n2s = calculate_gradient_contribution(offsets_2, gi2s, grad3)
n3s = calculate_gradient_contribution(offsets_3, gi3s, grad3)
return 23.0 * (n0s + n1s + n2s + n3s)
if __name__ == "__main__":
theano.config.mode = 'FAST_RUN'
theano.config.floatX = 'float32'
theano.config.openmp = True
theano.config.openmp_elemwise_minsize = 2000
# theano.config.compute_test_value = 'warn'
perm = T.vector('perm', dtype='int32')
# perm.tag.test_value = np_perm
grad3 = T.matrix('grad3', dtype='float32')
# grad3.tag.test_value = np_grad3
vertex_table = T.tensor4('vertex_table', dtype='int32')
# vertex_table.tag.test_value = np_vertex_table
vl = T.matrix('vl', dtype='float32')
# vl.tag.test_value = np.array([[0.5, 0.1, 1.7], [1.7732, 0.1461, 1.7]], dtype=np.float32)
output = matrix_noise3d(vl, perm, grad3, vertex_table)
simplex_noise = theano.function([vl, perm, grad3, vertex_table], output)
print("Compiled")
shape = (512, 512)
phases = 5
scaling = 200.0
input_vectors = get_input_vectors(shape, phases, scaling)
start_time = time()
raw_noise = simplex_noise(input_vectors, np_perm, np_grad3, np_vertex_table)
print("The calculation took " + str(time() - start_time) + " seconds.")
show(raw_noise, phases, shape)
perm[skewed_v_and[0] + perm[skewed_v_and[1] + perm[skewed_v_and[2]]]] %
12, perm[skewed_v_and[0] + vertices[0][0] +
perm[skewed_v_and[1] + vertices[0][1] +
perm[skewed_v_and[2] + vertices[0][2]]]] % 12,
perm[skewed_v_and[0] + vertices[1][0] +
perm[skewed_v_and[1] + vertices[1][1] +
perm[skewed_v_and[2] + vertices[1][2]]]] % 12,
perm[skewed_v_and[0] + 1 +
perm[skewed_v_and[1] + 1 + perm[skewed_v_and[2] + 1]]] % 12
])
n = np.zeros(4, dtype=np.float)
for i in range(4):
t = 0.5 + np.sum(-1 * offsets[i]**2)
if t >= 0:
t *= t
n[i] = t**2 * np.dot(grad3[gi[i]], offsets[i])
return 23.0 * np.sum(n)
if __name__ == "__main__":
shape = (512, 512)
phases = 5
scaling = 200.0
input_vectors = get_input_vectors(shape, phases, scaling)
raw_noise = np.empty(input_vectors.shape[0], dtype=np.float32)
start_time = time()
for i in range(0, input_vectors.shape[0]):
raw_noise[i] = np_noise3d(input_vectors[i])
print("The calculation took " + str(time() - start_time) + " seconds.")
show(raw_noise, phases, shape)
theano.config.openmp_elemwise_minsize = 2000
# theano.config.compute_test_value = 'warn'
perm = T.vector('perm', dtype='int32')
# perm.tag.test_value = np_perm
grad3 = T.matrix('grad3', dtype='float32')
# grad3.tag.test_value = np_grad3
vertex_table = T.tensor4('vertex_table', dtype='int32')
# vertex_table.tag.test_value = np_vertex_table
v_shape = T.vector('shape', dtype='int32')
# v_shape.tag.test_value = shape
v_phases = T.constant(phases, name='phases', dtype='int32')
# v_phases.tag.test_value = phases
v_scaling = T.constant(scaling, name='scaling', dtype='float32')
# v_scaling.tag.test_value = scaling
v_offset = T.vector(name='offset', dtype='float32')
# v_offset.tag.test_value = offset
vl = get_input_vectors(v_shape, v_phases, v_scaling, v_offset)
v_noise = matrix_noise3d(vl, perm, grad3, vertex_table)
v_image_data = calculate_image(v_noise, v_phases, v_shape)
simplex_noise = theano.function([v_shape, v_offset, perm, grad3, vertex_table], v_image_data)
print("Compiled")
num_steps_burn_in = 10
num_steps_benchmark = 20
for i in range(num_steps_burn_in):
image_data = simplex_noise(shape, offset, np_perm, np_grad3, np_vertex_table)
start_time = time()
for i in range(num_steps_benchmark):
image_data = simplex_noise(shape, offset, np_perm, np_grad3, np_vertex_table)
print("The calculation took %.4f seconds." % ((time() - start_time) / num_steps_benchmark))
show(image_data)
grad3 = T.matrix('grad3', dtype='float32')
# grad3.tag.test_value = np_grad3
vertex_table = T.tensor4('vertex_table', dtype='int32')
