Comparison of laptop processors (AMD and Intel). AMD Mobile Processors New Ryzen Laptops

Now smartphones, if desired, can process a mountain of information. The power of their processor is enough to solve absolutely any tasks. At the same time, modern chipsets consume a minimum amount of electricity, for which the improved technical process should be thanked. Our rating of processors for smartphones will tell you about the most powerful and interesting models. Devices based on them can be blamed for anything, but definitely not for lack of power!

Good to know!

Samsung Exynos 9820

• Year of issue: 2019
• Process technology: 8 nm
• Architecture: 2*Custom + 2*Cortex A75+ 4*Cortex A55
• Video accelerator: Mali-G76 MP12

Geekbench result: 4382/9570 points

An interesting fact was the appearance of the chipset from Samsung in third place in the most powerful processors. Previously, the company's solutions in synthetic tests were inferior not only to Qualcomm, but also to Huawei, but in this case, the numbers for the Koreans turned out to be higher.

The novelty is built on an 8-nanometer process technology, the cores are divided into 3 groups - two fourth-generation proprietary cores, two productive Cortex A75 and four energy-efficient Cortex A55. Like other manufacturers, the company focused on improving the performance of neural networks, the security of user data, as well as improving camera performance and supporting a large number of sensors. So the novelty can work with 5 cameras, including infrared for face scanning and a resolution of no more than 22 megapixels or two front cameras 16 Mp. In addition, the chipset supports 4K displays, shooting in 8K at 30 fps, 4K up to 120 fps.

Advantages:

• Instant camera settings for AR and VR.
• Support shooting in 8K.
• Work with 5 cameras.
• Economical battery consumption.
• Improved encoding of user personal data.
• 4K video processing up to 150 fps.
• Support for 4K displays.
• Working with UFS 2.1/3.0 SSDs.

Flaws:

• 8 nm manufacturing process - even with the most ideal optimization, these chipsets cannot compete with 7 nm models in energy efficiency.
• Of the 8 cores, only the first two can be called new, the remaining 6 cores migrated from previous processors without any changes, that is, it is difficult to fully consider the novelty as such - rather, it is an improved old processor.

Samsung S10, S10+, S10e

Huawei Kirin 980

• Year of issue: 2018
• Process technology: 7 nm
• Architecture: 2*Cortex-A76 + 2*Cortex-A76 + 4*Cortex-A55
• Video accelerator: Mali-G76 MP10

Geekbench result: 3390/10318 points

The new processor from Huawei was technically the first to be created using the 7 nm process technology, however, by introducing it to the world, the company was late with the release, giving the laurels to Apple. The cores are divided into three groups - two high-performance, two energy-efficient and four medium-performance. In order to use exactly the cores that are needed in a given task, the company introduced Flex-Scheduling technology, thanks to which performance increased by 37% compared to last year's Kirin 970. Not without a new graphics accelerator, which is not only more powerful, but also automatically increases the clock speed in games.

Just like Huawei's competitors improved the machine block, it is 120% faster than its predecessor. In this case, these are not just words. Every year, all manufacturers focus on cameras and the fact that through the work of AI, pictures really turn out better in any conditions is difficult to dispute. This is best seen in Huawei. Already with last year's Kirin 970, the company achieved first place in the most prestigious camera rating from DxOMark. With the new processor, we should expect to overcome the previous bar. A lot has been written about the photo capabilities of the chipset, or rather its neural unit. To simplify, Huawei does the following - it does not chase top performance in tests, but it makes really excellent chipsets for photos and videos, and again there will be no difficulties with launching games and any other tasks. In many cases, this is what users want, who look at real possibilities, not numbers.

Advantages:

• The best processor for photos and videos.
• Two modules of neural networks.
• Support for high-speed LPDDR4X RAM up to 16GB.
• Flex-Scheduling is a technology of "correct" selection of cores for specific tasks, which results in excellent energy efficiency and fast loading of any applications.
• HDR10+ support.
• Support for the new standard - Wi-Fi 802.11ay.
• Work with 48 MP cameras or dual 22 MP modules.
• 4K recording at 60 fps.

Flaws:

• The graphics coprocessor is weaker than that of competitors - this is not a minus for users, since there is GPU Turbo technology for automatic overclocking which makes up for the difference.
• The company used the "old" cores and improved them, that is, in fact, this is an updated processor, and not a fundamentally new development.
• No 5G support.

The most popular smartphones: View 20, Huawei P30, Huawei Mate 20

Mediatek Helio P90

• Year of issue: 2018
• Process technology: 12 nm
• Architecture: 2*Cortex-A75 + 6*Cortex-A55
• Video accelerator: PowerVR GM 9446

Geekbench result: 2025/6831 points

MediaTek has long been associated by users with processors for budget and occasionally mid-priced smartphones. The company is struggling to create a competitive flagship-level model, but it is not very successful. So MediaTek did not get into the top 10 most productive chipsets, but took 11th place with the Helio P90 released in 2019. The model has an eight-core structure, which, although it has a breakdown into two and six cores, but in fact they are all high-performance. Not surprisingly, the P90 was able to overtake the Snapdragon 710 described below, which has exactly the same cores, but with a six-core emphasis on energy efficiency.

In general, the new product from MediaTek is quite interesting - there is support for the fastest RAM up to and UFS 2.1 solid-state drives, like its competitors, it can work with a single 48 MP camera or a dual-module solution of 24 and 16 MP. An interesting feature is the support for displays with a resolution of 2520 * 1080 and an aspect ratio of 21:9. The chipset has three image processing units and an updated AI with support for AI Fusion is responsible for ensuring that tasks are distributed among all units, which increases the speed of data processing. One of the features of this technology is the adaptation of the screen in real time to the selected application - in particular, when making a video call and switching from full-screen video to preview, users will not notice any delays.

Advantages:

• Eight powerful cores for maximum performance.
• Updated AI block for working with photos.
• Support high-speed LPDDRX memory up to 8GB.
• Support for modern cameras up to 48 MP.
• Shooting slow motion video with fps 480 in HD format.

Flaws:

• Not the best energy efficiency.
• No shooting in 4K.
• Old generation graphics coprocessor.

The most popular smartphones: BV9800

Qualcomm Snapdragon 710

• Year of issue: 2018
• Process technology: 10 nm
• Architecture: 2*Cortex-A75 + 6*Cortex-A55
• Video accelerator: Adreno 616

Geekbench result: 1897 / 5909 points

A mid-range processor that took 12th place in the ranking of the most powerful chipsets. The model was the first in the 700th series. Prior to this, Qualcomm had a clear division: the 800 series is the flagship level with maximum features, the 600 series is the mid-level with stripped down GPU and CPU cores, and the 400 series is the budget line with a minimum of features. Processors of the 700th series, and in particular the Snapdragon 710, are all the current chips from the main line and at the same time a fairly affordable price.

The novelty runs on two high-performance cores and six power-efficient ones. Taking into account the new graphics system the model shows excellent performance in games and at the same time low power consumption. In addition, she knows how to process photos with high quality - reduces noise, supports two dual cameras up to 16 MP, and 4K video. The company did not forget about AI, in this case, the manufacturer did not waste time on trifles and supplied machine cores Hexagon 685, that is, the same as in 2018 - Snapdragon 845. The output turned out to be a fairly inexpensive chipset, which, if inferior to flagships, is quite a bit . For those who are looking for a mid-range smartphone with excellent performance, energy efficiency and photo processing, the Snapdragon 710 will be a real find.

