Created by: joshua-warburton
The current implementation of vector, fixed point ucvtf has an incorrect value for bit 24 and matching incorrect tests.
This patch corrects the implementation and adds a better test coverage for the instruction and its variants.
Issues: #2626