Crate micromeasure 

Microbenchmark harness for tiny operations where PMU behaviour matters.

This crate provides a reusable microbenchmark framework with:

• Performance counter integration (Linux only)
• Console output with Unicode tables
• Explicit report rendering and JSON persistence
• Structured, serializable comparison of persisted reports
• Portable raw-sample reports for external orchestrators
• Advisory or gating policy with terminal, Markdown, and JSON rendering
• Multi-suite file and directory validation and comparison
• Warm-up and calibration phases
• Progress indicators
• Generic table formatting

Re-exports

pub use bench::BenchContext;

pub use bench::BenchmarkCaseOrder;

pub use bench::BenchmarkRunner;

pub use bench::BenchmarkRuntimeOptions;

pub use bench::ConcurrentBenchContext;

pub use bench::ConcurrentBenchControl;

pub use bench::ConcurrentBenchmarkGroup;

pub use bench::ConcurrentSampleInfo;

pub use bench::ConcurrentSampleLifecycle;

pub use bench::ConcurrentSamplePhase;

pub use bench::ConcurrentWorker;

pub use bench::ConcurrentWorkerResult;

pub use bench::CounterValue;

pub use bench::DiagnosticError;

pub use bench::DiagnosticResult;

pub use bench::EnergyScope;

pub use bench::MeasurementBackend;

pub use bench::MeasurementDomain;

pub use bench::MemoryBandwidthScope;

pub use bench::MetricFormat;

pub use bench::MetricValue;

pub use bench::NoContext;

pub use bench::OperationReportedDeviceBackend;

pub use bench::PmuCounterProfile;

pub use bench::PmuScope;

pub use bench::Throughput;

pub use bench::WallClockBackend;

pub use bench::LinuxPerfBackend;

pub use bench::LinuxPerfThreadSet;

pub use table::Alignment;

pub use table::BorderColor;

pub use table::TableFormatter;

pub use bench::PerfCounters;

pub use perf_event;

Modules

bench

table

Macros

benchmark_main

Structs

BenchSampleResult

Per-sample result returned by a benchmark function that wants to report custom metrics in addition to the standard timing/throughput/PMU pipeline.

BenchmarkCaseIdentity

Stable semantic identity for one native benchmark case.

BenchmarkMainOptions

BenchmarkReport

Persisted benchmark report for serialization and optional comparisons.

BenchmarkResult

Collected benchmark result for session summary and JSON export

BenchmarkStats

ComparisonAnalysis

A comparison bundled with a reproducible policy evaluation.

ComparisonCaseSnapshot

Measurement evidence for one side of a comparison.

ComparisonOptions

Options controlling report and report-set comparison.

ComparisonReport

Serializable, policy-free relationship between two benchmark reports.

ComparisonStatistics

Stability evidence retained for one side of a matched or unmatched case.

ComparisonSummary

Counts summarizing a structured comparison.

EnvironmentComparison

Complete environment relationship recorded in a comparison artifact.

EnvironmentOverride

Operator-supplied justification for comparing non-identical environments.

Instant

A measurement of a monotonically nondecreasing clock. Opaque and useful only with Duration.

MatchedBenchmark

Structured comparison for a benchmark found on both sides.

MetricComparison

A custom metric present in both versions of a matched benchmark.

MetricSummary

Aggregated view of a custom per-sample metric, computed by [crate::bench::stats::aggregate_metrics] from Vec<MetricValue> per sample. One summary per (name, unit) pair across all samples of a benchmark.

NativeMeasurementProjection

Native-only measurements kept alongside the normalized primary value.

PolicyCaseEvaluation

Policy result for one matched case.

PolicyEvaluation

Complete policy evaluation for a comparison.

PolicySummary

Aggregate policy counts.

PrimaryMeasurement

A direction-aware primary value used for comparison.

RegressionPolicy

Configurable classification and regression-gating policy.

ReportContext

Run-level context retained in every new benchmark report.

ReportReference

Durable reference to one evidence document.

SampleMetric

Owned, serializable form of a per-sample crate::MetricValue.

SampleMetricSet

Custom metrics captured for one measured sample, preserved in execution order for drift and state-transition analysis.

SeriesCaseIdentity

Stable semantic identity for one external series case.

SeriesReport

A portable external report containing chronological raw samples.

SeriesResult

Raw observations for one externally orchestrated measurement.

SeriesValidationError

Semantic failure in an otherwise structurally readable series report.

SuiteComparisonAnalysis

Versioned combined result for one file pair or two report directories.

SuiteComparisonSummary

Aggregate counts across matched, added, and removed suites.

UnmatchedBenchmark

A benchmark present on only one side of a partial comparison.

Validity

Correctness status for a report or individual result.

WorkerCounterSummary

WorkerSummary

Enums

BenchmarkKind

ChangeClassification

Direction-aware material-change classification.

ComparisonCaseIdentity

Semantic identity for a native benchmark or external series case.

ComparisonError

Semantic failure while comparing two otherwise readable reports.

ComparisonExitStatus

Stable process-status contract for comparison frontends.

ComparisonPolicy

ComparisonSide

Which side of a comparison supplied a value or caused an error.

ContextError

Failure to load or validate an explicitly requested context document.

MeasurementDirection

Whether larger or smaller primary values represent an improvement.

MeasurementKind

The semantic kind of a comparison’s primary measurement.

PolicyError

Invalid regression-policy configuration.

PolicyFinding

Stability or correctness observation attached during policy evaluation.

ReportDocument

Loaded evidence plus the reference derived from its original bytes.

ReportDocumentType

Kind of evidence document referenced by a comparison.

ReportError

Failure to load or interpret a persisted report.

ReportInputRole

Role of an input while loading and validating a report set.

SuiteComparisonError

Failure while loading, validating, or comparing report inputs.

ValidityStatus

Whether evidence passed its independent correctness checks.

Constants

ANALYSIS_SCHEMA_VERSION

JSON schema emitted for a comparison plus its separate policy evaluation.

BASELINE_PATH_ENVIRONMENT

Environment variable selecting an exact baseline report.

COMPARISON_SCHEMA_VERSION

JSON schema emitted for structured comparison reports.

CONTEXT_FILE_ENVIRONMENT

Environment variable selecting an explicit report context document.

DEFAULT_MAXIMUM_CV_PERCENT

Default coefficient-of-variation threshold for a stability finding.

DEFAULT_MAXIMUM_OUTLIER_FRACTION

Default outlier-fraction threshold for a stability finding.

DEFAULT_MINIMUM_CHANGE_PERCENT

Default material-change threshold for advisory evaluation.

OUTPUT_PATH_ENVIRONMENT

Environment variable selecting an explicit JSON report destination.

REPORT_SCHEMA_VERSION

JSON document schema emitted by this version of micromeasure.

SERIES_DOCUMENT_TYPE

JSON document discriminator required on external series reports.

SERIES_SCHEMA_VERSION

JSON schema emitted and accepted for external series reports.

SUITE_ANALYSIS_SCHEMA_VERSION

JSON schema emitted for a combined suite comparison.

Functions

benchmark_filter_from_args

benchmark_filter_from_env

black_box

An identity function that hints to the compiler to be maximally pessimistic about what black_box could do.

compare_report_inputs

Compare one report file pair or two directories of reports.

compare_reports

Compare loaded evidence documents without terminal or filesystem effects.

render_markdown

Render Markdown suitable for a CI annotation or persisted summary.

render_terminal

Render a concise plain-text comparison for an interactive job log.

run_benchmark_main

validate_report_input

Validate one report file or every JSON report beneath a directory.