Computer memory has come a long way in the last 30 years. From a maximum of 2 MB in 1990 to over 64 GB in the 2020s. But why does RAM always come in powers of 2?

RAM comes in powers of 2 because computer memory and storage chips are made of tiny electronic switches with two possible states, on and off. As a result, the capacity of these chips is always measured in powers of 2. The actual size of memory is always lower than what the manufacturer reports.

This article will explain why RAM and computer storage always come in powers of 2. I’ll also explain to you the difference between computer memory and storage.

common ram amounts in powers of 2

But first, I’ll explain what RAM is, and how it works, so the explanation will hopefully be clearer as to why RAM comes in powers of 2.

What is RAM?

ram modules stacked together

“RAM” is an acronym for “random-access memory,” a type of computer memory that can be read in any direction from any point in a directory. Random access makes it much faster than other computer memory and storage types. It performs a similar function to human short-term memory.

Random-access memory, or “RAM,” is a type of volatile computer storage used to run surface-level operations of a computer.

Here, surface-level means the applications and other operations actively used on the computer. When RAM fills up, information is transferred to the computer’s storage drive.

Modern computers typically use two types of RAM, dynamic (DRAM) and static (SRAM). 

DRAM performs most of the surface computing duties. It stores more data and can be rewritten faster. DRAM also costs less than SRAM.

SRAM serves as cache and runs background operations. It stores less information, but it can be accessed much faster than DRAM. SRAM requires larger chips and is more costly.

SRAM is used for accessing smaller pieces of information that are needed more frequently. DRAM is used for everything else.

How does RAM work?

RAM chips are made of many binary cells, where a cell with a charge represents a “1,” and an uncharged cell represents a “0”. Each cell represents one binary digit or “Bit.” Modern computers arrange cells in rows of eight, representing one byte.

RAM is a binary system, meaning it only has two possible states—on and off. In computers, the “on” state is represented by a “1”, and the “off” state is represented by a “0”.

In RAM, the “on” state is represented by an electric charge in a single cell. Each cell represents a single digit of binary code or a “bit.” A sequence of eight bits in a “byte.”

Each byte has 28, or 256 possible configurations. In binary code, 16 bits —two bytes— equals a word. However, newer computers use 32-bit/4 byte words.

The exact mechanism used depends on the type of RAM. Dynamic RAM (DRAM) cells consist of microscopic transistors and capacitors.

The transistors decide where electrical power is sent, and the capacitor stores the charge.

DRAM cells are tiny, and thousands or millions of them can be stored on a chip the size of a postage stamp.

The micro capacitors in DRAM cells quickly lose their charge and must be refreshed frequently. That makes their access speed slower, but their rewrite speed faster.

Static RAM (SRAM) cells are made of clusters of four transistors. The destination of electrical power in SRAM cells is decided through Boolean logic, which is very complicated and outside of the scope of this article.

SRAM chips are larger and more expensive than those of DRAM.

SRAM cells hold onto their charge as long as the chip is powered. That makes their access speed faster but slows their rewrite speed.

Modern RAM cells are made of MOSFETs. “MOSFET” is the acronym for “metal-oxide-semiconductor field-effect transistor.”

Rather than being a physical switch, a MOSFET is an electronic switch triggered by applying an electric current. The MOSFET retains a small amount of electric charge, serving as a memory device.

Why RAM comes in powers of 2

Each RAM cell has two possible states, “on” and “off,” represented by “1” and “0” in binary code.

Each additional cell multiples the number of possible states by 2. As a result, RAM is always measured in powers of 2 and results in an even number.

A single RAM cell has two, or 21 possible states: “on” and “off.” Two RAM cells together have four, or 22 possible states: “off-off,” “on-off,” “off-on,” and “on-on.” Three cells together have eight possible states, and so on.

As a result of being built on binary circuits, the size of RAM is always an even number and a power of two. However, as we’ll discuss later, not all of the stated capacity is available for use.

