def test_return_stats():
A, _ = diagonal(100)
evals, evecs, stats = primme.eigsh(A, 3, tol=1e-6, which='LA',
return_stats=True, return_history=True)
assert(stats["hist"]["numMatvecs"])
svecs_left, svals, svecs_right, stats = primme.svds(A, 3, tol=1e-6,
which='SM', return_stats=True, return_history=True)
assert(stats["hist"]["numMatvecs"])
eval, evec = primme.eigsh(A, 1, which='LM', convtest=convtest_lm)
assert_allclose(eval, [99.], atol=.1)
# Sparse singular mass matrix
M = scipy.sparse.spdiags(np.asarray(range(99, -1, -1), dtype=np.float32), [0],
100, 100)
evals, evecs = primme.eigsh(A, 3, M=M, tol=1e-6, which='SA')
assert_allclose(evals, [0. / 99., 1. / 98., 2. / 97.], atol=1e-6 * 100)
print(evals)
# Sparse rectangular matrix 100x10 with non-zeros on the main diagonal
A = scipy.sparse.spdiags(range(10), [0], 100, 10)
# Compute the three closest to 4.1 singular values and the left and right corresponding
# singular vectors
svecs_left, svals, svecs_right = primme.svds(A, 3, tol=1e-6, which=4.1)
assert_allclose(sorted(svals), [3., 4., 5.], atol=1e-6 * 10)
print(svals) # [ 4., 5., 3.]
# Sparse random rectangular matrix 10^5x100
A = scipy.sparse.rand(10000, 100, density=0.001, random_state=10)
# Compute the three closest singular values to 6.0 with a tolerance of 1e-6
svecs_left, svals, svecs_right, stats = primme.svds(A,
3,
which='SM',
tol=1e-6,
return_stats=True)
A_svals = svals
print(svals)
print(stats["elapsedTime"], stats["numMatvecs"])