ARMv6 and ARMv7 are generations of architecture mobile processors companies ARM Limited based on 32-bit instructions.
ARM architecture quite common in a market that previously belonged exclusively to desktop processors of such popular architectures as Intel x86/64 and AMD64. Today, thanks to ARMv6 or ARMv7, the processor of modern TVs, home theaters and other familiar equipment can fit in your hand.
The main niche for ARM mobile architecture has become smartphones, tablets and other similar mobile devices. These days, 95% of smartphones are already running ARM architecture processors, as well as half of smart TVs and 90% hard drives. And due to their “survivability” on a single battery charge and acceptable performance, devices with ARM architecture processors on board replaced the entire line of “netbooks”, becoming tablets with docking stations, which gave the device almost a whole day of work instead of just a few hours as before and gave some jump in performance due to the low cost of the processors themselves, the presence of multi-core solutions and high overclocking potential.
Key features of these architectures:
- ARMv6 does not officially support Flash.(In any case, since mid-2012, Google completely abandoned Flash on the Android platform, so support for this technology is no longer relevant).
- ARMv7 is often found in multi-core mobile processors, while the sixth generation is limited to only one physical and logical core.
- Applications built for ARMv7 have a larger overall weight and require more dedicated RAM than similar programs that only work with ARMv6.
- ARMv7 processors are more powerful than the previous generation.
- Games and programs developed for ARMv6 are compatible with ARMv7 by default, but not vice versa.
- The fact that one or another application supports ARMv6 and ARMv7 at the same time does not always mean improved graphics performance on the latter architecture. In this case, we recommend looking towards processors from Nvidia and Tegra. They have a separate store with toys with higher detail and other graphic goodies that are not available on any other devices not running Tegra.
- The standard ARMv7 frequency of such processors is stated to be 1 GHz nominal and higher, which cannot be said about ARMv6.
- Games for armv7 significantly more than under armv6.
- Many popular video player applications (like armv6) requires downloading and installing an additional set of codecs for armv6 or armv7 processor architectures, without which you will not achieve hardware acceleration.
Frequently asked questions - answers:
I want to download the game, but the description contains a warning that this game is only compatible with ARMv7 or has two versions separately for both ARMv6 and ARMv7, respectively, what should I download?
Find out in any way known to you the exact name of the processor used in your device, and then find it on a specially designated page in Wikipedia and determine the version of architecture used, a clear example this time will be Snapdragon processors from the well-known company Qualcomm, whose page is located at the following link:
After installing an Android application from third-party resources, it refuses to launch, what should I do?
Make sure your version is operating system matches the compatible Android versions of this application, and also find out which generation of ARM architecture your processor corresponds to and, if it is ARMv7 and higher, then in 99.9% any relatively new program or the game must at least start until the license is verified, some technical characteristics and other device recognition data, and additional application cache data if necessary. In addition, it does not hurt to free RAM from active background processors prematurely if the free space does not match minimum requirement one game or another. We recommend keeping 256, or better yet 512 megabytes of free RAM.
Find today armv7 phones much easier than a couple of years ago, because... This microprocessor architecture has already reached the budget area of the mobile smartphone market, but for owners of “oldies” this article can really be useful.
Here we did not post the current list of devices different versions ARM, because this list is constantly updated and it’s simply impossible to keep track of it. We recommend that you immediately search for your device on the Wikipedia pages dedicated to one or another mobile processor.
The computer world is changing rapidly. Desktop PCs have lost first place in the sales rankings to laptops, and they are about to give the market to tablets and other mobile devices. 10 years ago we valued pure megahertz, true power and performance. Now, in order to conquer the market, the processor must be not only fast, but also economical. Many people believe that ARM is the architecture of the 21st century. Is this true?
New - well forgotten old
Journalists, following ARM PR people, often present this architecture as something completely new that should bury the gray-haired x86.
In fact, ARM and x86, on the basis of which the Intel, AMD and VIA processors installed in laptops and desktop PCs are built, are almost the same age. The first x86 chip was released in 1978. The ARM project officially started in 1983, but was based on developments that were carried out almost simultaneously with the creation of the x86.
The first ARMs impressed specialists with their elegance, but with their relative low performance they could not conquer a market that demanded high speeds and did not pay attention to efficiency. Certain conditions had to exist for ARM's popularity to skyrocket.
