BNS_NURATES: A GPU friendly library for neutrino opacities written in c++ with GPU support

BNS_NURATES requires the following dependencies: - A C++ compiler which supports c++17 - The GNU scientific library - Kokkos - OpenMP

Installation

BNS_NURATES is a header-only library, so it can be simply included into your code, provided the requirements listed above are satisfied. The various headers are in the include directory.

Stand alone build and testing

Clone the repository recursively to also clone Kokkos which is available as a submodule: git clone --recursive https://github.com/RelNucAs/bns_nurates.git cd bns_nurates Create a build directory: mkdir build cd build By default Kokkos and OpenMP are enabled. To compile for a CPU based system run cmake ../ and build nurates cmake --build . --target nurates If the code is intended for Nvidia GPUs (say, A100 in this case), enable CUDA during compilation cmake -DKokkos_ENABLE_CUDA=ON -DKokkos_ARCH_AMPERE80=ON ../

Similarly, to disable OpenMP, add the flag -DENABLE_OPENMP=OFF to cmake.

Using bns_nurates as a thorn for the Einsten Toolkit

bns_nurates is also meant for use as a thorn Weakrates2 for Cactus with THC. To do this, one must export these files:

python utils/export_thc.py /path/to/bns_nurates/ /path/to/destination/

Here /path/to/bns_nurates/ is the full path for the top-level bns_nurates directory and /path/to/destination/ is the full path to the src folder insider Weakrates2

Purpose of the library

BNS_NURATES can be exploited to compute both energy-dependent (spectral) and energy-integrated (gray) emissivities and opacities for the most relevant neutrino-matter interactions in BNS mergers. The definitions of such quantites can be found in [^fn1], which also discusses in more details the design of the library. A self-explanatory minimal working example is provided in the repo (mwe.cpp), which can be compiled within the build folder as follows cmake --build . --target MWE Thermodynamic conditions extracted from a BNS merger simulation (used for the postprocessing analysis in [^fn1]) are also provided in the inputs/BNS folder.

Python bindings

Bindings can be generated for calling BNSNURATES functions directly from Python. This is done through SWIG (tested version: 4.0.2). Once SWIG is installed on the system, enter the bindings/ folder cd bindings and run the following command ``` python3 setup.py buildext --inplace This creates a Python module namedbnsnuratescan that imported as import bnsnurates as bns To make the module available to your Python environment either add thebindings/folder to your PYTHONPATH or install the module.bindings/test_bindings.py``` is a minimal working example that can be used for testing the bindings.

Included neutrino-matter interactions

BNS_NURATES implements the following neutrino reactions (see also [^fn1]):

• Beta processes

  Implemented as in [^fn2], [^fn3] under the assumption of zero-momentum transfer and non-relativistic nucleons. We account for the impact from relativistic mean-field effects as in [^fn3], [^fn4], and the correction due to the sum of phase-space, recoil and weak magnetism effects as in [^fn5]. They include:

  • Neutrino absorptions on nucleons / $e^\pm$ captures ### $\nue + n \leftrightarrow e^- + p$ ### $\bar{\nu}e + p \leftrightarrow e^+ + n$
  • (Inverse) nucleon decay ### $\nue + n \leftrightarrow e^- + p$ ### $\bar{\nu}e + p \leftrightarrow e^+ + n$

• Pair processes

  • $e^+ e^-$ annihilation (implemented as in [^fn6]) ### $e^+ + e^- \leftrightarrow \nu + \bar{\nu}$
  • Nucleon-nucleon bremsstrahlung (implemented as in [^fn7], including the medium modification from [^fn8]) ### $N + N \leftrightarrow N + N + \nu + \bar{\nu}$

• Scattering processes

  • Isoenergetic scattering off nucleons (implemented as in [^fn2], including phase-space, recoil and weak magnetism effects as in [^fn5]) ### $\nu + N \rightarrow \nu + N$
  • Inelastic scattering off $e^\pm$ (implemented as in [^fn2], [^fn9]) ### $\nu + e^\pm \rightarrow \nu + e^\pm$

[^fn1]: L. Chiesa et al., Phys. Rev. D 111, 063053 (2025) [^fn2]: S. W. Bruenn, Astrophys. J. Suppl. Ser. 58, 771 (1985) [^fn3]: M. Oertel, A. Pascal, M. Mancini, and J. Novak, Phys. Rev. C 102, 035802 (2020) [^fn4]: M. Hempel, Phys. Rev. C 91, 055807 (2015) [^fn5]: C. J. Horowitz, Phys. Rev. D 65, 043001 (2002) [^fn6]: J. A. Pons, J. A. Miralles, and J. M. Ibanez, Astron. Astrophys. Suppl. Ser. 129, 343 (1998) [^fn7]: S. Hannestad and G. Raffelt, Astrophys. J. 507, 339 (1998) [^fn8]: T. Fischer, Astron. Astrophys. 593, A103 (2016) [^fn9]: A. Mezzacappa and S. W. Bruenn, Astrophys. J. 410, 740 (1993) [^fn10]:A. Burrows, S. Reddy, and T. A. Thompson, Nucl. Phys. A777, 356 (2006)