Microprocessor architecture: concept, types, advantages and disadvantages

The microprocessor consists of several units interconnected, each of them performs a specific function. The design and interconnection of these blocks is called architecture. The speed with which a computer can read instructions and perform appropriate calculations is determined by the operating frequency of the microprocessor. Manufacturers have made great strides in the development of architecture, which allows computers to become less dependent on frequency, that is, micro with low frequencies and speeds can perform more calculations and tasks. In the process of evolution of microprocessor architecture, it has evolved from single-core to multi-core, capable of processing several pieces of information simultaneously.

Microprocessor Generation History

Microprocessor Generation History




Fairchild Semiconductor, founded in 1957, invented its first integrated circuit in 1959, which marked the beginning of microprocessor history. In 1968, Gordan Moore, Robert Noyce, and Andrew Grove left the Children's Semiconductor Fair and founded their own company: Integrated Electronics (Intel). In 1971, the company invented the first Intel 4004.

Today there are several generations of microprocessor architectures:





  • 1st generation from 1971 to 1973. In 1971, the INTEL 4004 clocked at 108 kHz. During this period, other models on the market were used, including Rockwell International PPS-4, INTEL-8008 and National semiconductors IMP-16, they were not TTL-compatible devices.
  • 2nd generation from 1973 to 1978, implemented very efficient 8-bit microprocessors, such as Motorola 6800 and 6801, INTEL-8085 and Zilogs-Z80. Due to their ultrafast speed, they were expensive because they were based on the manufacture of NMOS technology, however, despite the price, they were very popular.
  • The 3rd generation microprocessor architecture was used from 1979 to 1980. In this period, the INTEL 8086/80186/80286 and Motorola 68000 and 68010 were developed. The speeds of these processors were four times better than their predecessors.
  • 4th generation from 1981 to 1995 - 32-bit microprocessors using HCMOS were developed. INTEL-80386 and Motorola 68020/68030 were popular processors.
  • The 5th generation started in 1995 and to the present. During this period, the market released the 64-bit architecture of the modern microprocessor, which include Pentium, Celeron, Dual and Quad Core.

Intel Celeron and Pentium Milestones

Intel Celeron and Pentium Milestones




Intel Celeron was introduced in April 1998 and belongs to a number of Intel X86 processors for personal computers. It is based on Pentium 2 and can run on all IA-32 computer programs.









Intel Celeron History:

  • 4/01/2000 - Intel Celeron 533.0 MHz;
  • 14/02/2000 - Mobile Intel Celeron 450/500 MHz;
  • 19/06/2000 - mobile Intel Celeron low voltage, 500.0 MHz;
  • 3/01/2001 - Intel Celeron, 800 MHz;
  • 2001 - Intel Celeron (1.2 GHz);
  • 2002 - Intel Celeron microprocessor architecture (1.3, 2.10, 2.20 GHz);
  • 2003 - Intel Celeron 2 / 2.55 GHz mobile processor.
  • 2004- Intel Celeron M 320 and 310 (1.3, 1.2 GHz);
  • 2008 - Celeron Core 2 DUO (Allendale).

Pentium was introduced on March 2, 1993. He changed the architecture of the Intel 486 microprocessor, the number 4 indicates the fourth generation microarchitecture in the history of microprocessors. Pentium belongs to the single-core Intel x 86, which is based on the fifth generation architecture. The name of this processor comes from the Greek word penta, which means "five."

The original Pentium processor was replaced by the Pentium MMX in 1996 and has a 64-bit data bus. A standard single transmission cycle can read or write up to 64 bits at a time. Burst read and write cycles are supported by Pentium processors. They are used for caching and transferring 32 bytes (Pentium cache line size) for 4 clock cycles. All cache operations are batch cycles for it.

CPU design

CPU design




The microprocessor architecture has many peripherals made on a single chip. It has an ALU (arithmetic logic unit), a control unit, registers, bus systems and clocks for performing computational tasks.

