jaxFMM: An Adaptive, GPU-Parallel Implementation of the Fast Multipole Method in JAX

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kraft, Robert, Bruckner, Florian, Suess, Dieter, Abert, Claas
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914164265451520
author Kraft, Robert
Bruckner, Florian
Suess, Dieter
Abert, Claas
author_facet Kraft, Robert
Bruckner, Florian
Suess, Dieter
Abert, Claas
contents We introduce jaxFMM, an open-source, adaptive, highly parallel point-charge Fast Multipole Method implementation for the Laplace kernel written in JAX. It is based on a non-uniform refinement strategy, which results in extremely concise and simple code. Benchmarks show that the algorithm performs well even for highly non-uniform charge distributions. JaxFMM already massively speeds up stray-field computations in micromagnetics and with JAX features like autodiff, novel applications such as inverse-design problems and machine-learning tasks can be tackled with ease in the future.
format Preprint
id arxiv_https___arxiv_org_abs_2511_15269
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle jaxFMM: An Adaptive, GPU-Parallel Implementation of the Fast Multipole Method in JAX
Kraft, Robert
Bruckner, Florian
Suess, Dieter
Abert, Claas
Computational Physics
65R20
We introduce jaxFMM, an open-source, adaptive, highly parallel point-charge Fast Multipole Method implementation for the Laplace kernel written in JAX. It is based on a non-uniform refinement strategy, which results in extremely concise and simple code. Benchmarks show that the algorithm performs well even for highly non-uniform charge distributions. JaxFMM already massively speeds up stray-field computations in micromagnetics and with JAX features like autodiff, novel applications such as inverse-design problems and machine-learning tasks can be tackled with ease in the future.
title jaxFMM: An Adaptive, GPU-Parallel Implementation of the Fast Multipole Method in JAX
topic Computational Physics
65R20
url https://arxiv.org/abs/2511.15269