Detailed explanation of the technical architecture of the four major mainstream CPU processors

Latest update time：2020-03-16

Reads：

RISC (Reduced Instruction Set Computer) is a microprocessor that executes fewer types of computer instructions. It originated from the MIPS host (i.e. RISC machine) in the 1980s. The microprocessors used in RISC machines are collectively called RISC processors.

This allows it to perform operations at a faster rate (more millions of instructions per second, or MIPS).

Because each type of instruction a computer needs to execute requires additional transistors and circuit elements, a larger computer instruction set makes the microprocessor more complex and slower to perform operations.

Performance feature 1: Due to the simplification of the instruction set, pipelines and common instructions can be executed by hardware;

Performance feature 2: A large number of registers are used, so that most instruction operations are performed between registers, which improves the processing speed;

Performance feature three: It uses a cache-host-external memory three-level storage structure to execute the data fetch and store instructions separately, so that the processor can complete as much work as possible without slowing down the processing speed due to accessing information from the memory. Among them, ARM/MIPS/PowerPC are all based on the architecture of reduced instruction set machine processors; X86 is based on the architecture of complex instruction set, and Atom is x86 or a simplified version of the x86 instruction set.

According to various news, the current status of Android in supporting various processors:

ARM+Android is the earliest developed and perfect support, mainly in the mobile phone market, netbook, smart phone and other markets;
X86+Android has a relatively complete development. There are netbooks with atom+Android, and they support
Atom+Android and Atom+Window7 dual systems;
MIPS+Android is currently in the process of porting and improving;
Powpc+Android is currently in the process of porting and improvement.

ARM series processors

The ARM architecture, formerly known as Advanced RISC Machine (Acorn RISC Machine), is a 32-bit reduced instruction set (RISC) processor architecture that is widely used in many embedded system designs.

Due to its energy-saving characteristics, ARM processors are very suitable for the mobile communications field, which is in line with its main design goal of low power consumption.

Today, the ARM family accounts for 75% of all 32-bit embedded processors, making it one of the most common 32-bit architectures in the world.

ARM processors can be found in many consumer electronic products, from portable devices (PDAs, mobile phones, multimedia players, handheld electronic games, and computers) to computer peripherals (hard drives, desktop routers), and even in military facilities such as missile onboard computers.

There are also some derivative products based on ARM design, and important products include Marvell's XScale architecture and Texas Instruments' OMAP series.

Advantages: low price and low energy consumption;

ARM licensing method: ARM itself does not rely on its own designs to manufacture or sell CPUs, but instead licenses the processor architecture to interested manufacturers.

ARM provides a variety of licensing terms, including pricing and distribution. For the licensee, ARM provides an integrated hardware description of the ARM core, including complete software development tools (compiler, debugger, SDK), and the right to sell silicon chips containing ARM CPUs.

For fabless licensees who wish to integrate ARM cores into their own chip designs, they usually only aim to obtain a production-ready IP Core certification.

For these customers, ARM releases the gate schematics of the selected ARM cores, along with abstract simulation models and test programs to aid design integration and verification.

Customers with more demands, including integrated device manufacturers (IDMs) and wafer manufacturers, choose to obtain processor intellectual property (IP) in the form of synthesizable RTL (register transfer level, such as Verilog).

With integrable RTL, customers have the ability to optimize and enhance the architecture. This approach allows designers to achieve additional design goals (such as high oscillation frequency, low power consumption, instruction set extension, etc.) without being restricted by an unchangeable circuit diagram.

Although ARM does not grant licensees the right to resell the ARM architecture itself, licensees can sell finished products (such as chip components, evaluation boards, complete systems, etc.) at will.

Merchant foundries are a special case because not only are they licensed to sell finished silicon containing ARM cores, they also typically reserve the right to re-manufacture the ARM cores for other customers.

x86 series/Atom processor

x86 or 80x86 is a general term for a microprocessor architecture first developed and manufactured by Intel.

The x86 architecture is primarily a variable instruction length CISC (Complex Instruction Set Computer).

Intel Atom (Chinese: 凌动, development code: Silverthorne) is an ultra-low voltage processor series from Intel. The processor is manufactured using a 45 nanometer process and integrates 47 million transistors. The L2 cache is 512KB, supports the SSE3 instruction set, and VT virtualization technology (some models).

Currently, there are 6 models in the Atom processor series, all of which belong to the Z500 series. They are Z500, Z510, Z520, Z530, Z540 and Z550. The lowest-end Z500 has a core frequency of 800MHz and a FSB of 400MHz. The highest-speed Z550 has a core frequency of 2.0GHz and a FSB of 533MHz.

Starting from Z520, all processors support Hyper-Threading technology, but it only increases power consumption by less than 10%. The dual-core version is the N series, which still uses the 945GC chipset.

The dual-core version will still support Hyper-Threading technology, so the system will show that there are 4 logical processors. The two cores of this version are not natively designed, but simply two single cores packaged together.

