RAM: What Is It?
Random Access Memory, or RAM for short, is the hardware in a computer device that stores the operating system (OS), open applications, and data that is now being used so the CPU can access them fast. RAM is a computer's primary memory. Compared to other types of storage, such an optical drive, solid-state drive (SSD), or hard disc drive (HDD), it is far faster to read from and write to.
RAM, or random access memory, is erratic. In other words, information is stored in RAM while the machine is running and is deleted when it is shut down. Upon restarting the computer, the operating system and other data are reloaded into RAM, often from an HDD or SSD.
Function of RAM
RAM is not suitable for storing persistent data due to its instability. A hard drive is comparable to a person's long-term memory, and random access memory (RAM) to their short-term memory. Although short-term memory is concentrated on tasks that require immediate attention, it can only hold a certain amount of information at once. Facts from the brain's long-term memory can be recalled to replenish an individual's short-term memory when it becomes overloaded.
This is also how computers operate. When RAM is full, the computer's CPU has to keep accessing the hard drive in order to replace the outdated data stored in RAM with fresh information. The computer's performance is slowed by this procedure.
RAM never runs out of memory, unlike a computer's hard disc, which can fill up with so much data that it can no longer hold any more. Nevertheless, RAM and storage memory combined may run out of space entirely.
How does RAM work?
When referring to RAM, the phrase "random access" refers to the ability to access any storage location—also referred to as any memory address—directly. The phrase "random access memory" was first used to describe the difference between offline memory and ordinary core memory.
The term "offline memory" usually applied to magnetic tape, from which a particular piece of data could only be retrieved by progressively identifying the address, commencing at the tape's beginning. RAM can store and retrieve data directly to and from designated areas because of the way it is arranged and managed.
In essence, RAM is comparable to a group of boxes, each of which has the capacity to store a 0 or a 1. You can get the address of each box by counting down the rows and across the columns. An array is a collection of RAM boxes, with each box referred to as a cell.
The RAM controller uses a small electrical wire etched into the chip to convey the column and row addresses down to a particular cell. In a RAM array, one address line corresponds to each row and column. Read data travels back on a different data line.
In contrast, a hard drive uses a magnetised surface that resembles a vinyl record to store data. As an alternative, an SSD uses nonvolatile memory chips to store data instead of RAM. They don't require continuous power, and if the power is cut off, the data won't be lost. Memory modules are composed of a collection of RAM microchips. These fit into motherboard slots in computers. A bus, which is a collection of electrical pathways, is utilised to link the CPU and motherboard slots.
Users may add RAM modules to most PCs up to a specific amount. A computer with more RAM reduces the amount of times the CPU needs read data from the hard drive, which is a slower process than reading data from RAM. Whereas store memory access time is measured in milliseconds, RAM access time is measured in nanoseconds.
What much of RAM is required?
The user's activity determines how much RAM is required. For instance, a PC with at least 16 GB of RAM is advised while editing videos, however more is preferable. Adobe recommends a machine with at least 3GB of RAM in order to run Photoshop CC on a Mac for picture processing. Even 8GB of RAM, though, might cause lag if the user is using other programmes at the same time.
RAM types
RAM is available in two main types:
- Dynamic Random Access Memory (DRAM) comprises the RAM found in most computers, and as was already said, for it to keep its stored data, the power must be on. An electrical capacitor stores the charge or lack thereof in each DRAM cell. To account for leakage from the capacitator, this data has to be continuously updated every few milliseconds using an electrical charge. Transistors act as gates, deciding whether the value of a capacitor is writeable or readable.
- Static Random Access Memory (SRAM) also need continuous electricity to store data, but unlike DRAM, it does not require frequent refreshing. In SRAM, the transistor functions as a switch, with one position designating 1 and the other position designating 0, in place of a capacitor retaining the charge. In contrast to dynamic RAM, which requires just one transistor per bit, static RAM requires several transistors to store a single bit of data. Because of this, SRAM chips cost significantly more and are bigger than a comparable quantity of DRAM.
SRAM, on the other hand, is substantially quicker and requires less energy than DRAM. Static RAM is mostly utilised in modest numbers as cache memory inside a computer's CPU due to price and speed differences.
RAM's past: RAM vs. SDRAM
Because the RAM microchips had a different clock speed than the computer's CPU, RAM was initially asynchronous. This became an issue when CPU power increased and RAM was unable to keep up with the processor's demands for data.
Clock speeds were synchronised with the advent of synchronous dynamic RAM (SDRAM) in the early 1990s. Computers were able to do jobs quicker by synchronising the inputs from the processor with the memory.
But the first single data rate SDRAM (SDR SDRAM) swiftly hit its limit. Double data rate synchronous Random Access Memory (DDR SRAM) was invented about 2000. This sent data at both the beginning and the conclusion of a single clock cycle.
DDR SDRAM has undergone three iterations: DDR2, DDR3, and DDR4. With each iteration, data throughput speeds have increased and power consumption has decreased. But because data is processed in greater chunks with each cycle of DDR, the versions have not been compatible with one other.
