• Processors with graphics core. Built-in and integrated graphics in the processor

    Compare Iris Pro 6200 and Radeon R7 with HD Graphics and discrete Radeon R7 250X

    The publication of our first article on desktop processors of the Broadwell family, among other things, caused a couple of fair comments regarding testing the graphics core in gaming applications. Indeed: there are tests, but for comparison only the HD Graphics 4600 GPU was taken, with which everything is clear. And here’s how the successes of Intel’s new “graphics top” look against the background AMD processors or inexpensive discrete video cards - from a practical point of view, the question is more important. Moreover, C-series processors are about 100 dollars more expensive than similar Haswell ones, and this is quite enough to purchase a Radeon R7 250X or something close, that is, not a very slow solution.

    Today we will solve all the questions.

    Test bench configuration

    CPUIntel Core i5-4690KIntel Core i5-5675CIntel Core i7-4770KIntel Core i7-5775C
    Kernel nameHaswellBroadwellHaswellBroadwell
    Production technology22 nm14 nm22 nm14 nm
    Core frequency, GHz3,5/3,9 3,1/3,6 3,5/3,9 3,3/3,7
    Number of cores/threads4/4 4/4 4/8 4/8
    L1 cache (total), I/D, KB128/128 128/128 128/128 128/128
    L2 cache, KB4×2564×2564×2564×256
    L3 (L4) cache, MiB6 4 (128) 8 6 (128)
    RAM2×DDR3-16002×DDR3-16002×DDR3-16002×DDR3-1600
    TDP, W88 65 84 65
    GraphicsHDG 4600IPG 6200HDG 4600IPG 6200
    Qty EU20 48 20 48
    Frequency std/max, MHz350/1200 300/1100 350/1250 300/1150
    PriceN/A(0)
    T-10887398
    N/A(0)
    T-12645002
    $412()
    T-10384297
    N/A(0)
    T-12645073

    There will be two pairs of Intel processors - to clearly understand where the Core i7 has preferences over the Core i5, and where one vanity of vanities and vexation of spirit. The comparison will be in gaming applications, of course, and with a discrete video card. We, however, have already investigated this issue, but there the i5 and i7 were of different frequencies, and today we have equalized them in this parameter. In principle, it would be possible to take Broadwell of the same frequency, but it is only available in the form of Xeon, i.e. it is not a mass solution. So there will be no direct intersections here - just both socket models for household use.

    CPUAMD A10-6800KAMD A10-7850K
    Kernel nameRichlandKaveri
    Production technology32 nm28 nm
    Core frequency std/max, GHz4,1/4,4 3,7/4,0
    Number of cores (modules)/threads2/4 2/4
    L1 cache (total), I/D, KB128/64 192/64
    L2 cache, KB2×20482×2048
    L3 cache, MiB- -
    RAM2×DDR3-21332×DDR3-2133
    TDP, W100 95
    GraphicsRadeon HD 8670DRadeon R7
    Number of GPs384 512
    Frequency std/max, MHz844 720
    Price$138()
    T-10387700
    $162()
    T-10674781

    We decided to take two AMD processors so that it would not be boring. In addition, here it is also interesting to evaluate the progress of graphics, and do not forget that the A10-6800K also has a twin brother in the form of the Athlon X4 760K. And which of the Atlons to choose when using a discrete video card (760K or 860K) is an interesting question from a practical point of view. Moreover, 760K will work on a board with a “regular” FM2. Could it be that the user was no longer satisfied with some old A6-5400K, and he decided to change the processor and add a discrete video card? Quite possibly. So let's see if in this situation it makes sense to change the motherboard.

    As for other testing conditions, they were equal, but not the same: the RAM operating frequency was the maximum supported according to the specifications, but they are slightly different. But its volume (8 GB) and system drive (Toshiba THNSNH256GMCT with a capacity of 256 GB) were the same for all subjects. All tests were carried out using the built-in video core (which all six processors have) and in conjunction with a discrete Radeon R7 250X.

    Testing methodology

    Since we have already established that a specific video card has very little effect on programs from the iXBT Application Benchmark 2015 set, we limited ourselves to game technique iXBT Game Benchmark 2015. All results were obtained in a resolution of 1920x1080 (Full HD) at minimum quality settings and at 1366x768 at maximum settings. Why this choice? Maximum settings at FHD resolution are too tough not only for integrated video adapters, but also for many inexpensive discrete solutions. But many people want to improve the quality - even at the cost of reducing the resolution. Moreover, the reduction is not always so radical - users still have old monitors on their hands, up to those that support a maximum of 1280x1024 pixels. So why not check out the “low” modes. In addition, with settings for maximum quality, the specific share of the load on the GPU increases, and today we are interested in GPUs. And even if they don’t cope with the job, it will be a stress test that well demonstrates the actual graphics capabilities.

    Minimum high resolution quality

    As you can see, HD Graphics in Haswell cannot cope with this task; you can already play on both A10s, but on the edge, and Broadwell with Iris Pro leaves no doubt. But if we talk about using a discrete video card, then all processors are equal. The price of the Athlon X4 is several times lower than that of any Core i7. The same state of affairs will be in other games with low requirements for processor performance, but high requirements for graphics.

    But WoT, however, is the exact opposite of what was formulated above - here graphics are needed insofar as. As long as it doesn't interfere. HD Graphics 4600 is obviously not enough. The rest are enough that when adding a discrete video card, performance does not increase, and may even decrease.

    Another processor-dependent game, which requires an HDG 4600 for the selected mode. However, faster graphics even with a weak processor allows you to achieve better results. And the discrete video adapter shows that the fourth level cache in some cases actually makes Broadwell-C a much faster solution than Haswell. However, there is little practical benefit from this - 200 or 300 frames is no longer important. Here, obviously, the quality needs to be improved, which we will do a little later.

    The game is hard on all systems, but especially on video cards. As you can see, only integrated Broadwell graphics, and in the older modification (GT3e), even allow you to play in this mode: Haswell GT2 is traditionally two times behind, and the best AMD IGPs are one and a half times behind. However, when using an inexpensive discrete video card, everyone suddenly becomes equal: both cheap Athlons (and disabling the graphics part in the A10 converts processors in exactly this way), and dear Core i7.

