https://github.com/f-koehler/itesol

WIP: Iterative eigensolvers for C++20, Python and CUDA
https://github.com/f-koehler/itesol

cpp20 cuda eigenvalues linear-algebra python

Last synced: about 2 months ago
JSON representation

WIP: Iterative eigensolvers for C++20, Python and CUDA

• Host: GitHub
• URL: https://github.com/f-koehler/itesol
• Owner: f-koehler
• License: mit
• Created: 2022-10-19T23:02:01.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2023-10-18T03:14:35.000Z (over 1 year ago)
• Last Synced: 2024-11-07T11:53:06.298Z (3 months ago)
• Topics: cpp20, cuda, eigenvalues, linear-algebra, python
• Language: C++
• Homepage:
• Size: 94.7 KB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE.txt

Awesome Lists containing this project

README

        
# itesol

:warning: This is work in progress software currently in an early design phase.

[![CI](https://github.com/f-koehler/itesol/actions/workflows/ci.yml/badge.svg)](https://github.com/f-koehler/itesol/actions/workflows/ci.yml)

[![GitHub](https://img.shields.io/github/license/f-koehler/itesol)](https://github.com/f-koehler/itesol/blob/main/LICENSE.txt)

A C++20 header-only library to implement iterative eigensolvers ([power method](https://en.wikipedia.org/wiki/Power_iteration), [Lanczos algorithm](https://en.wikipedia.org/wiki/Lanczos_algorithm), …).

By heavily relying on template metaprogramming we try to interface with various different linear algebra backends (Eigen, Blaze, BLAS/LAPACK, CUDA) and build algorithms from simple elementary building blocks that can be easily implemented for a new custom backend. 

A Python module is also worked on.

# Roadmap

## Algorithms

- [x] Power method

- [ ] Lanczos

- [ ] Implicitly restarted Lanczos

- [ ] Arnoldi iterations

- [ ] ARPACK wrapper

## Backends

- [x] Eigen3

- [ ] Blaze

- [ ] CUDA

- [x] BLAS/LAPACK (experimental)