased L3 cache latency and a slowed down memory controller. So the Core i7-980X is clearly aimed at those high-net-worth enthusiasts who gravitate towards new things primarily out of curiosity, rather than sound judgment. Pragmatists, even after the appearance of the Core i7-980X, will probably not lose interest in the existing quad-core processors, the performance of which is quite enough for everyday work and for modern 3D games. Moreover, the 32-nm process technology does not give any significant dividends: as tests have shown, the Core i7-980X has become only slightly more economical than the four-core LGA1366 predecessors, and its overclocking potential does not at all exceed the capabilities of 45-nm processors.

In general, truly innovative Intel processors, which may be of interest to a wide range of users, will have to wait at least until the beginning of 2011, when the microprocessor giant should bring to the market dual-core and quad-core products with the updated Sandy Bridge microarchitecture, for the manufacture of which 32- nm process technology. Regarding the novelties discussed in this article, I just want to say: "Nothing special."

Dual-core or quad-core, as easy as possible

Let's keep it simple. Here's everything you need to know:

There is only one processor chip. This chip can have one, two, four, six or eight cores.
Currently, an 18-core processor is the best you can get on consumer PCs.
Each "core" is a part of a chip that performs processing. Essentially, each core is a central processing unit (CPU).

Speed

Now, simple logic dictates that more cores will make your processor faster overall. But this is not always the case. This is a little tricky.

More cores give more speed only if the program can split its tasks among the cores. Not all programs are designed to separate tasks between cores. More on this later.

The clock speed of each core is also a decisive factor in speed, as is the architecture. A newer dual-core processor with a higher clock speed often outperforms an older quad-core processor with a lower clock speed.

Power consumption

More cores also result in higher processor power consumption. When the processor is on, it supplies power to all cores, not just the ones involved.

Chip manufacturers are trying to reduce power consumption and make processors more energy efficient. But the general rule of thumb is that a quad-core processor will draw more power from your laptop than a dual-core processor (and therefore drain your battery faster).

Heat generation

Each core affects the heat generated by the processor. Again, as a general rule, more kernels result in higher temperatures.

Because of this extra heat, manufacturers must add better radiators or other cooling solutions.

Price

More cores are not always higher than the price. As we discussed earlier, clock speed, architecture versions, and other considerations come into play.

But if all other factors are the same, then more cores will get a higher price.

All about software

Here's a little secret that processor manufacturers don't want you to know. It's not about how many cores you use, but what software you use on them.

Programs must be specially designed to take advantage of multiple processors. This "multi-threaded software" is not as common as you might think.

It is important to note that even if it is a multi-threaded program, what it is used for is also important. For example, the Google Chrome web browser supports multiple processes as well as Adobe Premier Pro video editing software.

Adobe Premier Pro offers different kernels to work with different aspects of your editing. Considering the many layers involved in video editing, this makes sense as each core can work on a separate task.

Likewise, Google Chrome offers different kernels to work in different tabs. But therein lies the problem. Once you open a web page in a tab, it is usually static after that. No further processing required; the rest of the job is to save the page to RAM. This means that even though the kernel can be used to bookmark the background, there is no need for it.

This Google Chrome example is an illustration of how even multi-threaded software can fail to give you big real performance gains.

Two cores don't double the speed

So let's say you have the right software and all your other hardware is the same. Will a quad-core processor be twice as fast as a dual-core processor? No.

Expanding cores does not address the software scaling problem. Scaling to cores - The theoretical ability of any software to assign the right tasks to the correct cores, so each core computes at its optimal speed. This is not what is actually happening.

In reality, tasks are broken up sequentially (as most multithreaded programs do) or randomly. For example, let's say you need to complete three tasks to complete an activity, and you have five such activities. The software tells core 1 to solve problem 1, while core 2 solves the second, core 3 to the third; meanwhile, core 4 is idle.

If the third task is the hardest and longest, then it would make sense for the software to split the third task between cores 3 and 4. But that's not what it does. Instead, although kernel 1 and 2 will complete the task faster, the action will have to wait for kernel 3 to complete and then compute the results of kernels 1, 2, and 3 together.

This is all a roundabout way of saying that software, as it is today, is not optimized to take full advantage of multiple cores. And doubling the cores is not equal to doubling the speed.

Where are more cores really going to help?

Now that you know what kernels do and their limitations in improving performance, you should ask yourself, "Do I need more kernels?" Well, it depends on what you plan to do with them.

If you play computer games often, then more cores on your PC will certainly come in handy. The vast majority of new popular games from large studios support multi-threaded architecture. Video games still largely depend on which graphics card you have, but a multi-core processor helps, too.

For any professional who works with video or audio software, more cores will be helpful. Most of the popular audio and video editing tools use multi-threaded processing.

Photoshop and design

If you are a designer, a higher clock speed and more processor cache will increase speed better the more cores. Even the most popular design software, Adobe Photoshop, supports single-threaded or light-threaded processes to a large extent. Multiple cores will not be a significant incentive for this.

Faster web browsing

As we said, having more cores doesn't mean faster web browsing. While all modern browsers support multiprocessing architecture, kernels will only help if your background tabs are processing power-intensive sites.

Office tasks

All core Office applications are single-threaded, so a quad-core processor will not increase speed.

Do you need more cores?

In general, a quad-core processor will be faster than a dual-core processor for general computing. Each program you open will run on its own kernel, so if the tasks are separated, the speeds will be better. If you use many programs at the same time, often switch between them and assign them your own tasks, choose a processor with a large number of cores.

Just know this: overall system performance is one area where there are too many factors. Don't expect magical performance gains by replacing just one component, even a processor.

The first computer processors with multiple cores appeared on the consumer market in the mid-2000s, but many users still do not quite understand what multi-core processors are and how to figure out their characteristics.

