Ethosoft
Open research projectAstrophysics · SPH · Bayesian inference

DART-RIFT

A deterministic, GPU-accelerated SPH shock-physics and Bayesian-inference pipeline for estimating which Dimorphos interior structures are supported by NASA DART impact observables.

Project record

This page reports the maturity and evidence documented in Ethosoft's working archive. It does not imply peer review, regulatory clearance, or clinical validation.

Methods and evidence

Open research projectAstrophysics · SPH · Bayesian inference

DART-RIFT

A deterministic, GPU-accelerated SPH shock-physics and Bayesian-inference pipeline for estimating which Dimorphos interior structures are supported by NASA DART impact observables.

Forward model
Tillotson equation of state, pressure-dependent strength, P-α porosity, Barnes–Hut self-gravity, and a resolved DART projectile on a PDS-derived Dimorphos shape.
Validation discipline
The repository documents analytic and experimental checks including Sod, Sedov–Taylor, Taylor-bar, basalt Hugoniot, conservation, determinism, and CPU↔GPU comparisons.
Inference plan
Hundreds of interior-parameter simulations train a surrogate model; Bayesian inference then produces a posterior distribution rather than a single asserted interior.

Evidence boundary: the repository describes a validated simulation foundation and an inference target. It explicitly does not claim a completed Dimorphos scientific result before Hera observations; impact runs and posterior conclusions remain future project phases.

Expanded project brief

A falsifiable inverse problem for Dimorphos.

DART-RIFT works backward from the NASA DART impact: instead of choosing one asteroid interior and reporting one answer, it samples uncertain interior parameters such as bulk density, porosity, cohesion, and coarse-block fraction, runs a forward impact simulation, and learns how those parameters map to observables. Bayesian inference over that surrogate then returns a posterior distribution showing which interiors the data support.

The repository's foundation is a deterministic, GPU-accelerated SPH engine with a Tillotson equation of state, pressure-dependent strength, P-α porosity, Barnes–Hut self-gravity, a resolved projectile, and a PDS-derived Dimorphos shape. Its engineering record emphasizes gates and negative evidence: analytic and experimental tests, bit-level CPU/GPU comparisons, clean-tree runs, and explicit statements that passing validation gates does not yet constitute a scientific result about Dimorphos.

Scientific boundary: the project is designed to lock its prior prediction before Hera observations and make it falsifiable. Forward-model error must remain part of the posterior; a fast but miscalibrated simulator would produce overconfident conclusions.