Advantages:

• Affordable price.
• Support for two cameras up to 16 MP.
• Low power consumption.
• Powerful AI for taking photos.
• Support for 4K c 30 fps and HDR.
• Working with biometric sensors.
• Support for Quick Charge 4+.

Flaws:

The most popular smartphones:samsung galaxy A8s, 16, Xiaomi Mi8 SE

Conclusion

It is worth noting that our rating does not include Snapdragon 845 and 660, Kirin 970, Apple A11, Exynos 8895, Helio X30 chipsets due to the fact that they were all released in late 2017 or early 2018. Despite their relevance, many readers are familiar with them, and there are a lot of smartphones based on them. For this reason, we chose powerful new items, which does not entail recommendations for buying devices based solely on them. But if you want a smartphone with the latest and most powerful chipset, then the models presented above are the best of their kind.

Excluded from selection

Samsung Exynos 8 Octa 8890

• Year of issue: 2016
• Process technology: 14 nm
• Architecture: Samsung Exynos M1 + ARM Cortex-A53 (ARMv8-A)
• Video accelerator: Mali-T880, 12 cores, 650 MHz

Geekbench result: 5940 points

If not best processor for a smartphone, then at least one of those who deserve this title. It is no accident that they are equipped with all variations of the South Korean Galaxy S7. Is it possible to reproach this flagship with a lack of power? The chipset easily digests 4K video at 60 fps. It consists of eight cores. The maximum frequency is 2290 MHz. But it rarely comes to raising it to such a level, since more low frequencies enough to solve most problems.

Unfortunately, the processor also has certain problems. It just so happened that South Korean chipsets are not endowed with the best video accelerator (GPU). Here, too, the Mali-T880, despite its 12 cores, works out strictly for the “good” rating, but nothing more. This is proved by tests in GFXBench, where the Samsung Exynos 8 Octa 8890 outperforms some other chipsets reviewed today in terms of graphics.

Advantages

• Video support in 2160p resolution at 60 fps;
• Not very big heating;
• Low power consumption;
• High marks in benchmarks.

Flaws

• The memory test shows not the highest results;
• The graphics accelerator could have performed better.

The most popular smartphones: Samsung Galaxy S7, Samsung Galaxy S7 Edge, Samsung Galaxy Golden 4

Qualcomm Snapdragon 820 MSM8996

• Year of issue: 2015
• Process technology: 14nm FinFET
• Architecture: Qualcomm Kryo
• Video accelerator: Adreno 530, 624 MHz

Geekbench result: 4890 points

Qualcomm does not have its own manufacturing facilities. However, it has many patents at its disposal. And with them, it is not difficult to develop a processor close to the ideal, after which it remains only to place an order for production from other companies. pleases with both computing power and graphics processing capabilities. Many flagships that were born in 2016 were equipped with this chipset. And none of their customers complained about the graphics in mobile games!

The chip consists of only four cores. However, this did not prevent him from scoring record scores in benchmarks - not least thanks to the graphics accelerator. The maximum frequency of this processor is 2150 MHz. At the hardware level, the chipset supports HDMI 2.0, USB 3.0, and Bluetooth 4.1. In a word, the processor could easily cope even with the tasks assigned to a laptop! It also features support for a camera with a resolution of up to 28 megapixels - that is why the company made its choice in favor of this processor, in flagship smartphones which such a sensor is present.

Advantages

• Very high resolution camera support;
• Capable of processing Full HD video at up to 240 fps;
• Support for 10-bit 4K video;
• Windows devices use DirectX 11.2;
• Very high clock frequency;
• Not very high energy consumption;
• High scores in benchmarks;
• The memory test produces high results;
• Excellent performance in games.

Flaws

• Sometimes it gets quite hot.

The most popular smartphones: Moto Z Force, Elite X3, ZenFone 3, 10, Samsung Galaxy S7, Samsung Galaxy S7 Edge, Sony Xperia X Performance, Sony Xperia XR, Xiaomi Mi5 Pro, Z11

HiSilicon Kirin 95

• Release year: 2016
• Process technology: 16 nm
• Architecture:
• Video accelerator: Mali-T880, 4 cores

Geekbench result: 6000 points

This chipset is made on a 16-nanometer process technology, which indicates its decent energy efficiency. The maximum frequency here is increased to 2.5 GHz. The creators had to take such a step because of the Mali-T880 graphics accelerator, which does not cope with its task in the best way.

The Chinese chipset consists of eight cores, four of which can be called auxiliary. Paired with a GPU, it is capable of playing 4K video at 60 fps. But only to reproduce - the processor is capable of creating a video recording on its own only in 1080p resolution. And this despite the fact that the chip even supports dual cameras, the total resolution of which is 42 megapixels. It is also capable of recognizing Bluetooth 4.2 and USB 3.0 modules.

Advantages

• Support for many modern wireless technologies;
• Almost record clock speed;
• No big problems with overheating;
• Can decode 4K video at 60fps;
• Supports dual high definition cameras.

Flaws

• The graphics accelerator shows poor results.

The most popular smartphones: Huawei P9, Huawei P9 Plus, Huawei Honor V8, Huawei Honor Note 8.

HiSilicon Kirin 950

• Year of issue: 2015
• Process technology: 16 nm
• Architecture: 4x ARM Cortex-A72 + 4x ARM Cortex-A53
• Video accelerator: Mali-T880, 4 cores, 900 MHz

Geekbench result: 5950 points

In 2015-2016, this processor was used by many Huawei smartphones. The chipset consists of eight cores, the power of four of them can reach 2300 MHz. It would seem that the result is quite good. But not everything is so clear. The weak point of the chip lies in the graphics accelerator. The first version of the Mali-T880 is used here as it. It copes with video decoding with dignity - in theory, you can even run 4K video at 60 frames / s. But in games, this GPU performs disgustingly, especially by the standards of flagships.

However, the computing power of this chipset cannot be faulted, which is why it got into our top processors. The product supports Bluetooth 4.2 and USB 3.0 standards, although the Chinese giant did not really produce smartphones with such high-speed interfaces, preferring to save money. Also, in theory, the processor copes with the data stream from, which has a total resolution of 42 megapixels.

Advantages

• Supports USB 3.0 and Bluetooth 4.2;
• High computing power;
• Support for modern memory formats;
• Not very expensive to manufacture;
• Decodes high definition video;
• Capable of handling a dual 42-megapixel camera.

Flaws

• The graphics accelerator could be much better;
• Cannot provide the camera with 4K video recording.

The most popular smartphones: Huawei Honor 8, Huawei Honor Note 8, Huawei Mate 8, Huawei Honor V8.

Apple A9X APL1021

• Year of issue: 2015
• Process technology: 16 nm
• Architecture: Apple Twister 64-bit ARMv8-compatible
• Video accelerator: PowerVR Series 7X 12 cores

Result in Geekbench : 5400 points

Why game developers are primarily focused on smartphones and Apple tablets? Really only their owners can afford to buy a toy? No, it's much easier. It is on this technique that the game shows itself best. The Apple A9X APL1021 processor is endowed with an almost perfect graphics accelerator that can handle absolutely any task! If desired, Apple could even implement 4K video recording at 60 fps!

As for computing power, everything is in order with it, although the processor still does not gain record points in benchmarks. It would seem that only two cores are used here. But this is enough for everyday tasks. Not least because of the better optimized operating system.

Advantages

• High power of two cores;
• Excellent 12-core graphics accelerator;
• Full support for 4K video at 60 fps;
• Support for many modern technologies;
• Recognizes modern memory formats.

