feat: add timer-overhead correction, saturation warning, and resolution diagnostic by jerome-benoit · Pull Request #571 · tinylibs/tinybench

jerome-benoit · 2026-05-30T01:22:03Z

Summary

Three cooperating timer diagnostics for sub-microsecond benchmarking, integrated in a single coherent code path:

subtractTimerOverhead — opt-in calibration of one timestamp provider call cost Ĉ, applied as max(0, raw - Ĉ) to non-overridden samples before statistics.
'warning' event — dispatched on both Bench and Task when a Task's measured samples are dominated by the timer resolution. Carries a TimerSaturationReason payload ('zero-dominated' | 'low-distinct' | 'zero-mad').
Task.detectedResolution — populated after each run with the smallest reproducibly observed positive sample among timer-measured samples.

Replaces #568, #569 and #570: their naïve composition would have the diagnostics operate on the corrected-and-clamped sample set, producing artificially small resolution estimates and false-positive saturation warnings.

Public API

import { Bench } from 'tinybench'

// Per-Bench (opt-in, default false)
const bench = new Bench({ subtractTimerOverhead: true })
bench.subtractTimerOverhead  // readonly boolean
bench.timerOverhead          // number | undefined — calibrated Ĉ in ms

// Per-Task diagnostic
const task = bench.getTask('foo')
task.detectedResolution      // number | undefined

// Saturation warning event with typed reason
bench.addEventListener('warning', evt => {
  evt.task    // Task
  evt.reason  // TimerSaturationReason | undefined
})

// Standalone helpers
import {
  calibrateTimerOverhead,
  classifyTimerSaturation,    // → TimerSaturationReason | undefined
  detectTimerSaturation,      // boolean wrapper around classifyTimerSaturation
  estimateResolution,
  medianAbsoluteDeviation,    // mad statistic from a sorted sample
} from 'tinybench'

// Timestamp provider objects (pass to calibrateTimerOverhead)
import {
  hrtimeNowTimestampProvider,
  performanceNowTimestampProvider,
} from 'tinybench'

import type {
  CalibrateTimerOverheadOptions,
  TimerOverheadEstimatorKind,  // 'median' | 'min' | 'p05'
  TimerSaturationReason,
} from 'tinybench'

`Task#processRunResult` ordering (cross-cutting fix)

isOverridden[] is collected in lockstep with latencySamples[] during measurement. The phase ordering preserves the alignment invariant and isolates user-supplied overriddenDuration values from timer diagnostics:

Apply overhead correction in-place on the collection-order array (skipping overridden indices). Alignment intact.
Build a measuredOnly view by filtering out overridden indices, before any sort.
Sort the working latencySamples array.
computeStatistics on the sorted (possibly corrected) samples.
estimateResolution + classifyTimerSaturation on the sorted measured-only subset (estimateResolution requires ascending-sorted input). With no overrides the subset is latencySamples itself; otherwise the filtered view is sorted first. An all-overridden run yields undefined resolution and no warning.
Dispatch the 'warning' event (carrying the reason, when a criterion fired), then 'cycle' and 'complete'.

isOverridden is allocated unconditionally so the measured-only filter is also active when subtractTimerOverhead: false (covers overriddenDuration users without overhead correction).

Constraints

subtractTimerOverhead + concurrency: 'task' is rejected. Asserted at construction and re-checked at the start of Bench.run() to guard against untyped (JS-side) mutation of the readonly concurrency field after construction.
When the timer is too coarse to resolve the call cost — fewer than half the calibration pairs produce a positive delta (C < R / 2, e.g. a Date.now-class timer with >= 1 ms resolution) — calibration returns 0 and the option becomes a no-op.

