Skip to content
@BenchCommons

Benchmarks Commons

Open benchmarks, methodologies and tools for the web.
README.md

BenchCommons

Open benchmarks, methodologies and tools. Every benchmark here is reproducible: the code, the inputs, the pinned versions and the raw results are public.

Repositories by role

Catalogs (catalog-*): the implementations that are measured.

Repository What it is
catalog-ssg 63 static site generators in 20 languages. Each one has a pinned Dockerfile, a canonical build.sh, the reference site and metadata in generators.json.
catalog-algorithms Classic algorithms implemented in hundreds of languages. It started as Hello World in 452 languages, and 83% of them are run in CI.
catalog-web-frameworks Web framework implementations. It continues the archived TechEmpower Framework Benchmarks.

Datasets (dataset-*): inputs for the benchmarks.

Repository What it is
dataset-names Given names and surnames worldwide, with country, frequency, gender and script, from openly licensed official sources.
dataset-commons Small shared datasets (texts, graphs, numbers) used by several catalogs.

Tools (tools-*, plus perfratio)

Repository What it is
tools-benchmark Builds each catalog implementation in Docker, runs it and records reproducible results. Today it runs SSGBerk, the static site generator benchmark.
perfratio A command-line benchmarking tool in Rust, built on hyperfine. It measures time, CPU, peak memory and RAPL energy, with in-depth statistics.
tools-dataset Builds, validates and exports the datasets as pinned, checksummed inputs.
tools-commons Shared libraries and utilities used by the tools: contract checkers and common CLI helpers.

Sites: website-results (benchmark results) and website-perfratio.

Governance:

  • .github: community files, templates, shared workflows and the AI agent rules.
  • .product: roadmap, cross-repo decisions and contracts.

Run a benchmark

git clone --recurse-submodules https://github.com/BenchCommons/tools-benchmark.git
cd tools-benchmark
./ssgberk --test hugo -nf 50          # one generator, 50 pages
./ssgberk --suite P --quality         # the small suite with the output audit

The methodology, suites and results format are specified in tools-benchmark/docs/specs.

Contributing

Read CONTRIBUTING. To add a generator, follow How to add a generator. Issues and pull requests are welcome in each repository.

Popular repositories Loading

  1. perfratio perfratio Public

    Command benchmarking across several dimensions (time, CPU, peak memory, RAPL energy) with in-depth statistics and ratios. A hyperfine fork.

    Rust 2

  2. catalog-algorithms catalog-algorithms Public

    Classic algorithm implementations across hundreds of programming languages, measured by BenchCommons with perfratio. Started as Hello World in 452 languages.

    Python 1

  3. catalog-ssg catalog-ssg Public

    Static site generator implementations measured by BenchCommons: a pinned Dockerfile, a canonical build script and metadata for each generator.

    Shell 1

  4. website-results website-results Public

    Results website for BenchCommons benchmarks.

    JavaScript 1

  5. tools-benchmark tools-benchmark Public

    BenchCommons toolset: builds each catalog implementation in Docker, runs it and records reproducible results. Derived from the TechEmpower toolset.

    Python

  6. dataset-names dataset-names Public

    Forked from Debdut/names.io

    Open worldwide dataset of given names and surnames, with country, frequency, gender and script, built from openly licensed official sources. A BenchCommons benchmark input.

    Python

Repositories

Showing 10 of 14 repositories
View all repositories

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Loading…

Most used topics

Loading…