The "nested" data set contains questions and answers from Stack Overflow.
The benchmarks are run with a larger heap of 4GB. For more details please refer to the nested track specification and have a look at our benchmarking methodology).