{"id":"arxiv/2610.28372","url":"https://arxiv.org/abs/2610.28372","tags":["math.NA"],"score":0,"title":"Fast and scalable multi-GPU sliced FFT solvers for asymptotically compatible discretizations of 3D nonlocal diffusion problems","domain":"math.NA","source":"arxiv","authors":["Yulong Li","Jiale Meng","Zhenlin Guo","Jiwei Zhang"],"reviews":0,"subject":"math.NA","summary":"Large-scale 3D peridynamic transient diffusion simulations are constrained by the computational cost and memory requirements of nonlocal interactions. We propose a fast and scalable multi-GPU sliced FFT algorithm based on the recently developed asymptotically compatible finite difference (ACFD) scheme. By exploiting the discrete operator's structure and finite interaction range, the algorithm combines FFT acceleration with local inter-GPU communication. It preserves the ACFD scheme while avoiding the global data redistribution required by full 3D FFTs. We build a matrix-free GPU conjugate…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-10-20","reactions":{}}