ScaMaC: A Scalable Matrix Collection

Warning

At present (version 0.8.2), the documentation is still incomplete.

Overview

The scalable matrix collection ScaMaC provides routines for the generation and exploration of sparse matrices of scalable size. It allows for the generation of various application and benchmark matrices, including several examples from real-world and scientific research applications. Through a unified interface, generation of small test matrices with a few hundred rows and columns is as straightforward as the generation of huge billion-dimensional matrices for distributed computing.

The ScaMaC consists of three components:

• the ScaMaC library with the top-level header scamac.h
• the ScaMaC toolkit with the top-level header scamactools.h
• the ScMaC toolkit application scamact

This documentation includes

• the general ScaMaC - a Scalable Matrix Collection README from the top-level source directory, with hints on installation and usage
• a complete list of the ScaMaC matrices and benchmark suites

Author

Andreas Alvermann (University of Greifswald)

Date

May 2019