    IN previous version Metro has a similar layout. True, here the A10 is already approaching the threshold of playability, but without stretching it, only Broadwell-C and the like are suitable. A discrete drive (even one as relatively weak as 250X) already depends on the performance of the processors. Another question is that there will still be enough “athlons”, and ten frames per second can be neglected.

    Once again, Hitman is similar to Metro 2033 with minor variations. For example, here two A10s behave very differently different generations even when using discrete data, i.e. optimization in Kaveri is not an empty phrase. However, no matter how you optimize it, the Core i5 is much faster. As for integrated solutions, here again only Broadwell-C is suitable without any stretch of the imagination - the others will have to reduce the resolution.

    A very difficult game that even Iris Pro can't handle! However, as we see, here even 250X is enough without much reserve - paired with slow processors, it is completely on the threshold of playability.

    As we've said many times before, Tomb Raider runs great on everything (or almost everything) in minimum mode. However, the new Broadwell still has something to praise for, since it is not so far behind a budget but discrete video card :)

    In this game, you can't go without discrete data. Moreover, what is curious is that the Iris Pro 6200, as usual, is twice as fast as the HDG 4600, but it is only slightly ahead of AMD solutions. Apparently, the main load is on shader and other units, and they cannot be accelerated using eDRAM. Let's see how this manifests itself when the quality increases.

    The new A10s are more or less enough, Broadwell-C is enough without a stretch, Haswell is nothing to catch here (except for the R-series, also equipped with a GT3e video core). But... but it will be cheaper to install a discrete video card.

    So what do we have in minimum quality mode? Broadwell-C handles almost all the games in our set, except one. The performance of Broadwell GT3e is approximately twice as high as that of Haswell GT2, and these solutions are one and a half times faster than integrated AMD graphics. But it’s better, of course, to use a discrete video card if possible - it may even work out cheaper. And always at least not slower.

    Low resolution but high quality

    A discrete video card allows you to play even when using an inexpensive processor; integrated graphics are still unusable. None.

    With great difficulty and strain, the Core i5-5675C reached 30 FPS. A cheaper combination of Athlon X4 760K or 860K and R7 250X easily scores almost 40. Comments are unnecessary.

    This is where the Iris Pro 6200 looks very good. Let discrete video card and a little faster, but not significantly. What's worse is that its use is not always possible, so the advent of powerful integrated video is a great boon for those in such circumstances.

    There are not enough junior discrete cards either, which means that integrated solutions can be forgotten in practice. From a theoretical point of view, what is interesting is that here they are quite close to each other, which is no wonder: when the main load falls on the GPU itself, no tricks in terms of memory performance will help.

    Everything is even more pronounced than in the previous case. The only interesting thing is that the HDG 4600 is faster than the Radeon HD 8670D. However, this is not practically significant.

    Again it can’t even cope discrete card, and its gap from integrated solutions increases to three to five times. With the minimum quality, let us remember, there were sometimes less than two. Those. the higher the GPU requirements, the more difference between integrated and discrete versions of the latter. Which is more than expected, but not taken into account by everyone.

    If you have a discrete video card, you can play, but an integrated one is not enough, even just any one. A similar picture was seen at minimum FHD settings, only here it became even more clear. But nothing surprising - in general, cards of a minimum Radeon R7 265 level and higher are desirable for this game. And there are not so few such games.

    If with minimal settings this game is very gentle on the video system, then increasing the quality can “bring to its knees” much more powerful solutions than we are considering today. Those. The room for maneuver here is huge, but only owners of discrete video cards can use it successfully.

    Sleeping Dogs behaves in a similar way, only the advantages of a discrete solution are even more visible. But the benefits from eDRAM disappear even more noticeably, since it doesn’t even come to texturing speed: the graphics processors themselves are still too weak. But they are weak in different ways, so the integrated Radeon R7 can even outperform the Iris Pro. In practice, however, this does not matter, since both are still too slow.

    And another similar case confirms the hypothesis stated above :)

    In general, as we see, attempts to use modes with high quality Pictures (even with reduced resolution) only on integrated graphics are usually doomed to fiasco.

    Total

    So what do we see? Modes low quality lend themselves well to modern integrated graphics. At least the best representatives of the latter. The idea with eDRAM is correct and logical - it helps alleviate the shortage bandwidth memory. Actually, thanks to this, the solutions of the Iris Pro line become the fastest in their class. Not necessarily Broadwell - Haswell is not much worse, but such modifications of the latter are not installed in the socket, which imposes its own specifics.

    But can gamers be satisfied with low-quality modes? Probably not. In any case, if modern games they are generally interesting to him - at minimum settings, “modernity” easily disappears, often reminiscent of a picture from ten years ago. Especially if you remember the high cost of Intel processors with GT3e - for this money you can buy something simpler, but with a good discrete video card. AMD solutions are much more affordable, and with an increase in image quality, the performance “sag” is weaker, since the graphics processors themselves are still more powerful (and eDRAM cannot fix this), but... But this does not change anything fundamentally - the final performance is still the same too low, so gamers don’t have to seriously rely on the graphics capabilities of the AMD APU.

    What awaits us in the near future? Processors in the Skylake line are predicted to eventually acquire graphics cores like GT4e, which will have more actuators than before (in fact, GT will also “grow up” with the usual numbers, but much less noticeably, but the appearance new modification directly hints at radical changes), and eDRAM. Moreover, DDR4 support will increase memory bandwidth - albeit not immediately, perhaps. However, it does not follow from this that even such processors will be able to cope with high-quality game modes from our methodology even at low resolutions - for this the performance must be increased by 3-5 times, which is unlikely to happen. They will be able to outperform junior discrete video cards more often, but mostly only in areas where either “it’s enough as it is” or “it’s still fundamentally not enough,” so the fact of greater or lesser performance in itself is not very important.

