def test_dirichlet():
obj = test('dirichlet', 'dirichlet')
obj._A_q1, obj._A_q2 = af.Array([100]), af.Array([100])
obj._A_q1 = af.tile(obj._A_q1, 1, 1, obj.q1_center.shape[2])
obj._A_q2 = af.tile(obj._A_q2, 1, 1, obj.q1_center.shape[2])
obj.dt = 0.001
obj.f = af.constant(0,
obj.q1_center.shape[0],
obj.q1_center.shape[1],
obj.q1_center.shape[2],
dtype=af.Dtype.f64)
apply_bcs_f(obj)
expected = af.constant(0,
obj.q1_center.shape[0],
obj.q1_center.shape[1],
obj.q1_center.shape[2],
dtype=af.Dtype.f64)
N_g = obj.N_ghost
# Only ingoing characteristics should be affected:
expected[:N_g] = af.select(obj.q1_center < obj.q1_start, 1, expected)[:N_g]
expected[:, :N_g] = af.select(obj.q2_center < obj.q2_start, 2,
expected)[:, :N_g]
assert (af.max(af.abs(obj.f[:, N_g:-N_g] - expected[:, N_g:-N_g])) < 5e-14)
assert (af.max(af.abs(obj.f[N_g:-N_g, :] - expected[N_g:-N_g, :])) < 5e-14)
def np2af(data, dtype=None):
# return af.Array(data.ctypes.data, data.shape, data.dtype.char)
print(type(data))
if dtype:
return af.Array(data.ctypes.data, data.shape, dtype)
else:
return af.Array(data.ctypes.data, data.shape, data.dtype.char)
def _split_fourier_cuda(self, signal: Signal, step):
'''
This function is called by split_fourier,and should not be used outside
:param signal: signal to traverse the span
:param step: the step of split fourier
:return: None
'''
af.set_backend('cuda')
freq = fftfreq(len(signal.data_sample[0, :]),
(signal.sps * signal.symbol_rate_in_hz)**(-1))
freq = af.Array(freq.ctypes.data, freq.shape, freq.dtype.char)
signal_x = np.asarray(signal.data_sample[0, :])
signal_y = np.asarray(signal.data_sample[1, :])
signal_x = af.Array(signal_x.ctypes.data,
signal_x.shape,
dtype=signal_x.dtype.char)
signal_y = af.Array(signal_y.ctypes.data,
signal_x.shape,
dtype=signal_y.dtype.char)
Disper = (1j / 2) * self.beta2 * (2 * np.pi * freq)**2 * step + (
1j / 6) * self.beta3 * (
(2 * np.pi * freq)**3 * step) - self.alphalin / 2 * step
dz_Eff = (1 - np.exp(-self.alphalin * step)) / self.alphalin
step_number = np.ceil(self.length / step)
for number in range(int(step_number)):
print(number)
if number == step_number - 1:
# dz = step
dz = self.length - (step_number - 1) * step
dz_Eff = (1 - np.exp(-self.alphalin * dz)) / self.alphalin
Disper = (1j / 2) * self.beta2 * (2 * np.pi * freq)**2 * dz + (
1j / 6) * self.beta3 * (
(2 * np.pi * freq)**3 * dz) - self.alphalin / 2 * step
signal_x, signal_y = self.linear(signal_x, signal_y, Disper)
energy = signal_x * af.conjg(signal_x) + signal_y * af.conjg(
signal_y)
signal_x, signal_y = self.nonlinear(energy, signal_x, signal_y,
dz_Eff)
signal_x, signal_y = self.linear(signal_x, signal_y, Disper)
signal_x_array = np.array(signal_x.to_list())
signal_y_array = np.array(signal_y.to_list())
signal_x_array = signal_x_array[:, 0] + 1j * signal_x_array[:, 1]
signal_y_array = signal_y_array[:, 0] + 1j * signal_y_array[:, 1]
signal.data_sample[0, :] = signal_x_array
signal.data_sample[1, :] = signal_y_array
def simple_array(verbose=False):
display_func = _util.display_func(verbose)
print_func = _util.print_func(verbose)
a = af.Array([1, 2, 3])
display_func(a)
display_func(a.T)
display_func(a.H)
print_func(a.shape)
b = a.as_type(af.Dtype.s32)
display_func(b)
print_func(a.elements(), a.type(), a.dims(), a.numdims())
print_func(a.is_empty(), a.is_scalar(), a.is_column(), a.is_row())
print_func(a.is_complex(), a.is_real(), a.is_double(), a.is_single())
print_func(a.is_real_floating(), a.is_floating(), a.is_integer(),
a.is_bool())
a = af.Array(host.array("i", [4, 5, 6]))
display_func(a)
print_func(a.elements(), a.type(), a.dims(), a.numdims())
print_func(a.is_empty(), a.is_scalar(), a.is_column(), a.is_row())
print_func(a.is_complex(), a.is_real(), a.is_double(), a.is_single())