At the turn of the eighties and nineties, with their relatively inexpensive oil, huge SUVs with powerful 6-liter engines were in demand. Few people were interested in electric cars. But in our time, when a barrel of oil costs more than $100, large cars with power-hungry engines are needed only by the rich; the rest are in a hurry to switch to economical cars. A similar thing happened with ARM. When the question of mobility and efficiency arose, architecture turned out to be in great demand.
"Risk" processor
ARM is a RISC architecture. It uses a reduced set of commands - RISC (reduced instruction set computer). This type of architecture appeared in the late seventies, around the same time when Intel offered its x86.
While experimenting with various compilers and microcode processors, engineers noticed that in some cases, sequences of simple instructions were executed faster than one complex operation. It was decided to create an architecture that would involve working with a limited set of simple instructions, the decoding and execution of which would take a minimum of time.
One of the first RISC processor projects was carried out by a group of students and teachers at the University of Berkeley in 1981. Just at this time, the British company Acorn faced the challenge of time. It produced BBC Micro educational computers based on the 6502 processor, which were very popular in Foggy Albion. But soon these home PCs began to lose to more advanced machines. Acorn was at risk of losing the market. The company's engineers, having become acquainted with student work on RISC processors, decided that creating their own chip would be quite simple. In 1983, the Acorn RISC Machine project was launched, which later became ARM. Three years later the first processor was released.
First ARM
He was extremely simple. The first ARM chips even lacked multiply and divide instructions, which were represented by a set of more simple instructions. Another feature of the chips was the principles of working with memory: all operations with data could be carried out only in registers. At the same time, the processor worked with the so-called register window, that is, it could access only a part of all available registers, which were basically universal, and their operation depended on the mode in which the processor was located. This made it possible to abandon the cache in the very first versions of ARM.
In addition, by simplifying instruction sets, architecture developers were able to do without a number of other blocks. For example, the first ARM completely lacked microcode, as well as a floating point unit - FPU. The total number of transistors in the first ARM was 30,000. In similar x86 there were several times, or even an order of magnitude more. Additional energy savings are achieved through conditional execution of commands. That is, this or that operation will be performed if there is a corresponding fact in the register. This helps the processor avoid “unnecessary movements”. All instructions are executed sequentially. As a result, ARM lost in performance, but not significantly, while gaining significantly in power consumption.
The basic principles of the architecture remain the same as in the first ARM: working with data only in registers, a reduced set of instructions, a minimum of additional modules. All this provides the architecture with low power consumption and relatively high performance.
In order to increase this, ARM has introduced several additional instruction sets in recent years. Along with the classic ARM, there are Thumb, Thumb 2, Jazelle. The latter is designed to speed up the execution of Java code.
Cortex - the most advanced ARM
Cortex – modern architectures for mobile devices, embedded systems and microcontrollers. Accordingly, CPUs are designated as Cortex-A, embedded – Cortex-R and microcontrollers – Cortex-M. All of them are built on the ARMv7 architecture.
The most advanced and powerful architecture in the ARM line is Cortex-A15. It is expected that mainly two- or four-core models will be produced on its basis. Cortex-A15 of all previous ARMs is closest to x86 in terms of the number and quality of blocks.
The Cortex-A15 is based on processor cores equipped with an FPU unit and a set of NEON SIMD instructions designed to speed up the processing of multimedia data. The cores have a 13-stage pipeline, they support free-order instruction execution, and ARM-based virtualization.
Cortex-A15 supports advanced memory addressing system. ARM remains a 32-bit architecture, but the company's engineers have learned to convert 64-bit or other advanced addressing into processor-friendly 32-bit. The technology is called Long Physical Address Extensions. Thanks to it, Cortex-A15 can theoretically address up to 1 TB of memory.
Each core is equipped with a first-level cache. In addition, there is up to 4 MB of distributed low-latency L2 cache. The processor is equipped with a 128-bit coherent bus, which can be used to communicate with other units and peripherals.
The cores that underlie Cortex-A15 are a development of Cortex-A9. They have a similar structure.
Cortex-A9, unlike Cortex-A15, can be produced in both multi- and single-core versions. The maximum frequency is 2.0 GHz, Cortex-A15 suggests the possibility of creating chips operating at a frequency of 2.5 GHz. Chips based on it will be manufactured using 40 nm and thinner technical processes. Cortex-A9 is produced in 65 and 40 nm process technologies.