    In general, progress in the field of integrated graphics is clearly visible. But so far, from a gamer’s point of view, it is still not enough to fundamentally change the state of affairs. Full-fledged gaming computer as before, you must have a discrete video card, moreover, more expensive than the processor. What, by the way, makes Broadwell-C a poor gaming solution in any case (even with a discrete graphics card) is that the benefits of L4 cache are not great enough to justify higher prices. If instead of 250X we used 290X (for example), they would be more noticeable, but all the same, it’s better to spend this money on a video card - the return will be much greater. In addition, the limited thermal package interferes - the Core i5 often turns out to be slightly faster than the Core i7, operating at a higher clock frequency, which is not even close when comparing 4690K and 4770K. In general, Broadwell-C is initially a niche solution, perfect for compact computers, but it has nothing special to do in a “regular” modular desktop: there is no need to “squeeze” into 65 W and you can use powerful video cards, or save a lot if high performance no video required.

    Tired of a long search for a new processor, viewing numerous reviews on forums and flipping through catalogs, the user can finally go to the website of the largest online store in Ukraine, Electronic World, at http://elmir.ua. Of course, he will be amazed not only by the fact that delivery to Kyiv, Kharkov and other cities is possible, not only by the low prices, but also by the wealth of choice that the store provides.

    Looking through all these clever processor specifications, the user may notice that some of them include such a parameter as an integrated GPU. At the same time, other processors this parameter may be absent. What is it and why might it be needed?

    Built-in GPU

    The fact is that some manufacturers, among other things, build a special graphics accelerator into their processors. Or the so-called graphics core. For example, if you buy an amd a6 processor, then the graphics core will be found in it. In others it may not exist.

    The role of the graphics core - the GPU - is exactly the same as that of any video card. It processes the image and displays it on the screen, but the purchase of a separate video card can be avoided, for example, in order to reduce the cost of the entire system as a whole.

    However, does this mean that in this case you can abandon the discrete video card altogether? The built-in graphics core with relatively high performance can really be used not only in office systems, but also in gaming entry level. However, in any case, the power of the GPU is often much lower than that of a separate video card.

    When it may be required

    Purchasing a system with a built-in video accelerator seems more rational at first glance. After all, the cost of a good gaming video card is sometimes several times higher than the cost of such a processor. However, you should not make hasty conclusions. A GPU in a processor can be useful in cases where:

    • the user assembles an office system, from which all that is required is to work with texts, spreadsheets and surf the Internet;
    • a graphics core with decent performance will replace a discrete video card for not too demanding gamers who are not interested in the latest gaming innovations;
    • the user wants to build a system with two graphics cards - built-in and discrete. In this case, one graphics chip will work while running “heavy” applications, and the second one - for example, built into the amd a6 - will come into play when it is necessary to process the needs operating system or office applications. This will achieve a balance between performance and energy consumption.

    Built-in GPU for both gamers and undemanding users plays important role.

    The quality of games, movies, watching videos on the Internet and images depends on it.

    Operating principle

    The graphics processor is integrated into the computer's motherboard - this is what integrated graphics looks like.

    As a rule, they use it to remove the need to install a graphics adapter -.

    This technology helps reduce the cost of the finished product. In addition, due to the compactness and low power consumption of such processors, they are often installed in laptops and low-power desktop computers.

    Thus, integrated graphics processors have filled this niche so much that 90% of laptops on US store shelves have such a processor.

    Instead of a regular video card, integrated graphics often use the computer's RAM itself as an auxiliary tool.

    True, this solution somewhat limits the performance of the device. Still, the computer itself and the graphics processor use the same memory bus.

    So this “neighborhood” affects the performance of tasks, especially when working with complex graphics and during gameplay.

    Species

    Integrated graphics have three groups:

    1. Shared memory graphics is a device based on shared memory management with the main processor. This significantly reduces cost, improves energy saving system, but degrades performance. Accordingly, for those who work with complex programs, integrated graphics processors of this type are most likely not suitable.
    2. Discrete graphics - a video chip and one or two video memory modules are soldered onto the motherboard. Thanks to this technology, image quality is significantly improved, and it also becomes possible to work with 3D graphics with the best results. True, you will have to pay a lot for this, and even if you are looking high powerful processor in all respects, the cost can be incredibly high. In addition, the electricity bill will increase slightly - the energy consumption of discrete GPUs higher than usual.
    3. Hybrid discrete graphics- a combination of the two previous types, which ensured the creation of the PCI Express bus. Thus, access to memory is carried out both through the soldered video memory and through the RAM. With this solution, manufacturers wanted to create a compromise solution, but it still does not eliminate the shortcomings.

    Manufacturers

    As a rule, large companies - , and - are engaged in the manufacture and development of integrated graphics processors, but many small enterprises are also involved in this area.

    This is not difficult to do. Look for Primary Display or Init Display First. If you don’t see something like that, look for Onboard, PCI, AGP or PCI-E (it all depends on the buses installed on the motherboard).

    By choosing PCI-E, for example, you enable the PCI-Express video card and disable the built-in integrated one.

    Thus, to enable the integrated video card, you need to find the appropriate parameters in the BIOS. Often the activation process is automatic.

    3 Great processor for gaming 4 Best price 5

    Computers have entered our lives so tightly that we already consider them something elementary. But their structure cannot be called simple. Motherboard, processor, RAM, hard drives: all these are integral parts of the computer. You can’t throw away this or that detail, because they are all important. But the most important role is played by the processor. It’s not for nothing that they call it “central”.

    The role of the CPU is simply enormous. It is responsible for all calculations, which means it depends on it how quickly you will complete your tasks. This could be surfing the web, drafting a document in text editor, photo editing, moving files and much, much more. Even in games and 3D modeling, where the main load falls on the shoulders of the graphics accelerator, the central processor plays a huge role, and if the “stone” is incorrectly selected, the performance of even the most powerful video card will not be fully realized.

    At the moment, there are only two major processor manufacturers in the consumer market: AMD and Intel. We will talk about them in the traditional ranking.

    The best inexpensive processors: budget up to 5000 rubles.

