cuDART: a GPU-accelerated ray tracing code for generating synthetic observations of relativistic astrophysical sources
Abstract
The generation of synthetic observations is required to compare simulations of astrophysical phenomena with real sources. We present a new open-source toolkit cuDART which allows meshed simulation data to be rapidly rendered into synthetic observations of optically thin emission, automatically accounting for a range of relativistic and geometric effects such as Doppler boosting. Unlike previous visualisation schemes which generally consider a single snapshot in time, cuDART is capable of accounting for the finite time delay across astrophysical distances by rendering data from multiple simulation epochs simultaneously. We demonstrate how failing to account for a finite light time delay can result in synthetic observations with incorrect observed motion, morphologies and fluxes. The code does not require a static pattern frame for integration, relaxing standard assumptions required for analytical observational estimates. The rendering processes is accelerated through hardware (using GPUs) and software (using the 3D differential digital analyser algorithm), allowing for rapid generation of observations from a large number of viewing orientations. In this paper we describe the physical motivation, document the code operation and test the output against theoretical relativistic and geometric expectations. cuDART is publicly available via GitHub and documented using ReadTheDocs.
Related
- Relativistic Spherical Accretion in Nonrotating Black Hole Spacetimes: Astrophysical Implications
- Charged-current neutrino opacity within the relativistic Hartree-Fock framework for astrophysical simulations of core-collapse supernovae and binary neutron star mergers
- Chana: A GPU-Accelerated Adaptive-Mesh Code for General Relativistic Magnetohydrodynamics with Multigroup M1 Neutrino Transport
- FUGUE: A New GPU-accelerated General Relativistic Magnetohydrodynamics Code with Weak-Field Multipole Self-Gravity and Quasi-Stationary Kerr Black Hole Evolution
- Probing the details of relativistic electrons with multifrequency observations of M87 black hole
BibTeX
@misc{whitehead2026cudart,
title = {{cuDART: a GPU-accelerated ray tracing code for generating synthetic observations of relativistic astrophysical sources}},
author = {Henry Whitehead and Emma L. Elley and Christoper N. Everett},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.09322},
eprint = {2609.09322},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.09322},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - cuDART: a GPU-accelerated ray tracing code for generating synthetic observations of relativistic astrophysical sources
AU - Henry Whitehead
AU - Emma L. Elley
AU - Christoper N. Everett
PY - 2026
DA - 2026/09/08
PB - arXiv
DO - 10.48550/arxiv.2609.09322
UR - https://arxiv.org/abs/2609.09322
AB - The generation of synthetic observations is required to compare simulations of astrophysical phenomena with real sources. We present a new open-source toolkit cuDART which allows meshed simulation data to be rapidly rendered into synthetic observations of optically thin emission, automatically accounting for a range of relativistic and geometric effects such as Doppler boosting. Unlike previous visualisation schemes which generally consider a single snapshot in time, cuDART is capable of accounting for the finite time delay across astrophysical distances by rendering data from multiple simulation epochs simultaneously. We demonstrate how failing to account for a finite light time delay can result in synthetic observations with incorrect observed motion, morphologies and fluxes. The code does not require a static pattern frame for integration, relaxing standard assumptions required for analytical observational estimates. The rendering processes is accelerated through hardware (using GPUs) and software (using the 3D differential digital analyser algorithm), allowing for rapid generation of observations from a large number of viewing orientations. In this paper we describe the physical motivation, document the code operation and test the output against theoretical relativistic and geometric expectations. cuDART is publicly available via GitHub and documented using ReadTheDocs.
ER -