Universal test failure? about nist-statistical-test-suite HOT 3 CLOSED

I'll have to go back and look. Thanks for pointing this out.

from nist-statistical-test-suite.

probably the insufficient number of data. as said in the NIST test suit doc (Appendix B Empirical Results for Sample Data), when running 1 Million bits in length, the Universal P-value is 0.669012.

The reason why NIST didn't include 1M data in the test suite, may be that the size will be too large. For such data length, it would be around 70MB for ASCII encoding 64 width per line, so you can imagine how large the data folder would be if you put all 1M data in the folder.

When I tested my own data, I encountered the same problem for small data set, but when I generate 1M data, the final report is OK.

Anyway, I didn't spend time investigating the algorithm in the code, will find some time to look into it.

from nist-statistical-test-suite.

@wanghs09 Shall we close this?

from nist-statistical-test-suite.