Cortex-A9, like Cortex-A15, is intended for use in high-performance smartphones and tablets, but it is not suitable for more serious applications, for example, in servers. Only Cortex-A15 has hardware virtualization, advanced memory addressing. Additionally, the NEON Advanced SIMD instruction set and FPU are optional in the Cortex-A9, while they are required in the Cortex-A15.
Cortex-A8 will gradually disappear from the scene in the future, but for now this single-core variant will find use in budget smartphones. The low-cost solution, ranging from 600 MHz to 1 GHz, provides a balanced architecture. It has an FPU unit and supports the first version of SIMD NEON. Cortex-A8 assumes a single technological process - 65 nm.
ARM of previous generations
ARM11 processors are quite common in the mobile market. They are built on the basis of the ARMv6 architecture and its modifications. It is characterized by 8-9-stage pipelines, Jazelle support, which helps speed up the processing of Java code, SIMD stream instructions, Thumb-2.
XScale, ARM10E, ARM9E processors are based on the ARMv5 architecture and its modifications. Maximum length the pipeline is 6 stages, Thumb, Jazelle DBX, Enhanced DSP. XScale chips have a second level cache. The processors were used in smartphones of the mid-2000s; today they can be found in some inexpensive mobile phones.
ARM9TDMI, ARM8, StrongARM - representatives of ARMv4, which has a 3-5 stage pipeline and supports Thumb. ARMv4, for example, could be found in the first classic iPods.
ARM6 and ARM7 belong to ARMv3. In this architecture, the FPU unit appeared for the first time; 32-bit memory addressing was implemented, and not 26-bit, as in the first examples of the architecture. ARMv2 and ARMv1 were technically 32-bit chips, but in reality only actively worked with a 26-bit address space. The cache first appeared in ARMv2.
Their name is legion
Acorn did not initially intend to become a player in the processor market. The task of the ARM project was to create a chip of its own production for the production of computers - it was the creation of PCs that Acorn considered its main business.
ARM went from being a development group to becoming a company thanks to Apple. In 1990, Apple, together with VLSI and Acorn, began developing a low-cost processor for the first pocket computer, the Newton. For these purposes, a separate company was created, which received the name of the internal project Acorn - ARM.
With the participation of Apple, an ARM6 processor was created, which is closest to modern chips from an English developer. At the same time, DEC was able to patent the ARM6 architecture and began producing chips under the StrongARM brand. A couple of years later, the technology was transferred to Intel as part of another patent dispute. The microprocessor giant has created its own analogue, the XScale processor, based on ARM. But in the middle of the previous decade, Intel got rid of this “non-core asset”, focusing exclusively on x86. XScale moved into the hands of Marvell, which already licensed ARM.
At first, ARM, which was new to the world, was not able to produce processors. Its management chose a different way of making money. The ARM architecture was simple and flexible. At first, the core did not even have a cache, so subsequently additional modules, including FPU, controllers were not closely integrated into the processor, but were, as it were, attached to the base.
Accordingly, ARM got its hands on an intelligent designer that allowed technologically advanced companies to create processors or microcontrollers to suit their needs. This is done using so-called coprocessors, which can expand the standard functionality. In total, the architecture supports up to 16 coprocessors (numbers from 0 to 15), but number 15 is reserved for the coprocessor that performs cache and memory management functions.
Peripherals connect to the ARM chip, mapping their registers to the memory space of the processor or coprocessor. For example, an image processing chip may consist of a relatively simple ARM7TDMI-based core and a coprocessor that provides HDTV signal decoding.
ARM began licensing its architecture. Other companies have already been implementing it in silicon, including Texas Instruments, Marvell, Qualcomm, Freescale, but also completely non-core ones like Samsung, Nokia, Nintendo or Canon.
The absence of its own factories, as well as impressive licensing fees, allowed ARM to be more flexible in developing new versions of the architecture. The company baked them like hot cakes, entering new niches. In addition to smartphones and tablets, the architecture is used in specialized processors, for example, in GPS navigators, digital cameras and video cameras. Industrial controllers and other chips for embedded systems are created on its basis.
The ARM licensing system is a real microelectronics hypermarket. The company licenses not only new but also legacy architectures. The latter can be used to create microcontrollers or chips for low-cost devices. Naturally, the level of licensing fees depends on the degree of novelty and complexity of the architecture variant of interest to the manufacturer. Traditionally, the technical processes for which ARM develops processors are 1-2 steps behind those considered relevant for x86. The high energy efficiency of the architecture makes it less dependent on the transition to new technological standards. Intel and AMD are striving to make thinner chips in order to increase frequencies and the number of cores while maintaining physical size and power consumption. ARM inherently has lower power requirements and also delivers higher levels of performance per watt.
