One of tools is currently failing on 1e9 due to lack of memory, 126GB is not enough to aggregate 50GB csv. This requires use bigger machine to run 1e9 grouping benchmark. Still memory usage should be investigated and probably added to report.
R's solution to call system("ps -o RSS") happens to be sub-optimal in terms of impact on benchmarks timings. It has been turned off as of now. We need to find reliable way to measure memory.
One of tools is currently failing on 1e9 due to lack of memory, 126GB is not enough to aggregate 50GB csv. This requires use bigger machine to run 1e9 grouping benchmark. Still memory usage should be investigated and probably added to report.