Text 20. Simple Hardware, Complicated Logic

The computer does only additions in a sequence determined by a logical control. This seems to imply that it is a simple machine. In a sense it is, since only a handful of different types of circuits are needed to build the most awesome machine in existence. What then produces the complication? The complication is produced by the many interconnections among those few basic circuits.

A digital computer processes information. In that sense each part of it can be said to be communicating with or “speaking” to other parts of the machine. For this purpose the machine uses a “language”. Whatever form this language may take, it must pos­sess rules indicating how subjects are connected. These connec­tives are similar to connectives used in ordinary speech. Some of these grammatical connectives are “if”, “all”, “then”, “also”, “and”, “or”, “not”, “only”... These connectives are the cement of the lan­guage. They indicate the relationships among the different ele­ments of a sentence. Because they are so important for indicating relationships, they are the fundamental building blocks of logic. Now here is the important consideration. All the possible connec­tives may be reduced to only three basic ones — “and”, “or”, and “not”. Suffice it to say here that if a circuit could be developed for each of the three logical connections, we would have the necessary cement for the machine language. Then we would only need cir­cuits to store the “positions” before and after being applied through the connective circuit. These connective and storage circuits are called logical elements.

Logical elements may take on many forms. Originally, relays were used for this purpose. Early in the development of electronic comput­ers, relays were replaced almost entirely by vacuum tubes. But a new era began when the magnetic core, the transistor, and many other solid state devices were developed. These tiny components enable man to reduce the size of electronic computers to the point where they became feasible for use. In addition, the fact that only a handful of basic circuits are needed to construct a computer encouraged the development of plug-in packages. These plug-in packages simplified the maintenance problem tremendously.

Notes

hardware [’haidwea] — апаратне забезпечення

complicated ['komplikeitid] — складний

logic ['lodjik] — логіка

In a sence it is ... the most — В деякому смислі.., це складна

awesome ['oisam] machine... машина

interconnection [^ntaka'nek/n] — взаємозв’язок

suffice it to say here — Достатньо сказати...

to apply O'plai] — вживати

feasible ['firzabl] — можливий

maintenance [‘meintinans] — підтримка

Text 21. Machine Language and Language Structure

A digital computer is composed of five functional units, each performing its own function. A machine language is needed for each functional unit to communicate with the others. This same language is required by man to instruct the machine what to do. Man translates his problems into machine language by means of a program of instructions.

The language structure of a digital computer is fairly simple. Bits are encoded into digits or characters which are, in turn, en­coded into words. There are two kinds of words, the instruction word and the data word. The instruction word is active. It oper­ates upon the passive data word, a group of instruction words forming a program.

In general the instruction word consists of two parts: the com­mand, portion and the address portion. The command specifies the operation to be performed by the machine. The address portion speci­fies the location where the operand or operands are stored.

Most of the operations of the computer are performed in the auto­matic unit. The command portion of the instruction causes the arith­metic unit to perform addition, subtraction, multiplication, division, square root; or to bring operands into the arithmetic unit; or to remove results from the arithmetic unit. Some commands rearrange the information stored in the arithmetic registers. Otiier commands allow the computer to choose among different sets of instructions according to some criterion such as whether the result of a previous operation is positive or negative.

The other part of the instruction word is the address. Instruction codes differ significantly according to the number of addresses speci­fied by the instruction. One-, two-, three- and four-addresses instruc­tions have been used in digital computers. These instructions are described in the following.

One-Address Instruction. When only one address is specified, the machine executes the given command upon the operand found in the location specified by the address. Since most arithmetic operations require at least two operands — augend and addend, multiplicand and multiplier — the computer must be built so that one operand is brought into the arithmetic unit during one in­struction, and the other operand is brought into the arithmetic unit during the following instruction; during the second instruc­tion the arithmetic operation is performed. The sequencing of the instruction words in the program is performed by a program counter. As soon as one instruction is completed, the program counter is advanced by one. The new reading of the program counter specifies the address of the next instruction to be executed. The machine thus automatically chooses instructions stored in succes­sive memory locations.

Two-Address Instruction. The two-address instruction consists of a command, an address, specifying the operand, and an address specifying the location of the next instruction word to be execut­ed. The two-address-instruction machine does not need a program counter.

Three-Address Instruction. The three-address instruction consists of a command and three addresses. The first two addresses specify two operands; the third address specifies the location in memory where the result should be stored. Machines using the three-address instruction need a program counter to sequence instructions in the program.

Four-Address Instruction. The four-address instruction has a com­mand and four addresses. The first two addresses specify two ope­rands. The third address specifies the location of the result. The fourth address indicates the location of the next instruction to be executed. Machines using the four-address instruction need no pro­gram counter.

The above descriptions are true for the majority of computers. However, computers have been built which interpret their address codes differently.

Notes

functional ['fAnk/anl] — функціональний

unit — елемент, блок, одиниця

to instruct [in'strAkt] — вчити

fairly ['feali] — досить

to encode [in'koud] — кодувати, шифровати

to specify ['spesifai] — точно визначати

operand — операнд, компонент операції

to execute ['eksikju:t] — виконувати

augend ['o:gand] — перший доданок

addend [o'dend] — додаток

sequence ['si:kwans] — послідовність

counter — лічильник

... the program counter is — ... лічильник команд просувається

advanced by one на одну поділку далі

to interpret [m'tarprit] — інтерпретувати