Fully homomorphic encryption (FHE) has recently attracted significant attention as both a cryptographic primitive and a systems challenge. Given the latest advances in accelerated computing, FHE presents a promising opportunity for progress, with applications ranging from machine learning to information security. We target the most computationally intensive operation in deep neural networks from a hardware perspective, matrix multiplication (matmul), and adapt it for execution on AMD GPUs. We propose a new optimized method that improves the runtime and complexity of ciphertext matmul by using FIDESlib, a recent open-source FHE library designed specifically for GPUs. By exploiting sparsity in both operands, our sparse matmul implementation outperforms its CPU counterpart by up to 3.0× and reduces the time complexity from cubic to semi-linear, demonstrating an improvement over existing FHE matmul implementations.
@article{arxiv.2604.11659,
title = {GPU Acceleration of Sparse Fully Homomorphic Encrypted DNNs},
author = {Lara D'Agata and Carlos Agulló-Domingo and Óscar Vera-López and Kaustubh Shivdikar and Ardhi W. B. Yudha and Ferhat Yaman and David Kaeli and José L. Abellán and Ian Colbert and José Cano},
journal= {arXiv preprint arXiv:2604.11659},
year = {2026}
}
Comments
Accepted to the 6th Workshop on Machine Learning and Systems (EuroMLSys) co-located with EuroSys '26