Implementing SVE in `AXPY` Kernels for `A64FX` and `Graviton3E`

[hideaki-motoki]

While SVE has been implemented in some kernels, it is not yet implemented for the basic AXPY routines. Adding SVE support should improve performance. I would like to work on implementing this for [SD]AXPY on A64FX and Graviton3E.