Using binary search?

[molpopgen]

The "fwdpp copy" uses a linear time algorithm to update positions within parent genomes during offspring genome generation.

It it possible that the linear method is hurting performance when genomes contain a large number of mutations. The only way to know is to try to mimic fwdpp more closely here.

[molpopgen]

Closing:

1. The relevant routines do use binary search.
2. The main reasons (that we know about) for the perf diffs are due to not having "compact mutations" implemented and using an inefficient method to remove fixations. The former was fixed in Add compact_mutations #12 and the latter is being addressed in Performance audit for "fwdpp copy" #11.