print_func(a.is_real_floating(), a.is_floating(), a.is_integer(),
a.is_bool())
a = af.Array(host.array("I", [7, 8, 9] * 3), (3, 3))
display_func(a)
print_func(a.elements(), a.type(), a.dims(), a.numdims())
print_func(a.is_empty(), a.is_scalar(), a.is_column(), a.is_row())
print_func(a.is_complex(), a.is_real(), a.is_double(), a.is_single())
print_func(a.is_real_floating(), a.is_floating(), a.is_integer(),
a.is_bool())
c = a.to_ctype()
for n in range(a.elements()):
print_func(c[n])
c, s = a.to_ctype(True, True)
for n in range(a.elements()):
print_func(c[n])
print_func(s)
arr = a.to_array()
lst = a.to_list(True)
print_func(arr)
print_func(lst)
print_func(a.is_sparse())
def main2():
data = af.Array([0, 1, 2, 2.5, 8, 9, 20, 21, 22])
labels = af.Array([0, 0, 0, 0, 1, 1, 2, 2, 2]).as_type(af.Dtype.u32)
num_labels = 2
# query = af.Array([1.5])
query = af.Array([1.5, 12])
clf = AfKNearestNeighbors(num_nearest=5, weight_by_dist=False)
# clf = AfKNearestNeighbors(num_nearest=data.dims()[0], weight_by_dist=True)
clf.train(data, labels)
outputs = clf.predict(query)
outputs_proba = clf.predict_proba(query)
print('outputs:', outputs)
print('outputs_proba:', outputs_proba)
print('ndarray version:\n', outputs_proba.to_ndarray())
def __reshape__(self, newshape, order='C'):
if (order is not 'C'):
raise NotImplementedError
if isinstance(newshape, numbers.Number):
newshape = (newshape, )
# Replace a possible -1 with the
if -1 in newshape:
newshape = list(newshape)
i = newshape.index(-1)
newshape[i] = 1
if -1 in newshape:
raise ValueError('Only one -1 allowed in shape')
newshape[i] = self.size / numpy.prod(newshape)
if self.size != numpy.prod(newshape):
raise ValueError('total size of new array must be unchanged')
if len(newshape) != 0:
# No need to modify the af_array for empty shapes
# af_shape = numpy.array(pu.c2f(newshape), dtype=pu.dim_t)
# s = arrayfire.Array()
# arrayfire.backend.get().af_moddims(ctypes.pointer(s.arr), self.d_array.arr, af_shape.size, ctypes.c_void_p(af_shape.ctypes.data))
# self.d_array = s
if tuple(newshape) == self.shape:
# No need to do anything
return
af_shape = numpy.array(pu.c2f(newshape), dtype=pu.dim_t)
s = arrayfire.Array()
arrayfire.backend.get().af_moddims(
ctypes.pointer(s.arr), self.d_array.arr, af_shape.size,
ctypes.c_void_p(af_shape.ctypes.data))
self.d_array = s
self._shape = tuple(newshape)
def forward(nn_np, data):
"""
Given a dictionary representing a feed forward neural net and an input data matrix compute the network's output and store it within the dictionary
:param nn: neural network dictionary
:param data: a numpy n by m matrix where m in the number of input units in nn
:return: the output layer activations
"""
nn = nn_np
nn['activations'] = []
afData = data.ctypes.data, data.shape, data.dtype.char
for i in range(data.shape[0]):
nn['activations'].append(
af.Array(data[i].ctypes.data, data[i].shape, data[i].dtype.char))
nn['zs'] = []
for w, s, b in map(None, nn['weights'], nn['nonlin'], nn['biases']):
#print af.transpose(nn['activations'][-1])
print nn['activations'][-1]
print ""
print w
z = af.dot(nn['activations'][-1], w)
nn['zs'].append(af.transpose(z))
nn['activations'].append(s[0](af.transpose(z)))
return nn['activations'][-1]
def f_left(f, t, q1, q2, p1, p2, p3, params):
k = params.boltzmann_constant
E_upper = params.E_band
T = params.initial_temperature
mu = params.initial_mu
t = params.current_time
omega = 2. * np.pi * params.AC_freq
q1_start = af.Array([domain.q1_start])
q2_start = af.Array([domain.q2_start])
q2_end = af.Array([domain.q2_end])
#x_start, y_start = coords.get_cartesian_coords(q1_start, q2_start)
#x_start, y_end = coords.get_cartesian_coords(q1_start, q2_end)
y_start = -24.
