Home Random Page


CATEGORIES:

BiologyChemistryConstructionCultureEcologyEconomyElectronicsFinanceGeographyHistoryInformaticsLawMathematicsMechanicsMedicineOtherPedagogyPhilosophyPhysicsPolicyPsychologySociologySportTourism






Devices of memory of the computer.

Computer memory (data storage device, storage device) – part of the computer, the physical device or the environment for the data storage, used in computation, during certain time. Memory, as well as the central processor is an invariable part of the computer from the 1940th. Memory in computing devices has hierarchical structure and normally assumes the use of several storage devices having different characteristics.

In personal computers «memory» is one of its types – dynamic memory with direct access (DRAM), – which are used as RAM of the personal computer now.

The problem of computer memory is storage in the cells of a condition of external influence, a data recording. These cells can fix the most various physical impacts. They are functionally similar to the ordinary electromechanical switch and information in them registers in the form of two accurately distinguishable conditions – 0 and 1 («is switched «off/is included»). Special mechanisms provide access (reading, any or consecutive) to a condition of these cells.

Process of access to memory is broken into the processes divided in time – record operation (a slang. an insertion, in case of record a constant memory) and reading operation; in many cases these operations occur under control of the separate specialized device — the memory controller.

Also the operation of deleting of memory is distinguished – entering (record) in cells of memory of identical values, usually 0016 or FF16.

The most known memories used in personal computers: modules of random access memory, hard disks (Winchesters), diskettes (floppy disks), CD - or DVD disks and also devices of a flash memory.

Memory functions

Computer memory provides the support of one of functions of the modern computer, – ability of long storage of information. Together with the central processor a memory are key links or architecture, – a principle put in a basis of the vast majority of modern computers.

Any information can be measured in bits and consequently irrespective of what physical principles and in what notation the digital computer (binary, figurative, decimal, etc.), numbers, text information, images, a sound, video functions and it is possible to present other types of data sequences of bit lines or binary numbers. It allows the computer to manipulate data under condition of sufficient capacity of system of storage (for example, for storage of the text of the novel of the average size it is necessary about one Mbyte)

The set of the devices intended for data storage, the most different physical effects based on use is created. The universal decision doesn't exist, everyone has advantages and the shortcomings therefore computer systems are usually equipped with several types of systems of the storage which main properties cause their use and appointment.

Physical bases of functioning

At the heart of work of a memory any physical effect providing reduction of system to two or steadier conditions can lie. In modern computer equipment physical properties of semiconductors when current passing via the semiconductor or its absence are treated as existence of logic signals 0 or 1. The steady conditions are defined by the direction of magnetization allow using of various magnetic materials for data storage. Existence or lack of a charge in the condenser also can be based on storage system. Reflection or light dispersion from a surface of CD, DVD or a Blu-ray-disk also allows storing information.



Classification of types of memory

It is necessary to distinguish classification of memory and classification of memories (STORAGE). The first classifies memory on the functionality, the second – on technical realization. Here the first is considered as hardware types of memory (realized on the STORAGE) and the structures of data realized in most cases programmatically.

Available operations with data

• Memory only for reading (read-only memory, ROM)

• Memory for reading/record

Also it is offered to refer memory to this or that look on characteristic frequency of its rewriting in practice: the types in which information often changes in the course of work belong to RAM, the types intended for storage of rather invariable data belong to ROM.

Volatility: non-volatile memory, volatile, static memory, dynamic memory (English dynamic storage)

Access method: consecutive access (SAM), direct access (RAM), purpose, buffer storage, temporary intermediate, cache memory, patch memory, control storage, shared memory, shared access memory.

Organization of address space: real (physical) memory, virtual memory, overlay able storage

Remoteness and availability to the processor: primary memory, processor registers (processor or register memory), a processor cache, secondary memory, tertiary memory

Management of the processor: directly operated (operatively available) memory, independent memory

Organization of data storage and algorithms of access to them

Classification of structures of data is the following:

· Addressed memory – addressing is carried out on location of dataassociative memory, content-addressable memory,

· CAM – addressing is carried out according to contents of data, instead of on their location.