GDDR SDRAM
Graphics double data rate (GDDR) SDRAM is utilised in video and graphics cards. The technique allows data to be transported at different moments throughout a CPU clock cycle, just as DDR SDRAM. Nevertheless, compared to DDR SDRAM, it operates at greater voltages and with less precise timing.
Tight access times are less important for parallel workloads like 2D and 3D video rendering, and GDDR can provide the greater speeds and memory bandwidth required for GPU performance.
Like DDR, GDDR has undergone many generations of development, with each iteration offering reduced power consumption and increased performance. The newest type of graphics memory is called GDDR6.
Virtual memory vs. RAM
Memory issues can arise on a computer, particularly when several programmes are open at once. Operating systems can create virtual memory to make up for deficiencies in actual memory.
Data is temporarily moved from RAM to disc storage via virtual memory, and contiguous addresses that include a programme and its data are formed by using both RAM's active memory and an HDD's inactive memory to expand virtual address space. A system may execute more complex programmes or several programmes at once using virtual memory, giving each the impression of having endless memory without the need to add extra RAM.
RAM can accommodate twice as many addresses as virtual memory. Virtual addresses are used to hold programme data and instructions until the programme is performed, at which point the addresses become real memory addresses.
Data must be transferred between virtual and physical memory, which can slow down a computer. This is one drawback of virtual memory. Programmes operate directly from RAM when working with just physical memory.
RAM vs. flash memory
Solid-state semiconductors make up both RAM and flash memory. However, because to variations in their manufacturing processes, performance requirements, and prices, they have distinct functions in computer systems. Storage memory is facilitated by flash memory. RAM serves as active memory, calculating using data that has been accessed from storage.
The requirement to delete data from NAND flash memory in whole blocks distinguishes it significantly from RAM. Because of this, it operates more slowly than RAM, where data may be removed bit by bit.
On the other hand, NAND flash memory is nonvolatile and less costly than RAM. It can store data even in the absence of electricity, unlike RAM. Due to its nonvolatility, slower speed, and cheaper cost, flash is frequently utilised as the storage memory in SSDs.
RAM vs. ROM
Computer memory that contains data that can only be read, not written to, is known as read-only memory, or ROM. Boot-up software is stored in ROM and is activated whenever a computer is turned on. In most cases, it cannot be changed or reprogrammed.
Since ROM data is nonvolatile, it remains intact even if the computer's power is switched off. Consequently, read-only memory is employed to store data permanently. However, data stored in Random Access Memory is only stored momentarily. RAM typically has many gigabytes of storage, whereas ROM typically has few megabytes.
Trends and future directions
Nonvolatile storage that has the ability to change the resistance of the solid dielectric substance that makes up its structure is known as resistive random access memory, or ReRAM. ReRAM devices have a memristor, which changes resistance in response to applied voltages.
Oxygen vacancies are physical flaws that ReRAM introduces into an oxide layer. Like the electrons and holes in a semiconductor, these vacancies represent two values in a binary system.
When compared to other nonvolatile storage technologies, such NAND flash, ReRAM offers a faster switching speed. In addition, it promises to use less power than NAND flash and have a high storage density. For memory in sensors used in industrial, automotive, and internet of things applications, this makes ReRAM a viable choice.
For years, vendors have battled to advance ReRAM technology and start producing chips. They are now being shipped by a few sellers.
Eventually, 3D XPoint technology—like Intel's Optane—might bridge the gap between NAND flash memory and dynamic RAM. In the transistor-less, cross-point architecture of 3D XPoint, memory cells and selectors are located where two perpendicular wires converge. Although 3D XPoint is nonvolatile memory, it is not as quick as DRAM.
3D XPoint technology falls between slower, less expensive NAND flash and faster, more expensive DRAM in terms of both price and performance. Technology may make it more difficult to distinguish between RAM and storage as it advances.
5G and the market for RAM
The JEDEC Solid State Technology Association released the Low Power Double Data Rate 5 (LPDDR5), or JESD209-5, in February 2019. In the end, LPDDR5 will function at an I/O rate of 6400 MT/s, which is 50% faster than LPDDR4's first iteration. This will improve memory efficiency and speed considerably for a wide range of applications. This covers portable computer devices including tablets, cellphones, and very tiny laptops.
In contrast to LPDDR4, which was released in 2014 with a data rate of 3200 MT/s, LPDDR5 was published with a data rate of 6400 MT/s.
Samsung Electronics started mass manufacturing the first 12-gigabit LPDDR5 mobile DRAM in the industry in July 2019. Samsung claims that it has been tuned to enable 5G and AI capabilities in next handsets.
RAM Price
Even though DRAM prices had dropped from their previous highs, they were still unstable by the summer of 2019. The volatility was caused by several factors, including:
- an abundance of supplies
- tensions in the markets of South Korea and Japan, which are home to Samsung and SK Hynix, the two biggest manufacturers of memory chips worldwide.
- the release of the LPDDR5, a next-generation mobile chip
- the growing use of 5G technologies
- an expected rise in the market for wearable technology and consumer gadgets connected to the Internet of Things (IoT), such cars
