Hardware and software architecture of IBM-compatible technologies 3.1. Microprocessors

The central processor is an integral part of any computer. This is usually a large integrated circuit, which is a silicon crystal in a plastic, ceramic or metal-ceramic package, on which are located the terminals for receiving and issuing electrical signals. CPU functions are performed by microprocessors. They perform calculations, transfer data between internal registers and control the course of the computational process. The microprocessor interacts directly with the OP and controllers of the motherboard. The main information carriers inside it are registers.

An integral part of the microprocessor are:

• an ALU consisting of several blocks, for example, an integer processing block and a floating point processing block;

• a control device that generates control signals for executing commands;

• internal registers.

The basis of each microprocessor unit based on the principle of the pipeline, which consists in the following. The implementation of each machine command is divided into separate stages, and the execution of the next program command can be started before the previous one is completed. Therefore, the microprocessor simultaneously executes several consecutive program commands, and the time to execute a block of commands is reduced several times. An architecture is called superscalar , which is based on the principle of a pipeline. This is possible if there are several processing units in the microprocessor.

The program may contain control transfer commands, the execution of which depends on the results of the preceding command execution. In modern microprocessors using the pipeline architecture provides mechanisms for predicting transitions. In other words, if in the queue of commands a com *** and a conditional transition appear, then it is predicted which comma *** will be executed next before determining the transition sign. The selected program branch is executed in the pipeline, but the result is recorded only after calculating the transition feature, then, when the transition is selected correctly. In case of a wrong choice of the program branch, the microprocessor goes back and performs the correct operations in accordance with the calculated transition flag.

Important characteristics of the microprocessor are:

• its speed, which largely depends on the microprocessor clock frequency;

• microprocessor architecture, which determines what data it can process, which machine instructions are included in the set of commands executed by it, how data is processed, what is the internal memory of the microprocessor.

The composition of the microprocessor may include a cache memory (super-efficient), which provides faster transfer of information than the OP. Distinguish between the first level cache, which is usually built into the same crystal and operates at the same frequency as the microprocessor; L2 cache - shared, when commands and data are stored together, and shared, when they are stored in different places.

When solving complex mathematical and physical problems, some computers provide for the use of a special device called a mathematical coprocessor. This device is a specialized integrated circuit operating in conjunction with the CPU and designed to perform mathematical operations with a floating point.