Risk

Runtime: opt-in for behaviour-changing pieces. Default-options benchmarks behave identically.
TypeScript: BenchEvents widened with 'warning'; exhaustive switch consumers under noFallthroughCasesInSwitch need a case 'warning': arm. BenchLike.timerOverhead is now readonly and optional. New types TimerSaturationReason, TimerOverheadEstimatorKind, CalibrateTimerOverheadOptions. New runtime exports calibrateTimerOverhead, classifyTimerSaturation, detectTimerSaturation, estimateResolution, medianAbsoluteDeviation, hrtimeNowTimestampProvider, performanceNowTimestampProvider.
Bundle: +0.7 kB gzipped vs main. The pre-existing size-limit budget of 12 kB is already exceeded on main (14.15 kB) — this PR does not change that situation.
No new runtime dependency.

JSDoc honesty

The subtractTimerOverhead JSDoc describes the max(0, x) clamp explicitly:

Two regimes (clean-shift X >> Ĉ vs sub-overhead X ≈ Ĉ)
rme deterministically inflates by factor M / (M − Ĉ) whenever Ĉ > 0
In the sub-overhead regime, p50, mad, aad collapse to 0 once cumulative mass at ≤ Ĉ reaches the relevant quantile
Three observable consequences of the clamp (latency.min == 0, throughput-mean substitution, 'zero-dominated' criterion ambiguity)

Verification

pnpm typecheck clean
pnpm lint clean
pnpm test — 237 tests pass across 64 test files
pnpm build — dist generated cleanly

…on diagnostic Three cooperating diagnostics for sub-microsecond benchmarking, integrated in a single coherent code path so that each one can rely on the others: - BenchOptions.subtractTimerOverhead (default false): when enabled, the cost of one timestamp provider call is calibrated once at construction time via the new exported calibrateTimerOverhead helper, then subtracted from each raw latency sample (clamped to zero) before statistics are computed. Samples returned by the task function via overriddenDuration are intentional user values and are skipped by the correction. - 'warning' event on BenchEvents and TaskEvents, dispatched on both the Bench and the Task instances when the latency samples of a task are dominated by the timer resolution. Detection uses three OR'd criteria computed by detectTimerSaturation: more than half zero samples, fewer than max(3, min(10, n/1000)) distinct values, or zero MAD with n > 100. An n < 10 guard prevents false positives on unit-style benchmarks. - Task.detectedResolution getter, populated after each run with the smallest strictly-positive sample value that appears at least twice (smallest reproducibly observed increment). Falls back to the strict minimum when no positive value repeats. A new estimateResolution helper is exported. calibrateTimerOverhead (utils.ts): - Subtracts in the provider's native type before converting to milliseconds (toMs(b - a) rather than toMs(b) - toMs(a)), preserving bigint precision on long-uptime hosts. - Discards a configurable warmup phase (default 64 pairs) so the JIT reaches its steady-state tier before measurements begin. - Returns 0 when fewer than half the back-to-back pairs produce a positive delta — in that regime the timer resolution exceeds the call cost and the positive deltas measure a tick boundary, not the call cost. - Configurable estimator: 'median' (default), 'min', or 'p05'. Task#processRunResult orders the diagnostics so they always reflect the raw, uncorrected measurements: 1. sortSamples on raw latencies 2. estimateResolution on raw sorted samples 3. when overhead correction is active, compute raw statistics, evaluate detectTimerSaturation against the raw distribution, then apply the correction in-place (skipping overridden samples) and re-sort only when overridden samples were skipped 4. compute the final (possibly corrected) statistics 5. when no correction was applied, evaluate detectTimerSaturation against the final samples (raw == final in this path) This consolidates three previously separate proposals (PRs tinylibs#568/tinylibs#569/tinylibs#570) into a single coherent change: composing them naively would have caused the diagnostics to operate on the corrected-and-clamped sample set, producing artificially small detected-resolution values and false-positive saturation warnings on benchmarks that activate overhead subtraction.

