TOPIC 5.1.: PROCESSOR COMPONENTS

THE CPU

      CPU stands for central processing unit.

      This is the real brain of the computer.

      All arithmetic is performed here, every program instruction is interpreted and executed here, and every item to be placed in memory or retrieved from memory passes through the CPU.

      However, the processor would not be much use without memory in which to hold the data and the programs which process the data.

      The CPU and the main memory of the computer are located in the same physical unit, which is often referred to as the processor unit.

INSIDE THE CPU

      A computer consists of the processor, memory, input and output units.

      The processor itself consists of two main components:

1.      The Arithmetic-Logic Unit (ALU) in which all arithmetic and logic operations are carried out.

2.      The Control Unit, which coordinates the activities taking place in the CPU, memory and peripherals, by sending control signals to the various devices.

      In addition, the CPU contains circuitry controlling the interpretation and execution of instructions.

      Special storage locations called registers are included in this circuitry to hold information temporarily while it is being decoded or manipulated.

      Some of these special purpose register are shown in the block diagram in figure 57.1.

MEMORY

      Computer memory, or storage, can be classified into two main divisions:

o       1. Main Store.

o       2. Backing Store.

      Main store, often called ‘storage’, consists of a large number of electronic circuits. If the computer is opened up a number of plastic boxes will be found, most of them housing memory circuits – up to two million in each. It has been predicted that by the end of the century it will be possible to store over a billion circuits on a chip. This is the only type of memory that the CPU can use directly, and therefore all information has to go through main store at some point.

      All other types of storage are known as ‘backing store’, or sometimes auxiliary storage, and this includes many devices ranging from paper cards with holes punched in them to laser-scanned compact disks, also known as CD ROM (Compact Disk Read Only Memory). By far the most common types of backing store in use today are magnetic disk and magnetic tape.

STORING CHARACTERS IN A COMPUTER’S MEMORY

      In all modern computers each memory circuit can be either ON, indicated by a 1, or OFF, indicated by a 0.

      A series of eight circuits in the states ON, ON, OFF, OFF, ON, OFF, ON, OFF can be depicted as: 1 1 0 0 1 0 1 0

      Data can be stored in memory character by character, with each character being represented by a ‘code’ consisting of 1s and 0s.

      Each 1 or 0 is called as a ‘bit’, short for ‘Binary digiT’.

BYTES

      Most computers use 8-bits codes to represent each character.

      8 bits together are known as one byte.

      This is sufficient to allow a unique code for all the usual characters with one bit, the leftmost bit, to spare.

      The spare bit may be used either as a ‘parity bit’ or to give an extra 128 characters.

      In this latter case, if the spare bit is 0, the code represents a normal character; if it is 1, the code represents a character from the extended character set.

      When you press a key on the keyboard, the corresponding 8-bit code is generated and sent to the computer.

      When a code is sent to the printer, the corresponding character is printed.

      One byte holds one character.

ASCII CODE

      Over the years, different computer designers have used different sets of codes for representing characters, which has led to great difficulty in transferring information from one computer to another.

      Most personal computers (PCs) nowadays use the ASCII code (American Standard Code for Information Interchange), but many mainframe computers use a code called EBCDIC (Extended Binary Coded Decimal Interchange Code).

PARITY

      For each byte of memory, most computers have at least one extra bit, called as a ‘parity bit’, which is used for error detection.

      A parity bit can detect if one of the bits in a byte has been inadvertently changed, perhaps because of a chip failure, static electricity or a voltage fluctuation.

      Computers use either even or odd parity.

      In an even parity machine, the total number of ‘on’ bits in every byte (including the parity bit) must be an even number.

      Parity is checked each time a memory location is used.

      When data is moved from one location to another in memory, the parity bits are checked at both the sending and receiving end, and if they are different or the wrong number of bits is ‘on’, an error message is displayed.

MEMORY ADDRESSING

      The memory of a computer can be thought of a series of boxes, each containing 8 bits (1 byte), and each with its own unique address, counting from zero upwards.

      The memory capacity of a computer is measured in thousand-byte units called kilobytes – although strictly speaking one kilobyte is equal to 1024, or 2¹⁰ bytes.

      For example, a PC with 640K of memory contains approximately 640,000 bytes but, if we wish to be exact, the actual size is 640 * 1024 = 655 360 bytes.

      Memory is also measured in megabytes (approximately 1 million bytes), or in the case of very large computers, gigabytes (1,000,000 Kbytes).

