Hardware Acceleration for HPS Algorithms in Two and Three Dimensions

Owen Melia, Daniel Fortunato, Jeremy Hoskins, Rebecca Willett · 2025

Read Hardware Acceleration for HPS Algorithms in Two and Three Dimensions by Owen Melia, Daniel Fortunato, Jeremy Hoskins, Rebecca Willett (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-17535.txt.

Description

Owen Melia, Daniel Fortunato, Jeremy Hoskins, and Rebecca Willett move a fast equation-solving method onto graphics processors. Their paper appeared in 2025. The HPS family builds direct solvers for the elliptic equations behind heat, flow, and electrostatics by patching together local solutions. The team wrote an open framework that runs these solvers massively in parallel on GPUs, with two redesigns for the hardware.

In two dimensions, a recomputation trick cuts expensive data transfers to the card. In three, an adaptive mesh scheme slashes peak memory. Built in a modern accelerated algebra package, the solver also plugs into automatic differentiation tools, a first for this class. Tests in two and three dimensions showed speed with high-order accuracy.

Catalog record

CREATOROwen Melia, Daniel Fortunato, Jeremy Hoskins, Rebecca Willett
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE110 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-17535

Browse the full library · About the Tower of Babel