Every Computer Memory Type Explained in 9 Minutes

Computer Science
96 followers
Apr 5, 2026
Different PC Memory Types Explained.
Ever wondered what all those memory types actually do inside your computer? This video breaks down every type of computer memory you need to know about. We start with RAM and explain the difference between DRAM and SRAM, then get into how CPU cache and registers make your processor fast. You'll also learn what VRAM does on your graphics card and why virtual memory kicks in when you run out of RAM. Then we go through the whole ROM family, from old school PROM and EPROM all the way to EEPROM and the flash memory that powers your SSD and USB drives today. If you want to finally understand how computer memory works from top to bottom, this one covers it all.
RAM. RAM stands for random access memory, and it is the main working memory of your computer.
Every time you open a program, load a file, or switch between browser tabs,
that data gets loaded into RAM so the processor can access it quickly.
RAM is volatile, which means everything stored in it disappears the moment you turn off your
computer. The random access part means the processor can jump to any piece of data in RAM
Older storage methods like magnetic tape had to be read in order from start to finish.
RAM does not have that limitation. The RAM sticks on your motherboard are almost always
a type called DRAM. The other main type is SRAM, and it serves a very different purpose.
DRAM. DRAM stands for dynamic random access memory, and it is the type used as main memory
in almost every computer, laptop, phone, and game console.
When people say my computer has 16 gigabytes of RAM, they are talking about DRAM. Each bit of
data in DRAM is stored using one capacitor and one transistor. The capacitor holds a small
electrical charge representing either a 1 or a 0. The problem is that capacitors leak charge
over time, so the data fades unless each cell is refreshed roughly every 64 milliseconds.
That constant refreshing is what makes it dynamic. This simple design makes DRAM very
dense and cheap to produce, but slower than SRAM because of the time spent refreshing.
The DDR4 and DDR5 sticks you by today are modern generations of DRAM.
SRAM. SRAM stands for static random access memory. Unlike DRAM, it does not use capacitors.
Instead, each bit is stored using a small circuit called a flip-flop made up of 4 to 6 transistors.
Because there are no capacitors leaking charge, SRAM does not need constant refreshing.
That is what makes it static. SRAM is significantly faster than DRAM,
but 4 to 6 transistors per bit means it takes up much more space and costs far more to manufacture.
Because of that, SRAM is never used as main memory. It is used where speed matters more
than capacity, most importantly inside the processor itself.
Cache memory. Cache memory is a small amount of very fast SRAM built into or right next to your
CPU. It stores the data the processor is most likely to need next, so it does not have to wait
for the much slower main RAM. Modern processors have 3 levels. L1 is the smallest and fastest,
sitting inside each CPU core, typically 32 to 128 KB. L2 is larger, usually 256 KB to 2 MB per core,
but slower. L3 is the largest, often 16 to 64 MB or more, and unlike L1 and L2,
it is shared across all cores. When the CPU needs data, it checks L1 first, then L2, then L3.
Finding the data in cache is called a cache hit. If the data is not in any cache,
the processor goes to the main RAM, which is much slower. That is called a cache miss.
Registers. Registers are the fastest and smallest memory in any computer. They are not a separate
chip. They are tiny storage locations built directly into the processor, right next to the
circuits that do the actual computing. A register holds the exact piece of data the CPU is working
on at that very moment. When the processor adds two numbers, those numbers are in registers.
Registers are measured in bits, not megabytes. A modern 64-bit processor has only around 16 to 32
general-purpose registers, each 64 bits wide. The full memory hierarchy flows like this.
Data is stored long-term on a drive, loaded into RAM when needed, pulled into cache,
and finally placed into a register, right before the CPU processes it.
VRAM. VRAM stands for Video Random Access Memory, and it is the dedicated memory on your graphics
card. While regular RAM serves the CPU, VRAM serves the GPU, the processor responsible for
rendering everything you see on your screen. VRAM stores textures, frame buffers, and 3D model data.
Having this memory separate from system RAM means the GPU can access its data without competing with
the CPU for bandwidth. Modern VRAM uses GDDR, which stands for Graphics Double Data Rate. It is
based on DRAM technology but redesigned to prioritize bandwidth, moving large amounts of
data at once, which is what GPUs need most. Unlike system RAM, VRAM is soldered onto the graphics
card and cannot be upgraded separately. Computers without a dedicated graphics card use an integrated
GPU, which shares system RAM instead. Virtual memory. Virtual memory is not a physical chip.
It is a system managed by the operating system that uses part of your storage drive as if it
were extra RAM. When your RAM fills up, the operating system moves the least recently used
data to a special file on your drive, called a page file on Windows, or a swap file on Linux.
When a program needs that data again, it gets swapped back into RAM. This lets the computer
handle more tasks than its physical RAM could support on its own. The trade-off is speed. Even
the fastest SSD is dramatically slower than RAM. Heavy reliance on virtual memory causes noticeable
slowdowns. The best fix is adding more physical RAM so the system does not need to swap as often.
ROM. ROM stands for read-only memory. Unlike RAM, ROM is non-volatile. It keeps its data even when
the computer is powered off. It was designed to store permanent instructions a device needs
every time it starts up. Traditional ROM is programmed during manufacturing and cannot be
altered afterward. Pure, unchangeable ROM is rare in modern devices because the need to update
firmware led to more flexible versions. PROM, EPROM, EEPROM, and Flash memory. Each making it
progressively easier to erase and rewrite stored data. PROM. PROM stands for programmable read-only
memory. Unlike basic ROM, it can be programmed after manufacturing, but only once. Inside a
PROM chip, there are tiny fuses, one for each bit. Programming means burning specific fuses
using a PROM programmer. Once a fuse is burned, it cannot be restored. If a mistake is made,
the chip is discarded. For this reason, PROM is also called OTP, one-time programmable. That
inability to correct errors led to the development of EEPROM. EEPROM. EEPROM stands for erasable,
programmable, read-only memory. And it solved the biggest problem with PROM, the fact that you could
not erase it. An EEPROM chip can be erased and reprogrammed, but the process is unusual. You
expose the chip to strong ultraviolet light for about 15 to 20 minutes through a small quartz
crystal window built into the top of the chip. The UV light clears the stored charges, resetting the
chip to a blank state so it can be reprogrammed. During normal use, the window is covered with a
sticker to prevent accidental erasure. EEPROM was a big improvement, but the slow process
and need to physically remove the chip drove the development of EEPROM. EEPROM. EEPROM stands for
electrically erasable, programmable, read-only memory. Instead of ultraviolet light, EEPROM can
be erased and rewritten using electrical signals while the chip is still installed in the device.
Another improvement is precision. EEPROM erases the entire chip at once. EEPROM lets you erase
and rewrite individual bytes without touching the rest. EEPROM is used to store small amounts
of data that need to survive a power loss but also need occasional updates. Things like BIOS
settings on a motherboard or firmware in car key fobs. Its main limitations are speed and cost.
It is slower and more expensive per bit than flash memory. Flash memory. Flash memory is a direct
descendant of EEPROM. It uses the same basic principle, storing data as electrical charges,
but engineered to be faster, denser, and cheaper. The key difference is that EEPROM erases one byte
at a time, while flash erases in large blocks, making it much faster for handling large amounts
of data. There are two main types. NAND flash is the most common, used in SSDs, USB drives,
SD cards, and phones. NOR flash is mostly used for firmware, because it allows code to run
directly from the chip. Flash is non-volatile, has no moving parts, and is more durable than
traditional hard drives. The full ROM family evolution goes. ROM to PROM, to EEPROM, to EEPROM,
to FLASH, each step making it easier and faster to erase and rewrite data.