MIPS series processors

MIPS is a very popular RISC processor in the world. MIPS means "Microprocessor without interlocked piped stages". Its mechanism is to use software methods to avoid data-related problems in the pipeline as much as possible.

It was first developed in the early 1980s by a research team led by Professor Hennessy of Stanford University. MIPS's R series is a RISC industrial product microprocessor developed on this basis. These series of products are used by many computer companies to form various workstations and computer systems.

MIPS Technologies is a well-known chip design company in the United States. It uses Reduced Instruction Set Computing Architecture (RISC) to design chips. Compared with the Complex Instruction Set Computing Architecture (CISC) used by Intel, RISC has the advantages of simpler design and shorter design cycle, and can apply more advanced technologies to develop faster next-generation processors.

MIPS is one of the earliest commercial RISC architecture chips. The new architecture integrates all the original MIPS instruction sets and adds many more powerful functions. MIPS only designs the CPU itself and then licenses the design to customers, allowing them to manufacture high-performance CPUs.

In 1984, MIPS Computer Corporation was founded and began designing RISC processors;
The R2000 processor was launched in 1986.
In 1992, SGI acquired MIPS Computer Corporation.
In 1988, the R3000 processor was launched.
In 1991, the first commercial 64-bit microprocessor, the R4000, was launched; subsequently, models such as the R8000 (in 1994), R10000 (in 1996) and R12000 (in 1997) were launched.
In 1998, MIPS separated from SGI and became MIPS Technologies; subsequently, MIPS's strategy changed to focus on embedded systems; 1998 - MIPS Technologies' stock was publicly listed on the NASDAQ Stock Exchange in the United States.
In 1999, MIPS released the MIPS32 and MIPS64 architecture standards, laying the foundation for the future development of MIPS processors. The new architecture integrates all the original NIPS instruction sets and adds many more powerful functions. MIPS has successively developed the high-performance, low-power 32-bit processor core MIPS324Kc and the high-performance 64-bit processor core MIPS64 5Kc.
In 2000, MIPS released versions for the MIPS32 4Kc and the 64-bit MIPS 64 20Kc processor core.
August 16, 2007 - MIPS Technology announced that the Chinese Academy of Sciences Institute of Computer Science's Loongson CPU had obtained all patents and bus and instruction set licenses for its processor IP.
December 20, 2007 - MIPS Technologies announced that ALi Corporation has licensed its customizable system-on-chip (SoC) core "MIPS32 24KEc Pro" designed for advanced multimedia.

PowerPC series processors

PowerPC is a central processing unit (CPU) with a reduced instruction set computing (RISC) architecture. Its basic design is derived from the IBM PowerPC 601 microprocessor POWER architecture. In the 1990s, IBM, Apple, and Motorola successfully developed PowerPC chips and manufactured multi-processor computers based on PowerPC. The PowerPC architecture is characterized by good scalability, convenience, and flexibility.

PowerPC 处理器有广泛的实现范围，包括从诸如 Power4 那样的高端服务器 CPU 到嵌入式 CPU 市场(任天堂 Gamecube 使用了 PowerPC)。PowerPC 处理器有非常强的嵌入式表现，因为它具有优异的性能、较低的能量损耗以及较低的散热量。除了象串行和以太网控制器那样的集成 I/O，该嵌入式处理器与“台式机”CPU 存在非常显著的区别。

Real-time DSP architecture

DSP is a kind of microprocessor, which has a very high processing speed. Because the occasions where this type of processor is applied require high real-time performance. For example, if the processing speed is not fast enough, you can only wait for the other party to stop talking before you can talk.

If both parties are talking at the same time, the signal connection can only be closed because the digital signal processing speed is not fast enough. Before the emergence of DSP, digital signal processing can only be done by MPU (microprocessor). However, the low processing speed of MPU cannot meet the requirements of high-speed real-time.

因此，直到70年代，有人才提出了DSP的理论和算法基础。那时的DSP仅仅停留在教科书上，即便是研制出来的DSP系统也是由分立元件组成的，其应用领域仅局限於军事、航空航天部门。

90年代DSP发展最快，相继出现了第四代和第五代DSP器件。现在的DSP属於第五代产品，它与第四代相比，系统集成度更高，将DSP芯核及外围元件综合集成在单一芯片上。这种集成度极高的DSP芯片不仅在通信、计算机领域大显身手，而且逐渐渗透到人们日常消费领域。

Disclaimer

All information and charts published on this platform are for reference only. The publication of these documents does not constitute an invitation or intention to acquire, purchase, subscribe, sell or hold any shares. The profits and losses caused by investors' financial, securities and other investment projects based on the information, materials and charts provided by this website have nothing to do with this website. In addition to original works, the articles, pictures, videos and music used on this platform belong to the original rights holders. Due to objective reasons, there may be cases of improper use, such as some articles or part of the quoted content of the articles failed to contact the original author in time, or the author's name and original source were marked incorrectly, etc., which is not a malicious infringement of the relevant rights of the original rights holder. Please understand and contact us for timely processing to jointly maintain a good network creation environment.