Draft: Extend pexp_complex for double
Reference issue
This MR follows the SIMD implementation campaign #2635. This MR involves the extension of pexp_complex
for double precision.
What does this implement/fix?
This MR is a draft (for the moment) implementing the extension of pexp_complex
for double precision.
Additional information
References values in NEON are wrong in some edge cases. I will think of a workaround, but we might start discussing about the presence of psincos_aux
, which is an auxiliary function I used to avoid duplicating the code of pexp_complex
for float
and double
.