When matter falls toward a black hole, it doesn’t go quietly. The superheated accretion disk blazes across the electromagnetic spectrum, from radio waves to X-rays, creating some of the most luminous objects in the universe. Understanding these cosmic lighthouses requires solving the complex physics of photon transport through the warped spacetime near black holes—a computationally demanding challenge that has traditionally required massive CPU resources and patience.
GPUmonty transforms this bottleneck by unleashing GPU parallelization on relativistic Monte Carlo radiative transfer. The code traces millions of photon packets through curved spacetime using the Kerr metric, modeling synchrotron emission from hot electrons and Compton scattering as photons navigate the extreme gravitational environment. With over 10x speedup compared to CPU-based codes like grmonty, researchers can now iterate rapidly through parameter space, testing different black hole spins, viewing angles, and accretion models against observational data from sources like Sgr A* and M87*.
This GPU-accelerated approach opens new possibilities for constraining black hole properties through electromagnetic observations and testing general relativity in the strong-field regime. The code interfaces seamlessly with GRMHD simulations from iharm3D and harm, making it an essential tool for the growing field of computational black hole astrophysics and multi-messenger astronomy campaigns.
⭐ Stars: 3
💻 Language: Cuda
🔗 Repository: black-hole-group/gpumonty