    4 Intel Celeron G3900 Skylake

    The most affordable Intel processor
    Country: USA
    Average price: 4,381 RUR
    Rating (2019): 4.5

    The rating opens with an extremely weak processor from the Celeron line. The G3900 model has two cores of the previous generation - Skylake, which, coupled with a frequency of 2.8 GHz, gives the lowest performance result. IN synthetic tests the processor shows a result approximately half that of the Core i3. But the price here is quite affordable - 4-4.5 thousand rubles. This means that this processor is perfect for assembling, for example, a simple office computer or a multimedia system for the living room. Overall, this model cannot be called bad. Still, the 14 nm process technology provides good energy efficiency, and the HD Graphics 510 graphics core is suitable for casual games.

    Advantages:

    • Lowest price in class
    • Perfect for office PC or HTPC

    Flaws:

    • Does not support Hyper-Threading technology

    3 AMD Athlon X4 845 Carrizo

    Best price
    Country:
    Average price: 3,070 ₽
    Rating (2019): 4.5

    The processors of the Athlon line belong to the budget class, which is clearly evident from the cost of the bronze medalist. But for a little over three thousand rubles you will get a very interesting stone. There are 4 cores (2 logical cores for each physical), made using a 28 nm process technology. Thanks to this, power consumption is low, and heat dissipation is quite low for AMD - only 65 W. True, you don’t have to be particularly happy about this because the multiplier is locked - you won’t be able to overclock the processor. Another disadvantage is the lack of a built-in graphics core, which means that when assembling an office PC or multimedia system, you will have to separately purchase a video card.

    Advantages:

    • Lowest price in class
    • Great performance for the price

    Flaws:

    • Lack of built-in graphics core
    • Unlocked multiplier

    2 AMD FX-6300 Vishera

    The only 6-core processor in its class
    Country: USA (Produced in Malaysia, China)
    Average price: 4,160 RUR
    Rating (2019): 4.6

    AMD's FX-6300 is the only processor in the category with six cores. Unfortunately, you can’t hope for high power in the budget class - the model is based on the 2012 Vishera core. In normal mode, the cores operate at 3.5 GHz, but, like many AMD CPUs, it overclocks well. Yes, judging by user reviews, the performance is sufficient even for games, but there are still a lot of disadvantages.

    One of the main ones is high energy consumption. Due to the use of inexpensive 32 nm process technology, AMD gets very hot and consumes a lot of electricity. We also note the lack of support for modern DDR4 RAM. Because of this, the processor can be recommended not for building a new PC, but for updating an old one without replacing the motherboard and other components.

    Advantages:

    • 6 cores. Perfect for performing several simple tasks at the same time.
    • Good overclocking potential
    • Low cost

    Flaws:

    • Poor energy efficiency
    • Aging platform

    At the moment there are only two players in the processor market - Intel and AMD. But this doesn’t make the choice any easier. To make the decision to buy a CPU from one manufacturer or another easier, we have highlighted for you several main pros and cons of the products of these companies.

    Company

    Pros

    Cons

    Under Intel is better programs and games optimized

    Lower power consumption

    Performance tends to be slightly better

    More high frequencies cache

    Work effectively with no more than two resource-intensive tasks

    Higher cost

    When the line of processors changes, the socket also changes, which means the upgrade is more complicated

    Lower cost

    Better price/performance ratio

    Work better with 3-4 resource-intensive tasks (better multitasking)

    Most processors overclock well

    Higher power consumption and temperatures (not entirely true of recent Ryzen processors)

    Worse program optimization

    1 Intel Pentium G4600 Kaby Lake

    Better performance
    Country: USA
    Average price: 7,450 RUR
    Rating (2019): 4.7

    We can recommend the good old Pentium for purchase in this category. This processor, like previous participants, is made using a 14 nm process technology, LGA1151 socket. Refers to one of last generations– Kaby Lake. There are, of course, only 2 cores. They operate at a frequency of 3.6 GHz, which causes the lag behind the Core i3 by about 18-20%. But this is not much, because the price difference is twofold! In addition to the frequency of the cores relative low power due to the small size of the L3 cache – 3071 KB.

    In addition to the excellent price-performance ratio, the advantages of this CPU include the presence of a built-in Intel HD Graphics 630 graphics core, which is more than enough for comfortable use of a PC without a discrete video card.

    Advantages:

    • Great price for this performance
    • Generation Kaby Lake
    • Good integrated graphics core

    The best mid-class processors: budget up to 20,000 rubles.

    5 Intel Core i3-7320 Kaby Lake

    The most affordable processor with integrated graphics
    Country: USA
    Average price: 12,340 ₽
    Rating (2019): 4.6

    Let's open the rating with the most affordable processor i-core lines. It is extremely difficult to call the model excellent in terms of price/quality ratio, because the cheaper Ryzen 3 even shows several best results in synthetic tests. However, the model that opens the TOP 5 can be safely chosen not only for an office system, but also for a gaming computer.

    There are only two physical cores, but these are modern 14 nm chips from one of the latest generations - Kaby lake. Frequency - 4100 MHz. This is a very shameful indicator. In addition, there is the possibility of overclocking. Considering the excellent energy efficiency and low heat generation - even with the supplied cooler, the temperature remains at 35-40 degrees during idle, and up to 70 degrees under load - you can safely increase the frequencies. Unlike competitors from AMD, Core i3 has a built-in graphics core, which allows it to be used in office system without discrete video card. But keep in mind that officially it only works on Windows 10

    Advantages:

    • Built-in graphics core
    • Overclocking capability
    • Low temperatures

    Flaws:

    • Poor performance for the price

    4 AMD Ryzen 3 1200 Summit Ridge

    Best price
    Country: USA (Produced in Malaysia, China)
    Average price: 6,917 RUR
    Rating (2019): 4.7

    Ryzen 3 – junior inexpensive new line AMD processors, designed to once again impose a fight on Intel. And the 1200 does the job perfectly. For 7 thousand rubles, the buyer receives a 4-core processor. Factory frequencies are low - only 3.1 GHz (in high performance mode 3.4 GHz), but the multiplier is unlocked, which means enthusiasts can easily make the “stone” a little faster.