# vertex_table.tag.test_value = np_vertex_table
v_shape = T.vector('shape', dtype='int32')
# v_shape.tag.test_value = shape
v_phases = T.constant(phases, name='phases', dtype='int32')
# v_phases.tag.test_value = phases
v_scaling = T.constant(scaling, name='scaling', dtype='float32')
# v_scaling.tag.test_value = scaling
v_offset = T.vector(name='offset', dtype='float32')
# v_offset.tag.test_value = offset
vl = get_input_vectors(v_shape, v_phases, v_scaling, v_offset)
v_noise = matrix_noise3d(vl, perm, grad3, vertex_table)
v_image_data = calculate_image(v_noise, v_phases, v_shape)
simplex_noise = theano.function(
[v_shape, v_offset, perm, grad3, vertex_table], v_image_data)
print("Compiled")
num_steps_burn_in = 10
num_steps_benchmark = 20
for i in range(num_steps_burn_in):
image_data = simplex_noise(shape, offset, np_perm, np_grad3,
np_vertex_table)
start_time = time()
for i in range(num_steps_benchmark):
image_data = simplex_noise(shape, offset, np_perm, np_grad3,
np_vertex_table)
print("The calculation took %.4f seconds." %
((time() - start_time) / num_steps_benchmark))
show(image_data)
offset = (0.0, 0.0, 1.7)
v_shape = tf.Variable([512, 512], name='shape')
v_phases = tf.Variable(5, name='phases')
v_scaling = tf.Variable(200.0, name='scaling')
v_offset = tf.Variable([0.0, 0.0, 1.7], name='offset')
v_input_vectors = get_input_vectors(v_shape, v_phases, v_scaling, v_offset)
perm = tf.Variable(np_perm, name='perm')
grad3 = tf.Variable(np_grad3, name='grad3')
num_steps_burn_in = 10
num_steps_benchmark = 20
vertex_table = tf.Variable(np_vertex_table, name='vertex_table')
raw_noise = noise3d(v_input_vectors, perm, grad3, vertex_table,
shape[0] * shape[1] * phases)
raw_image_data = calculate_image(raw_noise, phases, v_shape)
init = tf.initialize_all_variables()
input_vectors = get_input_vectors(shape, phases, scaling, offset)
noise = noise3d(input_vectors, np_perm, np_grad3, np_vertex_table,
shape[0] * shape[1] * phases)
image_data = calculate_image(noise, phases, shape)
sess = tf.Session()
sess.run(init)
for i in range(num_steps_burn_in):
raw_img = sess.run(image_data)
start_time = time()
for i in range(num_steps_benchmark):
raw_img = sess.run(image_data)
print("The calculation took %.4f seconds." %
((time() - start_time) / num_steps_benchmark))
# writer = tf.train.SummaryWriter("tf-logs/", sess.graph) # write logs for TensorBoard
show(raw_img.astype(np.uint8))
shape = (512, 512)
phases = 10
scaling = 200.0
offset = (0.0, 0.0, 1.7)
v_shape = tf.Variable([512, 512], name='shape')
v_phases = tf.Variable(5, name='phases')
v_scaling = tf.Variable(200.0, name='scaling')
v_offset = tf.Variable([0.0, 0.0, 1.7], name='offset')
v_input_vectors = get_input_vectors(v_shape, v_phases, v_scaling, v_offset)
perm = tf.Variable(np_perm, name='perm')
grad3 = tf.Variable(np_grad3, name='grad3')
num_steps_burn_in = 10
num_steps_benchmark = 20
vertex_table = tf.Variable(np_vertex_table, name='vertex_table')
raw_noise = noise3d(v_input_vectors, perm, grad3, vertex_table, shape[0] * shape[1] * phases)
raw_image_data = calculate_image(raw_noise, phases, v_shape)
init = tf.initialize_all_variables()
input_vectors = get_input_vectors(shape, phases, scaling, offset)
noise = noise3d(input_vectors, np_perm, np_grad3, np_vertex_table, shape[0] * shape[1] * phases)
image_data = calculate_image(noise, phases, shape)
sess = tf.Session()
sess.run(init)
for i in range(num_steps_burn_in):
raw_img = sess.run(image_data)
start_time = time()
for i in range(num_steps_benchmark):
raw_img = sess.run(image_data)
print("The calculation took %.4f seconds." % ((time() - start_time) / num_steps_benchmark))
# writer = tf.train.SummaryWriter("tf-logs/", sess.graph) # write logs for TensorBoard
show(raw_img.astype(np.uint8))