Memory (computing)

In computing, memory is the device that retains, memorises, or stores computer data for some period of time. Memory provides one of the main functions of modern computing: the storage of information and knowledge. It is one of the fundamental components of the computer, which interconnects the central processing unit (CPU, for the acronym in English of Central Processing Unit) and the input devices /output, implement the fundamentals of the computer model of the von Neumann architecture.

Today, "memory" often refers to a form of solid-state storage, known as RAM (random access memory)., and other times it refers to other forms of fast, but temporary storage. Similarly, it refers to forms of mass storage, such as optical disks, and types of magnetic storage, such as hard drives and other types of storage, slower than RAM memories, but more permanent in nature. These contemporary distinctions are helpful, because they are fundamental to computer architecture in general.

In addition, it reflects an important and significant technical difference between "memory" and "mass storage devices", which has been diluted by the historical use of the terms "primary storage" (sometimes "main storage"), for random access memories, and "secondary storage" for mass storage devices. This is explained in the following sections, where the traditional term “storage” is used as a subheading for convenience.

Storage hierarchy

The fundamental components of general-purpose computers are the CPU, storage space, and input/output devices. The ability to store the instructions that make up a computer program and the information that the instructions manipulate is what makes computers designed according to the stored-program architecture versatile.

A digital computer represents all information using the binary system. Text, numbers, images, sound and almost any other form of information can be transformed into a sequence of bits, or binary digits, each of which has a value of 1 or 0. The most common storage unit is the byte, equal to 8 bits. A certain piece of information can be manipulated by any computer whose storage space is large enough to accommodate the corresponding data or the binary representation of the information. For example, a computer with a storage space of eight million bits, or one megabyte, can be used to edit a short novel.

Various forms of storage have been invented based on various natural phenomena. There is no universally practical storage medium and all forms of storage have their drawbacks. Therefore, a computer system contains several types of storage, each with its individual purpose.

Primary Storage

Primary memory is directly connected to the computer's CPU. It must be present for the CPU to perform any function. Primary storage consists of the primary system memory; Contains running programs and the data they operate on. Information can be transferred very quickly (typically less than 100 clock cycles) between a microprocessor register and main storage locations. Random access memories based on solid-state electronics are used in modern computers, which are directly connected to the CPU via address, data, and control buses.

The main requirement of storage is that any of its locations must be directly addressable, that is, all data contained in memory must be able to be found based on its address. This is why processor registers cannot be considered primary storage. References to these are made by name, directly, and not by address. Registers represent the current state of the computation and the data used immediately, but cannot store a program (only point to the current execution location).

The vast difference in speed between the processor and primary memory gave rise to cache memory. This is very high speed memory, typically 10 to 100 times that of primary memory, and is used to improve CPU efficiency or performance. Some of the information in main memory is duplicated in the cache. Compared to the logs, the cache is slightly slower, but larger in capacity. However, it is faster, albeit of much smaller capacity, than main memory.

Some authors present cache memory as a separate hierarchy, however, since it is not directly addressable memory (it strictly keeps copies of the information available in main memory), it is common to present it as a functional part of primary storage.

Secondary storage

Secondary memory requires the computer to use its input/output channels to access information and is used for long-term storage of persistent information. However, most operating systems use secondary storage devices as a swap area to artificially increase the apparent amount of main memory in the computer (this utilization of secondary storage is called virtual memory). Secondary memory is also called "mass storage." A hard drive is an example of secondary storage.

Secondary or mass storage memory usually has a larger capacity than primary memory, but it is much slower. In modern computers, hard drives are often used as mass storage devices. The time required to access a given byte of information stored on a magnetic platter hard drive is a few thousandths of a second (milliseconds). Instead, the time to access the same type of information in random access memory (RAM) is measured in billionths of a second (nanoseconds).

This illustrates how significant the difference is between the speed of solid-state memory and the speed of rotating magnetic or optical storage devices: hard drives are on the order of a million times slower than memory (primary). Rotating optical storage devices (CD and DVD drives) are even slower than hard drives, although their access speeds are likely to improve with technological advances.

Therefore, the use of virtual memory, which is about a million times slower than "real" memory, significantly slows down any computer's performance. Many operating systems implement virtual memory using terms like "swap file" or "cache file." The main historical advantage of virtual memory is the price; virtual memory was much cheaper than real memory; that advantage is less relevant today. Even so, many operating systems continue to implement it, despite causing significantly slower performance.

Tertiary storage

The tertiary memory is a system in which an industrial robot —robotic arm—, will mount —connect— an offline mass storage medium (see the next point) as requested by the system computer operating. Tertiary memory is used in the area of industrial storage, scientific computing, in large computer systems, and in enterprise networks. This type of memory is something personal computer users never see firsthand.

