I also cannot understand how a paper whose main contribution is a set of benchmarks, does not actually make the source code to those benchmarks publicly available. Unbelievable that such a paper can pass peer review. Very unscientific.