Video format of the article "The Whole Truth About Multicore Processors"

Simple explanation of the question "what is a processor"

The microprocessor is one of the main devices in the computer. This dry official name is often abbreviated to simply "processor"). Processor - a microcircuit, comparable in area to a matchbox... If you like, a processor is like a motor in a car. The most important part, but not the only one. The car also has wheels, a body, and a turntable with headlights. But it is the processor (like the motor of the car) that determines the power of the "car".

Many people call a processor a system unit - a "box" inside which all the components of a PC are located, but this is fundamentally wrong. The system unit is a computer case along with all its constituent parts - hard drive, RAM and many other details.

Processor function - computing... It is not so important which ones. The fact is that all the work of a computer is tied exclusively to arithmetic calculations. Addition, multiplication, subtraction and other algebra - all this is done by a microcircuit called a "processor". And the results of such calculations are displayed on the screen in the form of a game, a Word file, or just a desktop.

The main part of the computer that deals with calculations - here, what is processor.

What is a processor core and multicore

From time immemorial processor "ages" these microcircuits were single-core. The kernel is, in fact, the processor itself. Its main and main part. Processors also have other parts - for example, "legs" - contacts, microscopic "electrical wiring" - but the very block that is responsible for calculations is called processor core... When the processors became quite small, the engineers decided to combine several cores at once inside one processor "case".

If we imagine a processor as an apartment, then the core is a large room in such an apartment. A one-room apartment is one processor core (a large room-hall), a kitchen, a bathroom, a corridor ... A two-room apartment is like two processor cores along with other rooms. There are three, four, and even 12-room apartments. Also in the case of processors: inside one crystal-"apartment" there can be several cores-"rooms".

Multicore Is the division of one processor into several identical functional blocks. The number of blocks is the number of cores within one processor.

Varieties of multi-core processors

There is a misconception: "the more cores a processor has, the better." This is how marketers who are paid to create these kinds of delusions try to present the case. Their task is to sell cheap processors, moreover, at a higher price and in huge quantities. But in fact, the number of cores is far from the main characteristic of processors.

Let's go back to the analogy between processors and apartments. A two-room apartment is more expensive, more convenient and more prestigious than a one-room apartment. But only if these apartments are located in the same area, are equipped in the same way, and they have similar repairs. There are weak quad-core (or even 6-core) processors that are much weaker than dual-core ones. But it's hard to believe it: still, the magic of large numbers 4 or 6 against "some" two. However, this is exactly what happens very, very often. It seems like the same four-room apartment, but in a dead state, without repair, in a completely remote area - and even at the price of a luxurious "kopeck piece" in the very center.

How many cores are there inside a processor?

For personal computers and laptops, single-core processors have not really been produced for several years, and it is very rare to find them on sale. The number of cores starts with two. Four cores - as a rule, these are more expensive processors, but there is a return on them. There are also 6-core processors, which are incredibly expensive and much less practical. Few tasks are capable of gaining performance on these monstrous crystals.

There was an experiment by AMD to create 3-core processors, but this is already in the past. It worked out pretty well, but their time has passed.

By the way, AMD also produces multi-core processors, but, as a rule, they are significantly weaker than competitors from Intel. True, their price is much lower. You just need to know that 4 cores from AMD will almost always be noticeably weaker than the same 4 cores from Intel.

Now you know that processors have 1, 2, 3, 4, 6 and 12 cores. Single-core and 12-core processors are rare. Tri-core processors are a thing of the past. Six-core processors are either very expensive (Intel) or not so strong (AMD) to overpay for the number. 2 and 4 cores are the most common and practical devices, from the weakest to the most powerful.

Multi-core processor frequency

One of the characteristics of computer processors is their frequency. Those same megahertz (and more often - gigahertz). Frequency is an important characteristic, but far from the only one.... Yes, perhaps not the most important one yet. For example, a 2GHz dual-core processor is a more powerful offering than its 3GHz single-core counterpart.

It is completely wrong to assume that the frequency of a processor is equal to the frequency of its cores multiplied by the number of cores. To put it simply, a 2-core processor with a core frequency of 2 GHz has a total frequency of 4 gigahertz by no means! Even the concept of "common frequency" does not exist. In this case, CPU frequency is exactly 2 GHz. No multiplication, addition, or other operations.

And again, let's "turn" the processors into apartments. If the height of the ceilings in each room is 3 meters, then the total height of the apartment will remain the same - all the same three meters, and not a centimeter higher. No matter how many rooms there are in such an apartment, the height of these rooms does not change. Also clock frequency of processor cores... It does not add up or multiply.

Virtual multicore, or Hyper-Threading

There are also virtual processor cores... Intel's Hyper-Threading Technology makes the computer “think” that there are actually 4 cores inside a dual-core processor. Much like a single hard drive is divided into several logical- local drives C, D, E and so on.

Hyper-Threading is a very useful technology in a number of tasks.... Sometimes it happens that the processor core is only half involved, and the rest of the transistors in its composition toss about idle. The engineers figured out a way to make these idlers work, too, by dividing each physical processor core into two “virtual” parts. As if a large enough room was divided into two by a partition.

Does it make any practical sense v-cores trick? Most often - yes, although it all depends on the specific tasks. It seems that there are more rooms (and most importantly, they are used more rationally), but the area of the room has not changed. In offices, such partitions are incredibly useful, in some residential apartments too. In other cases, there is no sense at all in blocking the room (dividing the processor core into two virtual ones).

Note that the most expensive and productive class processorsCorei7 without fail equipped withHyper-Threading... They have 4 physical cores and 8 virtual ones. It turns out that 8 computational threads work simultaneously on one processor. Less expensive but also powerful Intel class processors Corei5 consist of four cores, but Hyper Threading does not work there. It turns out that Core i5s work with 4 threads of computing.

