🎴 Flashcard Mode

With serial input and output through Reformat, different stages can work simultaneously - input reads, Reformat processes, output writes. This is pipeline parallelism where data flows through stages like a pipeline. No data parallelism (same data on different flows) or component parallelism (multiple instances).