    The transition to new chips not only improved performance, but also reduced power consumption, and also reduced temperatures to acceptable values. Due to the lack of built-in graphics chip we can recommend this processor only for budget game builds. Productivity is only slightly higher than the previous participant.

    Advantages:

    • Unlocked multiplier

    Flaws:

    • No built-in graphics chip

    3 Intel Core i5-7600K Kaby Lake

    Great processor for gaming
    Country: USA
    Average price: 19,084 ₽
    Rating (2019): 4.7

    Let's start with the fact that the i5-7600K is by no means an outsider. Yes, in terms of performance it is somewhat worse than the mastodons that you will see below, but for most gamers it will be enough. The processor has four Kaby Lake cores operating at 3.8 GHz (in reality up to 4.0 GHz with TurboBoost). There is also a built-in graphics core - HD Graphics 630, which means you can play even demanding games at minimum settings. With a normal video card (for example, GTX 1060), the processor reveals itself completely. In most games with FullHD resolution (the majority of gamers have these monitors) and high graphics settings, the frame rate rarely drops below 60 fps. Is anything else needed?

    Advantages:

    • Best price
    • Enough power for most gamers
    • Excellent graphics core

    2 AMD Ryzen 5 1600 Summit Ridge

    Best price/performance ratio
    Country: USA (Produced in Malaysia, China)
    Average price: 11,970 ₽
    Rating (2019): 4.8

    The second line of the TOP-5 mid-level processors is occupied by one of best processors in terms of price/performance ratio. With an average cost of only 12,000 rubles, in synthetic tests Ryzen 5 is able to compete with the well-known Intel Core i7-7700K standard settings(In PassMark 12270 and 12050 points respectively). This power is due to the presence of six Summit Ridge physical cores, made using a 12 nm process technology. The clock frequency is not a record - 3.6 GHz. Overclocking is possible, but in reviews users claim that at frequencies above 4.0-4.1 GHz the processor behaves unstable and gets very hot. With factory settings, idle temperatures remain at 42-46 degrees, in games 53-57 when using a standard cooler.

    Also, high performance is due to large cache volumes at all levels. The CPU supports the modern DDR4-2667 standard, which allows you to create on the basis of this processor great computers for playing at medium-high settings in FullHD.

    Advantages:

    • Excellent price/performance ratio
    • Heats up a little

    Flaws:

    • Low overclocking potential

    1 AMD Ryzen 7 1700 Summit Ridge

    The most powerful processor in its class
    Country: USA (Manufactured in Malaysia, China, China)
    Average price: 17,100 RUR
    Rating (2019): 4.8

    As expected, the processor from the top-line Ryzen 7 has the best performance in its class. Once again, we cannot help but remember the cost - for 17 thousand rubles we get power at the level of the top-end Core i7 of previous years. The processor includes eight cores, divided into two clusters. The standard clock speed is only 3.0 GHz, Ryzen 7 is guaranteed to overclock to 3.7, and with a little luck, up to 4.1 GHz.

    Like the previous representatives of the line, the leader is made using a 12 nm process technology, which allows for economical energy consumption. The situation with heat dissipation is good - in stress tests, temperatures remain at 70-75 degrees.

    Advantages:

    • High performance
    • There is an overclocking option
    • A fresh platform that will be supported for at least 4 years

    The best top processors

    3 Intel Core i7-7700K Kaby Lake

    The most popular top processor
    Average price: 29,060 ₽
    Rating (2019): 4.6

    More recently, the i7-7700K was the top processor in the Intel lineup. But technology is developing extremely quickly, and in 2018 it is difficult to recommend this particular chip for purchase. According to synthetic tests, the model clearly lags behind its competitors - in PassMark the CPU scores only 12 thousand points, which is comparable to modern mid-level processors. But these indicators are achieved on standard settings, when 4 physical cores operate at a frequency of 4.2 GHz, but the CPU can be easily overclocked to even higher frequencies, thereby increasing performance.

    Yes, the bronze medalist lags behind its competitors, but it costs at least half as much, and given its popularity, it is quite possible to find a good used processor. Also, the high prevalence and long-standing presence on the market allows you to find an affordable motherboard with the LGA1151 socket. In general, we have an excellent base for a powerful gaming system at a relatively low cost.

    Advantages:

    • Good price for this class
    • High performance
    • Great overclocking capabilities
    • High popularity

    Flaws:

    • Not entirely relevant in 2018

    2 Intel Core i9-7900X Skylake

    The most powerful processor in the Intel line
    Country: USA
    Average price: 77,370 RUR
    Rating (2019): 4.7

    Until recently, Intel's top line was the Core i7 series. But modern realities require more and more power. If you are not familiar with solutions, pay attention to the Core i9-7900X. The processor, already at a standard clock frequency, is capable of entering the TOP 10 most powerful CPUs. For example, in PassMark the model scores almost 22 thousand points - this is twice as much as the bronze medalist of the rating. At the same time, in reviews, users talk about trouble-free overclocking to 4.2-4.5 GHz with high-quality air cooling. Temperatures do not exceed 70 degrees under load.

    Such high performance is due to the use of 10 cores made using a 14 nm process technology. The model supports all the necessary modern standards and commands, which allows it to be used for any task.

    Advantages:

    • Highest performance
    • Excellent overclocking potential
    • Acceptable temperatures

    Flaws:

    • Very high cost
    • No solder under the cap.

    1 AMD Ryzen Threadripper 1950X

    The leader of the rating is crazy in everything - from the price of 65 thousand rubles to the incredible performance. In terms of power in synthetic tests, the model is slightly ahead of the previous participant. Internal structure however, it differs significantly. Threadripper uses 16 (!) cores. The clock speed is comparable to the Core i9 - 3400 MHz - but the overclocking capabilities are more modest. The “stone” operates stably at a frequency of 3.9 GHz; as rates increase, the necessary stability is lost.

    Such a large number of cores performs well in all tasks. But using a monster for games is not entirely reasonable - not all projects can reveal its potential. AMD will be useful for professional video editors, 3D designers, etc. - V professional software the increase in cores gives a noticeable increase in rendering speed.