pkg-pr-new · 2026-05-30T01:22:33Z

Open in StackBlitz

npm i https://pkg.pr.new/tinybench@571

commit: 739fbfa

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 6a71cb2d80

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Apply audit-driven fixes to PR tinylibs#571: * fix(task): correct overhead before sort to keep latencySamples aligned with isOverridden (collection order). Previous logic indexed isOverridden after sortSamples, corrupting both overriddenDuration preservation and measured-sample skip in mixed-mode tasks. * fix(task): run timer-saturation detection on a measured-only subset so constant overriddenDuration values cannot trigger a spurious low-distinct-count warning. * fix(bench): assert subtractTimerOverhead is incompatible with concurrency: 'task' (sequential calibration would not reflect per-iteration cost under concurrency). * fix(bench): normalize subtractTimerOverhead with ?? false instead of === true to accept any falsy default consistently. * fix(index): export detectTimerSaturation alongside the other timer diagnostics helpers. * docs(types): rewrite subtractTimerOverhead JSDoc with an honest treatment of the max(0, x) clamp and the two caveats (concurrency, overriddenDuration). Remove an orphan /** block. * test: rewrite the overriddenDuration warning test to assert warningCount === 0, matching the measured-only saturation behavior.

jerome-benoit

Cross-validated audit (8 oracle agents, 4 paired dimensions)

Findings classified by inter-pair convergence:

Convergent (≥2 pairs flagged the same item)

task.ts:648 — detectedResolution includes overriddenDuration values
task.ts:662 — Phase 6 computeStatistics discards 14 of 15 fields
bench.ts:215 — ?? false accepts truthy non-boolean from JS callers, inconsistent with sibling line 214
bench.ts:219 — assert is one-shot at construction; runtime concurrency mutation bypasses it
bench.ts:218 — assert message lacks remediation
task.ts:699 — 'warning' event payload carries no reason
types.ts:128 — BenchLike.timerOverhead required + non-readonly: breaking + unsafe
types.ts:204 — JSDoc enumeration omits aad/mad/min/max
tests — mixed overridden+measured run not exercised (the alignment invariant)
tests — concurrency:'task' + subtractTimerOverhead assert untested
tests — 'p05' estimator branch untested
utils.ts:488 — bigint cast misleads the type system

Single-agent / high-signal

task.ts:631 — clamp injects synthetic zeros affecting min, throughput, criterion A
types.ts:212 — rme deterministically inflates by M / (M − Ĉ) in the clean-shift regime, not "potentially"
types.ts:211 — p50 (and downstream mad/aad) can collapse to 0 when ≥ 50% samples clamp
types.ts:26 — 'warning' widening across three unions breaks exhaustive switch consumers
utils.ts:285 — detectTimerSaturation accepts unsorted input but correctness requires sorted
utils.ts:301 — distinct-count loop has no short-circuit at threshold
utils.ts:454 — TimerOverheadEstimator is a string literal union; suffix suggests a strategy
test/calibrate-timer-overhead.test.ts:57 — min ≤ median * 2 is mathematically trivial
test/detected-resolution.test.ts:59 — conditional if (resolution !== undefined) skips all assertions on regression
README.md — public API additions undocumented

Verified correct (sample)

Phase 1 alignment of latencySamples[i] with isOverridden[i] in collection order
measuredOnly === latencySamples reference-equality optimization
Aborted-run paths (no samples, partial samples)
estimateResolution edge cases ([0,0,0] → undefined, all-equal positives)
'warning' widening consistent across BenchEvents / BenchEventsWithTask / TaskEvents
Bigint precision preserved at runtime
Style and Biome conformance on new test files

Detail in the inline comments. The mixed-overridden test gap is the only blocker.

* Add `classifyTimerSaturation` returning a `TimerSaturationReason` (`'zero-dominated' | 'low-distinct' | 'zero-mad'`) and re-implement `detectTimerSaturation` as a boolean wrapper. * Tighten `detectTimerSaturation`/`classifyTimerSaturation` parameter type from `Samples` to `SortedSamples`. * Short-circuit the distinct-value loop once the threshold is reached. * Add `medianAbsoluteDeviation(SortedSamples)` helper. * Rename `TimerOverheadEstimator` to `TimerOverheadEstimatorKind`. * Replace `as unknown as number` with `as bigint` in `calibrateTimerOverhead`; document operator polymorphism. * Re-export `classifyTimerSaturation`, `medianAbsoluteDeviation`, `TimerSaturationReason`, `TimerOverheadEstimatorKind` from the package entry point.

