Added parallellism (with regards to rows of mutations) - which ended up resulting in a major refactor of the code, with movement of code all over the place, and multiple Trait inclusions.
The end result is an estimated 36% performance increase for 4 rows of 4 mutations each. Since the performance increase should only be reasonably associated with the number of rows - then 4 parallel rows coincide with a performance of 136% (meaning a ratio of 50:17, or 1:0.34)
Note: Mutations weren't parallellized because
Some of them couldn't be.
The ones that were presented a heavy performance decrease.
The latter could be coincidental, but I decided against implementing it at the end.
Added parallellism (with regards to rows of mutations) - which ended up resulting in a major refactor of the code, with movement of code all over the place, and multiple
Traitinclusions.The end result is an estimated 36% performance increase for 4 rows of 4 mutations each. Since the performance increase should only be reasonably associated with the number of rows - then 4 parallel rows coincide with a performance of 136% (meaning a ratio of 50:17, or 1:0.34)
Note: Mutations weren't parallellized because
The latter could be coincidental, but I decided against implementing it at the end.