Flaws

Apple iPad Pro

MediaTek MT6797 Helio X25

• Year of issue: 2016
• Process technology: 20 nm
• Architecture: 2x ARM Cortex-A72 + 4x ARM Coptex-A53 + 4x ARM Coptex-A53
• Video accelerator: Mali-T880MP4, 4 cores, 850 MHz

Geekbench result: 4920 points

A processor with a rather complex structure. It consists of ten nuclei belonging to two varieties. The two cores are the most powerful - they belong to the Cortex-A72 type, and their clock speed can reach 2500 MHz. The remaining computing cores belong to the Cortex-A53 type. At the same time, half of them are overclocked to a frequency of 2000 MHz, while the frequency of the rest is limited to 1550 MHz.

All this allows the processor to score a lot of points in benchmarks. And the result would be even higher if not for the graphics accelerator. This element here is seriously limited in its capabilities. Yes, it supports full 4K video work, including its creation, but only at 30 fps. And in games, the GPU copes with its task even worse. As for the rest of the characteristics, we should highlight the support for 32-megapixel cameras and Bluetooth standard 4.1. The maximum display resolution of a smartphone with this chipset can reach 2560 x 1600 pixels.

Advantages

• 32MP camera support;
• Very high computing power;
• Relatively low power consumption;
• Albeit limited, but support for 4K video;
• Low cost chipset.

Flaws

• The GPU performs poorly in games;
• No Bluetooth 4.2 support.

The most popular smartphones: Meizu Pro 6, K6000 Premium, Xiaomi Redmi Pro, Speed ​​8, Apollo.

Qualcomm Snapdragon 625 MSM8953

• Year of issue: 2016
• Process technology: 14 nm
• Architecture: ARM Cortex-A53 (ARMv8)
• Video accelerator: Adreno 506

Geekbench result: 4900 points

One of Qualcomm's most popular creations. They are endowed with a huge number of smartphones from the mid-budget and even top segments. The manufacturer did not bother with the architecture, endowing the chipset with eight identical cores. The maximum clock frequency is 2000 MHz, which ordinary user quite enough.

The graphics accelerator here is optimized for video content processing. Theoretically, a smartphone based on this processor is capable of playing and recording 4K video at 60 frames / s. But in games, some problems begin. Although their presence is surprising, because the GPU even has support for DirectX 12, which is activated on devices with Windows on board. The chipset also supports a dual camera, the total resolution of which does not exceed 24 megapixels. The only thing missing here is USB 3.0 support. However, the creators of smartphones do not like to build such high-speed connectors into their creations.

Advantages

• Dual camera supported;
• Well implemented fast charging technology;
• High power of all eight cores;
• Full support for 4K video content at 60 fps;
• Relatively low cost.

Flaws

• Camera resolution cannot exceed 24 MP;
• No support for Bluetooth 4.2;
• Display resolution cannot exceed 1920 x 1200 dots;
• In games, the chipset does not perform well.

The most popular smartphones: Huawei G9 Plus, ASUS ZenFone 3, Fujitsu Easy, Huawei Maimang 5, Vibe P2, Motorola Moto Z Play, Samsung Galaxy C7.

Qualcomm Snapdragon 620 APQ8076

• Year of issue: 2016
• Process technology: 28 nm
• Architecture: 4x ARM Cortex-A72 + 4x ARM Cortex-A53
• Video accelerator: Adreno 510

Geekbench result: 4886 points

This chipset is also known as the Snapdragon 652. This is one of the last processors that is still being manufactured on the 28nm process. The creators are not at all embarrassed by the relatively large size of the chip, since it is mainly built into tablets.

The processor consists of eight processing cores. The clock frequency of four of them can reach 1800 MHz. This is quite enough for the tablet to solve the main tasks without any thought. The chipset also includes the Adreno 510 graphics accelerator. There are no particular complaints about it, because no one will expect excellent graphics performance from a tablet. It should be noted that theoretically the chip supports video in 2160p resolution at 30 fps. It also boasts support for Bluetooth 4.1 and proprietary Quick Charge 3.0 fast charging technology.

Advantages

• Supports devices with large screen resolution;
• Great computing power;
• Albeit limited, but still support for 4K video;
• Built-in fast charging technology.

Flaws

• No support for Bluetooth 4.2;
• Still not the best graphics accelerator.

Most popular devices: Samsung Galaxy Tab S2 Plus 8.0, Samsung Galaxy Tab S2 Plus 9.7.

MediaTek MT6797M Helio X20

• Year of issue: 2016
• Process technology: 20 nm
• Architecture: 2x ARM Cortex-A72 + 4x ARM Cortex-A53 + 4x ARM Cortex-A53
• Video accelerator: Mali-T880MP4, 4 cores, 780 MHz

Geekbench result: 5130 points

Many mobile processors have four or even eight cores. In the case of the MediaTek MT6797M Helio X20, their number has been increased to ten. As a result, the performance of the chipset is very high. Especially in those applications where serious graphics processing is not required. It should be noted that only two computing cores are especially powerful here - their clock frequency reaches 2300 MHz. The remaining nuclei are divided into two groups. One is able to please with a frequency of 1850 MHz, while the other has this parameter fixed at 1400 MHz. But in any case, the result is very good, which is confirmed by synthetic tests, and by the smartphones themselves - the interface on them does not slow down at all thanks to the chipset.

As for the graphics accelerator, everything is much worse here. Theoretically, it copes with viewing and recording 4K video at 30 fps. But in games, the lack of power is immediately felt. Modern games on a smartphone with such a processor will work, but with simplified graphics. Especially if the device has a screen with Full HD resolution or higher. It should also be noted that the processor supports almost any mobile camera - as long as the resolution of the module does not exceed 32 megapixels.

• Year of issue: 2015
• Process technology: 28 nm
• Architecture: ARM Cortex-A72 + ARM Cortex-A53 (ARMv8)
• Video accelerator: Adreno 510

Geekbench result: 4610 points

There are two versions of the Qualcomm Snapdragon 620 processor, also known as the Snapdragon 652. The first is the MSM8976, which was released in 2015. A year later, a slightly more advanced version was released - APQ8076, which some received. The products are practically indistinguishable from each other. They have eight cores, half of which are capable of boosting the frequency up to 1800 MHz. Both processors are endowed with a far from ideal Adreno 510 graphics accelerator.

Qualcomm's creation is capable of supporting smartphones with a display resolution of no higher than 2560 x 1600 pixels. As for the camera, it is possible to process data coming from a dual module, the total resolution of which does not exceed 21 megapixels. Everything is in order with the module and with the ability to process data coming from dual-channel LPDDR3 memory.

Advantages

• High performance;
• View 4K video at 30 fps;
• The theoretical ability to record video in 1080p and 120 fps;
• Not very high cost;
• Support dual chambers;
• The screen resolution can reach 2560 x 1600 pixels.

Flaws

• Bluetooth 4.2 not supported;
• The maximum resolution of the camera cannot be very high.

The most popular smartphones: X6S A, Vivo X7, Vivo X7 Plus, LeEco Le2, G5 SE, R9 Plus, Samsung Galaxy A9 Pro (2016), ZTE Nubia Z11 Max, Xiaomi Mi Max

Sergey Pakhomov

Notebook sales have long surpassed desktop sales, and most home users today are focused on laptops. The retail network offers a huge variety of laptop models on both Intel and AMD platforms. On the one hand, such an abundance pleases the eye, and on the other hand, there is a problem of choice. As you know, computer performance is largely determined by the processor installed in it, but to understand modern families and legend processors is not so easy. And if everything is more or less clear with the designations of mobile processors from Intel, then AMD has a complete mess with this. Actually, it was this circumstance that prompted us to compile a kind of guide to AMD mobile processors.