How much RAM does a computer actually need?

open ram slots on a motherboard waiting for ram modules to be installed

The amount of RAM needed by a computer depends on what it is being used for. Most day-to-day computer operations can be performed with 8 GB of RAM. High-end computers used for gaming or video editing usually need from 16 GB to 32 GB.

In general terms, the more RAM a computer has, the faster it can run. Some applications use a surprising amount of memory—for example, Google Chrome.

Most modern video games advertise a system requirement of 16 GB of RAM. However, in practice, this is usual.

Many games will run on an 8 GB machine with minimal performance loss.

Is 8 GB of RAM enough?

8 GB is enough for most common tasks on a computer. 8GB is the minimum amount of memory needed for modern gaming. Heavier uses, like video editing, may require 16GB and over.

Gaming is a fair benchmark for computer activity.

Whereas word processing and web browsing use don’t use as much memory, computer-aided design (CAD) and other advanced forms of simulation software can use require a lot more.

Is 64 GB of RAM overkill?

64 GB of RAM is overkill for most applications, including gaming. Most current computer games are designed to run on 16 GB or 32 GB systems. 64 GB computers are typically used for editing high-resolution (4K) video and high-level professional simulation software.

Suppose you are editing the next IMAX nature documentary or simulating the effects of the next New Madrid earthquake on the continental USA.

In that case, a computer with 64 GB of RAM may be right for you.

Otherwise, you’ll be fine with 16GB.

What is the difference between computer memory and storage?

ram modules stacked in a pile with hard drive on top

Computer memory is volatile storage, while storage stores data for permanent access. The information stored in memory remains there as long as an electrical charge is applied. Computer memory is the equivalent of human short-term memory, and computer storage is the equivalent of long-term memory.

Computer storage, also called the computer’s “hard drive,” is analogous to human long-term memory. Computer storage is non-volatile and retains information after the device is turned off.

Where computer memory is used to perform surface-level operations, storage holds all the information needed to perform those tasks.

Without non-volatile computer storage, a computer’s operating system would have to be manually entered into memory before any operations could be performed.

Usually, computer memory is much faster than storage, but new technology is eroding this difference.

How does computer storage work?

a hard drive taken apart and placed on top of a electronics diagram

Computer storage uses binary language to encode electronic data into a non-volatile physical media, often referred to as a “hard drive.” Depending on the type, computer storage uses either magnetic fields, semiconductors, or laser beams to encode data onto the storage medium. 

The most common type of computer storage is the magnetic disk, commonly called the “hard disk drive,” “hard drive,” or “hard disk.”

Hard disk drives use microscopically focused magnetic fields to alter the orientation of particles in groves on the surface of a disk made of an exotic metal alloy.

The direction of the particles determines whether the data is encoded as a “1” or a “0.”

There are several other types of computer storage. The two most familiar to most people are optical discs (CDs and DVDs) and flash memory.

Optical discs use a laser to etch a microscopic mark into the surface of a disk. Materials used to produce optical disks are very chemically stable.

This means that data encoded on an optical disk should last several decades, provided the disc is stored properly.

Flash memory uses a special transistor called “floating-gate MOSFET” to store electronic data in a solid-state form.

Floating-gate or “FG” MOSFETs use a physical phenomenon called quantum tunneling to store electric charges in incredibly small spaces. 

The up-and-coming solid-state hard drive (SSD) relies on flash memory.

Solid-state hard drives can carry similar amounts of data as magnetic hard disk drives but load almost instantaneously and are much more durable.

The primary barrier to their commercial development has been their considerably higher costs when compared to magnetic disks.

Another type of computer storage is magnetic tape.

The magnetic tape uses a similar mechanism compared to hard disk drives, only the data is encoded onto a flexible tape impregnated with metallic particles.

Modern computers typically use several types of data storage arranged in a hierarchy of use.

