Using Float128 in RandomMatrices

Hi guys:

I am using the package RandomMatrices to generate Haar random unitary matrices. But recently I want to generate unitary matrices with higher precision. So I found the package Quadmath which provides Float128 object. But how to combines these two packages together? How to generate a Haar random unitary matrix with Float128? By the way, is there a way to use things like randn(Float128,(N,N))? It seems that this doesn’t work for Float128 in Quadmath

I’m not sure if Quadmath has the necessary support, but you can do this with DoubleFloats (which give you roughly 100 bits of precision and are usually about 2x faster).

julia> using DoubleFloats
julia> randn(Double64, 10, 10)
10×10 Matrix{Double64}:
...
1 Like

Thanks this sounds good.