cuPentBatch -- A batched pentadiagonal solver for NVIDIA GPUs

TL;DR

cuPentBatch is a GPU-optimized batched pentadiagonal solver designed for numerical PDE applications, outperforming standard NVIDIA solvers in relevant physical modeling scenarios.

Contribution

We developed cuPentBatch, a novel batched pentadiagonal solver optimized for NVIDIA GPUs, tailored for PDE applications with variable right-hand sides.

Findings

cuPentBatch outperforms NVIDIA's gpsvInterleavedBatch in relevant problems.

The solver is efficient for parameter studies involving PDEs.

Demonstrated improvements in computational speed and accuracy.

Abstract

We introduce cuPentBatch -- our own pentadiagonal solver for NVIDIA GPUs. The development of cuPentBatch has been motivated by applications involving numerical solutions of parabolic partial differential equations, which we describe. Our solver is written with batch processing in mind (as necessitated by parameter studies of various physical models). In particular, our solver is directed at those problems where only the right-hand side of the matrix changes as the batch solutions are generated. As such, we demonstrate that cuPentBatch outperforms the NVIDIA standard pentadiagonal batch solver gpsvInterleavedBatch for the class of physically-relevant computational problems encountered herein.