HONEI: A collection of libraries for numerical computations targeting multiple processor architectures

TL;DR

HONEI is an open-source library collection that enables efficient numerical computations across diverse hardware architectures like CPUs, GPUs, and Cell processors, simplifying development and achieving significant speedups.

Contribution

It introduces a hardware-abstracted library collection that facilitates portable and optimized numerical computations on multiple architectures, with demonstrated performance improvements.

Findings

Twofold speedup over naive C++ implementation using SSE backend

3-16 times faster execution on Cell and GPU architectures

Simplifies development of optimized hardware-specific kernels

Abstract

We present HONEI, an open-source collection of libraries offering a hardware oriented approach to numerical calculations. HONEI abstracts the hardware, and applications written on top of HONEI can be executed on a wide range of computer architectures such as CPUs, GPUs and the Cell processor. We demonstrate the flexibility and performance of our approach with two test applications, a Finite Element multigrid solver for the Poisson problem and a robust and fast simulation of shallow water waves. By linking against HONEI's libraries, we achieve a twofold speedup over straight forward C++ code using HONEI's SSE backend, and additional 3-4 and 4-16 times faster execution on the Cell and a GPU. A second important aspect of our approach is that the full performance capabilities of the hardware under consideration can be exploited by adding optimised application-specific operations to the HONEI libraries. HONEI provides all necessary infrastructure for development and evaluation of such kernels, significantly simplifying their development.