Extend `BenchEvent` with an optional `reason` payload symmetrical to `error`. The `reason` getter is typed as `TimerSaturationReason | undefined` for `'warning'` events and `undefined` for every other event type. * Move `TimerSaturationReason` from `utils.ts` to `types.ts` to align with the `Statistics`/`Samples` convention (types in `types.ts`, helpers in `utils.ts`). * Add a `'warning'` constructor overload accepting an optional reason. * Re-export `TimerSaturationReason` from the `./types` block in the package entry point.

…only samples * Compute `detectedResolution` from the measured-only subset (excluding `overriddenDuration` samples). A constant override value is no longer reported as the timer grain. * Allocate `isOverridden` unconditionally so the measured-only filter is also active when `subtractTimerOverhead` is disabled. * Replace Phase 6 `computeStatistics` recomputation with the dedicated `medianAbsoluteDeviation` helper. * Use `classifyTimerSaturation` and propagate the `TimerSaturationReason` onto the `'warning'` event payload. * Update `Task.detectedResolution` JSDoc to reflect the measured-only semantics; update the `#processRunResult` ordering description.

…ten options coercion * Coerce `subtractTimerOverhead` with `=== true`, matching the sibling `retainSamples` form. Truthy non-boolean values from JS callers are now rejected. * Re-state the constructor assert message in remediation form (action the user can take, not the internal cause). * Add the same assert at the start of `run()`. `concurrency` is documented as a post-construction-mutable field, so the constructor check alone leaves the mutation path uncovered. * Note the constraint and the dual enforcement in the `subtractTimerOverhead` field JSDoc.

Third-party `BenchLike` implementers can omit the field (semantically equivalent to the existing `undefined` sentinel that `Task` already handles). The `readonly` modifier matches the concrete `Bench.timerOverhead` declaration and forbids mutation through the interface, which `Task` reads on every cycle.

Rewrite the `subtractTimerOverhead` JSDoc with a mathematically grounded treatment: * Statistics list refers to all fields of `Statistics`; previously enumerated only seven of eighteen fields. * The `rme` inflation factor `M / (M − Ĉ)` is stated deterministically in the clean-shift regime, not hedged with 'potentially'. * The collapse of `p50`, `mad`, and `aad` to zero in the sub-overhead regime is named explicitly with the threshold. * Three observable consequences of the `max(0, …)` clamp are listed (`latency.min` may be 0; throughput substitutes the mean for clamped samples; criterion `'zero-dominated'` cannot distinguish clamped samples from genuine zeros).

… classifier * New `test/subtract-timer-overhead-alignment.test.ts` — exercises the Phase 1/2 alignment invariant on a heterogeneous run (alternating overridden + measured iterations) using a deterministic timestamp provider. Pins exact multiset counts so an off-by-one in the `isOverridden`/`latencySamples` index alignment fails the test. * `test/calibrate-timer-overhead.test.ts`: - Replace the loose `min ≤ median * 2` assertion with a deterministic estimator-ordering test using a scripted ascending-pair provider. - Add a deterministic `'p05'` test pinning the `max(0, ⌈n·0.05⌉ − 1)` index math at three sample sizes. - Add tests for the `subtractTimerOverhead` + `concurrency: 'task'` constructor assert and the equivalent `run()` runtime check. * `test/detected-resolution.test.ts`: replace the conditional `if (resolution !== undefined)` block with unconditional assertions. * `test/utils-detect-timer-saturation.test.ts`: add `classifyTimerSaturation` parallel coverage for each criterion (returning the precise reason string) plus the n<10 and healthy-spread negative cases. * New `test/warning-event-reason.test.ts` — verifies `BenchEvent.reason` carries the saturation reason for `'warning'` events and is `undefined` for other event types.

