def test_eig_dyn():
v = 0
for i in xrange(5):
n = 1 + int(mp.rand() * 5)
if mp.rand() > 0.5:
# real
A = 2 * mp.randmatrix(n, n) - 1
if mp.rand() > 0.5:
A *= 10
for x in xrange(n):
for y in xrange(n):
A[x, y] = int(A[x, y])
else:
A = (2 * mp.randmatrix(n, n) -
1) + 1j * (2 * mp.randmatrix(n, n) - 1)
if mp.rand() > 0.5:
A *= 10
for x in xrange(n):
for y in xrange(n):
A[x,
y] = int(mp.re(A[x, y])) + 1j * int(mp.im(A[x, y]))
run_hessenberg(A, verbose=v)
run_schur(A, verbose=v)
run_eig(A, verbose=v)
def test_eighe_randmatrix():
N = 5
for a in xrange(10):
A = (2 * mp.randmatrix(N, N) - 1) + 1j * (2 * mp.randmatrix(N, N) - 1)
for i in xrange(0, N):
A[i, i] = mp.re(A[i, i])
for j in xrange(i + 1, N):
A[j, i] = mp.conj(A[i, j])
run_eighe(A)
def test_eighe_randmatrix():
N = 5
for a in xrange(10):
A = (2 * mp.randmatrix(N, N) - 1) + 1j * (2 * mp.randmatrix(N, N) - 1)
for i in xrange(0, N):
A[i,i] = mp.re(A[i,i])
for j in xrange(i + 1, N):
A[j,i] = mp.conj(A[i,j])
run_eighe(A)
def test_svd_c_rand():
for i in xrange(5):
full = mp.rand() > 0.5
m = 1 + int(mp.rand() * 10)
n = 1 + int(mp.rand() * 10)
A = (2 * mp.randmatrix(m, n) - 1) + 1j * (2 * mp.randmatrix(m, n) - 1)
if mp.rand() > 0.5:
A *= 10
for x in xrange(m):
for y in xrange(n):
A[x, y] = int(mp.re(A[x, y])) + 1j * int(mp.im(A[x, y]))
run_svd_c(A, full_matrices=full, verbose=False)
def test_eighe_irandmatrix():
N = 4
R = 4
for a in xrange(10):
A = irandmatrix(N, R) + 1j * irandmatrix(N, R)
for i in xrange(0, N):
A[i, i] = mp.re(A[i, i])
for j in xrange(i + 1, N):
A[j, i] = mp.conj(A[i, j])
run_eighe(A)
def test_svd_c_rand():
for i in xrange(5):
full = mp.rand() > 0.5
m = 1 + int(mp.rand() * 10)
n = 1 + int(mp.rand() * 10)
A = (2 * mp.randmatrix(m, n) - 1) + 1j * (2 * mp.randmatrix(m, n) - 1)
if mp.rand() > 0.5:
A *= 10
for x in xrange(m):
for y in xrange(n):
A[x,y]=int(mp.re(A[x,y])) + 1j * int(mp.im(A[x,y]))
run_svd_c(A, full_matrices=full, verbose=False)
def test_eighe_irandmatrix():
N = 4
R = 4
for a in xrange(10):
A=irandmatrix(N, R) + 1j * irandmatrix(N, R)
for i in xrange(0, N):
A[i,i] = mp.re(A[i,i])
for j in xrange(i + 1, N):
A[j,i] = mp.conj(A[i,j])
run_eighe(A)