Features of NVIDIA, TI, Qualcomm, Marvell processors
By licensing ARM left and right, developers strengthened the position of their architecture at the expense of the competencies of their partners. A classic example in this case is NVIDIA Tegra. This line of systems-on-a-chip is based on ARM architecture, but NVIDIA already had its own very serious developments in the field of 3D graphics and system logic.
ARM gives its licensors broad discretion to redesign the architecture. Accordingly, NVIDIA engineers were able to combine in Tegra the strengths of ARM (CPU computing) and their own products - working with three-dimensional graphics, etc. As a result, Tegra has the highest 3D performance for its class of processors. They are 25-30% faster than PowerVR, used by Samsung and Texas Instruments, and are almost twice as fast as Adreno, developed by Qualcomm.
Other manufacturers of processors based on the ARM architecture are strengthening certain additional blocks and improving chips to achieve higher frequencies and performance.
For example, Qualcomm does not use the ARM reference design. The company's engineers seriously reworked it and called it Scorpio - it is the basis of Snapdragon chips. The design has been partly redesigned to accommodate more sophisticated technical processes than those provided by the standard IP ARM. As a result, the first Snapdragons were produced at 45 nm, which provided them with higher frequencies. And the new generation of these processors with a declared 2.5 GHz may even become the fastest among analogues based on ARM Cortex-A9. Qualcomm also uses its own Adreno graphics core, created on the basis of developments purchased from AMD. So in a way, Snapdragon and Tegra are enemies on a genetic level.
When creating Hummingbird, Samsung also took the path of optimizing the architecture. The Koreans, together with the Intrinsity company, changed the logic, thereby reducing the number of instructions required to perform certain operations. Thus, we managed to gain 5-10% of productivity. In addition, a dynamic L2 cache and the ARM NEON multimedia extension were added. The Koreans used PowerVR SGX540 as a graphics module.
Texas Instruments in its new OMAP series based on the ARM Cortex-A architecture has added a special IVA module responsible for accelerating image processing. It allows you to quickly process data coming from the sensor to the built-in camera. In addition, it is connected to the ISP and helps in video acceleration. OMAP also uses PowerVR graphics.
The Apple A4 has a large 512 KB cache, uses PowerVR graphics, and the ARM core itself is built on a variant of the architecture redesigned by Samsung.
The dual-core Apple A5, which debuted in the iPad 2 in early 2011, is based on the ARM Cortex-A9 architecture, just like the one previously optimized by Samsung. Compared to the A4, the new chip has double the amount of second-level cache memory - it has been increased to 1 MB. The processor contains a dual-channel RAM controller and has an improved video unit. As a result, it performs twice as well as the Apple A4 in some tasks.
Marvell offers chips based on its own Sheeva architecture, which, upon closer inspection, turns out to be a hybrid of XScale, once purchased from Intel, and ARM. These chips have a larger amount of cache memory compared to analogues and are equipped with a special multimedia module.
Currently, ARM licensees only produce chips based on the ARM Cortex-A9 architecture. At the same time, although it allows you to create quad-core variants, NVIDIA, Apple, Texas Instruments and others are still limited to models with one or two cores. In addition, the chips operate at frequencies up to 1.5 GHz. Cortex-A9 allows you to make two-GHz processors, but again, manufacturers are not trying to quickly increase frequencies - after all, for now the market will have enough dual-core processors at 1.5 GHz.
Processors based on Cortex-A15 should become truly multi-core, but even if they are announced, they are only on paper. Their appearance in silicon should be expected next year.
Modern ARM licensee processors based on Cortex-A9:
x86 is the main contender
x86 is a representative of CISC architectures. They use the full set of commands. One instruction in this case performs several low-level operations. The program code, unlike ARM, is more compact, but does not execute as quickly and requires more resources. In addition, from the very beginning, x86 were equipped with all the necessary blocks, which implied both their versatility and gluttony. Additional energy was spent on unconditional, parallel execution of commands. This allows you to achieve a speed advantage, but some operations are performed in vain because they do not satisfy the previous conditions.
These were the classic x86s, but starting with the 80486, Intel de facto created an internal RISC core that executed CISC instructions, previously decomposed into simpler instructions. Modern Intel and AMD processors have the same design.