    Advantages:

    • Relatively low price tag
    • High power
    • Excellent performance in professional programs

    Processors with integrated graphics have been fighting for a place in the sun for quite a long time and with varying success. However, initially no one imagined that graphics cores located on the same semiconductor chip with the CPU would be able to compete with discrete graphics cards. However, as semiconductor technologies improved, manufacturers learned to integrate fully-fledged graphics accelerators into processors, capable of accelerating both 3D graphics and video playback. high resolution, and video transcoding. All this became a completely natural and timely response to changes in the typical environment in which average computer users live. Three-dimensional graphics are used everywhere today, even on the Internet, and it’s impossible to ignore video content even if you want to.

    In addition, games have acquired serious importance and have become a full-fledged and popular form of mass leisure. The computer entertainment segment continues to grow at a rapid pace, but not all popular games make serious demands on the power of graphics accelerators. Network multi-user projects can also boast of wide distribution, the needs of which, at the current level of technology development, can be fully satisfied not only by traditional graphics cards, but also by integrated 3D accelerators. Therefore, the following statistics are not surprising: almost a third of personal computers currently sold do not have a discrete graphics accelerator at all. Moreover, a significant proportion of such systems are home computers purchased for entertainment.

    The power of the graphics core that can be built into the processor is limited by two factors: the size of the GPU semiconductor crystal and its heat dissipation. However, with the development of new production technologies and the introduction of modern graphic architectures, the scope of possibilities is gradually expanding. Now, with the widespread introduction of technological processes with 14-nm standards, it has become possible to combine a graphics accelerator with a central processor, occupying about 100 mm 2 on the chip. This is comparable to the area occupied by the GPUs of current discrete video cards in the “up to $100” price category. So it all comes down to this: modern processors with integrated graphics should be able to reach at least the performance level of the GeForce GT 1030.

    And these calculations don’t lie. The senior representative of the Raven Ridge family (this is the code name AMD named its new project - a Ryzen processor with an integrated graphics core of the Vega generation) promises a theoretical peak performance of 1.76 Tflops, which is comparable to the performance of not only GeForce GTX 1030, but also GeForce GTX 1050! However, you need to understand that in practice, the graphical performance of Raven Ridge, like any other processor with integrated graphics, is significantly limited by memory bandwidth. While budget discrete graphics cards get their own dedicated memory with a bandwidth of over 50-100 GB/s, integrated graphics are forced to make do with a common dual-channel memory controller shared with the processor, which usually offers significantly worse bandwidth, seasoned with higher latencies.

    In some situations, developers solve this problem by adding additional buffer memory to a processor with integrated graphics. For example, sensational Kaby Lake-G with Radeon RX Vega M graphics will contain their own 4 GB HBM2 video memory. Or another example: the most powerful Intel processors with an integrated video core that have been released to date, Skylake-R, are equipped with a 128 MB victim level 4 cache based on eDRAM.

    However, in the case of Raven Ridge, this approach does not work. Additional buffer memory entails an increase in the cost of the final product, and AMD's strategy is to use its new offerings to attack the lower market segment by offering good option for those users who assemble systems from inexpensive CPUs and budget GPUs. Therefore, Raven Ridge focuses on intensifying capabilities system memory. For the new processor with an integrated video core, AMD engineers optimized the existing DDR4 memory controller, added support for higher frequency modes and reduced latency. As a result, the company has a very interesting product that has no close analogues in its market niche.

    With the launch of the new Raven Ridge integrated processors, AMD continues its confident return to the CPU market as a full participant, which began last year. The Zen microarchitecture has already proven its viability as a foundation for performance chips, but now it should serve as the basis for low-cost mainstream integrated processors, into which AMD was able to package its best-to-date Vega graphics architecture. As AMD itself expects, with this step it will be able to easily “transplant” to its devices those users who have until now been content with discrete graphics cards with a price of less than $100. The goal is somewhat ambitious, but given the steps taken to achieve it, it is quite realistic.

    In addition, it was very fortunate that Raven Ridge came to the rescue at a very difficult time. The shortage of discrete graphics accelerators, provoked by crypto enthusiasts, is raging on the market, as a result of which you can buy even an entry-level video card today only at a noticeably inflated price. And this means that Raven Ridge can become a kind of “lifesaver” for those users who do not want to overpay exorbitantly for a video card and are either ready to be content with integrated solutions, or can afford to wait out troubled times with their help. Overall, there is a lot of interest in Raven Ridge for many reasons.

    Raven Ridge Formula: Zen + Vega

    To understand what Raven Ridge is, how AMD was able to put together two of its cutting-edge developments, and why it required almost a year of additional engineering effort, just look at what the semiconductor die of the new hybrid processors looks like. Here it is:

    You probably remember that all Ryzen processors released to date are based on a Zeppelin semiconductor crystal, which is assembled from two CCX (Core Complex) modules and the necessary wiring. Each CCX module has four computing cores with Zen microarchitecture and an 8 MB shared third-level cache. The modules are connected to each other and to “extra-core” controllers via a special Infinity Fabric bus, which is an improved version of HyperTransport. Thus, all Ryzens without integrated graphics, regardless of how many computing cores are available to the user, are based on a single eight-core chip with an area of ​​about 218 mm 2, which includes about 4.8 billion transistors.

    It is clear that it is difficult from a production point of view to further expand such a large crystal with a graphics core. Therefore, in order to release Raven Ridge, AMD engineers had to design a different crystal based on cores with Zen microarchitecture. In it, the graphics core took the place of the second quad-core CCX module. As a result, the Raven Ridge crystal area remained almost the same - it is 210 mm 2, and the number of transistors has increased slightly - to 4.94 billion.

    It took a lot of blood to get Raven Ridge into such a framework. AMD engineers intended to combine a fairly powerful version of the Vega graphics core with Zen computing cores. The company's past APUs, known under the code name Bristol Ridge, were equipped with an integrated graphics core with GCN 1.3 architecture (it, for example, was also used in the R9 Fury graphics cards) and in the maximum versions had a set of 512 stream processors. In Raven Ridge, which was initially positioned by AMD as products of a fundamentally different level, the power had to increase by a noticeable amount, so a very large GPU with 11 computing units (CUs) was inserted into the new semiconductor crystal, which in total corresponds to an array of 704 stream processors (SP).

