RAM Model of Computation

The Random Access Machine (RAM) is an abstract model of computation that abstracts the structure of real computers to focus on how Algorithms process information.

Major Components

Memory

Memory is modeled as an infinite sequence of cells (also called words), each storing a fixed-width integer.

• Each cell is directly addressable in $O(1)$ time — hence “random access”
• No distinction between cache, RAM, or disk at this abstraction level
• Practically, word size is assumed to hold values up to $O(\log n)$ bits

Processor

The model has a single processor that performs basic operations:

• Reading and writing to/from memory cells
• Arithmetic and logical operations
• Control flow (conditionals, jumps)

Key assumption: algorithms run sequentially, not in parallel. One operation at a time.

Instruction Set

All operations in the instruction set execute in $O(1)$ time:

Why This Model Matters

The RAM model lets us analyze algorithm complexity independently of hardware:

T(n) = number of RAM instructions executed
S(n) = number of memory cells used

This gives us the clean $O$-notation complexity we use in practice. The model is accurate enough for most algorithms but breaks down for:

• Cache-oblivious analysis — where memory hierarchy matters
• Parallel algorithms — where PRAM or BSP models are used
• Quantum algorithms — where quantum gates replace classical ops

Related

• Algorithm — what runs on the RAM model
• Computational Problem — what the algorithm solves