y_end = -4.6
contact_width = y_end - y_start #19.486
if (params.source_type == 'AC'):
vel_drift_x_out = -params.vel_drift_y_in / contact_width * np.sin(
omega * t)
elif (params.source_type == 'DC'):
vel_drift_x_out = -params.vel_drift_y_in / contact_width
else:
raise NotImplementedError('Unsupported source_type')
if (params.p_space_grid == 'cartesian'):
p_x = p1
p_y = p2
elif (params.p_space_grid == 'polar2D'):
p_x = p1 * af.cos(p2)
p_y = p1 * af.sin(p2)
else:
raise NotImplementedError('Unsupported coordinate system in p_space')
fermi_dirac_out = (1. / (af.exp(
(E_upper - vel_drift_x_out * p_x - mu) / (k * T)) + 1.))
if params.zero_temperature:
fermi_dirac_out = fermi_dirac_out - 0.5
f_left = fermi_dirac_out
af.eval(f_left)
return (f_left)
def testFromArrayfire(self):
a = af.Array([1, 2, 3, 4])
b = Array.from_arrayfire(a)
self.assertNotEqual(a.arr, 0)
self.assertNotEqual(b.arr_reference, 0)
self.assertNotEqual(a.arr, b.arr_reference)
np.testing.assert_array_equal(np.asarray(a.to_list()),
np.asarray(b.to_list()))
def simple_device(verbose=False):
display_func = _util.display_func(verbose)
print_func = _util.print_func(verbose)
print_func(af.device_info())
print_func(af.get_device_count())
print_func(af.is_dbl_supported())
af.sync()
curr_dev = af.get_device()
print_func(curr_dev)
for k in range(af.get_device_count()):
af.set_device(k)
dev = af.get_device()
assert(k == dev)
print_func(af.is_dbl_supported(k))
af.device_gc()
mem_info_old = af.device_mem_info()
a = af.randu(100, 100)
af.sync(dev)
mem_info = af.device_mem_info()
assert(mem_info['alloc']['buffers'] == 1 + mem_info_old['alloc']['buffers'])
assert(mem_info[ 'lock']['buffers'] == 1 + mem_info_old[ 'lock']['buffers'])
af.set_device(curr_dev)
a = af.randu(10,10)
display_func(a)
dev_ptr = af.get_device_ptr(a)
print_func(dev_ptr)
b = af.Array(src=dev_ptr, dims=a.dims(), dtype=a.dtype(), is_device=True)
display_func(b)
c = af.randu(10,10)
af.lock_array(c)
af.unlock_array(c)
a = af.constant(1, 3, 3)
b = af.constant(2, 3, 3)
af.eval(a)
af.eval(b)
print_func(a)
print_func(b)
c = a + b
d = a - b
af.eval(c, d)
print_func(c)
print_func(d)
print_func(af.set_manual_eval_flag(True))
assert(af.get_manual_eval_flag() == True)
print_func(af.set_manual_eval_flag(False))
assert(af.get_manual_eval_flag() == False)
display_func(af.is_locked_array(a))
def setup_mnist(frac, expand_labels):
root_path = os.path.dirname(os.path.abspath(__file__))
file_path = root_path + '/../../assets/examples/data/mnist/'
idims, idata = read_idx(file_path + 'images-subset')
ldims, ldata = read_idx(file_path + 'labels-subset')
idims.reverse()
numdims = len(idims)
images = af.Array(idata, tuple(idims))
R = af.randu(10000, 1)
cond = R < min(frac, 0.8)
train_indices = af.where(cond)
test_indices = af.where(~cond)
train_images = af.lookup(images, train_indices, 2) / 255
test_images = af.lookup(images, test_indices, 2) / 255
num_classes = 10
num_train = train_images.dims()[2]
num_test = test_images.dims()[2]
if expand_labels:
train_labels = af.constant(0, num_classes, num_train)
test_labels = af.constant(0, num_classes, num_test)
h_train_idx = train_indices.to_list()
h_test_idx = test_indices.to_list()
for i in range(num_train):
train_labels[ldata[h_train_idx[i]], i] = 1
for i in range(num_test):
test_labels[ldata[h_test_idx[i]], i] = 1
else:
labels = af.Array(ldata, tuple(ldims))
train_labels = labels[train_indices]
test_labels = labels[test_indices]
return (num_classes, num_train, num_test, train_images, test_images,
train_labels, test_labels)
def test_dx_dxi_analytical():
'''
Test to check the dx_dxi_analytical in wave equation module for an element
and compare it with an analytical value.