    As a result, it was not possible to leave one old CCX borrowed from Zeppelin untouched in Raven Ridge, providing the integrated processor with four computing cores and an 8-MB L3 cache. In pursuit of cost reduction, engineers had to cut it somewhat. As a result, the volume of third-level cache memory located in the Raven Ridge CCX module was halved - to 4 MB. True, its associativity did not change, which means that one should not count on a significant change in the speed characteristics of the L3 cache.

    Nevertheless, the fourfold reduction in the total volume of third-level cache compared to the “large Ryzen” still had an effect on its performance: latencies decreased slightly. Below, all this is demonstrated in the graphs, which show practically measured latencies of the memory subsystem of the quad-core Raven Ridge and the quad-core Ryzen 5 1500X processor, reduced to a single clock frequency of 3.8 GHz.

    L3 cache latency in Raven Ridge decreased by approximately 5 cycles. They turned out to be won back thanks to the simplification of the operating algorithms, which now do without supporting the coherence of parts of the cache memory located in different CCXs.

    Along the way, another interesting detail is revealed: the second-level cache also received a noticeable acceleration in Raven Ridge. Its latency dropped from 17 to 13 cycles, although the manufacturer did not advertise this change anywhere.

    Pointing to a change in the cache subsystem, AMD promises that reducing the size of the L3 cache in new processors should not adversely affect performance. The negative vector is compensated not only by the reduction in latencies, but also by the fact that Raven Ridge does not have to suffer from relatively slow inter-core connections between CCX, made due to the Infinity Fabric bus operating at the same frequency as the memory controller. Indeed, in the new processor design there is only one CCX module, and this internal bus connects it with the graphics core and other “extra-core” components, but does not in any way affect the exchange of data between the computing cores.

    This can be clearly seen if we compare practically measured delays during inter-core data exchange between Raven Ridge and Ryzen 5 1500X. Here Raven Ridge wins noticeably - for a quad-core processor, a design with one CCX looks more optimal.

    In addition to improvements to the cache system, Raven Ridge's memory controller has also been optimized. Firstly, it added official compatibility with DDR4-2933 modules, making Raven Ridge the first processor on the market to support such a fast JEDEC specification. Secondly, all other things being equal, Raven Ridge works with memory more efficiently than previous Ryzens. Tests indicate a decrease in latency that is not too dramatic, but still visible to the naked eye.

    True, a decrease in practical throughput can also be seen here, but this effect should rather be attributed to “dampness” Motherboard BIOS plat. After the release of Raven Ridge, motherboard manufacturers are again actively updating firmware, and new BIOS version really bring additional improvements to the performance of the Raven Ridge memory controller.

    Thus, the total changes in the Raven Ridge memory subsystem are varied, and the reduced L3 cache is unlikely to become a serious drawback of these processors. But he was not the only one to undergo resection at Raven Ridge. Another unit was also seriously cut down - the PCI Express graphics bus controller built into the processor. For connecting an external graphics card in Raven Ridge processors, a full-fledged PCI interface Express 3.0 x16 is not supported: instead it is proposed to use the truncated PCI Express 3.0 x8 bus. However, in the case of graphics cards not of the highest level this limitation It's unlikely to have any impact on performance, and the only thing worth keeping in mind is that Raven Ridge is not compatible with multi-GPU configurations.

    Raven Ridge also does not work with Dual Graphics technology, which was supported in previous generations of AMD APUs. It is impossible to “pair” the built-in Vega graphics core with an external video card of the same architecture into a single multi-GPU array directly using the graphics driver. However, the joint operation of integrated graphics and an external video card is still possible through mGPU technology, which is part of DirectX 12. In other words, the built-in Vega can still “help” an external accelerator, and it does not matter at all what kind of discrete video card is used, but it will work like this the bundle will be exclusively in DirectX 12.

    Ryzen 2000G Family: Ryzen 5 2400G and Ryzen 3 2200G

    AMD has released two variants of Raver Ridge for desktop systems. Both are based on the same design and are manufactured at GlobalFoundries using the 14nm (14LPP) process that is also used on the familiar Ryzen processors without integrated graphics. This means that, although the new hybrid products received model numbers from the 2000 series, the more advanced 12-nm process technology is not used for their production and they have nothing in common with the promising Zen+ generation processors, which are scheduled for release in April.

    The older desktop Raven Ridge is a quad-core Ryzen 5 2400G processor costing $169 with support for SMT technology and an integrated Vega 11 graphics core. Its younger brother, the Ryzen 3 2200G, is also a quad-core processor, but without SMT support and with a weaker Vega 8 graphics core. Read more The characteristics of the new processors can be found in the table, where we placed them next to the “classic” quad-core Ryzen 5 and Ryzen 3.

    Ryzen 5 2400G Ryzen 5 1500X Ryzen 5 1400 Ryzen 3 2200G Ryzen 3 1300X Ryzen 3 1200
    Codename Raven Ridge Summit Ridge Summit Ridge Raven Ridge Summit Ridge Summit Ridge
    Production technology, nm 14 14 14 14 14 14
    Cores/threads 4/8 4/8 4/8 4/4 4/4 4/4
    Base frequency, GHz 3,6 3,5 3,2 3,5 3,5 3,1
    Frequency in turbo mode, GHz 3,9 3,7 3,4 3,7 3,7 3,4
    XFR frequency, GHz - 3,9 3,45 - 3,9 3,45
    Overclocking Eat Eat Eat Eat Eat Eat
    L3 cache, MB 4 2×8 2×4 4 2×4 2×4
    Memory support DDR4-2933 DDR4-2666 DDR4-2666 DDR4-2933 DDR4-2666 DDR4-2666
    Integrated Graphics Vega 11 No No Vega 8 No No
    Number of stream processors 704 - - 512 - -
    Graphics core frequency, GHz 1,25 - - 1,1 - -
    PCI Express lanes 8 16 16 8 16 16
    TDP, W 65 65 65 65 65 65
    Socket Socket AM4 Socket AM4 Socket AM4 Socket AM4 Socket AM4 Socket AM4
    Official price $169 $174 $169 $99 $129 $109