The microprocessor is a single chip package in which a number of useful functions are integrated and made on a single silicon semiconductor chip. Its architecture consists of a central processor, memory modules, a system bus, and an input / output unit. The system bus connects the various units to facilitate the exchange of information. In addition, it consists of data, address and control buses for the proper exchange of data, which is the reason for the deep concept of microprocessor architecture.

The central processor consists of one or more arithmetic logic units (ALU), registers, and a control unit. Based on the registers, you can also classify the generation number. ALU calculates all arithmetic as well as logical operations on the data and determines the microprocessor size, for example, 16-bit or 32-bit.

The memory block contains the program, as well as data, and is divided into a processor, primary and secondary memory. An input and output unit connects similar peripheral communication devices with a microprocessor to receive and send information.

Micro types in systems

Micro types in systems




There are several types of microprocessor architectures available for use in various systems:

  1. Shared processors. This is an additional microprocessor that works next to the main one. It is designed and optimized for use in a specific task and increases processing speed due to simultaneous work with the main one. An example of this would be a mathematical coprocessor or a floating point accelerator.
  2. The scalar processor performs calculations for a single number or data set at a time, it is installed on most modern computers and is known as single-threaded with a single stream of instructions or, for short, SISD.
  3. The CPU array of the architecture of the modern microprocessor, also known as vector, allows one instruction to work simultaneously at several data points. Known as a SIMD processor with one plural data command, it is widely used in weather forecasting and airflow modeling.
  4. A parallel processor uses independent micros to work on the same program. The process under consideration is divided into tasks, each of which can be processed by any of them. All are coordinated by a complex operating system. Programs must be specially written for parallel processing, otherwise some cannot be completed until others are closed, depending on the result of the current process.

Microprocessors are divided into five types: CISC-Complex Instruction Set, RISC microprocessor of the reduced instruction set, ASIC specialized integrated circuit, superscalar processors and DSP digital signal microprocessors.

These processors are used for encoding and decoding video or for converting DACs (digital-to-analog) and A / D (analog-to-digital). They need a microprocessor that is excellent in math. The chips of this processor are used in RADAR, home theaters, SONAR, audio systems, set-top boxes and mobile phones.

Intel x86 evolution

Intel x86 evolution




Intel x86 architecture has evolved over the years. From 29,000 transistor microprocessors 8086, the quad-core Intel Core 2 processor contains 820 million transistors, and therefore the organization and production technology have changed dramatically.

Some of the highlights of x86 architecture development:

  1. 8080 is the world's first general-purpose microprocessor. It has a microprocessor memory architecture with an 8-bit data transfer path to the memory. It was used in the first personal computer.
  2. 8086 - a 16-bit machine, much more powerful than the previous one, had real mode and 1 MB of address memory. It had a wider data path: 16-bit and larger registers, as well as a cache or queue of commands that were selected before execution.
  3. 80286 - has an addressable memory of 16 MB and contains two modes: real and 16-bit mode of the first generation. It has a data transmission width of 16 bits and the software model is also 16 bits.

80286: 16-bit microprose

80286: microprose sor




Basically, this microprose is an extended version of 8086. Therefore, before you understand 80286, you need to have a minimal understanding of 8086. Intel 8086 is a 16-bit microprocessor designed for use as a CPU in microcomputers. The term 16 bits means that its arithmetic logical unit, internal registers, instructions are designed to work with 16-bit binary codes. It has a 20-bit address bus and a 16-bit data bus. Thus, this means that it can access any of 1,048,576 memory cells and can read data or write data to memory and ports 16 or 8 bits at a time.

The 80286 microprocessor architecture is specifically designed for multi-user and multi-tasking systems. It has four levels of memory protection and supports the operating system. Its performance is more than two times higher per cycle than that of its predecessors Intel 8086/8088. Sophisticated mathematical operations take less cycles compared to 8086. It eliminates bus multiplexing and has a linear address bus with 24 address lines that can directly transfer 16 MB of memory. This is supported by the memory management module, and through it it can give out 1 GB of memory, also known as virtual. The processor includes various built-in mechanisms that can protect the system software from user programs and restrict access to certain areas of memory.

