To this end, computations were run by increasing the pre-exponential factor in the Arrhenius equation for the catalytic reaction rate, thus simulating the effect of increasing catalyst activity.
For larger values, however, Ex Motif did not end computations, while SISMA_ Speller failed only when boxes were any of the known very hard PMP instances.
We have no results for the comparison of SISMA_ Speller and Ex Motif, because the latter never ended computation within this experimental settings.
Indeed SISMA_ Smile always outperformed Risotto when both tools ended computation, while Ex Motif outperformed SISMA_ Speller on input instances with very small values of ℓ, e, and b.
In the best case, Risotto ended computation in less than 10 seconds (with the exception of ten boxes, where the best run took 24 secs), while Risotto ended computation in less than 6 seconds for b ≤ 5 and in about 30 seconds for b > 5, never exceeding 35 seconds.
In the majority of tests, SISMA_ Smile ended computation before Risotto and it failed only once due to memory shortage in the first stage (because of too many simple motif occurrences).
Moreover, we separately computed the above measures by considering either all runs, or only runs where both tools ended computations.
In the end, the computation of this field on the entire configuration space is shown to be very efficient.
Combustibility cloud shows that there is still a stratification of combustibility in containment, though the hydrogen distribution is quite uniform at the end of computation.