    If we remember that Raven Ridge is based on a semiconductor chip with one CCX module, then it is absolutely clear that we cannot expect more powerful APU models from AMD in the foreseeable future. No Ryzen 7 with integrated graphics is simply possible. The Ryzen 5 2400G fully reveals the capabilities inherent in the developed design. This processor utilizes all four processor cores and SMT multi-threading technology, as well as the full set of 11 Compute Units (CUs) found in the Vega accelerator's embedded implementation. It is worth noting that as a result, the Ryzen 5 2400G turned out to be even more powerful than the mobile Ryzen 7 2700U, in which the graphics core operates only 10 out of 11 computing units.

    The set of 11 CUs available in the Ryzen 5 2400G is translated into 704 stream processors, which in quantitative terms is 38 percent greater than the arsenal possessed by solutions from the Kaveri, Carrizo and Bristol Ridge generations. This is accompanied by an approximately 13% increase in graphics frequency, an increased number of texturing units (from 32 to 44) and rasterization units (from 8 to 16), as well as a new generation of architecture. Vega belongs to the latest, fifth generation of GCN, while previously embedded video cores had a third-generation architecture. All this together should provide a significant superiority of the older new product over its predecessors in terms of performance.

    However, here it would be appropriate to recall again the existence of Kaby Lake-G with Radeon RX Vega M graphics. Raven Ridge, obviously, cannot compete with them in any of its manifestations. Due to the fact that in the Intel version of processors with Vega graphics, the video core is located on a separate semiconductor chip, it is much more powerful - it houses 24 computing units and 1536 stream processors. In addition, do not forget about the separate 4 GB HBM2 memory, which Intel also managed to fit into the processor package. Therefore, the scope of application for Ryzen and Kaby Lake-G with Vega graphics will be different. Intel's version is a premium and expensive product for laptops and ultra-compact desktop systems of the NUC class, while AMD is aiming for the mass segment.

    This is why it is noteworthy that the Ryzen 5 2400G received a recommended price of $169: this allows this processor become a direct and improved alternative to the Ryzen 5 1400. Obviously, the old version without graphics will now gradually leave the shelves, because the Ryzen 5 2400G is superior to the Ryzen 5 1400 in many ways basic parameters. In addition to the presence of a built-in GPU, it has a higher clock frequency (3.6 GHz versus 3.2 GHz - base and 3.9 GHz versus 3.4 GHz - turbo), there is support for faster DDR4-2933 memory and the situation with inter-core is much better interaction. In fact, the Ryzen 5 1400 may be more interesting only due to the more capacious L3 cache, but it is worth recalling that in this model it is also cut down from 16 to 8 MB. Thus, in the vast majority of scenarios, the Ryzen 5 2400G will be faster when used with an external graphics card.

    The Ryzen 3 2200G looks no worse than the $169 Ryzen 5 2400G in its niche. From the point of view basic characteristics This processor is a typical Ryzen 3: it has four processing cores without SMT and has a nominal frequency of 3.5 GHz with the ability to auto-overclock to 3.7 GHz. But to all this is added a relatively powerful Vega 8 graphics core, and the price is set at $99, which makes this proposal not only an attractive hybrid APU, but also the cheapest Ryzen in general. That is, even if we forget about the presence of good graphics in the Ryzen 3 2200G, it is unique in that it offers four powerful x86 cores at a price below $100. There are simply no other offers of similar generosity at the moment.

    As for the Vega 8 accelerator built into the Ryzen 3 2200G, this version of the GPU offers 512 stream processors, that is, it is at least not inferior to the graphics from APUs of previous generations, which AMD sold under the names A10 and A12 at a price significantly exceeding 100- dollar level.

    Despite the fact that Ryzen processors with Vega graphics received fairly high clock speeds, AMD managed to keep their heat dissipation within reasonable limits. The typical TDP of the Ryzen 5 2400G and Ryzen 3 2200G is 65 W, which is a big achievement compared to the fact that the company's fastest desktop APUs could previously have a TDP of 95 W. And even more than that, in Raven Ridge, when there is a simultaneous load on the computing and graphics parts of the processor, the frequency of both types of cores is not reset below the nominal values, as was common in APUs of previous generations. Even the older Ryzen 5 2400G can remain within the stated thermal package without any tricks.

    Separately, it should be mentioned that the management clock frequencies Raven Ridge uses updated Precision Boost 2 technology. It implements an improved and more aggressive algorithm, thanks to which the turbo mode in new processors with an integrated graphics core is turned on more often than before. In addition, when some cores are not fully loaded, intermediate frequencies between the base and maximum values ​​are more actively used. In other words, adjustment to a specific load in the Ryzen 5 2400G and Ryzen 3 2200G has become more sensitive than before.

    However, XFR technology, which made it possible to further increase the frequency in the case when the processor was operated in a favorable temperature conditions, is missing from Raven Ridge.

    You can install new processors from the Raven Ridge family on the same Socket AM4 motherboards that run other Ryzen motherboards. The only limitation is that compatible boards must use an updated BIOS: Raven Ridge requires versions built using AGESA 1.0.7.1 or later libraries. In other words, new CPUs with integrated graphics do not require any additional costs. They come to an already existing and widely used platform.

    Speaking about how attractive the combination of price and performance is with the new Raven Ridge desktops, we cannot ignore the fact that the boxed versions of the Ryzen 5 2400G and Ryzen 3 2200G come with a complete Wraith Stealth cooler, the cost of which is also included in the announced $169 and $99 .

    Of course, such a cooler is not related to highly efficient cooling solutions, but it will definitely cope with heat removal from 65-watt processors and will allow you to save an additional couple of tens of dollars when building a system on Raven Ridge. And even more than that, the capabilities of this cooler are probably enough for moderate overclocking.