|
Implement a benchmark of the entire process starting with an unsimplified MatchExpression and ending with a simplified MatchExpression: the entire simplifier encompassing conversion to DNF, finding prime implicants by Quine-McCluskey, Petrick's, and restoring the MatchExpression representation from maxterm/minterm bitsets. In doing so we want to answer two questions:
- Given an expression that is already in DNF, how does the performance scale as a function of the number of minterms and the number of actual input variables?
- Given an expression that is not already in DNF, how much does the explosion in expression size due to DNF normalization hurt us?
Based on what we find, in a future patch we might end up implementing a limit such that we bail out of boolean simplification if either the number of input variables (that is, leaf predicates with a unique hash) is too high, or if we generate too many minterms during conversion to DNF.
|