Processors Corei3- typical "middle peasants", both in price and in performance. They have two cores and no hint of Hyper-Threading. In total, it turns out that Corei3 only two computational threads. The same applies to frankly budget crystals. Pentium andCeleron... Two cores, no hyper-threading = two threads.

Does a computer need a lot of cores? How many cores does a processor need?

All modern processors are powerful enough for common tasks... Browsing the Internet, chatting on social networks and by e-mail, office tasks Word-PowerPoint-Excel: weak Atom, budget Celeron and Pentium are also suitable for this work, not to mention the more powerful Core i3. Two cores are more than enough for normal operation. A processor with a large number of cores will not bring a significant increase in speed.

For games, you should pay attention to processorsCorei3 ori5... Rather, gaming performance will not depend on the processor, but on the video card. Rarely does a game require the full power of a Core i7. Therefore, it is believed that games require no more than four processor cores, and more often two cores will do.

For serious work like special engineering programs, video encoding and other resource-intensive tasks really productive equipment is required... Often not only physical, but also virtual processor cores are involved here. The more computational threads, the better. And no matter how much such a processor costs: for professionals, the price is not so important.

Are there any benefits of multi-core processors?

Of course, yes. At the same time, the computer is engaged in several tasks - at least the operation of Windows (by the way, these are hundreds of different tasks) and, at the same time, playing a movie. Playing music and browsing the Internet. Text editor and music included. Two processor cores - which are, in fact, two processors, will cope with different tasks faster than one. Two cores will do this somewhat faster. Four is even faster than two.

In the early years of the existence of multicore technology, not all programs were able to work even with two processor cores. By 2014, the vast majority of applications are well aware and able to take advantage of multiple cores. The processing speed of tasks on a dual-core processor rarely doubles, but there is almost always a performance gain.

Therefore, the ingrained myth that, supposedly, programs cannot use multiple cores is outdated information. Once upon a time it was really so, today the situation has improved dramatically. The benefits of multiple cores are undeniable, this is a fact.

When a processor has fewer cores, it's better

You shouldn't buy a processor with the wrong formula “the more cores, the better”. This is wrong. First, 4, 6 and 8-core processors are significantly more expensive than their dual-core counterparts. A significant increase in price is not always justified in terms of performance. For example, if an 8-core is only 10% faster than a CPU with fewer cores, but it will be 2 times more expensive, then such a purchase is difficult to justify.

Secondly, the more cores a processor has, the more power-hungry it is. There is no point in buying a much more expensive laptop with a 4-core (8-thread) Core i7 if this laptop will only process text files, browse the Internet, and so on. There will be no difference with the dual-core (4 threads) Core i5, and the classic Core i3 with only two computational threads will not yield to its more eminent "colleague". And on battery power such a powerful laptop will work much less than the economical and undemanding Core i3.

Multi-core processors in mobile phones and tablets

The fashion for several computing cores within one processor also applies to mobile devices. Smartphones, along with tablets with a large number of cores, almost never use the full capabilities of their microprocessors. Dual-core mobile computers sometimes really work a little faster, but 4, and even more so 8 cores is an outright overkill. The battery is consumed completely godlessly, and powerful computing devices are simply idle. The bottom line is that multi-core processors in phones, smartphones and tablets are just a marketing tribute, not an absolute must. Computers are more demanding devices than phones. They really need two processor cores. Four won't hurt. 6 and 8 are overkill in common tasks and even in games.

How to choose a multi-core processor and not be mistaken?

The practical part of today's article is relevant for 2014. It is unlikely that in the coming years, something will change seriously. We will only talk about Intel processors. Yes, AMD offers good solutions, but they are less popular and more difficult to figure out.

Note that the table is based on processors from 2012-2014. Older designs have different characteristics. We also did not mention rare CPU variants, for example, a single-core Celeron (there are some even today, but this is an atypical option that is almost not available on the market). You should not choose processors solely by the number of cores inside them - there are other, more important characteristics. The table will only facilitate the choice of a multi-core processor, but a specific model (and there are dozens of them in each class) should be bought only after thorough acquaintance with their parameters: frequency, heat dissipation, generation, cache size and other characteristics.

CPU	Number of Cores	Computing streams	Typical area of application
Atom	1-2	1-4	Low-power computers and netbooks. Atom processors are designed to keep power consumption as low as possible. Their performance is minimal.
Celeron	2	2	The cheapest processors for desktops and laptops. The performance is sufficient for office tasks, but these are not gaming CPUs at all.
Pentium	2	2	As inexpensive and low-performance Intel processors as Celeron. An excellent choice for office computers. Pentiums are equipped with a slightly larger cache, and, sometimes, slightly higher performance compared to Celeron
Core i3	2	4	Two powerful enough cores, each of which is divided into two virtual "processors" (Hyper-Threading). These are already quite powerful CPUs at not too high prices. A good choice for a home or a powerful office computer without special demands on performance.
Core i5	4	4	Full-fledged 4-core Core i5 processors are quite expensive processors. Their performance is lacking only in the most demanding tasks.
Core i7	4-6	8-12	The most powerful but especially expensive Intel processors. As a rule, they rarely turn out to be faster than the Core i5, and only in some programs. There are simply no alternatives to them.

A short summary of the article "The Whole Truth About Multi-Core Processors." Instead of a synopsis

Processor core- its constituent part. In fact, a standalone processor inside the case. Dual core processor - two processors inside one.
Multicore comparable to the number of rooms inside an apartment. Two-room apartments are better than one-room apartments, but only with other things being equal (location of the apartment, condition, area, ceiling height).
The statement that the more cores a processor has, the better it is Is a marketing gimmick, a completely wrong rule. After all, an apartment is chosen not only by the number of rooms, but also by its location, repair and other parameters. The same goes for multiple cores inside a processor.
Exist "Virtual" multicore- Hyper-Threading technology. Thanks to this technology, each "physical" core is divided into two "virtual" ones. It turns out that a 2-core processor with Hyper-Threading has only two real cores, but these processors simultaneously process 4 computational threads. This is a really useful feature, but a 4-thread processor cannot be considered a four-core one.
For Intel desktop processors: Celeron - 2 cores and 2 threads. Pentium - 2 cores, 2 threads. Core i3 - 2 cores, 4 threads. Core i5 - 4 cores, 4 threads. Core i7 - 4 cores, 8 threads. Intel laptop (mobile) CPUs have a different number of cores / threads.
For mobile computers, energy efficiency (in practice, battery life) is often more important than the number of cores.

