LL perf: Upgrade the results' dataset to have more runs of longer duration from a quiet machine.
Stability under repetition worse than expected. LQ performing worse than expected. CFA-uCpp-STL trend about as expected.