I have a follow up question. I have not implemented the BNNS approach proposed above and it works. However, the calculation speed is still about 3 times as high when I use a similar approach in Python - without any particular special packages or approaches on Python side. I was hoping that Swift would generally be faster so wanted to get your view on whether maybe my approach is generally flawed and her any potential alternatives. I would love to continue with coding in Swift on the project but with runtime being 300% it is just too compelling to rather use Python. Let me explain what I am doing: I create a lot of (random) binary trees which contain arithmetic or logic operations in all non-leaf nodes and numeric arrays (all of the same length) on leaf nodes which are taken from return_data. I.e. my trees are representing (simple) mathematic formulas. I have tens of thousands such trees and need to efficiently evaluate them. I currently do this via a recursive function evaluate_signal. As said above, the same
Topic:
Programming Languages
SubTopic:
Swift
Tags: