Performance Portable SU(N) Lattice Gauge Theory Simulation with Kokkos
Abstract
The increasing diversity of high performance computing systems makes separate, architecture specific implementations of lattice gauge theory algorithms costly to maintain. We present \texttt{kwqft}, a performance portable Kokkos implementation of Wilson pure gauge Monte Carlo simulation for SU(N) Yang-Mills theory in an arbitrary number of space-time dimensions. The gauge group order N and the dimension D are compile time parameters. A single source targets the Serial, OpenMP, CUDA, HIP, and SYCL execution spaces, with MPI halo exchange overlapped with interior updates. The implementation reproduces the exact two-dimensional plaquette and published three and four dimensional values for gauge groups up to SU(17). On an NVIDIA A100 the Kokkos CUDA backend is competitive with a native CUDA code, SIMD acceleration improves the OpenMP path on Armv9 processors, and a large scale speedup is demonstrated for an SU(4) lattice on the LineShine supercomputer, currently ranked first on the TOP500 list.
Related
- ORB5X: a performance-portable global electromagnetic gyrokinetic PIC code in C++/Kokkos built using Agentic AI
- Tame the Umklapp Processes in Real-Time Lattice Simulation for Hydrodynamics: An Ising Field Theory Study
- Recovering What Breaks in a Trapped-Ion Quantum Simulation of a Lattice Gauge Theory
- Trapped-ion quantum simulation of (1+1)D SU(2) lattice gauge theory via loop-string-hadron formulation
- Spin Hamiltonian for Quantum Simulation of Minimally Truncated 3+1D SU(2) Lattice Gauge Theory and Testing Eigenstate Thermalization
BibTeX
@misc{sun2026performance,
title = {{Performance Portable $\mathrmSU(N)$ Lattice Gauge Theory Simulation with Kokkos}},
author = {Wei Sun},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.09766},
eprint = {2610.09766},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.09766},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Performance Portable SU(N) Lattice Gauge Theory Simulation with Kokkos
AU - Wei Sun
PY - 2026
DA - 2026/10/07
PB - arXiv
DO - 10.48550/arxiv.2610.09766
UR - https://arxiv.org/abs/2610.09766
AB - The increasing diversity of high performance computing systems makes separate, architecture specific implementations of lattice gauge theory algorithms costly to maintain. We present \texttt{kwqft}, a performance portable Kokkos implementation of Wilson pure gauge Monte Carlo simulation for SU(N) Yang-Mills theory in an arbitrary number of space-time dimensions. The gauge group order N and the dimension D are compile time parameters. A single source targets the Serial, OpenMP, CUDA, HIP, and SYCL execution spaces, with MPI halo exchange overlapped with interior updates. The implementation reproduces the exact two-dimensional plaquette and published three and four dimensional values for gauge groups up to SU(17). On an NVIDIA A100 the Kokkos CUDA backend is competitive with a native CUDA code, SIMD acceleration improves the OpenMP path on Armv9 processors, and a large scale speedup is demonstrated for an SU(4) lattice on the LineShine supercomputer, currently ranked first on the TOP500 list.
ER -