We can organize the data in three layers.

L1: raw output, one entry per global pointer
L2: data for times, states visited, new states, re-explored, dump time/size aggregated by algorithm and run
L3: time data is averaged

The complete data from layers L2 and L3 can be found here. A representative run from L1 (for spin) can be found here.

Note: the incremental algorithms were re-named for the paper.
- IncrFull is now called IncrFwd, and
- IncrSCC is now called IncrBack