Windows 8 and ARM
ARM and x86 today differ less than 30 years ago, but are still based on different principles, which separates them into different niches of the processor market. The architectures might never have intersected if the computer itself had not changed.
Mobility and cost-effectiveness came first, and more attention was paid to smartphones and tablets. Apple makes a lot of money from mobile gadgets and the infrastructure tied to them. Microsoft does not want to be left behind and has been trying to gain a foothold in the tablet market for the second year. Google is quite successful.
The desktop PC is becoming primarily a working tool; the niche of the household computer is occupied by tablets and specialized devices. In these conditions, Microsoft is going to take an unprecedented step. . It is not yet entirely clear what this will lead to. We will get two versions of the operating system, or one that will work with both architectures. Will Microsoft's ARM support kill x86 or not?
There is little information yet. Microsoft demonstrated Windows operation 8 on an ARM-powered device during CES 2011. Steve Ballmer showed that on an ARM platform with using Windows you can watch videos, work with images, use the Internet - Internet Explorer I even worked with hardware acceleration - connect USB devices, print documents. The most important thing in this demonstration was the presence Microsoft Office running on ARM without participation virtual machine. At the presentation three gadgets based on Qualcomm processors, Texas Instruments and NVIDIA. Windows had a standard “seven” shell, but Microsoft representatives announced a new, redesigned system kernel.
However, Windows is not only an OS made by Microsoft engineers, it is also millions of programs. Some software is critical for people in many professions. For example, the Adobe CS package. Will the company support an ARM-Windows version of the software, or will the new kernel allow Photoshop and other popular applications to run on computers with NVIDIA Tegra or other similar chips without additional code modifications?
In addition, the question arises with video cards. Nowadays, video cards for laptops are made by optimizing the power consumption of desktop graphics chips - they are architecturally the same. At the same time, now a video card is something like a “computer within a computer” - it has its own ultra-fast RAM and its own computing chip, which is significantly superior to conventional processors in specific tasks. It goes without saying that applications that work with 3D graphics have been appropriately optimized for them. Yes and various programs video editing and graphic editors(in particular Photoshop from version CS4), and more recently browsers also use hardware acceleration using GPUs.
Of course, in Android, MeeGo, BlackBerry OS, iOS and other mobile systems, the necessary optimization has been made for the various mobile (more precisely, ultra-mobile) accelerators on the market. However, they are not supported in Windows. Drivers, of course, will be written (and have already been written - Intel Atom Z500 series processors come with a chipset that integrates the PowerVR SGX 535 “smartphone” graphics core), but optimization of applications for them may be late, if at all.
Obviously, “ARM on the desktop” won’t really catch on. Perhaps in low-power systems on which they will access the Internet and watch movies. On nettops in general. So ARM is so far only trying to take aim at the niche that Intel Atom has occupied and where AMD is now actively pursuing with its Brazos platform. And she, apparently, will partially succeed. Unless both processor companies come up with something very competitive.
In some places, Intel Atom and ARM are already competing. They are used to create network storage data and low-power servers that can serve a small office or apartment. There are also several commercial projects of clusters based on cost-effective Intel chips. The characteristics of the new processors based on ARM Cortex-A9 allow them to be used to support infrastructure. So in a couple of years we may have ARM servers or ARM-NAS for small local networks, the emergence of low-power web servers cannot be ruled out.
First sparring
ARM's main competitor from the x86 side is Intel Atom, and now we can add the . A comparison of x86 and ARM was carried out by Van Smith, who created the OpenSourceMark, miniBench test packages and one of the co-authors of SiSoftware Sandra. Atom N450, Freescale i.MX515 (Cortex-A8), VIA Nano L3050 took part in the “race”. The frequencies of x86 chips were reduced, but they still had an advantage due to more advanced memory.
The results turned out to be very interesting. The ARM chip turned out to be as fast as its competitors in integer operations, while consuming less power. There is nothing surprising here. Initially, the architecture was both quite fast and economical. In floating point operations, ARM was inferior to x86. The traditionally powerful FPU unit found in Intel and AMD chips had an impact here. Let us remember that it appeared in ARM relatively recently. The tasks that fall on the FPU occupy a significant place in the life of a modern user - these are games, video and audio encoding, and other streaming operations. Of course, the tests conducted by Van Smith are no longer so relevant today. ARM has significantly strengthened weaknesses its architecture in versions Cortex-A9 and especially Cortex-A15, which, for example, can already execute instructions unconditionally, parallelizing the solution of problems.
