Skip to content
 
 

Latest commit

 

History

86 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 

Repository files navigation

# RSVDPACK baseline for Comparison in HH-BARRF

This repository is a modified and reduced version of **RSVDPACK / LowRankMatrixDecompositionCodes**, originally developed by Sergey Voronin as part of a collaboration with Per-Gunnar Martinsson.

Original repository:

https://github.com/sergeyvoronin/LowRankMatrixDecompositionCodes

This fork is used as the baseline implementation for experiments in:

> A GPU-Accelerated Blocked Adaptive Randomized Range Finder
Based on an Implicit Householder QR Decomposition

by Carolin Penke and Andreas Herten.

A guide on how to reproduce experiments in that paper is found here: https://gitlab.jsc.fz-juelich.de/penke3/reproduce-barrf-householder.

## Modifications

This version has been adapted for the experimental setup used in the paper.

Compared with the upstream repository:

* components not required for the experiments have been removed;
* Makefiles were added/adapted for the target systems;
* support for NVIDIA Performance Libraries (NVPL) was added;
* the retained CPU and GPU implementations were adapted as required for the benchmark setup.

Modified by Carolin Penke, 2026.

The underlying RSVDPACK implementation is used as the baseline method.
No changes were made to the baseline algorithm itself.

## Contents

* `lapack_code/` — CPU implementation used for the experiments
* `nvidia_gpu_cublas_code/` — CUDA/cuBLAS implementation used for the experiments

See the Makefiles in the respective directories for the supported BLAS/LAPACK configurations.

## License

The original RSVDPACK code is distributed under the **GNU General Public License, version 3 (GPLv3)**.

This modified version is distributed under the same license. See `LICENSE` for the full license text.

Copyright and authorship of the original code remain with the original authors. Modifications made in this fork are identified above.

About

RSVDPACK: Implementations of fast algorithms for computing the low rank SVD, interpolative and CUR decompositions of a matrix, using randomized sampling. Includes codes for single core (using GNU GSL), multi-core (using Intel MKL) and GPU (using NVIDIA CUDA/CULA) architectures.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages