⌚ benchmarks

Programming benchmarks

This repository contains various benchmarks on different programming languages for different algorithms. The primary goal of this repository is to provide a comparison between different programming languages and their speeds.

You can navigate to the website here, or by clicking the linked website on the repository.

Local Benchmarking

This repository also features the ability to benchmark all the algorithms locally.

Running Specific Benchmarks

Benchmarks offers two gradle tasks: compileBenchmark and runBenchmark, to run specific benchmarks.

./gradlew compileBenchmark runBenchmark -Planguage={language} -Pfile={file}

For example, to run count-1M for c using Clang, you would run:

./gradlew compileBenchmark runBenchmark -Planguage=c-llvm -Pfile=count-1M/main.c

Note

compileBenchmark is only required to be passed for compiled languages. Gradle will automatically skip interpreted languages passed, like python and ruby.

./gradlew runBenchmark -Planguage=ruby -Pfile=count-1M/main.rb

compileBenchmark

Any language with a compile command in the config.yml file needs to be compiled first before it can run with the run commmand.

To compile the code, you can run the compileBenchmark task in the root directory for the repository:

./gradlew compileBenchmark -Planguage={language} -Pfile={file}

The task accepts two parameters:

language - ID of the language in the config.yml file

file - File to compile relative to the benchmarks directory

For example, to compile the HTTP GET benchmark for Kotlin/Native, you would run:

./gradlew compileBenchmark -Planguage=kotlin-native -Pfile=http-get/main.kt

runBenchmark

After compiling the benchmarks, you can run them using the runBenchmark task:

./gradlew runBenchmark -Planguage={language} -Pfile={file}

Note

compileBenchmark is only required to be passed for compiled languages. You can use runBenchmark for interpreted languages:

./gradlew runBenchmark -Planguage=ruby -Pfile=count-1M/main.rb

The task accepts the same parameters as the compileBenchmark task.

For example, to run the HTTP GET benchmark for Kotlin/Native, you do:

./gradlew compileBenchmark runBenchmark -Planguage=kotlin-native -Pfile=http-get/main.kt

http-get measures in ms according to its benchmark config.yml, so the output will be in milliseconds.

Running All Benchmarks

Prerequisites

All command line tools present in config.yml must be installed. You can determine if you meet them by running the version, compile, and/or run commands in the terminal.

Here are some notable examples:

• python for Python
• java for Java
  • Using the JDK for Gradle requires Java 21 or higher.
  • Using the java command on the command line requires Java 8 or higher.
• kotlinc for Kotlin JVM
  • This requires the KOTLIN_HOME environment variable to be set to the Kotlin compiler directory (not the bin directory).
• kotlinc-native for Kotlin Native
  • This requires the KOTLIN_NATIVE_HOME environment variable to be set to the Kotlin Native compiler directory (not the bin directory).
• ruby for Ruby
• python for Python
• gcc/g++ and clang/clang++ for C/C++
• rustc for Rust
• go for Go
• node, deno and bun for JavaScript
• ruby for Ruby
• php for PHP
• zig for Zig

You can test if you have the necessary tools and setup by running the validate gradle task:

./gradlew validate

IntelliJ IDEA

The repository comes with two IntelliJ IDEA run configuration that allows you to run benchmarks directly from the IDE:

• benchmark - Runs and graphs the benchmarks
• all - Runs and graphs the benchmarks, then prepares and serves the website

CLI

To both run and graph the benchmarks, you can use the following command:

./gradlew benchmark

Benchmarks will be generated in JSON format in the benchmarks/output directory.

Interactive HTML graphs are available in the benchmarks/output/graphs directory.

After that, you can prepare and serve the website by running:

./gradlew preview

You can navigate to the website by visiting http://localhost:4000.

The preview output only creates benchmarks for your hosting operating system. For example, if you're benchmarking your computer on a Windows machine, clicking on macOS or Linux will report a 404 error. Otherwise, you can navigate through the website as if it were live.

Contributing

Contributors are always welcome.

If you would like to contribute to this repository, please read the CONTRIBUTING.md file.

License

This repository is licensed under the Apache License 2.0.

As per outlined in the license, you are free to use, modify, distribute, and sublicense this repository for personal or commercial use. However, you must include the original copyright and license in any copy of the software.

By using the software, locally or on its website, you agree to the terms and conditions outlined in the license.