There are two operating modes for 80286. Real address mode and protected virtual address mode. Basically, in this mode, one user does not interfere with another. Also, they cannot interfere with the operating system. These functions are called protection. 80286 contains four processing units:

  1. Bus block.
  2. Block instructions.
  3. Execution unit.
  4. Address block

While the current instruction is executing, BU pre-selects the s command and stores it in a queue of six bytes. Function U is to decrypt honed instructions and save a queue of three decoded instructions. The address block calculates the address of the memory or input / output devices that must be sent for reading and writing operations. All four blocks work in parallel inside the processor. This implementation of branch prediction in microprocessor architecture is called pipelining.

The next direction of evolution is the 80386 microprocessor - the first 32-bit Intel machine. Thanks to its architecture, it was able to compete with the complexity and power of microcomputers and mainframes introduced several years earlier. It was the first processor to support multitasking and contained 32-bit protected mode. It implements the concept of swap. It has an addressable physical memory of 4 GB and a data transfer width of 32 bits.

80486: Caching Technology

Caching technology




Later in 1989, the micro 80486 entered the market and introduced the concept of caching technology and instruction pipelining. It contained a write protection function and offered a built-in mathematical coprocessor that performs complex operations from the main processor.

Varieties of micro 4 generations:

  1. Pentium-using superscalar techniques was introduced when several commands began to run in parallel. Page Size Enhancement (PSE) has been added as a minor improvement in paging.
  2. Pentium Pro - uses register renaming, branch prediction, data flow analysis, speculative execution and other pipeline stages. Advanced microcode optimization techniques and a level 2 cache have also been added. It implements second-generation address translation, in which a 32-bit virtual address is translated into a 36-bit physical memory address.
  3. Pentium II. He was able to efficiently process video, audio and graphic data using Intel MMX family microprocessor architecture technology (multimedia dataset).
  4. Pentium III - Contains SMD (Streaming Extensions) instructions and supports 3D graphics software. It has a maximum processor clock speed of 1.4 GHz and contains 70 new instructions.
  5. Pentium IV implements third-generation address changes, which translates a 48-bit virtual memory address into a 48-bit physical memory address. It contains other floating point improvements for multimedia.
  6. Core is the first microprocessor of the Intel family architecture with a dual-core processor, which is an implementation of two processors on a single chip and has additional visualization technology.
  7. Core 2 extends the architecture to 64-bit, while the 2 Quad core provides four processors on a single chip. The set of registers, as well as addressing modes, are 64-bit. It contains an electronic circuit of 1.2 million transistors. Its operating frequency for its different versions is 25, 33, 66 and 100 MHz. It is 3-5 times faster than 80386. Basically, the chip is available in two versions: DX and SX. A version of the DX type is a 32-bit processor located in a 168-pin grid array and can work with clock frequencies from 25 to 66 MHz.

The basis of the 486th device

The concept of architectures into which microprocessors are divided is complex and includes such elements as a block diagram, means of access, bit depth of interfaces, data format and interrupts.

Important additional features of the 486 processor compared to the 386 are as follows:

  1. Built-in math coprocessor. In system 386, mathematics is implemented on an external device; therefore, such instructions in 486 are executed three times faster.
  2. 8 KB of code and data cache on the chip.
  3. High conveyor performance.
  4. Executive unit.
  5. Control device.
  6. Bus interface unit.
  7. Code prefetch block.
  8. Command decoding unit.
  9. Segmentation of n-units.
  10. Paging block.
  11. Cache block
  12. Floating point block.
  13. The code prefetch block contains a 32-byte queue for storing the extracted command codes.
  14. The control unit also contains a control ROM for storing microcodes. The address specified in the program is called a logical address. It also provides 4 levels of protection to isolate and protect tasks and the operating system from each other.
  15. Scalable microprocessor architecture - the concept implies work in windows.
  16. The paging module provides a call to an object within a segment.
  17. Physical adress. The actual capacities of RAM and ROM existing in a computer are known as physical memory.
  18. Segmentation and the paging unit is a memory management unit.