Long-awaited models for a mass platform, but already different

Some 15 years ago, the question of the number of cores in the central processors of typical personal computers simply did not arise - of course, there was only one core. True, there could be two processors themselves, although in those (and earlier) years this could not be called a cheap pleasure, and for most users it was even at least somewhat useful. In fact, the standard chicken-and-egg problem was observed: programmers did not take into account the possibility of having a second processor, since users rarely bought dual-processor computers, and rarely bought them precisely because there were practically no programs that could realize the potential of several computing devices. In certain areas, SMP configurations were quite appropriate, but they remained niche solutions - in fact, the most widespread operating systems of the Windows 9x line at that time did not support such "perversions" in principle.

That began to change in 2005, when both AMD and Intel began shipping dual-core processors, but the change did not come very quickly because there was still too little mainstream software to take full advantage of the new features. Of course, there was specialized software, and there were programs that could utilize a larger number of cores, but only in certain niches. However, the transition from one core to two was not even quantitative, but qualitative and when using mainly single-threaded software: the "extra" kernel remained free to ensure the normal functioning of the OS, so it became more difficult to "freeze" the computer even with "crooked" programs, which for many liked it. The beauty of the concept was spoiled by the fact that the first dual-core processor models were "gluing" of a pair of single-core ones, so that, other things being equal, they were more expensive or at comparable prices were not quite equal in technical characteristics (clock frequency, for example). This led to lower performance in mass software and, accordingly, the low popularity of dual-core processors in general. In general, it turned out to be such a kind of vicious circle.

They managed to "open" it in the second half of 2006 - when Intel presented the processors of the Core 2 Duo family. Firstly, they originally had a dual-core design, so the release of single-core models based on it was very limited and only affected the lowest segment (in other words, Celeron). Secondly, they themselves turned out to be very successful - both in desktop and mobile versions. At the same time, this led to a price war between AMD and Intel, as a result of which the prices of processors fell to the level we are used to today. In general, two cores have become the “norm of life”, which programmers have begun to take into account, albeit with a slight delay. But four cores could not become mass for a long time, although the company presented Core 2 Quad in the same year: they were spinning in the same vicious circle “no software - they don’t take it, and if they don’t take it, there is no software”. Only a few users had such software, and they greeted these quad-core processors warmly, thinking about more cores. Sometimes they even bought dual-processor systems from old memory :)

But for such products to become widespread, it was necessary to prepare the market, which is what Intel was doing. In particular, the first Core processors in late 2008 added Hyper-Threading support to the four cores, which allowed them to execute eight threads of code. In 2010, the first six-core processors appeared, quickly falling in price from $ 1000 (which is not so much - the price of extreme Core 2 Quad reached 1,500) to about $ 600. But all this preparation became especially noticeable in 2011 - with the release of Sandy Bridge for LGA1155. Then the company clearly limited the price niche of dual-core processors to $ 150, that is, they definitely did not get into expensive computers. And in general, the mass platform turned out to be "squeezed" by the bar in the region of $ 300 - quad-core Core i7 with HT were sold at these prices. In top-end systems, one could find, rather, six-core processors, which a little later (after the release of LGA2011-3) dropped in price to almost $ 400, that is, the difference became minimal. Well, in the most powerful systems, eight-core processors began to be registered - with a recommended price of "a piece of bucks", but not long before that, models with only four cores were sold at such (and even higher) prices.

In general, all these measures gradually led to the fact that the potential base for software capable of using eight or more threads of computation became large. AMD's efforts also made their contribution - the company tried to “show off its cores” more than once or twice in the competition (not very successfully, but largely because of the problems mentioned at the beginning). In addition, eight-core processors were firmly “registered” in game consoles, albeit with weak cores - and as a result, game engine developers were simply forced to parallelize the code to the maximum extent: it was impossible to “exit” on one or two fast threads due to their complete absence. As a result, they began to expect the next logical step from Intel - the introduction of at least six-core processors into the mass segment. Moreover, this event was expected along with the appearance of Skylake and the LGA1151 platform, that is, a couple of years ago, but it did not happen ...

Actually, at the beginning of 2015, the company made it clear that the distribution of roles and prices on the new platform will be exactly the same as on the previous LGA1150 and even on the LGA1155. Of course, this has frustrated many desktop users who have acquired a quad-core processor over the years and have started to think more. But "more" was available only on a more expensive platform, where some were forced to migrate. The rest did not see a way out of the impasse. Moreover, it was not traced later, when, a few months after the appearance of Skylake on the market, it became known that the next generation Core (Kaby Lake) would differ slightly from Skylake: no obvious changes should be expected either in terms of performance characteristics or in terms of the process technology. At the end of 2017, it was planned to supply 10-nanometer Cannonlake with unknown characteristics.