The range of AMD processors for laptops is more than diverse (see table). However, if we talk about modern processors, which it makes sense to focus on, then we can limit ourselves to considering only 45-nm processors of the Phenom II, Athlon II, Turion II, V-series, Sempron families with the following core code names: Champlain, Geneva and Caspian.

Processors codenamed Champlain were announced by the company as recently as May 2010, while 45nm processors codenamed Caspian were announced in September 2009.

The AMD mobile processor family includes both quad-core and triple-, dual- and single-core models.

Each processor core has a 128 KB L1 cache, which is divided into a dual-channel 64-kilobyte data cache and a dual-channel 64-kilobyte instruction cache. In addition, each processor core has a dedicated L2 cache of 512 KB or 1 MB.

But cache memory of the third level (L3) is mobile AMD processors deprived (unlike their desktop counterparts).

All AMD mobile processors feature AMD 64 technology (support for 64-bit computing). In addition, all AMD processors are equipped with MMX, SSE, SSE2, SSE3 and Extended 3DNow! instruction sets, Cool'n'Quiet power saving technologies, NX Bit virus protection and AMD Virtualization technology.

So, let's look at the families of modern AMD mobile processors in more detail. And we will start, of course, by considering a family of quad-core processors AMD Phenom II.

AMD's mobile quad-core processor family is the 900 series of Phenom II processors.

All 900-series Phenom II processors have a 2 MB L2 cache (512 KB per processor core) and an integrated DDR3 memory controller. In addition, all of these processors use 128-bit FPUs. The differences between the quad-core Phenom II 900-series processors are clock speed, power consumption, and supported memory. For its processors, AMD specifies another rather strange and, in our opinion, absolutely illogical characteristic - Maximum processor-to-system bandwidth (MAX CPU BW). We are talking about the total bandwidth of all buses between the processor and the system, or rather, the total bandwidth of the HyperTransport (HT) bus and the memory bus. If, for example, the processor works with DDR3-1333 memory, then throughput the memory bus is 21.2 GB/s (in dual channel mode). Further, if the bandwidth of the HyperTransport (HT) bus is 3600 GT/s, which corresponds to a bandwidth of 14.4 GB/s, then we get that the total bandwidth of the HyperTransport bus and the memory bus will be 35.7 GB/s. Of course, it would be more logical to indicate in the processor specification the maximum memory frequency that the processor supports, but ... that is, that is. Fortunately, knowing the bandwidth of the HyperTransport bus and such a parameter as MAX CPU BW allows you to unambiguously determine the maximum memory frequency supported by the processor.

So, back to the Phenom II 900-series quad-core processor family. This family is headed by Phenom II X920 Black Edition (BE) with an unlocked multiplier. This processor has the highest clock speed (2.3 GHz) in AMD's quad-core mobile processor family and is the hottest processor with a power consumption of 45 watts. The HyperTransport bus bandwidth is 3600 GT/s and the MAX CPU BW setting is 35.7 GB/s. As you can easily calculate, this means that the integrated DDR3 memory controller supports memory with a maximum frequency of 1333 MHz (in dual-channel mode).

Two more models of quad-core AMD mobile processors are Phenom II N930 and Phenom II P920. The Phenom II N930 has a 2GHz clock speed and 35W power consumption, while the Phenom II P920 model has a 1.6GHz clock speed and 25W power consumption. Both processor models have a HyperTransport bus bandwidth of 3600 GT/s, however the Phenom II N930 processor supports DDR3-1333 memory while the Phenom II P920 processor only supports DDR3-1066 memory.

AMD's tri-core mobile processor family is the 800 series of Phenom II processors. There are only two tri-core mobile processors available today: the Phenom II N830 and Phenom II P820, both with 1536KB L2 cache (512KB per processor core) and an integrated DDR3 memory controller. The difference between these models lies in the clock speed, power consumption, and the maximum frequency of supported DDR3 memory. Thus, the Phenom II N830 processor operates at a clock frequency of 2.1 GHz with a power consumption of 35 W, and the maximum frequency of DDR3 memory supported by the processor is 1333 MHz. The Phenom II P820 processor runs at 1.8GHz with 25W power consumption and supports DDR3-1066 memory.

In passing, we note that if the letter “P” is present in the marking of AMD processors, then this means that the power consumption of the processor is 25 watts. The presence of the letter "N" indicates the power consumption of the processor at 35 watts, and the letters "X" indicate 45 watts.

The Phenom II dual-core processor family is the 600 series. Two models are presented in this series today: Phenom II X620 BE and Phenom II N620. Both have 2 MB L2 cache (1 MB per core) and 3600 GT/s HT bus bandwidth. At the same time, both processor models support DDR3-1333 memory (MAX CPU BW is 35.7 GB / s). The difference between the processors is that the Phenom II X620 BE has a power consumption of 45W and a clock speed of 3.1GHz. In addition, this processor has an unlocked multiplier. The Phenom II N620 processor with a power consumption of 35 W has a clock frequency of 2.8 GHz.

Finishing the review of mobile processors of the Phenom II family, we note once again that it includes four, three- and two-core processors with 128-bit FPU, the power consumption of which can be 45, 35 or 25 W. All these processors have HT 3600 GT/s bus bandwidth and support DDR3 memory with a maximum frequency of 1333 or 1066 MHz. The size of the L2 cache depends on the number of processor cores and is 512 KB per processor core (for four and three-core models) or 1 MB (for dual-core models).

The next family of 45nm mobile processors based on the Champlain core is the Turion II dual-core processor family, which is represented by two models: Turion II N530 and Turion II P520. These processors differ from each other only in clock speed and power consumption. The Turion II N530 has a clock speed of 2.5GHz and power consumption of 35W, while the Turion II P520 has a clock speed of 2.3GHz and power consumption of 25W. In all other respects, the characteristics of these processors are the same. So, both models are equipped with 128-bit FPU, have a 2 MB L2 cache (1 MB per core), and the HT bus bandwidth is 3600 GT/s. In addition, both processor models support DDR3-1066 memory. Note that dual core processors The Turion II 500-series families are almost identical in their characteristics to the dual-core models of the Phenom II 600-series processors. The differences are only in the clock frequency and the maximum frequency of the supported memory. Actually, it is not very clear why these two processor models needed to be separated into a separate Turion II family, because they could be attributed to the Phenom II dual-core processor family.

AMD's next family of dual-core mobile processors based on the Champlain core is the Athlon II family, which is also represented by two models: Athlon II N330 and Athlon II P320. These processors are really very different from the dual-core Phenom II and Turion II processors. First of all, they cut the L2 cache to 1 MB (512 KB per core). In addition, these processors have 64-bit FPUs, and the HT bus bandwidth is 3200 GT/s. In addition, these processors only support DDR3-1066 memory. The differences between the Athlon II N330 and Athlon II P320 models themselves are clock speed and power consumption.

Single-core mobile processors based on the Champlain core are represented by the V-Series, which today includes only one model - the V120 with a clock speed of 2.2 GHz and a 512 KB L2 cache. This processor is endowed with 64-bit FPUs, and the bandwidth of the HT bus is 3200 GT/s. In addition, the V120 processor supports DDR3-1066 memory, and its power consumption is 25W. In general, the V120 processor is a single-core version of the Athlon II P320 processor in terms of its characteristics.

