You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
the different implementations were tested on a fixed set of 8 hermitian negative semidefinite matrices
the were randomly generated. the MAT file is available here.
the functions to generate the matrices were:
functiontest_chbv_real(n::Int, p=0.1)
D =diagm(-rand(n))
T =sprandn(n, n, p)
H = T * D * T'# random negative semidefinite symmetric matrix
vec =randn(n)
return H, vec
endfunctiontest_chbv_complex(n::Int, p=0.1)
D =diagm(-rand(n))
T =sprandn(n, n, p) +sprandn(n, n, p)*im
H = T * D * T.'# random negative semidefinite hermitian matrix
vec =randn(n) +randn(n)*im
return H, vec
end