Offline Storage

Offline storage (off-line) is a system where the storage medium can be easily removed from the storage device. These storage media are often used for data transport and archiving. In modern computers, floppy disks, optical discs, and flash memories, including USB drives, are commonly used for this purpose. There are also USB hard drives that can be quickly connected. Offline storage devices used in the past are magnetic tapes in many different sizes and formats, and removable batteries of Winchester disks.

Network Storage

Network storage is any type of computer storage that includes accessing information over a computer network. Arguably, network storage makes it possible to centralize "control of information" in an organization and reduce duplication of information. Network storage includes:

• Network-related storage is a secondary or tertiary memory that resides in a computer that another one can access through a local area network, an extensive area network, a virtual private network or, in the case of online file storage, internet.
• Computer networks are computers that do not contain secondary storage devices. Instead, documents and other data are stored on a network device.

Memory characteristics

The division between primary, secondary, tertiary, and offline is based on memory hierarchy or distance from the CPU. There are other ways to characterize the different types of memory.

Volatility of information

Photo of DDR-type RAMs installed in your socket.

• Volatile memory requires constant energy to keep the information stored. Volatile memory is usually used only in primary memories. RAM is a volatile memory, as it loses information in the lack of electricity.
• The non-volatile memory will retain the stored information even if it does not receive electric current constantly, as is the case of the ROM memory. It is used for long-term storage and is therefore used in secondary, tertiary and offline memories.
• Dynamic memory is a volatile memory that also requires that periodically the stored information be refreshed, or read and rewritten without modifications.

Sequential or random access to information

Depending on the ability to access contiguous information or not, it can be classified as:

• Random access means that you can access any location of memory at any time at the same time interval, usually small.
• Sequential access means that accessing an information unit will take a variable time interval, depending on the information unit that was read before. The device may need to look (position the reading/writing head of a disk correctly), or spin (waiting the proper position to appear below the reading/writing head in a medium that rotates continuously).

Ability to change information

• The reading/writing memories or changeable memories allow the information to be rewritten at any time. A computer without a reading/writing memory as a main memory would be useless for many tasks. Modern computers also usually use reading/writing memories as a secondary memory.
• La only reading memory (Read-Only Memory, ROM) retains the information stored at the time of manufacture and
  • memory Unique Multiple Reading (Write Once Read Many, WORM) allows the information to be written once at some point after the manufacture. There are also immutable memories, which are used in tertiary and offline memories. An example is CD-ROM.
• The slow writing and fast reading memories are reading/writing memories that allow the information to be rewritten multiple times but with a much lower writing speed than reading. An example is CD-RW.

Addressing information

• In the addressable location memory, each information unit accessible individually in memory is selected with its numerical memory address. In modern computers, the addressable location memory is usually limited to primary memories, which are read internally by computer programs as the directionable location is very efficient, but difficult to use for humans.
• In the file system memories, the information is divided into computer files variable length and a specific file is located in directories and file names "readable by humans". The underlying device remains of directionable location, but the computer operating system provides the abstraction the file system to make the operation more understandable. In modern computers, secondary, tertiary and offline memories use file systems.
• In the addressable content memories (content-addressable memory), each readable information unit is individually selected with a hash value or a short identifier without relation to the memory address in which the information is stored. Addressable content memory can be built using software or hardware; hardware option is the fastest and most expensive option.

Memory capacity

Higher capacity memories are the result of rapid evolution in semiconductor material technology. The first chess programs ran on machines that used magnetic base memories. In the early 1970s, memories made by semiconductors appeared, such as those used in the IBM 370 series of computers.

The speed of computers increased, multiplied by approximately 100,000 and the memory capacity grew in a similar proportion. This fact is particularly important for programs that use transpose tables: as the speed of the computer increases, memories of proportionally greater capacity are needed to hold the extra number of positions that the program is looking for.

Processor capacity is expected to continue to increase in the coming years; It is not an abuse to think that memory capacity will continue to grow impressively. Larger capacity memories can be used by programs with larger Hash tables, which will keep the information permanently.

• Minicomputers: they are characterized by a regular basic configuration that can be composed of a monitor, disk drives, printer, etc. Its memory capacity varies from 16 to 256 KiB.

• Macrocomputers: are those that within their basic configuration contain units that provide massive information capacity, terminals (monitors), etc. Its memory capacity varies from 256 to 512 KiB, it can also have several megabytes or even gigabytes according to the needs of the company.

• Microcomputers and personal computers: with the advance of microelectronics in the 1970s it was possible to include all the components of the central processor of a computer in a single integrated circuit called microprocessor. This was the basis for creating computers that were called microcomputers. The origin of microcomputers took place in the United States from the marketing of the first microprocessors (INTEL 8008, 8080). In the 1980s the real massive explosion of personal computers began (Personal Computer PC) IBM. This machine, based on the microprocessor INTEL 8008, had interesting features that made its field of operations wider, especially because its new standard operating system (MS-DOS, Microsoft Disk Operating System) and a better optical resolution, made it more attractive and easy to use. The personal computer has gone through several transformations and improvements known as XT(Extended Technology), AT(Advanced Technology) and PS/2...