All AMD mobile processors reviewed by us are 2010 processors (they were announced by the company in May), focused on performance and universal laptops as well as laptops entry level. However, AMD's range also includes low-power processors - they are focused on ultra-thin laptops and netbooks. Also announced in May, these dual-core and single-core 45nm processors are codenamed Geneva and feature the Turion II Neo, Athlon II Neo, and V-Series.

Dual-core processors of the Turion II Neo series (Turion II Neo K665, Turion II Neo K625) have a power consumption of 15 W, dual-core and single-core processors of the Athlon II Neo series (Athlon II Neo K325, Athlon II Neo K125) have a power consumption of 12 W, but the power consumption of a single-core processor V105 is only 9 watts.

Turion II Neo dual-core processors feature 128-bit FPUs and 2MB L2 cache (1MB per core). The bandwidth of the HT bus is 3200 GT/s.

The Athlon II Neo series processors have 64-bit FPUs and 1 MB L2 cache per core, while the HT bus bandwidth is 2000 GT/s. Well, the single-core processor V105 differs (except for the clock frequency) from the single-core processor Athlon II Neo K125 by the L2 cache cut in half.

Note that all Geneva processors support DDR3-1066 memory in dual-channel mode.

In addition to Champlain and Geneva mobile processors, AMD also has other mobile 45nm processors in its product range. We are talking about processors codenamed Caspian, which were announced in September 2009 and have not yet become obsolete. Mobile processors Caspian is represented by the Turion II and Turion II Ultra dual-core processor families, the Athlon II dual-core processor family, and the Sempron single-core processor family.

All dual-core Caspian processors have a power consumption of 35W, and single-core processors have a power consumption of 25W. In addition, all Caspian processors only support DDR2-800 memory (in dual-channel mode).

The Turion II and Turion II Ultra processor families are equipped with 128-bit FPUs, and the HT bus bandwidth is 3600 GT/s. The difference between Turion II Ultra and Turion II processors is that Turion II Ultra processors have a 2 MB L2 cache (1 MB per core), while Turion II processors have 1 MB cache (512 KB per core).

The processors of the Athlon II and Sempron families have 64-bit FPUs and 512 KB L2 cache per core. In addition, the HT bus bandwidth for these processors is 3200 GT/s.

This article will compare laptop processors from two leading semiconductor manufacturers - Intel and AMD. The products of the first of them are equipped with an improved processor part and have a higher level of performance in this regard. In turn, AMD solutions boast a more productive graphics subsystem.

Division into niches

Comparison and Intel for laptops will be most optimally performed in three niches:

• Budget-class processors (they are also the most affordable).
• Mid-range CPUs that combine both a high level of performance and acceptable energy efficiency.
• Chips with the highest level of performance. In this case, speed, autonomy and energy efficiency fade into the background.

If in the first two cases AMD can provide a worthy alternative to Intel, then the premium segment has been dominated by the latter company for quite a long time. The only hope in this regard is new processor solutions based on the Zen architecture, which AMD should introduce next year.

Entry level Intel products

Until recently, this niche from Intel was occupied by products of the Atom line. But now the situation has changed and entry-level laptops are now based on processors. The most modest products of this class include only 2 cores, and the most advanced ones - 4. The following models are relevant for the 3rd quarter of 2016, which are shown in Table 1.

Table 1 - Current CPU models from Intel for entry-level mobile PCs.

There are essentially no cardinal differences between these CPU models. They aim to solve the most simple tasks and have a minimum level of performance. Also, this manufacturer of semiconductor solutions has a strong point in the processor part, but the integrated graphics subsystem is very weak. Another strength of these products is the high degree of energy efficiency and thus improved autonomy.

Mid-range solutions from Intel

"Kor i3" and "Kor i5" are mid-range Intel processors for laptops. Comparison of their characteristics indicates that the first family is closer to entry-level solutions, and the second - under certain circumstances, can compete with the most productive chips from this company. Detailed specifications of the specified family of products are given in Table 2.

Table 2 - Intel processor parameters for mid-range laptops.

The characteristics of the CPU of this class are almost identical. The key difference is the improved energy saving of the 7U54. As a result, autonomy in this case will also be better. Otherwise, there are no significant differences between these processors. The price of all chips of this family is the same - $281.

Premium laptop processors from Intel

For latest generation laptops, indicates that the highest performing solutions are i7 family CPUs. Moreover, in architectural terms, they practically do not differ from middle-class products. Even the models of video cards in this case are the same. But a higher level of performance compared to mid-range processors is provided by higher clock frequencies and an increased size of volatile memory of the 3rd level. The main parameters of the chips of this family are shown in Table 3.

Table 3 - Main characteristics of i7 family CPUs.

The difference between the two products is that the latter has improved energy efficiency, but the performance will eventually be lower.

AMD entry-level mobile processors

For laptops of the two leading manufacturers of this product, it indicates that Intel, as noted earlier, has a better processor part, and AMD has an integrated graphics subsystem. If an improved video system is a priority in a new laptop, then it is better to pay attention to laptops from a second manufacturer. Specific chip models technical specifications are shown in Table 4.

Table 4 - Most recent AMD processors for entry level laptops.

For the most part, these chips have almost identical technical parameters. The key difference here is only in the frequency range and the model of the integrated built-in accelerator. It is based on these parameters that you need to make a choice. If you need maximum autonomy, then choose products with lower performance. If autonomy comes to the fore, then you will have to sacrifice dynamism for this.

AMD chips for organizing mid-range laptops

FX-9XXXP and A1X-9XXXP are for laptops. Comparison of their characteristics with entry-level products indicates that they already have 4 computing units versus 2 that are available in entry-level products. Also in this case, it can compete with entry-level discrete accelerators. But the weak processor part is the factor today, which significantly reduces the performance of laptops based on these chips. Therefore, you can look in their direction only when, at the minimum cost of a mobile computer, you need the fastest graphics subsystem. The main specifications of this CPU family are shown in Table 5.

Table 5 - CPU settings from AMD for mid-range laptops.

It is most difficult to compare laptop processors in the entry-level segment. On the one hand, Intel solutions in this case have a lower cost and an improved processor part. In turn, AMD offers mobile PCs with an improved graphics subsystem. It is based on the last parameter that it is recommended to buy when choosing an entry-level laptop Pavilion 15-AW006UR from HP. Other things being equal, with competing solutions, the video card in this case will have a certain performance margin, and the processor does not lose much to the CPU from Intel. As a mid-level mobile PC, it is recommended to choose the Aspire E5 - 774 - 50SY from Acer. It has an i5 - 7200U chip installed, which is only slightly inferior to the flagship products. Yes, and other technical specifications are at an acceptable level, as for a middle-class laptop. Comparison of laptop processors in the niche of the most productive solutions indicated that it is best to buy mobile computers based on 7th generation i7 chips. The most affordable, but at the same time very equipped version of the laptop, is the IdeaPad 510-15 IKB from Lenovo. It is he who is recommended to buy when choosing the most productive mobile PC. At the same time, the price is quite democratic for such a class of devices, and the equipment is excellent.

