How many things live inside any steady flow? Exactly how fast they arrive times how long each one stays. You found it in the pipeline — five in flight because one enters per tick and each stays five ticks. Here it is with knobs. Watch latency and throughput pull apart while the law holds.
They're different numbers people constantly confuse. Deepening the pipe (more W) makes each instruction take longer but finishes just as many per tick. Widening it (more λ) finishes more per tick with no change to how long any single one takes. Pipelining buys throughput without touching latency — that's the whole trick.
At steady state arrivals equal departures, exactly like current at a wire node. Every instruction that ever entered is either retired or still in flight: entered = retired + L. Your screenshot proved it — cycle 72, retired 67, in flight 5. Nothing made, nothing lost.
It looks parallel — many busy at once — but it's a series chain you kept full by overlapping. No pipeline: latency W, one in flight. Pipelined: same latency, W in flight, W× the throughput, reusing the same stages. Truly parallel means cloning the chain.
The same L = λW counts water in a pipe, cars on a one-in-one-out road, electrons in a wire segment, patients in a ward, and a checkout queue. It assumes nothing about what flows — only that the system is stable and things don't pile up forever.
Trained as a physicist, became the first of a new discipline. Little earned his S.B. in physics from MIT in 1948, edited the campus humor magazine, then hitchhiked the country doing odd jobs — sheep-ranch labour, car valet — before a stint as a General Electric engineer sent him back to MIT for graduate school in 1951.
From whom: he started in physics but switched, at the urging of Philip M. Morse — the founder of operations research in the United States — into that brand-new field. Under Morse he became, in 1955, by MIT and INFORMS accounts the first operations-research PhD in the country. His thesis sat in both physics and OR.
Under what discipline — and on what machine: his dissertation, Use of Storage Water in a Hydroelectric System, used dynamic programming to optimise water held behind dams, and ran on MIT's Whirlwind I — likely the first non-military application of dynamic programming. The law that describes instruction pipelines was born from one of the earliest uses of a digital computer to solve a real optimisation problem.
The law itself came later: his general proof of L = λW was published in 1961. He spent 1957–62 at Case Institute of Technology, joined MIT's Sloan School in 1962, became an Institute Professor in 1989, and is also counted a founder of marketing science.
A resonance for this series: his wife, Elizabeth Alden, took her MIT physics PhD in 1954 on the domain-wall dynamics of barium titanate — a ferroelectric dielectric — under Arthur von Hippel. The high-k dielectric world from the board sheets and the queuing law from this one met at the same kitchen table.