Building and Testing parrot

April 22, 2026 ยท View on GitHub

This document describes how to build, test, and generate documentation for the parrot project.

Prerequisites

  • CMake (version 3.10 or higher)
  • C++ compiler with C++20 support
  • NVIDIA CUDA Toolkit 13.0 or later
  • NVIDIA GPU with compute capability 7.0 or higher
  • Python 3 with pip (for documentation, optional)
CUDA Installation (Click to expand if you need to install CUDA)

The most common setup issue is the missing NVCC compiler. Here's how to install CUDA properly:

  1. Download the CUDA Toolkit from NVIDIA Developer
  2. Follow the CUDA Installation Guide for your platform

Option 2: NVIDIA HPC SDK

  1. Download from NVIDIA HPC SDK Downloads
  2. Follow the HPC SDK Installation Guide

Environment Configuration

Add these to your ~/.bashrc, ~/.zshrc, or equivalent:

# CUDA installation paths (adjust based on your installation)
export CUDA_HOME=/usr/local/cuda
export PATH=$CUDA_HOME/bin:$PATH
export LD_LIBRARY_PATH=$CUDA_HOME/lib64:$LD_LIBRARY_PATH

# For NVIDIA HPC SDK installations
# export CUDA_HOME=/opt/nvidia/hpc_sdk/Linux_x86_64/26.3/cuda
# export PATH=/opt/nvidia/hpc_sdk/Linux_x86_64/26.3/compilers/bin:$PATH

Reload your shell configuration:

source ~/.bashrc  # or ~/.zshrc

Verification

Check that CUDA is properly installed:

# Check NVCC compiler
nvcc --version

# Check NVIDIA driver
nvidia-smi

Troubleshooting "Could not find NVCC compiler"

If CMake cannot find the NVCC compiler:

  1. Set CMAKE_CUDA_COMPILER manually:

    cmake -DCMAKE_CUDA_COMPILER=/usr/local/cuda/bin/nvcc ..
# or for HPC SDK:
cmake -DCMAKE_CUDA_COMPILER=/opt/nvidia/hpc_sdk/Linux_x86_64/26.3/compilers/bin/nvcc ..

  2. Check NVCC is in PATH:

    which nvcc

  3. Install NVIDIA drivers if nvidia-smi doesn't work:

    # Ubuntu/Debian
sudo apt-get install nvidia-driver-580  # or latest version

# CentOS/RHEL/Fedora
sudo dnf install akmod-nvidia

CUDA Architecture Configuration

The project automatically detects your GPU architecture, but you can also configure it manually:

Auto-detection (Default)

cmake ..

This will automatically detect the GPU architecture on your system.

Manual Architecture Selection

# For specific architecture (e.g., RTX 8000 with sm_75)
cmake .. -DCUDA_ARCH=75

# For Ampere GPUs (e.g., RTX 30xx series, A100, laptop GPUs)
cmake .. -DCUDA_ARCH=89

# For multiple architectures (fat binary)
cmake .. -DCUDA_ARCH=75,89

# For all common modern architectures
cmake .. -DCUDA_ARCH=ALL

Building the Project

  1. Create a build directory:

    mkdir -p build
cd build

  2. Configure with CMake:

    cmake ..

  3. Build the project:

    cmake --build . -j$(nproc)

Running Tests

The project uses doctest for unit testing with both comprehensive and individual test executables.

Run All Tests

# Run all tests via CTest
ctest

# Or run the main test executable directly
./parrot_tests

Run Individual Test Categories

# Basic operations
./test_basic

Building Documentation

The project uses Doxygen and optionally Sphinx for documentation.

Prerequisites

Install required Python packages (optional, for Sphinx):

pip install sphinx sphinx-rtd-theme breathe

Generate Documentation

  1. Generate Doxygen XML documentation:

    doxygen Doxyfile

  2. Build Sphinx documentation:

    cd docs
sphinx-build -b html . build/html

    The generated HTML documentation will be available in docs/build/html/.

Viewing Documentation

Open the HTML documentation in your browser:

open docs/build/html/index.html