Results

Comparison of processors for laptops of the two leading chip manufacturers today clearly and clearly indicates that the leading positions in most cases are occupied by products from Intel. AMD, in turn, is significantly behind its direct competitor. The only market segment where parity is still maintained is entry-level mobile products, where AMD has a worthy alternative. In all other cases, it would be more correct to purchase laptops based on the CPU from Intel. The current situation can be drastically changed by the release of processors based on the Zen architecture in 2017. But whether AMD can do it, time will tell. Now, in the niche of mid-range and premium mobile PCs, it is most correct to rely on solutions from Intel. Although they are somewhat overpriced, the level of performance more than compensates for this shortcoming.

AMD at a special event before CES 2018 released new mobile processors and announced desktop chips with integrated graphics. And Radeon Technologies Group, a structural subdivision of AMD, announced Vega mobile discrete graphics chips. The company also revealed plans to transition to new process technologies and future-oriented architectures: Radeon Navi graphics and Zen+, Zen 2 and Zen 3 processors.

New processors, chipset and cooling

First desktop Ryzen with Vega graphics

Two models of desktop Ryzen with integrated Vega graphics will go on sale on February 12, 2018. The 2200G is an entry-level Ryzen 3 processor, while the 2400G is a mid-range Ryzen 5 processor. Both models dynamically boost clocks by 200 and 300 MHz from base frequencies of 3.5 GHz and 3.6 GHz, respectively. In fact, they replace the ultra-budget Ryzen 3 1200 and 1400 models.

The 2200G has only 8 graphics units, while the 2400G has 3 more. The frequency of graphics cores 2200G reaches 1100 MHz, and 2400G - more than 150 MHz. Each graphic block contains 64 shaders.

The cores of both processors bear the same code name as mobile processors with integrated graphics - Raven Ridge (lit. Raven Mountain, a rock in Colorado). However, they plug into the same AMD AM4 LGA socket as all other Ryzen 3, 5 and 7 processors.

Reference: Sometimes AMD refers to processors with integrated graphics as non-CPU (Central Processing Unit, English Central processing unit), but APU (Accelerated Processor Unit, English. Accelerated processing unit, in other words, a processor with a video accelerator).
AMD desktop processors with integrated graphics are marked with a G at the end, after the first letter of the word graphics ( English graphic arts). Mobile processors and AMD and Intel are marked with the letter U at the end, according to the first letter of the words ultrathin ( English ultra-thin) or ultra-low power ( English ultra-low power consumption) respectively.
At the same time, you should not think that if the model numbers of the new Ryzen start with the number 2, then the architecture of their cores belongs to the second generation of the Zen microarchitecture. This is not so - these processors are still in the first generation.

New mobile Ryzen with Vega graphics

Last year, AMD already brought the first mobile Ryzen to market, codenamed Raven Ridge. The entire Ryzen mobile family is designed for gaming laptops, ultrabooks, and tablet-laptop hybrids. But there were only two such models, one at a time in the middle and older segments: Ryzen 5 2500U and Ryzen 7 2700U. The junior segment was empty, but right at CES 2018 the company fixed this - two models were added to the mobile family at once: Ryzen 3 2200U and Ryzen 3 2300U.

AMD VP Jim Anderson Demonstrates Ryzen Mobile Family

The 2200U is the first dual-core Ryzen CPU, while the 2300U is quad-core as standard, however both run on four threads. At the same time, the base frequency for the 2200U cores is 2.5 GHz, and for the lower 2300U - 2 GHz. But with increasing loads, the frequency of both models will rise to one indicator - 3.4 GHz. However, laptop manufacturers can lower the power ceiling, because they also need to calculate energy costs and think over the cooling system. There is also a difference between the chips in the cache size: the 2200U has only two cores, and therefore there is half the cache of levels 1 and 2.

The 2200U has only 3 graphics units, but the 2300U has twice as many, as well as processor cores. But the difference in graphic frequencies not so significant: 1,000 MHz versus 1,100 MHz.

The first mobile Ryzen PRO

For the second quarter of 2018, AMD is scheduled to release mobile versions Ryzen PRO, enterprise-grade processors. Mobile PRO specs are identical to consumer versions, with the exception of the Ryzen 3 2200U, which didn't get a PRO implementation at all. The difference between desktop and mobile Ryzen PRO is in additional hardware technologies.

Ryzen PRO processors are complete copies of regular Ryzen, but with additional features

For example, TSME, hardware encryption of RAM on the fly, is used for security (Intel has only software resource-intensive SME encryption). And for centralized management of the fleet of machines, the open standard DASH (Desktop and mobile Architecture for System Hardware, English mobile and desktop architecture for system devices) is available - support for its protocols is built into the processor.

Laptops, ultrabooks, and hybrid laptops with Ryzen PRO should primarily be of interest to companies and government agencies that plan to purchase them for employees.

New AMD 400 series chipsets

The second generation of Ryzen relies on the second generation system logic: The 300th series of chipsets is replaced by the 400th. The AMD X470 was expected to be the flagship of the series, and later simpler and cheaper chipsets, such as the B450, will be released. The new logic has improved everything related to RAM: reduced access latency, raised the upper frequency limit and added headroom for overclocking. Also in the 400th series, the USB bandwidth increased and the power consumption of the processor improved, and at the same time its heat dissipation.

But the processor socket has not changed. The AMD AM4 desktop socket (and its AMD FP5 mobile non-removable variant) is a particular strength of the company. The second generation has the same connector as the first. It will not change in the third and fifth generations either. AMD has promised in principle not to change AM4 until 2020. And in order for the motherboards of the 300th series (X370, B350, A320, X300 and A300) to work with the new Ryzen, you just need to update the BIOS. Moreover, in addition to direct compatibility, there is also a reverse one: old processors will work on new boards.

Gigabyte at CES 2018 has even shown a prototype of the first motherboard based on the new chipset - X470 Aorus Gaming 7 WiFi. This and other boards on X470 and lower chipsets will appear in April 2018, simultaneously with the second generation of Ryzen on the Zen + architecture.

New cooling system

AMD also introduced a new AMD cooler Wraith Prism (English prism of anger). While its predecessor, the Wraith Max, was illuminated in solid red, the Wraith Prism features motherboard-controlled RGB lighting around the fan perimeter. The blades of the cooler cooler are made of transparent plastic and are also highlighted in millions of colors. Fans of RGB lighting will appreciate it, and haters can simply turn it off, although in this case the point of buying this model is leveled.

Wraith Prism- full copy Wraith Max, but backlit with millions of colors

The rest of the specs are identical to the Wraith Max: direct contact heatpipes, software airflow profiles in overclocked mode, and near-silent 39dB operation under standard conditions.

There is no word yet on how much the Wraith Prism will cost, whether it will be bundled with processors, or when it will be available for purchase.

New laptops on Ryzen

In addition to mobile processors, AMD is also promoting new laptops based on them. In 2017, HP Envy x360 models came out on mobile Ryzen, Lenovo IdeaPad 720S and Acer Swift 3. Acer Nitro 5, Dell Inspiron 5000 and HP series will be added in the first quarter of 2018. All of them work on last year's mobile Ryzen 7 2700U and Ryzen 5 2500U.

The Acer Nitro family is a gaming machine. The Nitro 5 line is equipped with 15.6-inch IPS displays with a resolution of 1920 × 1080. And some models will add a discrete Radeon RX 560 graphics chip with 16 graphics units inside.

The Dell Inspiron 5000 line of laptops offers models with 15.6-inch and 17-inch displays, equipped with or hard drives or solid state drives. Some models of the line will also receive a discrete Radeon 530 graphics card with 6 graphics units. This is a rather strange configuration, because even the integrated graphics of the Ryzen 5 2500U have more graphics units - 8 pieces. But the advantage of a discrete card may be in higher clock speeds and separate graphic memory chips (instead of the operational memory section).

