This project generates artificial matrices, given specific structural features (parameters), that can be used for sparse-matrix benchmarks. The features that have to be provided to the generator are the following :
- number of rows
- number of columns
- average nonzeros per row
- standard deviation of nonzeros per row
- distribution of nonzeros per row (currently "normal" and "gamma" distributions are supported)
- seed for the random number generator (need to reproduce results on the same artificial matrices across different platforms)
- placement of nonzeros within row (currently "random", "diagonal" and "simple" are supported)
- bandwidth of matrix, which confines the column range that a nonzero can be placed at
- coefficient of skewness, which indicates the imbalance of row length
- average number of neighboring nonzeros within row (neighbors : maximum distance 1 for column values)
- similarity between adjacent rows (based on column position of nonzeros)
The output is a struct with the CSR-representation (row-pointer, column-index and values) of the generated sparse matrix and the above mentioned structural features.
An example of the usage of the generator can be found at artificial_matrix.c (called via bash script run.sh)