What is a memory.

A processor cannot store all the information it needs while running a program, thus it depends on memory. Memory can be thought of as a series of cells, with each cell storing a piece of information. That piece of information is known as a byte. Memories are often rated in terms of their information storage capacity, which is usually measured in millions of bytes.

Memories are generally classified by the nature of the accessibility of the information and are divided into two major categories: random-access memory (RAM) and read-only memory (ROM).

Random-access memory (RAM) is a type of memory that can be read from or written to. Most RAM is actually dynamic RAMor DRAM (pronounced dee-ram). It is dynamic because it is constantly being refreshed with electrical pulses. Dynamic RAM loses its contents without constant electrical refreshing. There is a type of RAM, called static RAM, that needs very little power for refreshing, but it is not as widely used because it is more costly to produce. Because most RAM is dynamic, it is said to be volatile; that is, it loses its contents when the electric power is shut off.

Random access means that the cells of a RAM memory are organized so that the access time for any one cell is the same as that for any other cell. This is accomplished by arranging the memory cells in a two-dimensional array, in much the same way that a spreadsheet is arranged as a grid of rows and columns. Then a memory cell can be accessed by supplying an address that contains the row and column numbers of that cell.

Read-only memory (ROM) is a type of memory that can be read from but cannot be written to or altered. ROM is static or nonvolatile; that is, it retains its contents when the electric power is shut off. ROM can be used to store programs that perform specialized routines. For example, a portion of ROM might contain a program that performs division, so that when a program executes a DIVIDE instruction, it is actually executed using instructions in the ROM memory.

A bus is a circuit that provides a path for transferring information between two or more devices. In a bus-oriented computer, the processor, memory, and the input and output devices are connected by a set of three buses. When instructions and data share the same memory, they travel to and from the processor and memory over a common data bus. A second bus, called an address bus, carries the memory addresses in which specific data needed for instructions can be found. A third bus, called a control bus, is used to signal the various parts to transfer information and to signal the processor when the transfer is completed.

Memory management. In addition to physically larger and faster memories, several techniques permit memory to be used more advantageously. One such technique is virtual memory, which creates the illusion of a computer with more memory than it actually has by moving some data to disk. When a program needs to be larger than the size of available memory, the program is divided into a number of pieces. To run such a program, the first piece is brought into memory and executes until another piece of the program is needed. Then the second piece is brought in, and so on. The pieces into which a program is divided are called segmentsandpages. A segment is a variable-length piece of memory that a program allocates, whereas a page is a fixed-length piece of memory.

Segmentationis a memory management technique for allocating and addressing memory. Segments also provide structure (a program can consist of a collection of segments) and protection (other programs cannot access those segments). A segment is a convenient unit to swapinto andout of memory, but because segments are of variable length, managing them in memory and on disk is complex and time-consuming. To overcome these problems, an additional technique called paging is used.

Pagingis a memory management technique for mapping virtual addresses on disk to real addresses in memory. A program or segment that appears to be a single large piece of memory may actually be scattered throughout real memory in pages. The system handles segmentationandpaging in such a way that the distinction between memory and the disk is invisible to the program.

Caching. A cache (pronounced cash) is SRAM. It is a high speed memory that speeds up the processing of memory to the CPU. It attempts to predict what instruction is about to be used.

It is a bridge between the CPU and slower main memory. Cache is available in two different types, asynchronous and synchronous. Asynchronous cache is not synchronized with the processor. This means the processor has to wait for the data in cache to be ready before it can be retrieved. Synchronous cache is synchronized with the processor, and therefore there is no delay between the cache and the processor. In general the difference would be very hard to notice, but go with synchronous if you have the choice.

The CPU cannot retain much data because of the limited capacity of its internal registers. This means that the processor must continually exchange data with main memory. This process slows system operation considerably particularly if the CPU is operating faster internally than it does externally, for example, a 486DX/4 operates at 99MHz internally but memory access is performed at 33MHz. A technique called Caching is used to solve this problem.

There are two basic approaches to caching:

1. Internal Cache: An additional storage area is included inside the microprocessor chip itself so access is at the internal speed of the processor. This additional (cache) memory is controlled by circuitry called a cache controller. When the processor requires data from main memory, a copy is placed in the internal cache. If the processor requires this data again, it can be fetched from the cache instead of from main memory. The cache controller also fetches additional (sequential) data along with the requested data in anticipation that it will be eventually needed by the processor. This technique, called read-ahead caching - speeds memory access considerably.