Several months passed, and the plans changed again: it turned out that there would be another version of the processors, and still using the 14 nm process technology - once again improved, but still quite old, since the first Broadwells based on it were released for another three years ago (of course, these were mobile processors - less mass markets, including desktop, usually receive new models with some delay). And most importantly, the older Coffee Lake models were supposed to get exactly the six cores they were looking for and the LGA1151 execution that was already familiar by that time - what was expected from Skylake the fall before last. At the same time, prices should have remained unchanged, that is, all families for the first time since 2011 had to "move down" one notch. In any case, according to the first assumptions, Core i5 should have received Hyper-Threading, and Core i3 - four cores (the "2 + HT" configuration remained only for Pentium, that is, it "went" to the segment below $ 100, and it is already did, starting with laptop Broadwell and desktop Kaby Lake). Then it turned out that all the same, the Core i5 will be six-core. Here, perhaps, the information available to Intel about AMD Ryzen has already affected: both the level of performance and the number of cores. Moreover, we recall (and we will tell someone for the first time), AMD Ryzen is not only the maximum eight cores, but also models for the mass (including mobile) market with four cores paired with a video core. True, these processors did not come out on time (they were expected this summer), but these are minor technical details. In fact, Coffee Lake is focused on the same niches and has a similar configuration (that is, with an integrated GPU), so giving all models six cores is very convenient for competition. Moreover, four cores with Hyper-Threading support Intel managed to "cram" into a 15 W thermal package - such are Kaby Lake-R, also belonging to the eighth generation and using similar optimizations, not only Core i7, but also Core i5. It is clear that AMD's video core will be (most likely) more productive, but the processor component is of interest to many users no less, if not more. After all, for those interested in graphics, there are discrete graphics cards - IGP will always lag behind them anyway. So from this side, everything is logical.

But with the "usual performance of LGA1151" everything turned out to be not so smooth at all. For obvious reasons, new processors demanded new chipsets - everyone, in general, has long been accustomed to such a situation. But the fact that the new chipsets will turn out to be incompatible with the old processors is something that everyone has lost the habit of since the days of LGA775. And even then, quite often "official incompatibility" in practice turned into "unofficial compatibility". Will it work this time? It is difficult to reject such a possibility yet, but at the moment old processors are physically installed in new boards, but they cannot work. At the same time, there are no completely new chipsets of the 300 series either, there is only the Z370, which is completely similar to the former Z270 - this is the top "caliph for an hour", since next year it should be replaced by the Z390 with USB 3.1 Gen2 support and other improvements. A little earlier, other models of the new family of chipsets should come out, including the inexpensive B360 or H310, which for some time will be sorely lacking for the younger Core i3-8100: the idea of installing an inexpensive non-overclocking processor on a board with an expensive overclocking chipset looks strange. However, the new Core i3 does not fall into the first wave of shipments, but this also applies to some extent to the Core i5-8400. In general, imbalances are possible on the market for the first time, so a pair of an old "expensive" processor and an old cheap motherboard may cost the buyer less than a new "cheap" processor for which no corresponding motherboards have been released yet. This must be taken into account by those who are going to buy new Intel solutions as soon as they become available. Well, we will now check how they work.

Testbed configuration

CPU	Intel Core i5-8600K	Intel Core i7-8700K
Kernel name	Coffee lake	Coffee lake
Prospect technology	14 nm	14 nm
Core frequency, GHz	3,6/4,3	3,7/4,7
# Of cores / threads	6/6	6/12
L1 cache (sum), I / D, KB	192/192	192/192
L2 cache, KB	6 × 256	6 × 256
L3 cache, MiB	9	12
RAM	2 × DDR4-2666	2 × DDR4-2666
TDP, W	95	95

So far, we have got the best pair - the Core i5-8600K and the i7-8700K, which have unlocked multipliers, so the Z370 chipset may come in handy. In principle, these processors differ from each other in the same way as before: i5 have slightly lower official frequencies and lack Hyper-Threading support. That's all. Both models have six physical cores, plus a dual-channel memory controller with support for DDR4-2667 and an old video core, which, although now called UHD Graphics 630, is similar to HD Graphics 630 in Kaby Lake (and it does not differ too much from HD Graphics 530 since Skylake ). However, we will not touch the video core today - all tests were performed with a discrete graphics card based on GTX 1070.

CPU	Intel Core i5-7600K	Intel Core i7-7700K
Kernel name	Kaby lake	Kaby lake
Prospect technology	14 nm	14 nm
Core frequency, GHz	3,8/4,2	4,2/4,5
# Of cores / threads	4/4	4/8
L1 cache (sum), I / D, KB	128/128	128/128
L2 cache, KB	4 × 256	4 × 256
L3 cache, MiB	6	8
RAM	2 × DDR4-2400	2 × DDR4-2400
TDP, W	91	91
Price	T-1716356460	T-1716356308

Without fail, we need to compare the new processors with their immediate predecessors of the seventh generation: Core i5-7600K and i7-7700K. It is easy to see that this is almost the same - only four cores, not six. A familiar (and even boring) configuration for six years.

CPU	Intel Core i7-6800K	Intel Core i7-7800X
Kernel name	Broadwell-e	Skylake-X
Prospect technology	14 nm	14 nm
Core frequency, GHz	3,4/3,6	3,5/4,0
# Of cores / threads	6/12	6/12
L1 cache (sum), I / D, KB	192/192	192/192
L2 cache, KB	6 × 256	6 × 1024
L3 cache, MiB	15	8,25
RAM	4 × DDR4-2400	4 × DDR4-2666
TDP, W	140	140
Price	T-13974485	T-1729322998

We took four more processors from the recent testing of HEDT platforms: the Core i7-6800K was recently the cheapest six-core Intel processor, and now it is being replaced by the i7-7800X (a direct comparison of this with the i7-8700K, we think, is generally very interesting). Due to the specifics of the platform, these subjects today will work with twice the memory capacity relative to other test participants, which, however, is not so important in practice (but it should be mentioned).