· Pushdown storage – stack realization.

· Matrix storage – cells of memory are located so that access to them is carried out on two or more coordinates.

· Object storage – the memory, which control system is focused on storage of objects. Thus each object is characterized by type and the amount of record

· Semantic storage – data take place and written off according to some structure of conceptual signs.

Physical principles

This classification repeats the corresponding classification of the storage: semiconductor storage, magnetic storage, optical storage, laser storage, and magneto optics storage, spin Torque Transfer Random Access Memory, STT-RAM, and capacitor storage

Molecules chalcogenide uses the change of a phase status chalcogenide – the substance capable under the influence of heating and electric fields to transfer from a nonconducting amorphous status (1) in carrying out crystalline (0). The diodes of vertical type and three-dimensional crystalline structure are applied. It doesn't require preliminary deleting old data before record new, it doesn't require power supply for saving of the statusphase change memory, PCM condensers submission of electrical tension on DRAM plates.

Molecules chalcogenide uses change of a phase status chalcogenide – the substance capable under the influence of heating and electric fields to transfer from a nonconducting amorphous status (1) in carrying out crystalline (0). The diodes of vertical type and three-dimensional crystalline structure are applied. It doesn't require preliminary deleting old data before record new, doesn't require power supply for saving of the status.

Varieties of semiconductor memory: NOR, NAND, NVRAM, SRAM, DRAM, FB-DIMM, EEPROM, Flash

Varieties of a magnetic memory: mmagnetic tape memory; mmagnetic disk memory, plated wire memory, core storage.

Varieties of an optical memory: phase Change Rewritable storage, out-of-date and experimental types, fcaustic storage, mmagnetic racetrack memory, hholographic storage, cryogenic storage, fferroelectric RAM, FeRAM, eelectrostatic storage, molecular storage.

Other terms: multibank memory, logic-in-memory, multiport storage memory, multilevel memory, parallel storage, page memory

Memory hierarchy of a computer.

Hierarchy of the modern computers has two principles: principle of locality of addresses and ratio cost/productivity. The principle of locality of addresses says that the majority of programs fortunately don't execute appeals to all the commands, data is equiprobable and there is a preference of the address space.

The memory hierarchy of the modern computers is built at several levels, and higher level is less on volume, quicker and has big cost in terms of byte than lower level. Levels of hierarchy are interconnected: all data at one level can be also found at lower level, and all data at this lower level can be found at the following lower layer and so on while we won't reach the hierarchy base.

The memory hierarchy normally consists of many levels, but during every moment of time we deal only with two neighboring levels. The minimum information unit which can either be present or absent in two-level hierarchy is called the unit. The size of the unit can be either fixed or variables. If this size is fixed, the memory size is the multiple to the size of the unit.

The successful or unsuccessful appeal to higher level is called respectively the hit (hit) or miss (miss). Hit – is the appeal to object in memory which is found at higher level while the miss means that it isn't found at this level. The share of hits (hit rate) or coefficient of hits (hit ratio) is a share of the addresses found at higher level. Sometimes it is represented in percent. The share of misses (miss rate) is a share of addresses which aren't found at higher level.

As increase of productivity is the principal reason of appearance of memory hierarchy, frequency of hits and misses is the important characteristic. Address time in case of hit (hit time) is time of the appeal to higher level of hierarchy which includes, in particular, and time necessary for determination, whether the address hit or a miss. Losses on a miss (miss penalty) are time for unit substitution in higher level on the unit from lower level plus time for transfer of this unit to the required device (it is normal in the processor). Losses on a miss include further two components: time of access (access time) – time of the appeal to the first word of the unit in case of a miss, and time of transfer (transfer time) – an extra time for transfer of the remained words of the unit. Time of access is connected to a time delay of memory of lower level while time of transfer is connected to channel bandpass range between devices of memory of two adjacent levels.

 


Date: 2015-12-24; view: 258


<== previous page | next page ==>
Input-output and interruptions. | Organization of the main memory and operation
doclecture.net - lectures - 2014-2017 year. Copyright infringement or personal data (0.006 sec.)