'''
threshold = 1e-14
nodes = af.Array([2, 6])
check_analytical_dx_dxi = af.sum(
af.abs(wave_equation.dx_dxi_analytical(nodes, 0) - 2)) <= threshold
assert check_analytical_dx_dxi
def af_gaussianDerivative1D(s,order):
radius = int(round(3*s + 0.5*order))
size = radius*2+1
csize = ct.c_int(size)
csigma = ct.c_double(s)
d_k = af.Array()
af.safe_call(af.backend.get().af_gaussian_kernel(ct.pointer(d_k.arr),csize,1,csigma,0))
if order == 1:
afx=af.range(size)-radius
return -afx/s/s*d_k
if order == 2:
afx=((af.range(size)-radius)**2-s*s)
return afx/(s**4)*d_k
return d_k
def test_gauss_weights():
'''
Test to check the gaussian weights calculated.
'''
threshold = 2e-8
analytical_gauss_weights = af.Array([0.23692688505618908, 0.47862867049936647,\
0.5688888888888889, 0.47862867049936647, \
0.23692688505618908
]
)
calculated_gauss_weights = lagrange.gaussian_weights(5)
assert af.max(af.abs(analytical_gauss_weights - calculated_gauss_weights))\
<= threshold
def test():
x1 =np.arange(2000)
x = af.Array(x1.ctypes.data, x1.shape, x1.dtype.char)
alpha = 2/3
s_old = x[0]
a = time.time()
for i in range(1, 2000):
s = alpha * x[i] + (1- alpha) * s_old
s_old = s
#time.sleep(1)
b = time.time()
print(b - a)
return s
def visitbatches(nn, batches, labelBatches, errlist, it=1000):
afBatches = []
for group in batches:
gp = []
for item in group:
tmp = []
tmp.append(item[0])
tmp.append(item[1])
gp.append(af.Array(tmp))
afBatches.append(gp)
for c in range(it):
#print "len of things " + str(len(nn)) + " " + str(len(batches[0]))
nnDif.master_node(nn, afBatches, labelBatches)
def forward(self, end):
if len(self.layers) is 0:
return self.data
out = af.Array()
data = self.data
for key in self.layers.viewkeys():
if key in self.params:
data = self.layers[key](image = data,
weights = self.params[key]['weights'],
biases = self.params[key]['biases'])
else:
data = self.layers[key](image = data)
if key is end:
return data
return data
def imag(self):
ret_type = numpy.real(numpy.zeros((), dtype=self.dtype)).dtype
shape = list(self.shape)
if not numpy.issubdtype(self.dtype, numpy.complexfloating):
return afnumpy.zeros(self.shape)
shape[-1] *= 2
dims = numpy.array(pu.c2f(shape), dtype=pu.dim_t)
s = arrayfire.Array()
arrayfire.backend.get().af_device_array(
ctypes.pointer(s.arr), ctypes.c_void_p(self.d_array.device_ptr()),
self.ndim, ctypes.c_void_p(dims.ctypes.data),
pu.typemap(ret_type).value)
arrayfire.backend.get().af_retain_array(ctypes.pointer(s.arr), s.arr)
a = ndarray(shape, dtype=ret_type, af_array=s)
ret = a[..., 1::2]
ret._base = a
ret._base_index = (Ellipsis, slice(1, None, 2))
return ret
def to_arrayfire(self):
""" Creates an Arrayfire array from this KHIVA array. This need to be used carefully as the same array
reference is oging to be used by both of them. Once the Arrayfire array is created, the destructor of
the KHIVA array is not going to free the allocated array.
