MUMPSpy

A python wrapper for the sparse direct solver.

The wrapper allows:

• real and complex factorization of symmetric or non-symmetric matrices
• single and double precision
• Schur complement calculation

Requirements

• MUMPS - MUltifrontal Massively Parallel sparse direct Solver
• mpi4py - Python bindings for MPI

Ubuntu/Debian users can use the following command to install the required packages:

apt-get install python-mpi4py libmumps-dev

The cmake version of MUMPS could be also used. See https://github.com/scivision/mumps for details and CI file for some syntax.

Installation

  pip install mumpspy

Usage

import mumpspy
import numpy as np
import scipy.sparse as sp

row = np.array([0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 3, 3])
col = np.array([0, 1, 2, 3, 0, 1, 2, 3, 0, 1, 0, 1])
val = np.array([1, 2, 2, 1, 1, 3, -1, 2, 1, 1, 3, 1], dtype='d')
b = np.array([15, 12, 3, 5], dtype='d')

A = sp.coo_matrix((val, (row, col)), shape=(4, 4))

solver = mumpspy.MumpsSolver()  # initialize solver, real-valued system
solver.set_mtx(A)  # set sparse matrix
x = solver.solve(b)  # solve system for a given right-hand side
print(x)

Additional examples of syntax can be found in test file.

Testing

pytest can be used to run testing and coverage in the package.

Compatibility

Tested for the following MUMPS library versions (see Actions for current working versions):

• 4.10.0
• 5.0.2
• 5.1.2
• 5.2.1
• 5.4.1
• 5.5.x
• 5.6.x
• 5.7.x
• 5.8.2