Technologies, devices and media

Semiconductor memory

Semiconductor memory uses semiconductor-based integrated circuits to store information. A semiconductor memory chip can contain millions of tiny transistors or capacitors. There are semiconductor memories of both types: volatile and non-volatile. In modern computers, main memory consists almost exclusively of dynamic, volatile semiconductor memory, also known as dynamic random access memory or more commonly RAM. With the turn of the century, there has been a steady growth in the use of a new type of non-volatile semiconductor memory called flash memory. Such growth has occurred mainly in the field of offline memories in home computers. Non-volatile semiconductor memories are also being used as secondary memories in various advanced electronic devices and specialized and non-specialized computers.

Magnetic memory

Magnetic memories use different patterns of magnetization on a surface covered with a magnetized layer to store information. Magnetic memories are non-volatile. The information is reached using one or more read/write heads. Because the read/write head only covers a part of the surface, magnetic storage is sequentially accessed and must search, circle, or both. In 'modern computers', the magnetic surface is one of these types:

• Disk, used for offline memory.
• Hard drive, used for secondary memory.
• Magnetic tape, used for tertiary and offline memory.

In 'early computers', magnetic storage was also used as main memory in the form of drum memory, core memory, core string memory, thin film memory, Twistor memory, or bubble memory. Also, unlike today, magnetic tapes used to be used as secondary memory.

Optical Disc Memory

Optical disc memories store information using tiny holes etched with a laser into the surface of a circular disc. The information is read by illuminating the surface with a laser diode and observing the reflection. Optical discs are nonvolatile and random access. The following formats are in common use:

• CD, CD-ROM, DVD: Memories of simply reading, used for massive distribution of digital information (music, video, software).
• CD-R, DVD-R, DVD+R: Unique writing memories used as tertiary and offline memory.
• CD-RW, DVD-RW, DVD+RW, DVD-RAM: Slow writing memory and fast reading used as tertiary and offline memory.
• Blu-ray: Optical disk format designed to store high quality video and data. For its development, the BDA was created, in which, among others, Sony or Phillips.
• HD DVD

The following formats have been proposed:

• HVD
• Discs Dual phase change.

Magneto-optical disk memory

Magneto-optical disks are optical memory disks where information is stored in the magnetic state of a ferromagnetic surface. Information is read optically and written by combining magnetic and optical methods. Magneto-optical disk memories are of the non-volatile type, sequential access, slow writing and fast reading. It is used as tertiary and offline memory.

Other initial methods

Cards pierced in a Jacquard loom.

Punch cards were first used by Basile Bouchon to control textile looms in France. In 1801 Bouchon's system was perfected by Joseph Marie Jacquard, who developed an automatic loom, known as the Jacquard loom. Herman Hollerith developed punch card data processing technology for the 1890 United States Census and later founded the Tabulating Machine Company, one of the forerunners of IBM. IBM developed punch card technology as a powerful tool for business data processing and produced an extensive line of recording machines that used punched paper for data storage and automatic processing.

By the 1950s, IBM cards and IBM recording machine units had become indispensable to American industry and government. During the 1960s, punched cards were gradually replaced by magnetic tape, although their use was very common until the mid-1970s with the advent of magnetic disks. Information was recorded on the cards by punching holes in the paper or card. The reading was done by electrical (later optical) sensors where a particular location might or might not be holed.

To store information, Williams tubes used a cathode ray tube and Selectron tubes used a large vacuum tube. These primary memory devices were short-lived on the market as the Williams tube was unreliable and the Selectron tube was expensive.

Delay line memory used sound waves in a substance such as Mercury to store information. Delay line memory was dynamic volatile memory, sequential cycle read/write. It was used as main memory.

Other proposed methods

Phase change memory uses the phases of a phase change material to store information. This information is read by observing the variable electrical resistance of the material. Phase change memory would be non-volatile read/write memory, random access could be used as primary, secondary and offline memory. Holographic memory optically stores information within crystals or photopolymers. Holographic memories can use the entire volume of the storage medium, unlike optical disc memories, which are limited to a small number of layered surfaces. Holographic memory could be non-volatile, sequential access, and both write-once and read/write. It can be used both as secondary memory and offline.

Molecular memory stores information in polymers that can store spikes of electrical charge. Molecular memory can be especially interesting as a main memory.

Recently it has been proposed to use the spin of an electron as memory. It has been shown that it is possible to develop an electronic circuit that reads the spin of the electron and converts it into an electrical signal.^{[citation needed]}