Parallelism
In plain terms
Actually executing multiple computations at the same instant.
Going deeper
Parallelism splits work to run simultaneously across multiple cores or machines for speed. It differs from concurrency, which is about structure, not literal simultaneity.
Under the hood
Amdahl's law limits speedup by the serial fraction; data and task parallelism exploit multicore CPUs, GPUs, and clusters.
Examples
GPU computing, multithreaded number crunching, MapReduce, SIMD instructions.
Think of it like...
Like many chefs each cooking a different dish at once, versus one chef juggling several.