Prospects for ARM
So which architecture should you choose in the end, ARM or x86? It would be most correct to bet on both. Today we live in conditions of reformatting of the computer market. In 2008, netbooks were predicted to have a bright future. Cheap compact laptops were supposed to become the main computer for most users, especially against the backdrop of the global crisis. But then the economy started to recover and the iPad came out. Now tablets are declared kings of the market. However, the tablet is good as an entertainment console, but not very convenient for work, primarily due to touch input - writing this article on an iPad would be very difficult and time-consuming. Will tablets stand the test of time? Perhaps in a couple of years we will come up with a new toy.
But still, in the mobile segment, where high performance is not required, and user activity is mainly limited to entertainment and not related to work, ARM looks preferable to x86. They provide an acceptable level of performance as well as great time battery life. Intel's attempts to bring Atom to fruition have so far been unsuccessful. ARM sets a new benchmark for performance per watt. Most likely, ARM will be successful in compact mobile gadgets. They can also become leaders in the netbook market, but here everything depends not so much on processor developers, but on Microsoft and Google. If the first implements normal ARM support in Windows 8, and the second brings Chrome OS to fruition. So far, the smartbooks offered by Qualcomm have not made it into the market. X86-based netbooks survived.
According to ARM, a breakthrough in this direction should be made by the Cortex-A15 architecture. The company recommends dual- and quad-core processors based on it with a frequency of 1.0-2.0 GHz for home entertainment systems that will combine a media player, 3D TV and Internet terminal. Quad-core chips with a frequency of 1.5-2.5 GHz can become the basis of home and web servers. Finally, the most ambitious use case for Cortex-A15 - infrastructure wireless networks. Chips with four or more cores and a frequency of 1.5-2.5 GHz can be used here.
But for now these are just plans. Cortex-A15 was introduced by ARM in September last year. Cortex-A9 was shown by the company in October 2007, two years later the company presented the A9 variant with the ability to increase the frequency of the chips to 2.0 GHz. For comparison, NVIDIA Tegra 2 - one of the most popular solutions based on Cortex-A9 - was released only in January last year. Well, users were able to touch the first gadgets based on it after another six months.
The work PC segment and high-performance solutions will remain with x86. This will not mean the death of the architecture, but in monetary terms, Intel and AMD should prepare for the loss of part of the income that will go to ARM processor manufacturers.
In 2011, ARM Limited announced a new family of processors called ARMv8. And in 2013, Apple released the first ARMv8 processor - the A7 single-chip system, which is used in the iPhone 5S, iPad Air and iPad mini Retina. The ARMv8 architecture received a 64-bit instruction set, but this is far from its only advantage over its predecessor ARMv7. Read the article about how 64-bit ARMv8 processors are designed and what they are like.
You can read about the history of ARM architecture, the specifics of the activities of ARM Limited and the generations of ARMv5, ARMv6 and ARMv7 processors in the article. And about popular models of ARMv7 chips produced by Qualcomm, NVIDIA, Samsung, Apple, MediaTek, etc. are described in detail in articles and.
ARMv8 innovations
The updated architecture of the ARMv8 family of processors was dubbed AArch64. It received a 64-bit instruction set and the ability to work with a large amount of RAM (4 GB or more). Of course, compatibility with 32-bit applications (AArch32) is provided. Other important innovations of ARMv8 were:
- 31 registers general purpose, each 64 bits long, whereas SP and PC are not general purpose registers. The higher the register bit depth, the more number can be stored in them. And the greater the number of registers, the more data is placed in them at the same time. As a result, a larger amount of data can be processed in one instruction and the entire algorithm will execute faster;
— translation of virtual addresses from a 48-bit format works using LPAE mechanisms borrowed from ARMv7;
— a new set of instructions with a fixed length. The instructions are 32 bits in size and many are the same as AArch32 instructions, although there are fewer conditional instructions;
— the number of 128-bit registers (compatible with 64-bit registers) available to SIMD NEON and VFP coprocessors has been increased from 16 to 32, and new cryptographic instructions AES and SHA have been added. The SIMD NEON instruction set accelerates media and signal processing applications. In turn, VFP is responsible for low-power calculations on floating point numbers;
— support for calculations on double-precision floating point numbers and the IEEE 754 standard, which is a generally accepted format for representing floating point numbers used in software implementations of arithmetic operations.