Price cuts for all Ryzen processors

Plans for 2020: Navi graphics, Zen 3 processors

2017 was a turning point for AMD. After years of trouble, AMD has completed development of the Zen core microarchitecture and released its first generation of CPUs: the Ryzen, Ryzen PRO and Ryzen Threadripper PC processor families, the Ryzen and Ryzen PRO mobile family, and the EPYC server family. In the same year, the Radeon group developed the Vega graphics architecture: Vega 64 and Vega 56 video cards were released on its basis, and by the end of the year, Vega cores were integrated into Ryzen mobile processors.

Dr. Lisa Su, CEO of AMD, assures that the company will release 7nm processors before 2020

The novelties not only attracted the interest of fans, but also captured the attention of ordinary consumers and enthusiasts. Intel and NVIDIA had to hastily retort: ​​Intel released six-core processors coffee lake, the unplanned second “so” of the Skylake architecture, and NVIDIA expanded the 10th series of video cards based on the Pascal architecture to 12 models.

Rumors about AMD's future plans have been accumulating throughout 2017. So far, Lisa Su, CEO of AMD, has only noted that the company plans to exceed the 7-8% annual rate of productivity gain in the electronics industry. Finally, at CES 2018, the company showed a roadmap not just until the end of 2018, but up to 2020. The basis of these plans is the improvement of chip architectures through the miniaturization of transistors: a progressive transition from the current 14 nanometers to 12 and 7 nanometers.

12nm: Second Gen Ryzen on Zen+

The Zen+ microarchitecture, the second generation of the Ryzen brand, is based on the 12nm process technology. Actually new architecture- This is a modified Zen. The technological production norm of GlobalFoundries factories is being transferred from 14nm 14LPP (Low Power Plus, English low power consumption plus) to 12nm norm 12LP (Low Power, English low power consumption). The new 12LP process technology should provide chips with a 10% performance boost.

Reference: The GlobalFoundries factory network is a former AMD manufacturing facility that was spun off in 2009 into a separate company and merged with other contract manufacturers. In terms of contract manufacturing market share, GlobalFoundries shares second place with UMC, significantly behind TSMC. Chip developers - AMD, Qualcomm and others - order production from both GlobalFoundries and other factories.

In addition to the new process technology, the Zen + architecture and chips based on it will receive improved AMD Precision Boost 2 (exact overclocking) and AMD XFR 2 (Extended Frequency Range 2) technologies. Precision Boost 2 and a special modification of XFR - Mobile Extended Frequency Range (mXFR) can already be found in Ryzen mobile processors.

The Ryzen, Ryzen PRO and Ryzen Threadripper family of PC processors will be released in the second generation, but there is no information about the update of the generations of the Ryzen and Ryzen PRO mobile family, and the server EPYC yet. But it is known that some models of Ryzen processors from the very beginning will have two modifications: with and without graphics integrated into the chip. Ryzen 3 and Ryzen 5 entry and mid-range models will come in both variants. And the high level Ryzen 7 will not receive any graphic modification. Most likely, the code name Pinnacle Ridge (literally, a sharp crest of a mountain, one of the peaks of the Wind River ridge in Wyoming) is assigned to the architecture of the cores for these particular processors.

The second generation of Ryzen 3, 5 and 7 will start shipping in April 2018 along with the 400 series chipsets. And the second generation of Ryzen PRO and Ryzen Threadripper will be late until the second half of 2018.

7nm: 3rd generation Ryzen on Zen 2, Vega discrete graphics, Navi graphics core

In 2018, the Radeon Group will release discrete Vega graphics for laptops, ultrabooks, and laptop tablets. AMD does not share specific details: it is known that discrete chips will work with compact multilayer memory such as HBM2 (integrated graphics use RAM). Separately, Radeon emphasizes that the height of the memory chips will be only 1.7 mm.

Radeon executive showing Vega integrated and discrete graphics

And in the same 2018, Radeon will transfer graphics chips based on the Vega architecture from the 14 nm LPP process technology to 7 nm LP directly, completely jumping over 12 nm. But first, the new graphics units will ship only for the Radeon Instinct line. This is a separate family of Radeon server chips for heterogeneous computing: machine learning and artificial intelligence - the demand for them is provided by the development of unmanned vehicles.

And already at the end of 2018 or the beginning of 2019, ordinary consumers will wait for Radeon and AMD products on the 7-nanometer process technology: processors on the Zen 2 architecture and graphics on the Navi architecture. Moreover, the design work for Zen 2 has already been completed.

AMD partners are already getting acquainted with chips on Zen 2, which will create third-generation Ryzen motherboards and other components. AMD is gaining such pace due to the fact that the company has two "jumping" teams to develop promising microarchitectures. They started with parallel work on Zen and Zen+. When Zen was completed, the first team moved on to Zen 2, and when Zen+ was completed, the second team moved on to Zen 3.

7nm plus: fourth generation Ryzen on Zen 3

While one department at AMD is solving the problems of mass production of Zen 2, another department is already designing Zen 3 on a technology standard designated as "7nm+". The company does not disclose details, but according to indirect data, it can be assumed that the technical process will be improved by supplementing the current deep ultraviolet lithography (DUV, Deep Ultraviolet) with a new hard ultraviolet lithography (EUV, Extreme Ultraviolet) with a wavelength of 13.5 nm.

GlobalFoundries has already installed new equipment for the transition to 5nm

Back in the summer of 2017, one of the GlobalFoundries factories purchased more than 10 lithographic systems from the TWINSCAN NXE series from the Dutch ASML. With the partial use of this equipment within the same 7 nm process technology, it will be possible to further reduce power consumption and increase chip performance. There are no exact metrics yet - it will take some more time to debug new lines and bring them to acceptable capacities for mass production.

AMD expects to start selling 7nm+ chips from processors based on the Zen 3 microarchitecture by the end of 2020.

5nm: fifth and next generations of Ryzen on Zen 4?

AMD has not yet made an official announcement, but we can safely speculate that the next frontier for the company will be the 5 nm process technology. Experimental chips at this rate have already been produced by the research alliance of IBM, Samsung and GlobalFoundries. Crystals based on the 5 nm manufacturing process will no longer require partial, but full-fledged use of hard ultraviolet lithography with an accuracy of more than 3 nm. This resolution is provided by the models of the TWINSCAN NXE:3300B lithographic system purchased by GlobalFoundries from ASML.

A layer one molecule thick of molybdenum disulphide (0.65 nanometers) exhibits a leakage current of only 25 femtoamperes/micrometer at 0.5 volts.

But the difficulty also lies in the fact that the 5 nm process will probably have to change the shape of the transistors. Long-established FinFETs (fin-shaped transistors, from English fin) may give way to promising GAA FETs (gate-all-around transistor form). It will take several more years to set up and deploy mass production of such chips. The consumer electronics sector is unlikely to receive them before 2021.

Further reduction of technological norms is also possible. For example, back in 2003, Korean researchers created FinFET at 3 nanometers. In 2008, the University of Manchester created a nanometer transistor based on graphene (carbon nanotubes). And in 2016, Berkeley Lab research engineers conquered the sub-nanometer scale: both graphene and molybdenum disulfide (MoS2) can be used in such transistors. True, at the beginning of 2018, there was still no way to produce a whole chip or substrate from new materials.

