eorcalc 0.1.0

EoR ionization diffusion calculations and power spectrum analysis with GPU acceleration

EoRCaLC - Epoch of Reionization Calculator

EoRCaLC is a Python package for calculating ionization diffusion and power spectrum analysis during the Epoch of Reionization (EoR). It features GPU acceleration support for high-performance cosmological simulations.

Features

• 🚀 GPU Acceleration: High-performance calculations using CuPy (CUDA 12.x)
• 📊 Power Spectrum Analysis: Matter power spectrum calculations via CAMB
• 🌌 Mass Functions: Halo mass function computations with custom transfer functions
• 🔬 Ionization Physics: Ionization diffusion and recombination modeling
• ⚡ Optimized Algorithms: Efficient numerical integration and interpolation

Installation

Prerequisites

• Python 3.8 or higher
• CUDA 12.x compatible GPU (for GPU acceleration)
• CUDA Toolkit 12.x installed

Basic Installation

pip install -e .

GPU Support (CUDA 12.x)

For GPU acceleration with CUDA 12.x:

pip install -e ".[gpu]"

Or install CuPy separately:

pip install cupy-cuda12x

Development Installation

pip install -e ".[dev]"

Dependencies

Core Dependencies

• numpy (≥1.20.0): Numerical computing
• scipy (≥1.6.0): Scientific computing and integration
• astropy (≥4.0): Astronomical calculations and units
• camb (≥1.3.0): Cosmological matter power spectrum
• massfunc (≥0.0.10): Halo mass function calculations
• cupy-cuda12x (≥12.0.0): GPU acceleration (optional)
• filelock (≥3.0.0): File locking for concurrent access
• cosfunc: Cosmological functions (local dependency)

Usage

Basic Example - CPU Version

from eorcalc.ioninti import Ion

# Initialize ionization calculator
z = 7.0  # Redshift
ion_calc = Ion(
    z=z,
    fesc=0.2,          # Escape fraction
    A2byA1=0.1,        # Transfer function parameter
    ktrans=200,        # Transition scale [Mpc^-1]
    alpha=2.0,         # Power law index
    beta=0.0           # Secondary index
)

# Calculate ionization properties
# ... your calculations here ...

GPU-Accelerated Version

from eorcalc.ioninti_gpu import Ion

# Same interface but with GPU acceleration
ion_calc = Ion(
    z=7.0,
    fesc=0.2,
    A2byA1=0.1,
    ktrans=200,
    alpha=2.0,
    beta=0.0
)

# GPU-accelerated computations
# ... your calculations here ...

Ionization Diffusion

from eorcalc.iondiff import Ion

# Full ionization diffusion calculations
ion_diff = Ion(
    z=7.0,
    fesc=0.2,
    kakaka=0.7e-28,     # Recombination coefficient
    xi_ion=10**25.6,    # Ionizing efficiency
    A2byA1=0.1,
    ktrans=200,
    alpha=2.0,
    beta=0.0
)

# Perform diffusion calculations
# ... your calculations here ...

Power Spectrum Analysis

from eorcalc.powerspec import MassFunctions

# Initialize cosmological calculator
cosmo = MassFunctions(
    A2byA1=0.1,
    kMpc_trans=200,
    alpha=2.0,
    beta=0.0
)

# Calculate mass functions and power spectra
z = 7.0
M = 1e10  # Solar masses
sigma = cosmo.sigma_M(M, z)
mass_func = cosmo.massfunc(M, z)

Module Overview

eorcalc.ioninti

CPU-based ionization calculations with standard numerical methods.

eorcalc.ioninti_gpu

GPU-accelerated ionization calculations using CuPy for enhanced performance.

eorcalc.iondiff

Ionization diffusion modeling including recombination effects.

eorcalc.powerspec

CAMB-based matter power spectrum and halo mass function calculations.

eorcalc.special

Special functions and utilities:

• qion_sb99: Ionizing photon production from Starburst99 models
• interp1d_gpu: GPU-accelerated interpolation
• xim: Ionization fraction calculations
• fstar: Star formation efficiency

Project Structure

EoRCaLC/
├── eorcalc/
│   ├── __init__.py
│   ├── iondiff.py       # Ionization diffusion (CPU)
│   ├── ioninti.py       # Ionization initial (CPU)
│   ├── ioninti_gpu.py   # Ionization (GPU accelerated)
│   ├── powerspec.py     # Power spectrum & mass functions
│   └── special.py       # Special functions & utilities
├── pyproject.toml       # Project configuration
├── LICENSE             # MIT License
└── README.md           # This file

Performance Tips

1. GPU Memory: Ensure sufficient GPU memory for large-scale simulations
2. Batch Processing: Process multiple redshifts in batches for efficiency
3. Caching: Enable CAMB result caching for repeated calculations
4. Precision: Use float32 on GPU for memory-intensive tasks if precision allows

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Citation

If you use this code in your research, please cite:

@software{eorcalc2025,
  author = {Hajime Hinata},
  title = {EoRCaLC: Epoch of Reionization Calculator},
  year = {2025},
  url = {https://github.com/SOYONAOC/IonDiff}
}

Contact

• Author: Hajime Hinata
• Email: [e-mail]
• Repository: https://github.com/SOYONAOC/IonDiff
• Issues: https://github.com/SOYONAOC/IonDiff/issues

Acknowledgments

This package builds upon:

• CAMB: Cosmological Boltzmann code
• CuPy: NumPy-compatible array library for GPU
• Astropy: Community Python library for Astronomy
• massfunc: Halo mass function library

Version History

v0.1.0 (2025-10-02)

• Initial release
• CPU and GPU implementations
• Power spectrum analysis
• Ionization diffusion modeling