ARM Limited reference cores
The first ARMv8 processor cores developed directly by ARM Limited were Cortex-A53 and A57. The A53 core is a mid-range solution with a performance of 2.3 DMIPS/MHz, which is approximately halfway between the current Cortex-A7 (1.9 DMIPS/MHz) and A9 (2.5 DMIPS/MHz). While the A57 occupies the upper segment, because its speed (4.1 DMIPS/MHz) exceeds that of both 32-bit flagships: Cortex-A15 (3.5 DMIPS/MHz) and A17 (4 DMIPS/MHz).
In addition to licensing reference processor cores, ARM Limited sells extended licenses that allow chipmakers to modify the ARM architecture at their discretion. For example, Apple, Qualcomm and NVIDIA have such licenses. Therefore, nothing prevents processor manufacturers from creating their own solutions based on ARMv8, which are significantly different from the reference Cortex-A53 and A57.
Apple A7
The first and so far only 64-bit ARM processor that is already used in smartphones and tablets is the Apple A7. It is built on Apple's proprietary Cyclone architecture, compatible with ARMv8. This is the company's second internally developed processor architecture; the first was Swift (A6 and A6X chips, ARMv7 family).
The A7 single-chip system has only two processor cores (frequency up to 1.4 GHz), but there is a PowerVR G6430 graphics accelerator with four core clusters. The performance of the A7 chip in processor-dependent tasks has increased by about one and a half times compared to the A6, while in various graphics tests the increase is from two to three times.
But the theoretical ability to work with a large amount of RAM thanks to the 64-bit architecture of the A7 processor of the device under iOS control They don’t feel it yet. For iPhone 5s, iPad Air and iPad mini Retina only 1 GB of RAM; and hardly in the new generation of mobile Apple devices The amount of RAM will more than double.
Qualcomm Snapdragon 410, 610, 615, 808 and 810
Following Apple, Qualcomm hastened to announce its 64-bit ARM processors, with five models at once. True, so far none of them are used in commercial smartphones or tablets. Most likely, the heyday of the era of 64-bit Android devices will take place in early 2015 at CES and MWC.
The Snapdragon 410 single-chip system (MSM8916) is the youngest of the announced 64-bit Qualcomm line. It includes four Cortex-A53 cores with a frequency of 1.2 GHz, an Adreno 306 graphics accelerator and, most interestingly, a navigation module with support for GPS, GLONASS and even Chinese satellite networks. They plan to use Snapdragon 410 in inexpensive smartphones based on Android, Windows Phone and Firefox OS.
The same four Cortex-A53 cores as the 410 contain the Snapdragon 610 (MSM8936) chip, only it has improved graphics Adreno 405. While the Snapdragon 615 (MSM8939) is similar to the 610 graphics, but the Cortex processor cores are It has twice as many A53s – eight Cortex-A53s.
Unlike the 410, 610, 615 models, made using a 28 nm process technology, the Snapdragon 808 (MSM8992) and 810 (MSM8994) chips will be produced using advanced 20 nm technology standards. They are both built according to the big.LITTLE scheme: two (model 808) or four (810) powerful Cortex-A57 cores and four energy-efficient Cortex-A53. Graphics are provided by Adreno 418 and Adreno 430 respectively. In addition, the older Snapdragon 810 has a built-in LPDDR4 RAM controller.
But the main question is: when exactly will Qualcomm introduce its own processor architecture based on ARMv8, as it did with Scorpion and Krait (modified ARMv7)?
MediaTek MT6732, MT6752, MT6795
MediaTek could not remain on the sidelines of the 64-bit race for long, either; in just a few years it had transformed from a small manufacturer of processors for Chinese iPhone clones into one of the world's largest chipmakers, albeit a factoryless one. However, Apple and Qualcomm do not have their own either.
MediaTek MT6732 and MT6752 single-chip systems should compete with Snapdragon 610 and 615 chips. They have four and eight Cortex-A53 processor cores (frequency 1.5 and 2 GHz, respectively) and the same Mali-T760 graphics (developed by ARM Limited). The older MT6795 chip was the answer to the Snapdragon 810: big.LITTLE architecture, four Cortex-A57 and A53 cores with a frequency of 2.2 GHz, as well as a PowerVR G6200 graphics accelerator.
NVIDIA Tegra K1 (Project Denver)
NVIDIA has decided to convert its existing Tegra K1 chip to a 64-bit processor architecture. Its graphics component was already perhaps the best among its competitors - the GK20A with 192 Kepler cores, 365 GFLOPS performance and support for PC graphics standards DirectX 11.2 and OpenGL 4.4 (and not their mobile counterparts).