CPU	AMD Ryzen 5 1600X	AMD Ryzen 7 1800X
Kernel name	Ryzen	Ryzen
Prospect technology	14 nm	14 nm
Core frequency, GHz	3,6/4,0	3,6/4,0
# Of cores / threads	6/12	8/16
L1 cache (sum), I / D, KB	384/192	512/256
L2 cache, KB	6 × 512	8 × 512
L3 cache, MiB	16	16
RAM	2 × DDR4-2667	2 × DDR4-2667
TDP, W	95	95
Price	T-1723154074	T-1720383938

And a couple of AMD models. The Ryzen 5 1600X used a discrete graphics card to compete directly with the Core i5-7600K and now has to fight the i5-8600K. Ryzen 7 1800X, strictly speaking, does not directly intersect with anyone. But the younger Ryzen 7 1700, unfortunately, never got into our hands, so it's enough to evaluate the ends of the range - both he and the 1700X in performance should be just somewhere between 1600X and 1800X. By the way, the 1700X, by the way, as we know, practically does not differ in performance from the 1800X, but it consumes more power - so it costs less for a reason. In general, we can assume that we gave AMD a slight head start by taking the Ryzen 7 1800X, and also testing both processors with slightly overclocked memory - DDR4-2933 instead of the standard 2667 MHz.

Testing methodology

Methodology. Here we briefly recall that it is based on the following four whales:

Methodology for measuring power consumption when testing processors
Methodology for monitoring power, temperature and processor load during testing
Methodology for measuring performance in games sample of 2017

Detailed results of all tests are available as a complete table with results (in Microsoft Excel 97-2003 format). Directly in the articles, we use already processed data. This is especially true for application tests, where everything is normalized relative to the reference system (AMD FX-8350 with 16 GB of memory, a GeForce GTX 1070 video card and a Corsair Force LE 960 GB SSD) and is grouped according to the scope of the computer.

iXBT Application Benchmark 2017

Eight cores is, of course, eight, but Intel's new six-cores are not too far behind the Ryzen 7 1800X, but are cheaper. Especially good, of course, is the i7-8700K, which works even a little faster than the 7800X. In principle, the i5-8600K did not disappoint us either: it easily bypassed the Core i7-7700K. True, it still lags behind the Ryzen 5 1600X, but this is not the same defeat that was observed in the case of the i5-7600K. By the way, it is worth paying attention to the fact that the advantage over its predecessor is more than one and a half times, that is, we are talking not only about an additional pair of cores. And the Core i7 was also "scaled" almost linearly.

The layout almost repeats itself, only here the Core i7-8700K has not lagged behind 1800X either. Excellent result in the upper segment! And worse - on average: The Ryzen 5 1600X continues to be attractive when used with a discrete graphics card. On the other hand, you can count on the fact that after the appearance of inexpensive motherboards, some Core i5-8400 will be perfect for those who do not need fast graphics - they, in fact, will have no one to compete with in this situation :)

As we already know, in this group, an increase in the number of cores from six to eight does not have a very large effect, and the benefit from SMT (of course) is minimal in such conditions. Therefore, today's newbie couple can simply be considered the winners.

Photoshop continues to freak out: the program clearly does not like not only the absence of Hyper-Threading, since the performance of the Core i5-8600K is only at the level of the i5-7400, not even the 7600K. The other two programs in the group "pull" the newcomer higher, but still we get a great illustration of how software problems can ruin anything. But the Core i7-8700K has no such problems, so in the overall standings it was second only to the i7-7800X.

And again threads are everything, so the Core i5-8600K failed to catch up with the Core i7-7700K. On the other hand, it is cheaper - it is possible for it :) But to lag behind the Ryzen 5 1600X, and even so noticeably, of course, it was not worth it, but the laws of physics are difficult to break. Quality doesn't always outweigh quantity, and the Core i7-8700K only looks like the fastest six-core processor (which it is). No more. But no less.

There is a feeling that the four-channel memory controller “played” once - in any case, it is difficult to explain such success of the i7-6800K with anything else. But the i7-8700K lags behind it slightly, but the Ryzen 7 1800X itself, which closes the top three, is quite noticeable. This program may have room for better performance with new processors, which will allow the i7-7800X and Ryzen to perform better. However, and so the state of affairs with archiving is favorable for beginners, although they are not too ahead of their immediate predecessors.

The main thing in this group is a noticeable increase in performance compared to its predecessors, and at the same prices. A very good level, although not a record one, but six cores are not the maximum by today's standards. But with such proximity to the mass price segment, the result is exactly what a record.

In general, a very serious application, especially in the case of the new Core i7, which can perfectly compete with both Ryzen 7 and the "namesakes" for the HEDT platform. Core i5 is a little less happy, but it is already reaching the level of recent Core i7s and noticeably outperforms its predecessor. At the same time, the new Core i5 is not supposed to lag behind the Ryzen 5 1600X. And the problem is not only in Photoshop - in many other programs the situation is similar. However, the presence of an integrated video core allows assembling small and energy-efficient (and inexpensive) computers on the new Core i5, while Ryzen is more difficult with this. But if you still need to use a discrete video card, AMD remains superior in this segment, and you don't have to buy a 1600X - you can overclock a very inexpensive 1600X. But the state of affairs from above has been radically corrected in favor of Intel.

Energy consumption and energy efficiency

However, performance and price are not the only characteristics of the processor, and in terms of energy consumption, the Core i5-8600K just looks great: it is almost identical to its predecessor. The power consumption of the Core i7-8700K is slightly higher than we would like.

This is especially noticeable if we evaluate only the energy consumption of the processor, without taking into account the platform: after all, a hundred watts for mass solutions is too much. Maybe Intel tried to squeeze the maximum performance out of the top model (it's no secret that such processor races of flagships are carefully studied by those who will buy only Celeron anyway), or maybe we came across a not very successful copy. But in general, we would like more ... More precisely, less: the result of the new flagship is only at the level of Ryzen 5 1600X, which is not bad for AMD, but not for Intel. However, at least the new product cannot be compared with the i7-7800X - and that's good.

