[C++] Improve the performance of the binary round kernel

[EpsilonPrime]

Describe the enhancement requested

The binary round kernel has some of the performance issues of the unary round kernel namely non-floating type arguments are first cast to double before implementation.

The other area where performance may be improved is in the handing of a scalar version of the number of significant digits argument (the second argument). The current implementation expects two arrays which may (depending on the engine) result in duplicating the data. Falling back to the unary version of round in such situations may be a potential solution to such a behavior.

Component(s)

C++

[thisisnic]

This issue has been marked as stale because it has had no activity in the past 2 years. Please remove the stale label or comment below, or this issue will be closed in 14 days.