SortingNetworks.jl

Sort 1..25 values with conditional swaps.  

Sorting networks that process 1,2,..25 values are given.
These sorting networks should minimize conditional swaps.
The first sixteen are known to minimize conditional swaps.

Values may be given with the args swapsort(x1::T, .., xn::T)::NTuple{n,T}
Values may be given using a tuple swapsort(tup::NTuple{n,T})::NTuple{n,T}
These sorts return a tuple of values sorted in ascending order.

Pkg.add("SortingNetworks")

using SortingNetworks

jumble = (5,3,1,4,2)
sorted = (1,2,3,4,5)

ascending = swapsort(5,3,1,4,2)
ascending == sorted

ascending == swapsort(jumble)
ascending == sorted

With v0.6-dev, timing sort([a,b,..]) relative to swapsort(a,b,...),
I got 15x for 4 Ints, 11x for 8 Ints, 5.75x for 16 Ints

Jan Weidner provided provably correct code to test each implementation.