But from the Core i5-8600K, we would like higher performance, since now the energy efficiency of a new pair of processors is approximately equal. And yet, Core i5 has it a little better, which also indirectly hints at certain problems with this Core i7 model (or our copy) - earlier the use of SMT improved it, and not vice versa. However, this is nit-picking - all the same, both of these processors are absolute leaders among those tested at the moment. And there are no competitors ... :)

iXBT Game Benchmark 2017

Today we will once again present all the diagrams first, and then - a general commentary for them.

As you can see, the results of all subjects fall into a very small range - which is expected. There are a couple of games where the Core i5-7600K lags behind its rivals (in one it is very noticeable), but it is the only "only" quad-core processor here, and even with a high core frequency, this may sometimes not be enough. However, more often than not, the difference, if there is, is small. It is clear that when using a more powerful video card, such situations can occur more often, but there are not so many more powerful video cards, and against the background of their prices, the savings on the processor looks strange - unless, of course, this is a faithful overclocked Core i5-2500K, which has been for many years he coped with any games and with any video card without any questions at all :) And only today, perhaps, the gamer will want to change it - there is already something for it.

Total

Summing up our testing, we can say: the new processors turned out to be successful, they can be used wherever their predecessors worked, the price has practically not changed. From the objective disadvantages - the power consumption of the Core i7-8700K could be lower. But it is clear that this can be easily "cured" by lowering frequencies, so that on the basis of this crystal it is possible even tomorrow to release notebook processors that are applicable not only in bulky "gaming" models. And this is also a plus, and for Intel, perhaps even more significant than the good results of desktop modifications. In fact, nothing fundamentally new has happened to the market of desktop processors, because six-core models have been here, and for a long time. Now they have dropped a little more - that's all. Here is a laptop (full-fledged, not obscure DTR modifications based on desktop or server processors) with a six-core processor - a new product that can change the market somewhat.

Coffee Lake's downside is the emergence of two incompatible LGA1151 platforms. And if compatibility is not too bad in one direction (except for the owners of two-year-old motherboards who have been cynically cut off the possibility of inexpensive upgrades), then in the other ... In fact, it turns out that for the new platform there are currently not only inexpensive motherboards, but also cheap processors. And the transfer of the same Pentiums to a new version will most likely “hit” the shipments of the old one. In general, this is a problem about which, as it seems to us, major manufacturers have already expressed their dissatisfaction to Intel. No other problems have been found at the moment. These are the processors that many have been waiting for - and now, finally, they waited :) We just think that if these processors came out instead of Kaby Lake, there would be more satisfied ones, even with the same compatibility problems (or rather, its absence) between the two versions of the platform ...

The battle between the two eternal rivals - the manufacturers of central processing units - continues. Shortly after Intel announced the new six-core Intel Core series processors for the consumer segment, AMD released its six-core AMD Phenom II X6 processor, proving that six cores can cost no more than $ 300. all the best from the previous series, as well as a new technology called Turbo CORE. We will talk about the new processor, its technical characteristics and innovations, as well as the test results in this article.

The new AMD Phenom II X6 processors are based on the Thuban core, while the K10.5 architecture remains the same. Unlike Intel, AMD went its own way: increasing the Phenom II X4 by two cores and thus turning it into a Phenom II X6, did not increase the L3 cache in the processor. This made it possible to reduce the total number of transistors and not go beyond the thermal package, without changing the 45-nm technological process.

The new AMD Phenom II X6 series of processors today offers the user a choice of four six-core processors with support for the new Turbo CORE technology. The first and weakest model is AMD Phenom II X6 1035T (2.6 GHz up to 3.0 GHz), followed by AMD Phenom II X6 1055T, clocked at 2.8 GHz with the ability to increase the frequency of individual cores to 3.2 GHz Turbo CORE. The AMD Phenom II X6 1075T processor is clocked at 3 GHz, up to 3.4 GHz with Turbo CORE enabled. The latest processor in this line - AMD Phenom II X6 1090T - is AMD's highest performing processor in the consumer segment at the time of writing. Its nominal clock speed is 3.2 GHz, up to 3.6 GHz. It comes with a multiplier unlocked, allowing it to be overclocked to high frequencies. There are rumors on the World Wide Web about plans to release a more powerful AMD Phenom II X6 1095T processor, which have not yet been confirmed.

AMD Phenom II X6 1090T processor

The AMD Phenom II X6 1090T is based on the Thuban core found in the Phenom II X4 quad-core processors, but is complemented by AMD Turbo CORE technology. According to its technical data, this function is the antipode of Cool'and'Quiet technology, which lowers the clock frequency of processor cores when there is no load on them. The new technology allows to increase the clock speed of active processor cores (no more than three) if the other cores (three or more) are not loaded. In this case, the frequency increase factor is chosen so that the processor does not go beyond the TDP package during operation. A kind of analogue of TurboBoost technology, which Intel uses in its processors. And if Intel's TurboBoost technology is more transparent (its operation can be seen using any system processor monitoring utility, for example CPU-Z), then for AMD processors with Turbo CORE, an increase in frequency can only be detected using a special AMD OverDrive utility. Unlike Intel, AMD Phenom II X6 processors do not have any special control chips that monitor the processor temperature and current consumption in real time. The principle of operation of Turbo CORE technology is quite simple: as soon as three or more processor cores turn out to be in an energy-saving state with the frequency reduced to 800 MHz within the Cool'and'Quiet technology, the processor raises the frequency of active cores by 400 MHz, that is, the multiplier increases by two. At the same time, to ensure stability of operation at an increased frequency, the processor voltage is automatically increased from 1.3 to 1.475 V (in our testing). According to AMD's announcement, the new Turbo CORE technology will be used in the next processors of this and other Phenom II X4 processor lines. That is, the company is betting on this technology, since, according to AMD, it allows you to get an increase in the performance of applications that do not support multi-core. This is a very extensive segment of software, because until now, no more than 30% of programs provide full support for multicore. Others either use it ineffectively, or only one core is enough for them. In general, parallelization support is a topic for a separate article, and therefore we will not be distracted. Let's just note that the introduction of TurboBoost and Turbo CORE technologies by processor giants speaks volumes. The technical characteristics of the AMD Phenom II X6 1090T processor are shown in table. one .

