D

distributed-computing

ScaLAPACK for Python (scalapy) provides a high-level Python interface to distributed dense linear algebra using MPI, BLACS, and ScaLAPACK. It enables efficient computation on large matrices using a block-cyclic distribution, while exposing a NumPy-like API based on the DistributedMatrix abstraction.

This fork modernizes and stabilizes the original scalapy project with: • Support for Python 3.x on modern Linux systems • Clean MPI + ScaLAPACK detection (OpenMPI + Netlib ScaLAPACK) • Reproducible builds using a known-good NumPy / Cython / mpi4py stack • Documentation, examples, and automated tests • High-level wrappers for SVD, QR, LU, Cholesky, eigenproblems, and DMD workflows • Efficient distributed transposes via the new p?lapv2 wrapper

The project makes it possible to write dense parallel linear-algebra pipelines in Python while executing BLAS/PBLAS/ScaLAPACK kernels across an MPI communicator.

This is the central "hub" project for the Mutually Assured Accountability (MAA) Framework. This repository contains the master README, which provides the unified vision, architecture, and links to all component projects.