A week ago, AMD held a small presentation dedicated to the new Ryzen Mobile APUs, formerly known under the code name Raven Ridge. The speaker, however, as usual, first lamented the current situation in the world of processors. Like, Moore's law is no longer so strictly enforced and everyone has already gotten used to "5-7% growth per year" (it is known in whose garden this stone is). And even in desktops, where there are no special restrictions, five years ago a competitor's mass processor had 4 cores (and 8 threads) with a frequency of about 3.5 GHz, and until recently, all the same 4C / 8T, but at about 4 GHz. It was only this year that the competitor changed tactics, offering more cores for the same price as before. In the mobile segment, in this sense, it was even worse until this fall - configuration stability is no longer a sign of skill. Lack of competition is bad for the market and end users. However, we have heard all this from AMD before.

On the left is the CCX block of Zen cores, on the right is the GPU block (blue)

The company itself has been developing new cores (CPU and GPU) for the past four years, and, according to AMD, it is important that they tried to make them as scalable as possible. Powerful server solutions and desktop systems are made on the same basis, and now also mobile ones - for laptops. As a matter of fact, AMD Ryzen Mobile 7 2700U and 5 2500U are one CCX for four Zen cores (8 threads), Radeon Vega graphics and a slightly modified Infinity Fabric bus. The latter combines the CPU, GPU, memory controller, display and multimedia units, as well as the peripheral controller. Basic version both chips have a TDP of 15 W, but system manufacturers, with AMD's approval, can independently configure a TDP in the range from 12 (9 is indicated in the table, but 12 was repeatedly announced) to 25 W - everything will depend on the quality of the cooling system. These settings are not available to the user.

At the microarchitecture level, the new APUs do not differ much from the desktop versions of the chips and . The changes concern those areas that are critical specifically for the mobile segment. Developers, for example, cut L3 caches down to 4 MB, just to keep the die size down. HBM for the GPU also had to be abandoned - the video memory is cut off from the main DDR4. The specific volume depends on the laptop OEM. For tests (benchmarks are given below), AMD used configurations with 256 MB of video memory, but in general there will be options for 512-1024 MB, fortunately relatively large volume RAM in modern laptops is no longer a rarity. And yes, the overall performance of the complex will again partially depend on the frequency of the RAM.

The DDR4-2400 memory controller has also remained almost unchanged: it is dual-channel here, but for some ultraportable solutions AMD insists on using a single-channel configuration - in this case, the difference in graphics performance will be about 20-40%. ECC is supported, but we are unlikely to see it in laptops. The differences between AMD Ryzen Mobile 7 2700U and 5 2500U are not that big. The older model has a base and boost frequencies of 2.2 and 3.8 GHz, respectively, and the younger one has 2.0 and 3.6 GHz. The 2500U has eight 1.1GHz Radeon Vega CUs, while the 2700U has ten of them running at 1.3GHz. Yes, only two APU models will be available for now, but next year AMD promises to significantly increase their number. The crystal has an area of ​​209.78 mm 2 and contains approximately 4.95 billion transistors. The manufacturing process is 14 nm.

However, some important changes in the new chipsets are worth mentioning. The Precision Boost crystal dynamic frequency control technology has acquired the number 2 in the title. It still changes frequencies in 25 MHz steps, but in this case, such a step is used in both the GPU and the CPU. Besides, a new version copes better with multi-threaded workloads - the main limiting factor in the case of laptops will be cooling efficiency rather than power limit. In addition, the Mobile XFR subsystem has appeared in the new APUs - it also additionally increases the turbo frequency above par, but here its task is to keep the established overclocking as long as possible. The exact amount of frequency increase, the number of activated cores and specific models of APUs with mXFR have not been announced, but it is reported that this technology is designed more for high-performance laptops with good cooling.

However, some additions are also provided in the power subsystem. There are thousands of separate sensors (and regulators) in the crystals that measure voltages directly at the transistor blocks, and with millivolt accuracy. That is, data on the state of external VREGs is no longer so important. There was already voltage regulation for individual Zen cores, and now it has been added for the GPU. It is curious that an AMD representative claims that the worst load scenario, when the peak occurs simultaneously on the CPU and GPU, allegedly does not occur in practical work scenarios. This, of course, is debatable. Nevertheless, the main task in the case of APUs is the correct and fast distribution of power between the graphics and processor parts, depending on which of them really needs it. Actually, the main innovation in the APU is the LDO controllers built into the GPU. It is argued that no one has such an effective implementation of this technology now.

The new internal LDOs unified for the CPU / GPU, as AMD itself says, allow in the case of the APU to reduce the current requirements by 36%, while increasing the maximum current for CPU power supply or GPU - in fact, you can either make a more powerful solution, leaving the same power system, or, conversely, reduce it, but maintain performance. In any case, the energy efficiency of the final solution increases, because the dynamic distribution of frequency and power, depending on the load, occurs both between the CPU cores and between the graphics and central processors. However, the specific details of the distribution algorithm are not disclosed. On the other hand, not only the algorithm is important, but also the speed of switching between different states of the CPU / GPU and their number, which, in particular, is necessary for more efficient use of the laptop battery.

In the new APUs, the GPU has a special mode in which the power consumption of the card is reduced by 95%. It is activated when literally nothing happens on the screen, that is, a static picture is displayed, for example, if the user simply moved away from the PC for a while. There is a similar state for CPU cores. The transition between the main states in both cases takes 100 microseconds or less (typical value - 50 microseconds), and for deep sleep mode - up to 1.5 ms. In addition, the internal components of the APU are conventionally divided into two zones with different power policies, which also contributes to energy efficiency. The Infinity Fabric bus carries data from various internal sensors and controllers.

Also, the developers note the small thickness of the finished product - only 1.38 mm. Previously, as stated, not all ultrabooks were able to place the existing chips just because of their thickness. As for the GPU, it is worth noting the presence of FreeSync 2 technology. AMD will try to ensure that manufacturers add support for it to the displays of their laptops whenever possible. The graphics card itself supports multi-monitor configurations, 4K and HDR image output. Right now, together with Microsoft, PlayReady support is being prepared, which is necessary for correct operation some video streaming services. Well, in general, AMD continues to adhere to long term strategy 25×20, which was announced in 2014. According to her, by 2020 the overall performance of the APU should increase 25 times compared to the 2014 models.

Unfortunately, during the presentation, AMD did not present the full characteristics of new products (for example, there is no data on integrated controllers for peripherals), showing only some benchmarks. We note a few important points in them. Firstly, in some cases, the comparison is not with competitor solutions, but only with AMD products on the old platform. Secondly, where such a comparison still exists, an eighth generation chip was used with the same nominal TDP of 15 W, which was available on the market (and there are still few of them). Thirdly, various acceleration technologies or any other “cheating” were not involved, including, for example, laptop tests in a pre-chilled room. Below in the gallery are the results of the tests, as well as comments and notes to them.

AMD Ryzen Mobile Benchmarks

Best of all, new items show themselves in multi-threaded applications, as well as in software that actively uses the graphics subsystem. AMD notes that now on ultra-thin laptops, for example, you can safely process video and graphics without worrying too much about the autonomy of the device. And of course, for them, according to the company, a new niche appears - games. Naturally, heavy gaming monsters will feel uncomfortable here, but popular eSports projects work well with acceptable resolution and graphics quality. By the way, options with Dual Graphics are not yet expected, instead, developers can use DirectX 12 tools to share the resources of different GPUs.