本页尚未翻译,因此以英文显示。 参与翻译

Fastest and slowest request

Toss-up async

Requirement

A health check probes eight endpoints through an async measure call and reports the fastest and the slowest latency in milliseconds. The fixed samples are in the code; both versions must print the two lines shown under Expected output.

Expected output
fastest: 12 ms
slowest: 210 ms

Side by side

RxDart

FxDart

Why they differ

Reductions are where push and pull converge: to know the extreme you must see the whole sequence, so both min and max are terminal and both return a Future. RxDart's versions also accept an optional comparator for elements that aren't Comparable; FxDart keeps min and max as numeric terminals and covers keyed cases with minBy/maxBy. On plain integers the two calls are word-for-word identical — the FxDart side just lifts the pulls into an async chain with toAsync first.

The mirrored wrinkle is that each reduction consumes the source. A Dart stream is single-subscription, so asking for min and then max means two subscriptions — hence the latencies() factory on the RxDart side. The FxDart side has the same shape for the same reason: a terminal call drains the chain, so the second reduction pulls a fresh one. Both therefore measure twice (collecting into a list first is the shared alternative), and neither model has an edge worth claiming: a tie, decided by which way the surrounding code already flows.

Benchmark

Apple M1 Max, 32 GB RAM · Dart 3.12.2 (AOT-compiled) · 2026-08-18

Async case: the headline scale is N = 10,000, not 1,000,000. Every element costs an event-loop turn on both sides, so a million real awaits would measure Dart's event loop for minutes — not the pipeline. Delays are zero-length and the example's concurrency limit is kept; what the bars compare is the pipeline machinery.

N = 100

Time Tie

RxDart 707 µs
FxDart 628 µs

Peak memory Tie

RxDart 16.5 MB
FxDart 16.6 MB

N = 10,000

Time FxDart wins

RxDart 65.2 ms
FxDart 61.4 ms

Peak memory Tie

RxDart 24.1 MB
FxDart 24.3 MB

Bars are medians of repeated timed iterations in fresh processes per side (small N is batched for timer resolution). Sides within 5% of each other — or within 0.6 ms, a difference no person can perceive — count as a tie; close relative races are re-measured up to 5 runs. In an app, anything under a few milliseconds is invisible to the user regardless of which bar is shorter. Memory is peak process RSS. The Dart VM and the dataset are identical on both sides, so the difference between the two bars is what the pipeline itself holds onto.