To introduce metrics of "sorting" in algorithm analyzer

[alex-kuzmin-hg]

Jasper:
https://swirldslabs.slack.com/archives/C03DU0QGE2G/p1738954651649259?thread_ts=1738096891.401219&cid=C03DU0QGE2G
"would also be nice to have summary section top 10 best improvements and 10 worst regressions or sorting at least. I worry how useful the data is as it feels like small changes in data processed could cause big changes. Like are we measuring the exact same N million transactions in identical order? How consensus groups into rounds based on network conditions could have a big effect. I worry how we eliminate noise not hiding any real insights."

For myself:

1. introduce sorting for columns
2. introduce sorting on same weight function