usymqr_test.jl

using IterativeSolvers
include("usymqr.jl")
using Base.Test
using LinearMaps

function unsymmetric_problem(T, n)
    A = rand(T, n, n)
    x = ones(T, n)
    b = A * x
    A, x, b
end

srand(123)
n = 100

T = Float32
A, x, b = unsymmetric_problem(T, n)
tol = sqrt(eps(real(T)))
x0 = rand(T, n)
x1, hist1 = usymqr(A, b, maxiter = 10n, tol = tol, log = true)
hist1.isconverged
hist1[:resnorm]
norm(b - A * x1) / norm(b) ≤ tol
isa(hist1, ConvergenceHistory)

# @testset "Unsymmetric Matrix{$T}" for T in (Float32, Float64)
#     A, x, b = unsymmetric_problem(T, n)
#     tol = sqrt(eps(real(T)))
#     x0 = rand(T, n)
#
#     x1, hist1 = usymqr(A, b, maxiter = 10n, tol = tol, log = true)
#
#     @test isa(hist1, ConvergenceHistory)
#     @test norm(b - A * x1) / norm(b) ≤ tol
#     @test hist1.isconverged
# end
#
# @testset "SparseMatrixCSC{$T}" for T in (Float32, Float64)
#     A = let
#         B = sprand(n, n, 2 / n)
#     end
#
#     x = ones(T, n)
#     b = A * x
#     tol = sqrt(eps(real(T)))
#
#     x_approx, hist = usymqr(A, b, maxiter = 10n, tol = tol, log = true)
#
#     @test norm(b - A * x_approx) / norm(b) ≤ tol
#     @test hist.isconverged
# end