RISC Blocks

RISC Blocks




RISC stands for Computer with a reduced set of instructions and is a type of architectural processor design strategy. The architecture of the RISC microprocessor refers to the method of planning and assembling the processor and may relate either to hardware or to software that is closest to the silicon on which it runs. The instruction set architecture (ISA) defines the underlying software.

The hardware architecture of a computer requires code that breaks down instructions into 0 and 1 and which a PC can understand, also known as machine code. The processor architecture may be completely different, and ISA software will reflect this. The difference between them can be found in how tasks are performed, for example, in register processing, interruption, memory addressing, external inputs and outputs.

In other words, machine code for one will not work on another. For example, the desktop version of Windows will not work on a smartphone because the architecture is different. While Microsoft has been encouraging the possible integration into a single OS for desktops, laptops, and tablets, starting with the introduction of Windows 8.

There are several types of processor architectures and associated ISAs. Some examples of RISC are ARM, MIPS, SPARC, and PowerPC. Modern processors are highly integrated and faster, so the RISC instruction sets are becoming more complex to take advantage of advanced technology.

CISC architecture

CISC architecture




In order to answer the question of what is meant by the architecture of the CISC microprocessor, you need to consider the CISC approach to the number of instructions that are executed. Its main task in this matter is to minimize the volume of the program, while sacrificing the number of cycles. Computers based on the CISC architecture are designed to reduce memory costs. Large programs need more memory, which increases the cost. , , .

MUL CISC. .

, :

  1. — , , . : ( ) . — , . — .
  2. — . , , , , .
  3. The processor performance is set by the basic law and depends on the number of instructions, CPI (cycles per instruction) and the cycle time.

Advantages and disadvantages

Advantages and disadvantages




According to the formats of the used commands (instructions), one can distinguish the main types from the classification of microprocessor architectures: RISC and CISC. The advantages of the RISC architecture are that it has a set of instructions, so higher-level language compilers can create more efficient code. This allows freedom of use of space on microprocessors due to simplicity. Many RISC processors use registers to pass arguments and store local variables. Functions use only a few parameters, and processors cannot use call instructions and, therefore, use a fixed-length method that is easy to translate.

Operation speed can be maximized, and runtime can be minimized. Fewer training formats, multiple instruction numbers, multiple addressing modes, and good scalability are required. The concept of scalable microprocessor architecture implies the use of register windows, which provides a convenient mechanism for transferring parameters between programs and returning results. A similar mechanism is implemented in SPARC.

The disadvantages of the RISC architecture are due to the fact that basically the performance of RISC processors depends on the programmer or compiler, therefore compiler knowledge plays a vital role when changing CISC code to RISC code. When CISC is rearranged into RISC code, called a code extension, the size will increase. And the quality of this extension will again depend on the compiler, as well as on the set of machine commands. The first-level cache of RISC processors is also a drawback. These processors have a large cache memory on the chip itself. They need very fast memory systems to give instructions.

The advantages of the CISC architecture — the ease of microcoding new instructions — have allowed developers to make CISC machines more compatible. As each became more advanced, fewer instructions could be used to accomplish tasks.

Disadvantages of the CISC architecture:

  1. Machine performance slows down due to the fact that the time spent by different instructions will be different.
  2. Only 20% of existing instructions are used in a typical programming event, although in reality there are various specialized instructions that are often not used.
  3. Condition codes are set by CISC instructions as a side effect of every command that takes time for this installation, and since the subsequent command changes the bits of the status code, the compiler must check the bits of the condition code before this happens.

Thus, we can draw the line that the architecture of the instruction set is an environment that provides communication between the programmer and the equipment. Part of execution and data copying, deletion or editing are user commands for a particular type of microprocessor architecture.




All Articles