      These measures can be abbreviated to Kb, Mb and Gb.

THE DATA BUS AND THE ADDRESS BUS

      The CPU is connected to main memory by two sets of wires, or buses.

      When the CPU whishes to access a particular memory location, it sends this address to memory on the address bus.

      The data in that location is then returned to the CPU on the data bus.

      Microcomputer buses can typically transfer 16 or 32 bits at a time.

      A 16-bit bus, for example, has 16 lines and can transmit 16 bits simultaneously.

      The larger the number of bits that can be transferred by the bus, the faster the computer can transfer data.

      The number of lines in the address bus determines how many locations can be assessed.

      With a 16-bit address bus, the highest location that can be addressed is 65 535 (1111111111111111 in binary).

      In the 1970s when IBM brought out their PC, the designers thought it extremely unlikely that a PC would ever have more than 1Mb of memory, and that therefore a 20-bit address bus would be quite adequate.

      Unfortunately time was to prove them wrong, and a major effort had to go into designing new machine architecture to cope with over 1Mb of memory.

WORD SIZE

      The word size of a computer is the number of bits that the CPU can process simultaneously, as opposed to the new bus size which determines how many bits are transmitted together.

      Processors can have 8, 16, 32 or 64 bit word sizes (or even larger), and the word size will be one of the factors which determines the speed of the computer.

      The faster, more powerful PCs have 32-bit processors, and most mainframes have 32-bit words.

      Supercomputers may use a 64-bit or 128-bit word size.

RANDOM ACCESS MEMORY

      Memory may be one of two types, known as RAM and ROM respectively.

      RAM stands for Random Access Memory, which means that any memory location can be directly accessed by referring to its address.

      Data and programs are transferred into and out of RAM, and the data in RAM is processed according to the instructions in the program.

      There are two types of RAM memory chips: static RAM and dynamic RAM.

      Static RAM retains its information as long as the power is switched on.

      Dynamic RAM chips are smaller and simpler in design, but generally slower in operation, and the information slowly leaks away and has to be refreshed by special regenerator circuits in the CPU.

      The main memory of most computers uses dynamic RAM chips, with the more expensive and faster static RAM being used for cache memory.

      Both these types of memory lose all information as soon as the power is switched off.

MEMORY MAPPING

      Some RAM is given a special function.

      For example, part of RAM is reserved for the operating system, and some is reserved to hold addresses and parameters used, say, when calling subroutines.

      This is called a stack. This allocation of memory to different functions is referred to as memory mapping, and a typical memory map for a microcomputer is shown in figure 51.2.

READ ONLY MEMORY

      ROM (Read Only Memory) retains information even when the power is switched off, but the information has to be permanently recorded into the circuits when the chip is manufactured, and cannot be changed.

      ROM is used to store some of the operating system commands, and the instruction set of the computer.

      In special purpose computers used in video recorders, washing machines and cars, the program instructions are stored in ROM.

      It is also possible to buy Erasable Programmable Read Only Memory Chips (EPROM), which can be cleared by subjecting them to ultraviolet light, and then reprogrammed.

CACHE MEMORY

      A cache is a small fast memory usually between 1Kb and 256Kb, too expensive to be used for the whole of RAM that acts as an intermediate store between the CPU and the main memory, and is used to improve the overall speed of the computer.

      Cache memory can return data to the CPU faster than main memory because:

o       It is closer to the CPU;

o       It may use a dedicated control bus;

o       It may use higher-speed components;

      In systems which use cache memory, the most active portions of a program are kept in cache, from where instructions and data can be retrieved faster than ordinary RAM.

      Computers using cache memory are exploiting two principles:

o       An instruction or item of data that has been accessed once is likely to be accessed again the near future. (This is known as temporal locality).

o       If an item (instruction or data) is addressed at a given memory location, it is likely that other items with addresses near to this one will be accessed in the near future (spatial locality).

      There are good reasons to expect temporal locality. For example, the existence of loops in a program means that the same instruction swill have to be fetched from memory over and over again, and furthermore the data that is operated on may well be the same in many of the instructions.

      Cache memory can also be used as a temporary storage for part of a disk.

      Disk drives are mush slower at storing and retrieving information than memory, so the information that is most likely to be required next from disk is stored in cache from where it can be quickly retrieved.

RESOURCE: P M Heatcote, [A Level Computing, 3rd. Edition], Letts Educational Ltd., 1996.