:return: an Arrayfire Array
"""
try:
import arrayfire as af
except ModuleNotFoundError:
logging.error(
"In order to use `to_arrayfire()` function, you need to install the Arrayfire Python library"
)
sys.exit(1)
result = af.Array()
result.arr = self.arr_reference
self.arrayfire_reference = True
return result
def af_gauss_sigmas_sep(data,sigmas):
d_kernels = []
for s in sigmas:
csize = ct.c_int(int(3*s + 0.5)*2+1)
csigma = ct.c_double(s)
d_k = af.Array()
af.safe_call(af.backend.get().af_gaussian_kernel(ct.pointer(d_k.arr),csize,1,csigma,ct.c_double(0.0)))
d_kernels.append(d_k)
out = []
for d in data:
d_img = af.np_to_af_array(d)
for d_k,i in zip(d_kernels,range(len(d_kernels))):
res = af.convolve2_separable(d_k, af.transpose(d_k), d_img)
# create numpy array
out.append(res.__array__())
return out
def simple_device(verbose=False):
display_func = _util.display_func(verbose)
print_func = _util.print_func(verbose)
print_func(af.device_info())
print_func(af.get_device_count())
print_func(af.is_dbl_supported())
af.sync()
curr_dev = af.get_device()
print_func(curr_dev)
for k in range(af.get_device_count()):
af.set_device(k)
dev = af.get_device()
assert (k == dev)
print_func(af.is_dbl_supported(k))
af.device_gc()
mem_info_old = af.device_mem_info()
a = af.randu(100, 100)
af.sync(dev)
mem_info = af.device_mem_info()
assert (mem_info['alloc']['buffers'] == 1 +
mem_info_old['alloc']['buffers'])
assert (mem_info['lock']['buffers'] == 1 +
mem_info_old['lock']['buffers'])
af.set_device(curr_dev)
a = af.randu(10, 10)
display_func(a)
dev_ptr = af.get_device_ptr(a)
print_func(dev_ptr)
b = af.Array(src=dev_ptr, dims=a.dims(), dtype=a.dtype(), is_device=True)
display_func(b)
af.lock_device_ptr(b)
af.unlock_device_ptr(b)
def array(object: tp.Union[tp.Sequence,
np.ndarray,
pyarray.array],
dtype: tp.Optional[np.generic] = None,
copy: bool = True,
order: tp.Optional = None,
subok: bool = False,
ndmin: int = 0) \
-> ndarray:
"""
Create an array.
"""
if order:
raise ValueError('order != None is not supported')
# if dtype is not None:
# raise ValueError("parameter dtype != None is not supported")
af_array = None
shape = None
if isinstance(object, np.ndarray):
if object.ndim == 0:
# object = [np.asscalar(object)]
object = np.reshape(object, (1, ))
af_array = af.to_array(object)
shape = object.shape
elif isinstance(object, list):
np_array = np.array(object, dtype=dtype)
af_array = af.to_array(np_array)
shape = np_array.shape
else:
af_array = af.Array(object)
result = ndarray(af_array, shape)
if dtype is not None and result.dtype != dtype:
result = result.astype(dtype)
return result
def master_node(nn, data, labels):
nabla_w = []
nabla_b = []
l = 1
for i in range(len(data[0])):
for n in range(len(nn)):
r = forward(nn[n], data[n][i])
tmp = []
tmp.append(labels[n][i][0])
delta = d_cost(r, af.Array(tmp))
w, b = gradient(nn[n], delta)
nabla_w += w
nabla_b += b
nabla_w = [x / len(nn) for x in nabla_w]
nabla_b = [x / len(nn) for x in nabla_b]
for net in nn:
backprop(net, nabla_w, nabla_b)
def simple_index(verbose=False):
display_func = _util.display_func(verbose)
a = af.randu(5, 5)
display_func(a)
b = af.Array(a)
display_func(b)
c = a.copy()
display_func(c)
display_func(a[0, 0])
display_func(a[0])
display_func(a[:])
display_func(a[:, :])
display_func(a[0:3, ])
display_func(a[-2:-1, -1])
display_func(a[0:5])
display_func(a[0:5:2])