An SSD might be reserved for the computer’s software and a few demanding programs, while an HDD may be used to store files like images, videos, and music.

What is the hierarchy of computer storage?

The purpose of the storage hierarchy is to reduce the bandwidth and increase the speed of electronic circuits. Each type of storage is used for different tasks at different times in the operation of a computer.

Primary storage

The first level of computer storage is primary storage. Primary storage is another term for computer memory used for surface-level operations. Dynamic and static RAM, read-only memory (ROM), and register memory make up primary computer memory.

Read-only memory (ROM) contains physically hardwired data into the computer. It typically carries the instructions a computer needs to operate on a basic level (the operating system).

It is not necessarily permanent, and newer forms of ROM can be changed via software updates.

Register memory is the fastest type of computer memory with the lowest capacity. It carries the instructions for the computer to start up and locate other forms of memory.

Secondary Storage

Secondary computer storage is where the computer stores data that is not immediately needed.

This would be the hard disk drive and other internal bulk data storage forms, such as SSDs. Secondary storage is slower but non-volatile, which means it can store data permanently.

Tertiary Storage

The next level below secondary storage is tertiary storage. Tertiary storage is external bulk data storage not necessary for the operation of a computer.

This might look like tapes and optical discs, and are usually reserved for large-scale operations.

Tertiary storage devices are typically connected to computers via physical data cables or server farms.

A newer example of tertiary storage is cloud storage—data is stored on several other devices and retrieved over the internet. 

Offline Storage

The lowest level of data storage is offline storage. Offline storage is removable external bulk storage, not under the control of the central processing unit (CPU).

When connected to a computer, offline storage devices behave like tertiary storage. External hard drives, flash drives, and rewritable optical discs are all forms of offline storage.

Offline storage has two important uses: data transfer and disaster protection.

Even in the days of 5G and high-speed DSL, many files are too large to transfer over the internet and must be loaded onto external storage devices and physically moved from computer to computer.

External offline storage devices are much less likely to be damaged or destroyed by a computer’s CPU, memory, or internal storage failure.

As a result, they are commonly used for backup drives, holding copies of an existing computer’s operating system and file. 

Cloud storage may eventually eliminate the need for physical backup drives. However, file transfer speeds over the internet are still much slower than USB connections. 

Why is the size of memory and storage lower than advertised?

The size of memory and storage is lower than what’s advertised by the manufacturer because they are marketed according to their decimal capacity, but computers use a binary system. When measured with a binary system, the number is lower. The device’s software also takes up space.

The prefixes used to describe large numbers of objects (kilo, mega, giga, etc.) are part of a base 10 or decimal numeral system.

Numbers are counted from 0 to 9, and then the counting sequence is repeated, beginning at 10.

Computer memory and storage are based on a base 2 or binary numeral system, where the only digits are “0” and “1”. 

Conveniently, the equivalents between the decimal and binary numbers are close enough that people often don’t notice, but there’s still a difference.

Computer memory and storage are sold under the base 10 prefixes for convenience, but computers measure them in base 2. 

In the decimal system, 1 GB is advertised as 1,000,000,000 bytes. However, in the binary system that computers use, 1 GB is actually 1,073,741,824 bytes.

This small difference builds up as the size of the storage or memory increases, which is why a hard drive advertised as 500 GB has actually 465 GB of real capacity.

In addition, due to included software that handles operations between your computer and the memory or storage medium, about 20% of the device’s capacity may be occupied straight out of the box. 

Storage media like external hard drives and flash drives can be reformatted to free up more capacity, and volatile memory can be freed up as needed.

You can check which apps are taking up your RAM through software like Task Manager.

Conclusion

RAM and other computer memory and storage forms are built around binary electronic circuits where numbers are in base 2 instead of base 10.

As a result, their capacity is always measured in powers of 2 and always even numbers.

However, their capacity is advertised using the decimal system, which is why there’s often a mismatch.