…, and timer diagnostics * New 'Timer Overhead Correction' section covers `subtractTimerOverhead`, the calibration helper, and the `concurrency: 'task'` and sub-overhead caveats. * New 'Per-Sample Override' section documents `overriddenDuration` (previously absent from the README despite being supported in code). * New 'Timer Diagnostics' section covers `Task.detectedResolution` and the `'warning'` event with its `TimerSaturationReason` payload, plus pointers to the standalone helpers. * Extend the `BenchEvents` listener example with a `'warning'` listener that reads `evt.reason`.

`computeStatistics` and `absoluteDeviationMedian` are not re-exported from the package entry point, so `{@link …}` references to them trigger `typedoc --treatWarningsAsErrors`. Switch them to plain backticked code references; `{@link}` is preserved only for symbols listed in the public exports.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 2f8232ac5b

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f9a0a6aff6

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 1376205e25

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 718cff9f78

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

The README timer-overhead example and `calibrateTimerOverhead` require a `TimestampProvider` object, but none was exported from the package entry.

`estimateResolution` scans the sorted sample array for the first repeated positive value instead of building a value-keyed Map over every sample, removing an O(distinct) allocation on the default post-processing path. Its signature is tightened to `SortedSamples` (matching its `classifyTimerSaturation` / `detectTimerSaturation` siblings) and resolution is computed after the working-array sort. Also corrects the `detectedResolution` getter doc (strict-min fallback; corrected-samples note under `subtractTimerOverhead`) and the `#processRunResult` phase-ordering doc (`'warning'` is dispatched before `'cycle'`/`'complete'`).

Cover the `zero-mad` n=100/101 boundary and the distinct-count ceiling of 10 at n=10000.

Calibration returns 0 when fewer than half the pairs yield a positive delta (C < R/2), not at a fixed 1 ms resolution.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 48ba856040

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

…ssert message The identical guard message in the constructor and `run()` is now a single module-level constant.

Branch on the named `hasAnyOverridden` boolean rather than the `measuredOnly === latencySamples` array identity; behaviour-identical, self-documenting.

The `CalibrateTimerOverheadOptions` fields count back-to-back call pairs, not latency samples; rename disambiguates from the codebase-wide `Samples` meaning (unreleased option, no published break). Also document the deliberate p05 nearest-rank choice vs `quantileSorted`.

…ions `concurrency` and `threshold` are `readonly`; the example now sets them at construction instead of mutating them after the fact.

The `run()` re-check exists to catch untyped mutation of the `readonly` `concurrency` field after construction, not a supported reconfiguration path.

Use optional chaining at both guard sites (dropping a non-null-assertion eslint-disable), and fix the #processRunResult param doc: isOverridden is undefined on the error / no-valid-samples path, not when overhead correction is disabled.

Return early on the degenerate zero-pairs input so the coarse-timer guard (deltas.length * 2 < pairs) covers every remaining case, removing the redundant empty-deltas check.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: d5822c0407

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

`pairs < 0` fell through the coarse-timer guard and produced `NaN` (median) or `undefined` (min/p05); `pairs <= 0` now returns 0, restoring the graceful behaviour of the removed empty-deltas check. Covered by a regression test.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 19358ab91c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

The timestamp provider is still invoked around the task function; only the measured value is discarded and overhead correction skipped.

…erOverhead `{ pairs: Infinity }` / `{ warmupPairs: Infinity }` (and NaN/non-integer) hung or produced NaN; `Number.isInteger` guards coerce them to the existing no-op (return 0 / skip warmup). Covered by regression tests.

…lean[] Replace the always-allocated `isOverridden: boolean[]` (one slot per sample, even with no override) with a `Set<number>` of overridden collection-order indices, populated only on override. Eliminates an O(n) parallel array on the default path; alignment invariant and concurrency safety preserved (index is taken synchronously right after the sample push).

