Fast GPU Linear Algebra via Compile Time Expression Fusion

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Curtin, Ryan R., Edel, Marcus, Sanderson, Conrad
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914504490614784
author Curtin, Ryan R.
Edel, Marcus
Sanderson, Conrad
author_facet Curtin, Ryan R.
Edel, Marcus
Sanderson, Conrad
contents We describe the Bandicoot GPU linear algebra toolkit, a C++ based library that prioritises ease of use without compromising efficiency. Bandicoot's API is compatible with the popular Armadillo CPU linear algebra library, enabling easy transition for existing CPU-based codebases. Unlike other GPU-focused toolkits, Bandicoot uses template metaprogramming to generate fused GPU kernels directly at compile time, yielding efficient kernels that are often able to saturate memory bandwidth. This removes the need for runtime overhead or JIT infrastructure. Empirical results show that Bandicoot outperforms (sometimes by considerable margins) commonly-used linear algebra toolkits including PyTorch, TensorFlow, and JAX.
format Preprint
id arxiv_https___arxiv_org_abs_2604_22242
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Fast GPU Linear Algebra via Compile Time Expression Fusion
Curtin, Ryan R.
Edel, Marcus
Sanderson, Conrad
Mathematical Software
68N99, 65Y04, 65Y15, 65F45
G.4; G.1.3; D.2.3; F.2.1
We describe the Bandicoot GPU linear algebra toolkit, a C++ based library that prioritises ease of use without compromising efficiency. Bandicoot's API is compatible with the popular Armadillo CPU linear algebra library, enabling easy transition for existing CPU-based codebases. Unlike other GPU-focused toolkits, Bandicoot uses template metaprogramming to generate fused GPU kernels directly at compile time, yielding efficient kernels that are often able to saturate memory bandwidth. This removes the need for runtime overhead or JIT infrastructure. Empirical results show that Bandicoot outperforms (sometimes by considerable margins) commonly-used linear algebra toolkits including PyTorch, TensorFlow, and JAX.
title Fast GPU Linear Algebra via Compile Time Expression Fusion
topic Mathematical Software
68N99, 65Y04, 65Y15, 65F45
G.4; G.1.3; D.2.3; F.2.1
url https://arxiv.org/abs/2604.22242