CosmoSim-ΛCDM

A high-performance 3D ΛCDM N-body simulator written in C++20. This code implements a Barnes–Hut tree for gravity in comoving coordinates with cosmological expansion, OpenMP parallelization, and CSV snapshot output. It’s intended as a foundation for extending into TreePM solvers, MPI scaling, and cosmological initial condition generation.

---

Features

• 3D N-body simulation in a periodic cubic box.
• ΛCDM expansion: configurable Ωm, ΩΛ, Ωr; H(a) = H0√(Ωm a⁻³ + Ωr a⁻⁴ + ΩΛ).
• Barnes–Hut octree for O(N log N) force evaluation with Plummer softening.
• KDK leapfrog integrator in comoving coordinates with Hubble drag.
• OpenMP parallelism for multi-core CPU acceleration.
• Snapshot output as CSV files (snapshots/a_xxxxxx.csv) containing positions and velocities.

---

Build

Requirements:

• C++20 compiler (GCC ≥ 10, Clang ≥ 12)
• OpenMP (optional, but recommended for performance)

g++ -O3 -march=native -fopenmp -std=c++20 cosmosim.cpp -o cosmosim

---

Run

Default parameters are built into the executable. You can override them via CLI arguments:

./cosmosim [N=20000] [steps=3000] [theta=0.6] [soft=0.04] [L=100.0] [a_start=0.02] [a_end=1.0] [dump_every=50]

Example:

./cosmosim 100000 4000 0.6 0.04 200 0.02 1.0 40

This produces snapshots in ./snapshots/ as CSV files, e.g.:

snapshots/a_00200.csv
snapshots/a_00400.csv
...

---

Visualization

You can view snapshots in:

• ParaView or VisIt (CSV import, scatter plot).
• Custom Python scripts (e.g., using matplotlib or plotly).

Example Python quickview:

import pandas as pd, matplotlib.pyplot as plt

snap = pd.read_csv("snapshots/a_01000.csv", comment='#')
plt.scatter(snap.x, snap.y, s=1, alpha=0.5)
plt.gca().set_aspect('equal')
plt.show()

---

Roadmap

• Add TreePM long-range solver with FFTW.
• Implement Zel’dovich approximation initial condition generator.
• Replace CSV with HDF5 snapshot format.
• Add MPI domain decomposition for scaling across nodes.
• Validation: compare power spectrum, 2-pt correlation, halo mass function against CAMB/CLASS.

---

Disclaimer

This is a research-grade toy code, not a replacement for production cosmology simulators like GADGET, RAMSES, or ENZO. It’s designed for learning, experimentation, and extension.

---

License

MIT License. See LICENSE for details.