We have provided a synthetic data generator for testing purposes (requires python 2.7+). To see detailed instructions, run

python app/matrixfact/scripts/generate_data.py -h

For example, you can create a 1000 by 1000 matrix “data” in the petuum-java root directory by running:

python app/matrixfact/scripts/generate_data.py 10 100 0.1 test-matrix