Threads vs Processes

Threads vs Processes

Step 1 of 9

address spaces1

A process owns its memory. Threads live inside one process and share every byte of it.

The Idea

A process owns its memory. Threads live inside one process and share that memory, so starting one is cheap and handing over data costs nothing. Processes are walled off: sending work means copying it, but one crash cannot corrupt a sibling.

Real-World Example

Housemates share a kitchen; neighbours have their own. Housemates pass the salt across the counter in a second, and one of them leaving the gas on ruins everyone's evening. Borrowing sugar next door means packaging it and walking it to the door.

The Tradeoff

Threads are cheap and fast to coordinate, and every shared variable is now a bug waiting for the wrong interleaving. Processes buy isolation with memory duplication and serialisation on every message. In Python the GIL settles most arguments: I/O-bound work wants threads, CPU-bound work wants processes.

Your turn

Put the steps in the right order.

1. 1Each child runs in its own interpreter, on its own core↑↓
2. 2The parent process starts, holding the whole input batch↑↓
3. 3Results are serialised and sent back to the parent↑↓
4. 4The batch is split and copied into each child process↑↓

Check order

Mini quiz

1 / 3

Two threads in one process share:

• ANothing at all
• BThe same memory
• CSeparate address spaces
• DOne CPU core only