2. Secondary (External) Cache: Recall, a PCs main memory is DRAM which has relatively slow access time. When memory can't supply data fast enough to the CPU, the processor must pause for one (or more) clock cycle causing non-productive wait states. A single wait state slows data transfer rates by 50%. To minimise this problem, most modern computers install secondary cache memory consisting of the faster SRAM devices. An external cache controller chip is required to manipulate the secondary cache.

Today, most computer systems use a memory hierarchy, that is, a series of levels of memory ranked one above the another on the basis of speed of access. Visualize a pyramid with a processor at the very top. At the next lower level of the hierarchy is the cache. At the next lower level is RAM memory, and beneath that is a RAM disk—a reserved section of relatively slow RAM that is used to store information normally located on an even slower disk. At the bottom of the pyramid is disk storage.

The memory hierarchy works as follows. A large program is divided into pages. The cache memory stores the most often used pages, the RAM memory and RAM disk will store the next most often used pages, and the least used pages are stored on disk. The combination of techniques can speed up processing because the number of times the program has to fetch instructions and data from RAM and disk memory is reduced.

2.4 Give synonyms of the following words.

Divide, major, need, alter, perform, instruction, connect, complete, permit, project, frequently, to fetch.

2.5 Complete the sentences using the words from the box. You can use some words more than once.

  a cache memory ROM RAM virtual memory  

1. ­­___________ is a type of chip into which instructions are permanently embedded at the time of manufacture.

2. ­___________ can be written to or read from and is often used for temporary storage of programs and data during processing.

3. ­___________ is a special technique for expanding the capacity of main memory.

4. A very high –speed memory called ­___________may have access times of less than 25 nanoseconds.

5. Information stored in ­___________ is lost when the computer is turned off.

2.6 Give English equivalents of the following words and word – combinations.

Ряд комірок, керування пам’яттю, сторінкова організація, віртуальна пам’ять, двовимірний масив, завантажувати, енергонезалежний, розвантажувати, тривалість доступу, прямий/довільний доступ, спільна шина даних, енергозалежний, доступність інформації, виконувати спеціальні (стандартні) програми, створювати ілюзію, переміщати дані на диск.

2.7 Translate into English in writing.

Так як розміри транзисторів в BiCMOS – структур у мікросхемах нових поколінь постійно зменшуються у розмірах, то і шар диоксиду кремнію повинен ставати все тоншим і тоншим. Проте товщина ізолюючого шару неухильно наближається до межі, за якою суттєво погіршується електричні якості цього традиційного ізолятора напівпровідникових інтегральних схем.

Найбільш вдалу пропозицію для вирішення цієї проблеми зробили спеціалісти фізичної дослідницької лабораторії (PSRR, Physical Sciences Research Lab Motorola) – використовувати матеріал, яких за своїми електричними властивостями схожий на діоксид кремнію, але має набагато менші фізичні розміри. Цей матеріал – титанат стронцію відноситься до класу мінералів – перовскітів (perovskites).

Він має діелектричні показники, які перевищують діоксид кремнію більше ніж в 10 разів.

У лабораторії цей матеріал одержують штучно, нарощуючи атоми шар за шаром, що дає можливість одержати чисті і майже ідеальні кристали перовскітів.

Унікальні діелектричні характеристики дозволяють у 3-4 рази зменшити товщину елементарних транзисторів кристалу в порівнянні з використанням діоксиду кремнію.

Перспективи використання перовскітів дуже широкі. На їх основі можна проектувати напівпровідникові мікросхеми широкого вжитку з дуже низьким енергоспоживанням.

2.9 Answer the questions.

1. What is memory?

2. What are the main characteristics of ROM?

3. Why is RAM said to be volatile?

4. In what way are the cells of RAM organized?

5. What are the functions of data, address and control buses?

6. What do we call a memory which creates the illusion of a computer with more memory than it really has?

7. What are the techniques for memory management? What is the difference between them?

8. What type of memory holds the most –used data?

9. How does a memory hierarchy work?

Дата добавления: 2015-01-03