Closed RamType0 closed 2 years ago
I think I have to say no to this, unfortunately.
Forcing all users to do this would be a large ergonomics burden on users, and adding it as an alternative API would require a very large amount of boilerplate that would be too large to maintain and document.
I appreciate the suggestion though!
@RamType0, what was your motivation for suggesting this?
@RamType0, what was your motivation for suggesting this?
Zero initializing output
and scratch
before we pass them to this function is unnecessary cost.
I think using &mut [MaybeUnInit<Complex<T>>]
is better for scratch
at least.
Because the initial value of these would not affect the result , we could use
&mut [MaybeUnInit<Complex<T>>]
instead of&mut [Complex<T>]
.