Open
Description
Currently vDSP_vsortD
, which is supposed to sort a vector, can only sort it if it's size n
is 0 < n < Int64.max
, instead of the API's required 0 < n < UInt64.max
due to not being able to use UInt
as pointer index. We need to work around this.
Also, we're currently using LAPACK.dlasrt_
to sort it if the size n
is 0 < n < Int32.max
, and uses a custom QuickSort implementation for Int32.max <= n < Int64.max
. Once #1 is solved we should switch to using this as default for performance reasons