Instead of four 32-bit Cortex-A15 cores (plus a fifth energy-efficient core), the updated Tegra K1 single-chip system will receive two ARMv8-compatible cores of NVIDIA Project Denver proprietary architecture. The processor clock speed will increase to 2.5 GHz, and the cache size will also increase. Interesting fact: Tegra K1 graphics are about fifty times more powerful than Tegra 2.
Conclusions
ARMv8 architecture processors are capable of processing significantly more data in one clock cycle. This improves both overall processor performance and performance per watt. Considering the limitations of technological standards (maximum permissible clock frequency), switching to ARMv8 is the only possible way to increase the performance of mobile processors without going beyond reasonable limits of power consumption and heating.
Naturally, only those applications for iOS and Android that are able to use all the resources of the new processors will benefit from the ARMv8 architecture. Optimization of programs for a new architecture can be either manual or automatic, at the compiler level.
The first Android device with a 64-bit ARM processor and 4 GB of RAM is a phablet Samsung Galaxy Note 4 (. And the second, perhaps, will be the HTC series tablet computer.
Previously, smartphones were only used ARM architecture, but now Intel is already on the verge of mass production of mobile chips with x86 architecture. Which is better: ARM or x86?
Introduction and general concepts.
The x86 architecture, which is now used in almost all computers, is the CISC architecture. This means that such processors will have the following properties:
- non-fixed command length value;
- arithmetic operations are encoded in one instruction;
- a small number of registers, each of which performs a strictly defined function.
ARM uses an advanced RISC architecture. The main features of this approach are:
- load/store architecture;
- no support for non-linear (not word-aligned) memory access (now supported in ARMv6 processors with some exceptions);
- uniform 16x32-bit register file;
- fixed instruction length (32 bits) to simplify decoding by reducing code density. Later, Thumb mode increased code density;
- single-cycle execution.
If you try to run a program written specifically for a set of commands of one architecture on another processor, you may not get the desired result.
Computing power
Historically, the x86 architecture was developed with increased power in mind. Each new generation of processors became significantly more powerful, which led to rapid growth computer technology. The frequency increased, the technological process decreased, and the structure of the processor itself improved.
For a long time, energy efficiency remained a secondary consideration, while power was the main focus. The turning point occurred not so long ago, since the popularization of laptops.
Portable machines had to have a long run time.
On the contrary, the ARM architecture was used initially in portable devices, which gave it low power consumption and low power levels. A breakthrough in development has occurred in the last five years.
Modern smartphones already require a fairly high level of computing, and they also need to run for a sufficient amount of time on battery power.
If we compare energy efficiency indicators, ARM processors have an indicator of 2 TDP (a value indicating how much thermal power the cooling system of a processor or other processor should be designed to remove). semiconductor device. For example, if a CPU cooling system is rated for a 30W TDP, it should be able to dissipate 30W of heat under some given "normal conditions", and the most efficient Atom processors are around 5 TDP. This means that the most low-demand Intel processors still require twice as much power as ARM competitors.
If we talk about performance, x86 is clearly ahead of ARM. Even if you look at , you can see that single-core x86 is faster than dual-core ARM. It is also worth considering that this is the first Intel processor model in an engineering sample. Further power will only increase.
Popularity and licensing
Intel is very jealous of its x86 architecture, so no one except itself and AMD can produce x86 processors.
The situation with ARM is different. Anyone can buy a license and create their own processors, as Qualcomm, Samsung, Apple, NVIDIA and other companies do. Currently, AMD has no plans to release mobile processors, so Intel will become a monopoly on x86 processors for smartphones and tablets, which is not very good for the development of the architecture. There is serious competition in the ARM market, which leads to improvements in products from all manufacturers.
On the other hand, the Intel brand has better recognition than Qualcomm, Cortex, etc. Therefore, a buyer coming to the store and seeing the inscription “Intel inside” may prefer this device to competitors.
Conclusion
At the conclusion, the winner is usually announced, but not in this case. I believe that x86 and ARM architectures are not entirely correct to compare. Each one is good at something. In the future, the user will choose not only between the mobile OS, the manufacturer and the quality of individual components, but also between the processor architecture. Different architectures are suitable for different purposes and this must be taken into account. Although there are no Intel Medfield battery life tests yet, ARM will be ahead in this test. At the same time, ARM will not catch up with x86 in terms of pure power.
