sTiles

High-performance tile-based framework for sparse Cholesky factorization, triangular solves, and selected inversion of symmetric positive-definite matrices, spanning the full spectrum from very sparse to fully dense.

Developed at King Abdullah University of Science and Technology (KAUST).

Documentation

Please follow esmail-abdulfattah.github.io/sTiles for installation, downloads, and examples for Python, R, Julia, and C/C++.

This repository

This repository holds the language bindings and the project website. The libstiles engine itself is distributed as prebuilt binaries for Linux (x86_64, arm64), macOS (Apple Silicon, Intel), and Windows (x86_64), published as release assets. None of the packages needs a compiler or a build step. The Python and Julia packages fetch the binary for your platform on first use; the R package asks first, and installs it on request with sTiles::sTiles_install_library().

pip install sTiles
install.packages("sTiles")
pkg> add https://github.com/esmail-abdulfattah/sTiles:julia

Contact

Esmail Abdul Fattah, esmail.abdulfattah@kaust.edu.sa