We cannot ignore the announcement of the new AMD Leo platform, which should be a continuation of the Dragon platform, which combines the highest-performance processor, high-performance video subsystem and the most functional AMD chipset. The new platform should include a six-core AMD Phenom II X6 processor, an AMD Radeon HD5800 series graphics card (s), and an AMD 890FX chipset. So far, there has been no official announcement of this platform.

But back to the processor in question. The AMD Phenom II X6 1090T came to our test lab as an engineering sample, so it's not clear yet what packaging it will ship to the end user. The appearance of the processor remains the same, only the inscription has been updated - AMD Phenom X6.

To see how Turbo CORE works, the latest version of AMD OverDrive 3.2.1 was installed. To load the processor cores, we used our own development of our laboratory, which is used when testing coolers. The processor was gradually loaded with several threads. When running one, two or three load streams, the OverDrive utility displayed a very interesting result (Fig. 1).

Unlike Intel processors, where each thread is directed to a separate core, this model takes a different approach. Each thread is evenly distributed between the processor cores, that is, first a part of the code is executed on one core, then on the other, etc. As a result, a smooth heating of the processor is achieved, and the clock frequency of all cores, without exception, varies from 800 MHz to 3.645 GHz. Such a picture of work is observed when the processor is loaded with one, two or three threads.

When increasing to four threads (Fig. 2), the Turbo CORE technology is disabled, and the frequency of all processor cores, without exception, becomes the nominal 3.2 GHz. Today it is difficult to say how justified such an approach is when implementing this technology.

Testing methodology

To test this processor, we were provided with a Gigabyte 890GPA-UD3H motherboard based on the latest AMD 890GX system chip. Since this board, like all modern models, supports DDR3 memory, two Kingston KVR1333D3N8K2 memory modules were installed in it, each with a capacity of 1 GB. The operating system was a 32-bit version of Microsoft Windows 7. The testing methodology for this processor is no different from that described in detail in the article "New version of the test script ComputerPress Benchmark Script v.8.0" and published in the November issue of the magazine last year ... Table 2 shows the execution time of test tasks in seconds for the assembled stand and the reference PC used by us for comparison. In addition, using the utilities from the AMD Phenom II X6 CPU Cooler Test Kit, the 1090T was stress-tested to determine thermal performance. Note that during testing we used a stock cooler for AMD processors.

Test results

Based on those given in table. 2 test results, it can be argued that this processor has 33% lower performance than the reference system. Fields are highlighted in red, where the processor lags behind by more than a minute when performing a task, and in green, those tests in which the result of the new processor approaches the reference values. Recall that we used a stand based on an Inte Core Extreme I7-965 processor and a Gigabyte GA-EX58-UD7 board as a reference PC. According to our classification, the result obtained can be characterized as quite expected. Since AMD has been pursuing a policy of developing mid-range and budget-class processors for a long time, you should not expect very high performance from the new processor. However, AMD decided to take an important step towards users by making six-core processors available with a fairly high performance. As you can see from the table. 2, in most tests the new processor loses to its competitor. However, in the Adobe Soundbooth CS4 benchmark, when editing the audio stream, this processor outperformed the Intel Core Extreme I7-965.

As for the heat dissipation tests, here the new processor can pleasantly impress the user. When all cores were idle, the processor temperature did not exceed 25 ° C. With all cores at maximum load, the temperature rose by only 20 ° C and stabilized at around 45 ° C. This is a very decent result considering the six cores of the processor combined with the 45 nm process technology.

conclusions

Compared to the previous high-performance Phenom II X4 models of the previous generation, the novelty has a number of important advantages. The first is, of course, two additional cores, which gives a certain performance boost when working with multithreaded applications. The second plus is the low power consumption and heat dissipation for the 45nm technological process. The third benefit is undoubtedly the introduction of the new Turbo CORE technology, which can increase processor performance when working with single-threaded applications. However, the most important advantage of the new AMD processors is the company's pricing policy, which continues to make affordable, high-tech, but at the same time efficient processors available to users. The official MSRP of the highest performing Phenom II X6 1090T model is set at up to $ 300 - which means that multi-core architecture will be available to the user like never before.

The whole truth about multi-core processors. Six-core Intel Core i5 and Core i7 processors (Coffee Lake) for the "new" LGA1151 Six-core processor

Core i7-980X Extreme Edition in detail

L3 cache and memory subsystem

Turbo Boost and Hyper-Threading Technologies

How we tested

Performance

power usage

Overclocking

conclusions

Other materials on this topic

Dual-core or quad-core, as easy as possible

Speed

Power consumption

Heat generation

Price

All about software

Two cores don't double the speed

Where are more cores really going to help?

Photoshop and design

Faster web browsing

Office tasks

Do you need more cores?

Video format of the article "The Whole Truth About Multicore Processors"

Simple explanation of the question "what is a processor"

What is a processor core and multicore

Varieties of multi-core processors

How many cores are there inside a processor?

Multi-core processor frequency

Virtual multicore, or Hyper-Threading

Does a computer need a lot of cores? How many cores does a processor need?

Are there any benefits of multi-core processors?

When a processor has fewer cores, it's better

Multi-core processors in mobile phones and tablets

How to choose a multi-core processor and not be mistaken?

A short summary of the article "The Whole Truth About Multi-Core Processors." Instead of a synopsis

Testbed configuration

Testing methodology

iXBT Application Benchmark 2017

Energy consumption and energy efficiency

iXBT Game Benchmark 2017

Total

AMD Phenom II X6 1090T processor

Testing methodology

Test results

conclusions