Skip to main content
. 2014 Sep 4;10(9):e1003794. doi: 10.1371/journal.pcbi.1003794

Table 3. Runtimes and errors of the vectorised Inline graphic-leap method compared to the SSA.

Simulation method Runtime (hours) Total error
SSA 67.4 0.201
Inline graphic-leap, Inline graphic 44.6 0.173
Inline graphic-leap, Inline graphic 6.7 0.175
Inline graphic-leap, Inline graphic 2.9 0.189
Inline graphic-leap, Inline graphic 0.9 0.214
Inline graphic-leap, Inline graphic 0.7 0.312

The errors are calculated by subtracting the weight of each point of the PDF (that is, value multiplied by bin width) from the corresponding point of the SSA PDF. The error in the SSA row is the SSA self-distance, i.e. the error between two different SSA simulations. These simulations are of uncoupled niche lineages only, hence the SSA can be regarded as the true solution.