Scaling Fourier-Based Sparse Matrix Analysis on GPUs

DOI10.48550/arXiv.2609.20483MaRDI QIDQ7353383arXivFDO

Authors Ruifeng Zhang, Xipeng Shen, Sai Krishna Teja Varma Manthena, Jiajia Li

Publication date 17 September 2026

Published in arXiv

Tldr We design and implement a lossless Binary-Sparse FFT (BS-FFT) for GPU acceleration, along with two compression methods: Elastic BS-FFT, which samples frequency grids while reusing BS-FFT pipeline, and density-map-based spatial compression; these techniques significantly reduce memory usage and computation time while preserving spectral patterns.

Full work available at URL https://doi.org/10.48550/arXiv.2609.20483

arXiv classification

cs.PF

Cites work

• A parallel non-uniform fast Fourier transform library based on an "exponential of semicircle" kernel
• A short note on an inequality between KL and TV
• An Algorithm for the Machine Calculation of Complex Fourier Series
• cover
• cuFINUFFT: a load-balanced GPU library for general-purpose nonuniform FFTs
• Title not available (Why is no real title available?)
• Introduction to probability
• Iterative Methods for Sparse Linear Systems
• Modeling the shape of the scene: A holistic representation of the spatial envelope
• Multi-scale attributed node embedding
• Neue Begründung der Theorie quadratischer Formen von unendlichvielen Veränderlichen.
• On Choosing and Bounding Probability Metrics
• The University of Florida sparse matrix collection

This page was built for publication: Scaling Fourier-Based Sparse Matrix Analysis on GPUs

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7353383)