…rkSync The identical error-XOR-samples return union was duplicated across both methods; hoist it to a single module-local `BenchmarkResult` type (mirrors the local-type convention in utils.ts). Type-only; no behavior change.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 344fa4b5b5

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector · 2026-07-22T18:41:43Z

+  classifyTimerSaturation,
+  detectTimerSaturation,
+  estimateResolution,


Export a way to construct sorted samples

For TypeScript consumers performing the advertised custom analysis on their own data, these helpers require the branded SortedSamples type, so even a sorted number[] is rejected by the compiler. The only function that establishes this brand is sortSamples, but it is not exported from the package root; consequently classifyTimerSaturation, detectTimerSaturation, and estimateResolution require an unsafe cast unless samples came directly from a retained task result. Export sortSamples or change the public helpers to accept and validate ordinary arrays.

Useful? React with 👍 / 👎.

…ssRunResult The 'Ordering' JSDoc block still referenced the removed `isOverridden` array (and array-indexed it, which a Set is not); align the wording with the `overriddenIndices` Set introduced in 15f30d9.

The tracked field was renamed to `overriddenIndices` (a Set) in 15f30d9; this aligns the last remaining `isOverridden` reference (a test title).

jerome-benoit requested a review from Uzlopak as a code owner May 30, 2026 01:22

This was referenced May 30, 2026

feat: add opt-in subtractTimerOverhead option #568

Closed

feat: dispatch warning event on timer saturation #569

Closed

feat: expose detected timer resolution per task #570

Closed

chatgpt-codex-connector Bot reviewed May 30, 2026

View reviewed changes

Comment thread src/task.ts Outdated

Comment thread src/task.ts Outdated

jerome-benoit commented May 30, 2026

View reviewed changes

jerome-benoit added 14 commits May 30, 2026 15:32

Merge branch 'main' into feat/timer-diagnostics-and-overhead-correction

411c1d7

Merge branch 'main' into feat/timer-diagnostics-and-overhead-correction

2a5a701

Merge branch 'main' into feat/timer-diagnostics-and-overhead-correction

b7e7e42

Merge branch 'main' into feat/timer-diagnostics-and-overhead-correction

ce3c2c4

Merge branch 'main' into feat/timer-diagnostics-and-overhead-correction

2f8232a

chatgpt-codex-connector Bot reviewed Jun 5, 2026

View reviewed changes

Comment thread README.md

jerome-benoit added 2 commits June 7, 2026 19:31

Merge branch 'main' into feat/timer-diagnostics-and-overhead-correction

a3b9a37

Merge branch 'main' into feat/timer-diagnostics-and-overhead-correction

f9a0a6a

chatgpt-codex-connector Bot reviewed Jun 9, 2026

View reviewed changes

Comment thread src/task.ts

Comment thread src/bench.ts Outdated

Merge branch 'main' into feat/timer-diagnostics-and-overhead-correction

1376205

chatgpt-codex-connector Bot reviewed Jun 14, 2026

View reviewed changes

Comment thread src/task.ts

Merge branch 'main' into feat/timer-diagnostics-and-overhead-correction

c53307d

chatgpt-codex-connector Bot reviewed Jun 25, 2026

View reviewed changes

Comment thread src/bench.ts

jerome-benoit added 7 commits June 25, 2026 23:46

Merge branch 'main' into feat/timer-diagnostics-and-overhead-correction

42fd0be

Merge branch 'main' into feat/timer-diagnostics-and-overhead-correction

0a221a3

Merge branch 'main' into feat/timer-diagnostics-and-overhead-correction

eadeb79

fix(index): export hrtimeNow and performanceNow timestamp providers

8600893

The README timer-overhead example and `calibrateTimerOverhead` require a `TimestampProvider` object, but none was exported from the package entry.

