Parlo: A Performant, Parallel Utility Library for Go

Parlo is a Go library that provides utility functions for efficiently working with slices, maps, and channels.

View all benchmark results at: https://mahdi-shojaee.github.io/parlo/

Key Advantages:

• Parallel Processing: Parlo leverages Go's concurrency features to provide parallel versions of several functions. This allows you to utilize multiple CPU cores and significantly improve performance for large datasets.
• Generics: Parlo utilizes Go's generics system (available in Go 1.18 and later) to eliminate the need for reflection in many functions. This translates to better type safety and improved performance compared to reflection-based approaches.

Current Features:

• Slices: Sequential and parallel versions of Sort, IsSorted, Min, Filter, etc.

(Note: The list of features is subject to change based on development progress.)

Installation:

go get -u github.com/mahdi-shojaee/parlo

Usage:

package main

import (
  "fmt"

  "github.com/mahdi-shojaee/parlo"
)

func main() {
  data := []int{5, 2, 4, 1, 3}

  // Sequential Sort
  parlo.Sort(data)
  fmt.Println("Sequential:", data)

  // Parallel Sort
  data = []int{5, 2, 4, 1, 3}
  parlo.ParSort(data)
  fmt.Println("Parallel:", data)
}

Parallel Functions

All parallel versions of functions are prefixed with Par, indicating they utilize multi-core processing for better performance. For example, ParMap, ParFilter, and ParSort are the parallel counterparts of their sequential versions.

Benchmarking

Parlo includes a comprehensive benchmarking system to measure and compare the performance of sequential and parallel functions.

Defining Benchmarks

Benchmarks are defined in the charts/benchmark/benchconfigs.yml file. For detailed instructions on adding new benchmarks, please refer to the Benchmark Configuration Guide.

Running Benchmarks

To run benchmarks, use the following command from the charts directory:

cd charts
go run main.go [options] [benchmark names...]

Options:

• -benchtime=<duration>: Set the benchmark time for each test (default: 1s)

Note: Some benchmarks like Sort may require longer benchmark times for accurate results. Adjust the -benchtime flag accordingly when running these benchmarks.

Examples:

# Run all benchmarks
go run main.go -all

# Run Sort and SortFunc benchmarks with a 4s benchmark time
go run main.go -benchtime=4s Sort SortFunc

# Generate the `benchmark-results.js` file for the viewer without running any benchmarks
go run main.go -build

Viewing Benchmark Results

After running benchmarks, results are saved as JSON files in the charts/benchmark-results-viewer/benchmark-results/ directory.

To view the results:

1. Run go run main.go -build to generate the benchmark-results.js file.
2. Open charts/benchmark-results-viewer/index.html in a web browser.

You can also view the latest benchmark results online at: https://mahdi-shojaee.github.io/parlo/

The viewer provides interactive charts comparing the performance of sequential and parallel functions across different input sizes and scenarios.

Contributing:

We welcome contributions to Parlo! Feel free to open issues, submit pull requests, or reach out for discussions.

License:

Parlo is licensed under the MIT License.

Stay Updated:

Follow this repository for updates and new feature announcements.