cord-path

Crates.iocord-path
lib.rscord-path
version1.0.0
created_at2025-08-08 08:29:40.569035+00
updated_at2025-08-08 08:29:40.569035+00
description2 dimensional coordinate based path finder
homepage
repositoryhttps://github.com/HardBoss07/cord-path
max_upload_size
id1786415
size57,580
M4tt3o (HardBoss07)

documentation

README

cord-path

A high-performance CLI tool for finding an optimal travel path through a set of 2D points. cord-path leverages NVIDIA CUDA to accelerate distance matrix calculations, making it ideal for large datasets where speed is critical.

Requires: NVIDIA driver 580.88 or later.

🗺️ Features

  • CUDA-Accelerated Pathfinding: Utilizes NVIDIA CUDA for extremely fast distance matrix computations.
  • Optimized Routes: Employs the Nearest Neighbor heuristic combined with 2-opt optimization for efficient and improved path results.
  • Flexible Coordinate Support: Handles both positive and negative coordinates.
  • Customizable Start: Optional feature to specify a custom start position.
  • Batch Processing: A quiet mode is available for use in scripts, suppressing all console output.
  • CSV Output: Easily save the final, optimized path to a CSV file.

⬇️ Installation

Download the latest pre-compiled binary from the Releases page. Place the executable in a directory that is included in your system's PATH.

Linux

chmod +x cord-path
sudo mv cord-path /usr/local/bin/

Windows (PowerShell)

Move-Item .\cord-path.exe "C:\Program Files\cord-path\"
# You may need to add "C:\Program Files\cord-path\" to your system's PATH

🚀 Usage

cord-path -f <file.csv> [options]

Required Argument

Flag Description
-f, --file <FILE> Path to a CSV file containing coordinates. Each line must be in the format: x,y

Optional Arguments

Flag Description
-x, --start-x <X> Optional starting X coordinate (supports negative values).
-y, --start-y <Y> Optional starting Y coordinate (supports negative values).
-o, --output <FILE> Path to the output CSV file where the calculated path will be saved.
-q, --quiet Suppresses all console output. Only works when used with --output.
-h, --help Displays the help message and exits.
-V, --version Displays the version number and exits.

📄 Examples

1. Simple run (console output only)

cord-path -f points.csv

Output:

Path length before 2-opt: 51234.77
Path length after 2-opt: 49201.12
Path order:
Point 0: (100, 250)
Point 1: (150, 200)
...

2. With a custom starting position

cord-path -f points.csv -x 200 -y -500

3. Outputting to a CSV file

cord-path -f points.csv -o result.csv

4. Silent mode for scripts

cord-path -f points.csv -o result.csv -q

This command will save the path to result.csv without printing any information to the console.

📂 CSV File Format

The input file must be a plain CSV with two integer or float values per line, representing the X and Y coordinates.

100,200
150.5,220.8
-50,400
...

⚙️ Performance Notes

  • This tool's performance is dependent on the CUDA kernel used for distance calculations.
  • An NVIDIA driver version 580.88 or newer is mandatory.
  • For optimal performance, ensure your GPU is configured for high-performance mode.

🛠️ Compile from Source

Prerequisites

  • Rust (latest stable): https://rustup.rs
  • NVIDIA CUDA Toolkit installed and configured
  • NVIDIA driver 580.88+
  • A C++ compiler compatible with your CUDA version

Build Steps

  1. Clone the repository:

    git clone [https://github.com/HardBoss07/cord-path.git](https://github.com/HardBoss07/cord-path.git)
cd cord-path

  2. Build the CUDA static library:

    cd cuda
nvcc -O3 --compiler-options '-fPIC' -c kernels.cu -o kernels.o
ar rcs libcuda_kernels.a kernels.o
cd ..

  3. Build the Rust CLI:

    cargo build --release

The final executable binary will be located at: target/release/cord-path

⚖️ License

This project is licensed under the GNU General Public License v3.0 (GPL-3.0).

You can read the full license text in the LICENSE file.

Commit count: 0

cargo fmt