Large dataset and improved script

[qzhu2017]

@mmukta
I just uploaded a large dataset in dataset/references-aspirin.xml

I also optimized the script for ntest.py, the new script should save some time from reinitialize the worker object. You can test it on the cluster now.