test(utils): lock timer-saturation classifier thresholds

80e4ac5

Cover the `zero-mad` n=100/101 boundary and the distinct-count ceiling of 10 at n=10000.

docs(readme): clarify the coarse-timer no-op condition

48ba856

Calibration returns 0 when fewer than half the pairs yield a positive delta (C < R/2), not at a fixed 1 ms resolution.

chatgpt-codex-connector Bot reviewed Jul 22, 2026

View reviewed changes

Comment thread src/utils.ts

Comment thread src/task.ts Outdated

Comment thread README.md Outdated

jerome-benoit added 7 commits July 22, 2026 18:49

refactor(bench): hoist duplicated subtractTimerOverhead/concurrency a…

fdb108c

…ssert message The identical guard message in the constructor and `run()` is now a single module-level constant.

refactor(task): use hasAnyOverridden instead of reference-identity check

264c751

Branch on the named `hasAnyOverridden` boolean rather than the `measuredOnly === latencySamples` array identity; behaviour-identical, self-documenting.

docs(readme): configure concurrency and threshold via constructor opt…

35fce79

…ions `concurrency` and `threshold` are `readonly`; the example now sets them at construction instead of mutating them after the fact.

docs(bench): clarify the run() re-assert guards JS-side mutation

94c0d8c

The `run()` re-check exists to catch untyped mutation of the `readonly` `concurrency` field after construction, not a supported reconfiguration path.

refactor(utils): guard pairs===0 explicitly in calibrateTimerOverhead

d5822c0

Return early on the degenerate zero-pairs input so the coarse-timer guard (deltas.length * 2 < pairs) covers every remaining case, removing the redundant empty-deltas check.

chatgpt-codex-connector Bot reviewed Jul 22, 2026

View reviewed changes

Comment thread src/utils.ts Outdated

chatgpt-codex-connector Bot reviewed Jul 22, 2026

View reviewed changes

Comment thread src/utils.ts

jerome-benoit added 4 commits July 22, 2026 20:20

docs(readme): clarify overriddenDuration does not bypass the timer

eca70bf

The timestamp provider is still invoked around the task function; only the measured value is discarded and overhead correction skipped.

chatgpt-codex-connector Bot reviewed Jul 22, 2026

View reviewed changes

jerome-benoit added 2 commits July 22, 2026 20:45

docs(task): finish isOverridden -> overriddenIndices rename in #proce…

26d8507

…ssRunResult The 'Ordering' JSDoc block still referenced the removed `isOverridden` array (and array-indexed it, which a Set is not); align the wording with the `overriddenIndices` Set introduced in 15f30d9.

test(task): rename stale isOverridden test title to overridden samples

739fbfa

The tracked field was renamed to `overriddenIndices` (a Set) in 15f30d9; this aligns the last remaining `isOverridden` reference (a test title).

jerome-benoit merged commit 0647d27 into tinylibs:main Jul 22, 2026
20 of 21 checks passed

jerome-benoit deleted the feat/timer-diagnostics-and-overhead-correction branch July 22, 2026 18:56

This was referenced Jul 22, 2026

docs: add FAQ.md (measurement precision, JIT deopt, cross-tool differences) #609

Merged

docs(examples): add convert() callback example for bench.table() #610

Merged

Uh oh!

Uh oh!

Conversation

jerome-benoit commented May 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Public API

Task#processRunResult ordering (cross-cutting fix)

Constraints

Risk

JSDoc honesty

Verification

Uh oh!

pkg-pr-new Bot commented May 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Uh oh!

jerome-benoit left a comment

Choose a reason for hiding this comment

Cross-validated audit (8 oracle agents, 4 paired dimensions)

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector Bot Jul 22, 2026

Choose a reason for hiding this comment

jerome-benoit commented May 30, 2026 •

edited

Loading

`Task#processRunResult` ordering (cross-cutting fix)

pkg-pr-new Bot commented May 30, 2026 •

edited

Loading