The SuiteSparse Matrix Collection (formerly known as the University of Florida Sparse Matrix Collection), is a large and actively growing set of sparse matrices that arise in real applications. The Collection is widely used by the numerical linear algebra community for the development and performance evaluation of sparse matrix algorithms. It allows for robust and repeatable experiments: robust because performance results with artificially-generated matrices can be misleading, and repeatable because matrices are curated and made publicly available in many formats. Its matrices cover a wide spectrum of domains, include those arising from problems with underlying 2D or 3D geometry (as structural engineering, computational fluid dynamics, model reduction, electromagnetics, semiconductor devices, thermodynamics, materials, acoustics, computer graphics/vision, robotics/kinematics, and other discretizations) and those that typically do not have such geometry (optimization, circuit simulation, economic and financial modeling, theoretical and quantum chemistry, chemical process simulation, mathematics and statistics, power networks, and other networks and graphs). We provide software for accessing and managing the Collection, from MATLAB, Mathematica, Fortran, and C, as well as an online search capability. Graph visualization of the matrices is provided on this site via graphviz. The Collection is maintained by Tim Davis, Texas A&M University (email: email@example.com), Yifan Hu, Yahoo! Labs, and Scott Kolodziej, Texas A&M University.
Click here for an easy-to-understand description of the SuiteSparse Matrix Collection and its images. The page includes a video of a talk I gave at the Harn Museum at the University of Florida, for general audiences.
Sample Gallery of the SuiteSparse Matrix Collection:
Click on an image above for more details on each matrix. The images were created by Yifan Hu at Yahoo! Labs, with a graph drawing program that can generate truly beautiful drawings large graphs, based solely on the connectivity (that is, a sparse matrix). Take a look at his drawings of the matrices in the Collection, which are also mirrored here on the web page for each matrix.
Full details of the Collection are provided in our paper entitled The University of Florida Sparse Matrix Collection, by T. Davis and Y. Hu, published in the ACM Transactions on Mathematical Software, Vol 38, Issue 1, 2011, pages 1:1-1:25 (click here for PDF). Please cite that paper when using this Collection. For additional background, see Duff, I.S, Grimes, R. G, and Lewis, J. G, Sparse matrix test problems, ACM Trans. on Mathematical Software, vol 15, no. 1, pp 1-14, 1989, which describes the Harwell-Boeing Collection. See also how to cite LAW group.
Archival data for reproducible research
The Collection serves a vital role in the sparse matrix algorithms community, as a benchmark for algorithmic testing and development. Results in journal articles that use these matrices can be repeated by other researchers. The Collection also appears widely in other repositories or citation indices:
To access the Collection
There are several ways to download matrices from this Collection.
- Via your web browser. Click on the Index tab in the top menu bar on this page.
- Via MATLAB with the ssget function.
- Via a stand-alone Java GUI.
- Via Julia.
See the Interfaces tab at the top of this page for details on ssget, the Java GUI, and the Julia interface. Matrices are provided in three formats: MATLAB *.mat file, Rutherford-Boeing, and Matrix Market. The Rutherford-Boeing format is described in a document on the Rutherford-Boeing Sparse Matrix Collection. See the Matrix Market for a description of the Matrix Market format.
To submit matrices to the Collection:
Click on the Submit Matrix tab in the top menu bar on this page to submit matrices to the Collection.
Other Computed Data
While the Collection contains a variety of precomputed data about each matrix, others have gone further and computed even more interesting properties for matrices from the Collection (or subset thereof). Here are a few:
- Khalid Theeb and Dr. Mary Hall of the University of Utah have computed a variety of metrics for a selection of matrices from the Collection, including average, maximum, and standard deviation of nonzeros per row, density, diagonal fill ratio, and plots of eigenvalues.
- Dr. Rob Bisseling's research group at Utrecht University has computed optimal solutions to the edge bipartitioning problem for a subset of matrices from the Collection.
You can submit more metrics, properties, or other work based off of the Collection to Tim Davis at firstname.lastname@example.org.