Skip to content

Latest commit

 

History

History
53 lines (32 loc) · 2.43 KB

README.md

File metadata and controls

53 lines (32 loc) · 2.43 KB

Mongoose

Build Status codecov

Mongoose is a graph partitioning library. Currently, Mongoose only supports edge partitioning, but in the future a vertex separator extension will be added.

Prerequisites and Dependencies

Mongoose requires CMake 2.8 and any ISO/IEC 14882:1998 compliant C++ compiler. Mongoose has been tested to work with GNU GCC 4.4+ and LLVM Clang 3.5+ on Linux, and Apple Xcode 6.4+ on macOS.

Installation

Mongoose uses CMake. To build Mongoose, follow the commands below:

git clone https://github.com/ScottKolo/Mongoose
cd Mongoose
make         # Builds Mongoose (uses CMake) and runs the demo

Then to install, do

sudo make install 

After compilation, the Mongoose demo can be run from the build directory using ./bin/demo.

Usage

You can use Mongoose in one of three ways:

  1. The mongoose executable. Once built, the mongoose executable will be located in build/bin/mongoose. This executable can read a Matrix Market file containing an adjacency matrix and output timing and partitioning information to a plain-text file. Simply call it with the following syntax: mongoose <MM-input-file.mtx> [output-file]
  2. The C++ API. A static library is built at Lib/libmongoose.a. Include the header file Include/Mongoose.hpp and link to this library to access the C++ API.
  3. The MATLAB API. From MATLAB, navigate to the Mongoose/MATLAB/ directory and build the Mongoose MEX functions by calling mongoose_make. This will build Mongoose, run a demo, and allows access to the MATLAB API.

For more details about the specific APIs and their available functionality, see the Mongoose user guide located at Doc/UserGuide.pdf.

Credits

The following people have made significant contributions to Mongoose:

  • Nuri Yeralan, Microsoft Research
  • Scott Kolodziej, Texas A&M University
  • Tim Davis, Texas A&M University
  • William Hager, University of Florida

License

Mongoose is licensed under the GNU Public License, version 3. For commercial license inquiries, please contact Dr. Tim Davis at davis@tamu.edu. The specific text of the license can be found in Doc/License.txt.