Scientific Context

🌌 Scientific Context

GRANITE v0.6.8 | ← Roadmap | Simulation Health & Debugging →

The scientific motivation, the astrophysical problem, and GRANITE's place in the NR landscape.

---

1. The Problem GRANITE Is Built to Solve

The universe's most energetic events are the coalescence of compact objects — black holes and neutron stars — driven by gravitational radiation. These events produce:

• Gravitational waves (detectable by LIGO, Virgo, LISA, PTAs)
• Electromagnetic counterparts (X-ray, optical, radio)
• Neutrino emission
• Nucleosynthesis of heavy elements (r-process)

The physics is governed by four deeply coupled systems:

System	Governs
Spacetime metric (Einstein equations)	Curved geometry of space and time
Matter fields (GRMHD)	Gas, plasma, nuclear matter
Electromagnetic field	Magnetic fields in disc and jets
Radiation field	Neutrinos and photons

Binary black hole mergers — where items 2–4 are negligible — are handled exquisitely by existing codes (SpECTRE, GRChombo, Einstein Toolkit). GRANITE's scientific domain begins where these codes reach their limits: multiple compact objects simultaneously, in the presence of magnetized matter, radiation, and neutrino physics.

---

2. Why Existing Codes Are Insufficient

Code	Strength	Key Gap
Einstein Toolkit	Mature, community-tested	Modular thorn system creates coupling barriers; N>2 BH not standard
GRChombo	Clean C++17, AMR, BBH/BNS	No M1 radiation, no neutrino leakage
SpECTRE	Spectral accuracy	GRMHD partial, no radiation, massive engineering overhead
AthenaK	Excellent GRMHD, GPU-native	No dynamic spacetime (metric fixed or GRFFE only)
GRANITE	All of the above in one framework	Under active development — v0.6.8 validated BBH; full multi-physics v0.7+

The fundamental difficulty is not implementing any single piece of physics — it is coupling them correctly in a single evolution loop: CCZ4 metric updated at every RK3 stage, GRMHD reading it, radiation coupling to matter, all with AMR tracking moving horizons.

---

3. The B5_star Flagship Scenario

The B5_star benchmark is GRANITE's primary scientific motivation. No existing code can simulate it at full physics fidelity.

3.1 Configuration

• 5 supermassive black holes, each 10⁸ M☉, arranged in a regular pentagon at 1 pc radius
• 2 ultra-massive stars, each ~4300 M☉, at the center
• Full physics: CCZ4 + GRMHD + M1 + neutrino leakage + 12 AMR levels
• Evolution time: 10⁵ years

3.2 Physical Sequence

Phase I — Stellar Disruption (t ~ 10³–10⁴ yr):
Central stars are tidally disrupted by the SMBH quintuple. TDE flares observable in X-ray and UV. Disrupted matter forms accretion structures around the nearest SMBHs.

Phase II — SMBH Inspiral (t ~ 10⁴–10⁵ yr):
Gravitational radiation and dynamical friction drive the SMBHs inward. Mergers occur sequentially. Each merger produces a burst of gravitational waves in the PTA/LISA frequency band.

Phase III — Remnant Evolution:
Final merged SMBH accretes residual matter and launches MHD jets. Total mass radiated: ΔM ~ η M_total c² with η ~ 0.05.

3.3 Multi-Messenger Signatures

Messenger	Observable	Band	Detector
Gravitational waves	Inspiral chirp	10⁻⁹–10⁻⁷ Hz	PTA (NANOGrav, EPTA, PPTA)
Gravitational waves	Ringdown	10⁻⁴–10⁻³ Hz	LISA
X-ray	TDE flares	0.1–10 keV	Chandra, XMM, eROSITA
Radio	MHD jet	GHz	VLBI, SKA
Neutrinos	Accretion disc	MeV	IceCube

3.4 Computational Requirements

Parameter	Value
AMR levels	12
Grid cells at peak	~10¹⁰ across all levels
RAM estimate	~2 TB
CPU-hours estimate	~5 × 10⁶
GPU-hours (H100)	~5 × 10⁵

Scaling path: 128³ desktop → GPU porting (vast.ai H100) → 256–512³ cluster → 12-level flagship.

---

4. Analytic Pre-Validation (NRCF / PRISM)

Before the full simulation, key observables are pre-validated against analytic estimates from predecessor frameworks NRCF and PRISM:

Observable	NRCF/PRISM Estimate	Tolerance
Peak GW frequency	From chirp mass formula	±10%
Total GW energy radiated	η M c² (η ~ 0.05)	±15%
Merger shock temperature	T ~ GM μ / (k_B r_merger)	±20%
Final spin parameter	a/M ≲ 0.7 (NR bound)	±5%

---

5. Physical Unit System

GRANITE uses geometrized units: G = c = 1, M_total = 1.

Quantity	Code	M = 1 M☉	M = 10⁸ M☉
Length	M	1.48 km	1.48 × 10¹³ cm ≈ 1 AU
Time	M/c	4.93 μs	493 s
GW frequency	c/M	203 kHz	~2 mHz

Critical: 1 km = 1.0 × 10⁵ cm (not RSUN_CGS). See Known Fixed Bugs: TOV.

---

See also: Physics Formulations | Benchmarks & Validation | Roadmap