When a [CPU](/Computer_Architecture/CPU/CPU_architecture.md) executes a program, it needs a place to store the program's **instructions** and **related data**.
There are two types of memory: SRAM and DRAM. Both types of RAM memory are _volatile_ : the memory is only retained whilst the computer has a power supply and is wiped when the computer is rebooted. This contrasts with the memory of the harddisk which is non-volatile and is retained after a reboot.
Programs that are executing are loaded into memory because the chips that comprise memory can read and store data much faster than the harddisk. It would be possible to run a program from the harddisk but it would be 500 - 1000 times slower than memory.
> a **capacitor** is an electronic component that stores electrical energy in an electrical field. A device which can accumulate and release electrical charge.
In a DRAM cell, each bit of data is stored as a charge in a capacitor. The presence of charge represents a '1' bit and the absence of charge represents a '0' bit. Each of these cells is paired with a [transistor](/Electronics_and_Hardware/Digital_circuits/Transistors.md) that controls the reading and writing of data.
However capacitors lose [charge](/Electronics_and_Hardware/Analogue_circuits/Current.md) over time due to leaks. As a result DRAM is memory that needs to be refreshed (recharged) frequently. For this reason and because it only uses one transistor and capacitor per bit, DRAM is the less expensive form of volatile memory.
SRAM (Static Random Access Memory) is also volatile memory but, in terms of the electronics, it is different in its implementation. In contrast to DRAM it doesn't use capacitors. As a result the transistors do not leak and therefore do not need to be refreshed, hence why SRAM is _static_ and DRAM is _dynamic_.
SRAM uses [flip flops](/Electronics_and_Hardware/Digital_circuits/Flip_flops.md) to store the bits. It also uses multiple transistors per bit. This makes it faster than DRAM but more expensive. DRAM is at least ten times slower than SRAM.
Each single-bit cell in the grid can be identified using two dimensional coordinates. The coordinates are the location of that cell in the grid. Handling one bit at a time isn't very efficient so RAM accesses multiple grids of 1-bit memory cells in parallel. This allows for reads or writes of multiple bits at once, such as a whole byte.