idx = af.Array(host.array("i", [0, 3, 2]))
display_func(idx)
aa = a[idx]
display_func(aa)
a[0] = 1
display_func(a)
a[0] = af.randu(1, 5)
display_func(a)
a[:] = af.randu(5, 5)
display_func(a)
a[:, -1] = af.randu(5, 1)
display_func(a)
a[0:5:2] = af.randu(3, 5)
display_func(a)
a[idx, idx] = af.randu(3, 3)
display_func(a)
a = af.randu(5, 1)
b = af.randu(5, 1)
display_func(a)
display_func(b)
for ii in ParallelRange(1, 3):
a[ii] = b[ii]
display_func(a)
for ii in ParallelRange(2, 5):
b[ii] = 2
display_func(b)
a = af.randu(3, 2)
rows = af.constant(0, 1, dtype=af.Dtype.s32)
b = a[:, rows]
display_func(b)
for r in range(rows.elements()):
display_func(r)
display_func(b[:, r])
a = af.randu(3)
c = af.randu(3)
b = af.constant(1, 3, dtype=af.Dtype.b8)
display_func(a)
a[b] = c
display_func(a)
solver_method_in_p = 'FVM' reconstruction_method_in_q = 'weno5' reconstruction_method_in_p = 'weno5' riemann_solver_in_q = 'upwind-flux' riemann_solver_in_p = 'upwind-flux' # Dimensionality considered in velocity space: p_dim = 1 # Number of devices(GPUs/Accelerators) on each node: num_devices = 1 # Constants: mass = af.Array([1, 100], (1, 2)) boltzmann_constant = 1 charge = af.Array([-1, 1], (1, 2)) # Initial Conditions used in initialize: rho_background_e = 1 rho_background_i = 1 temperature_background_e = 2.5 temperature_background_i = 1 # Parameter controlling amplitude of perturbation introduced: alpha = 0.01 # Time parameters: N_cfl = 0.1
def mode1_fdtd(Ez, Bx, By, length_domain_x, length_domain_y, ghost_cells, Jx,
Jy, Jz, dt):
forward_row = af.Array([1, -1, 0])
forward_column = af.Array([1, -1, 0])
backward_row = af.Array([0, 1, -1])
backward_column = af.Array([0, 1, -1])
identity = af.Array([0, 1, 0])
""" Number of grid points in the field's domain"""
(x_number_of_points, y_number_of_points) = Ez.dims()
""" number of grid zones from the input fields """
Nx = x_number_of_points - 2 * ghost_cells - 1
Ny = y_number_of_points - 2 * ghost_cells - 1
""" local variables for storing the input fields """
Ez_local = Ez.copy()
Bx_local = Bx.copy()
By_local = By.copy()
"""Enforcing BC's"""
Ez_local = periodic_ghost(Ez_local, ghost_cells)
Bx_local = periodic_ghost(Bx_local, ghost_cells)
By_local = periodic_ghost(By_local, ghost_cells)
""" Setting division size and time steps"""
dx = np.float(length_domain_x / (Nx))
dy = np.float(length_domain_y / (Ny))
""" defining variables for convenience """
dt_by_dx = dt / (dx)
dt_by_dy = dt / (dy)
""" Updating the Electric field using the current too """
Ez_local += dt_by_dx * (af.signal.convolve2_separable(identity, backward_column, By_local)) \
- dt_by_dy * (af.signal.convolve2_separable(backward_row, identity, Bx_local)) \
- dt*(Jz)
# dEz/dt = dBy/dx - dBx/dy
""" Implementing periodic boundary conditions using ghost cells """
Ez_local = periodic_ghost(Ez_local, ghost_cells)
""" Updating the Magnetic fields """
Bx_local += -dt_by_dy * (af.signal.convolve2_separable(
forward_row, identity, Ez_local))
# dBx/dt = -dEz/dy
By_local += dt_by_dx * (af.signal.convolve2_separable(
identity, forward_column, Ez_local))
# dBy/dt = +dEz/dx
""" Implementing periodic boundary conditions using ghost cells """
Bx_local = periodic_ghost(Bx_local, ghost_cells)
By_local = periodic_ghost(By_local, ghost_cells)
af.eval(Ez_local, Bx_local, By_local)
return Ez_local, Bx_local, By_local
def mode2_fdtd(
Bz,
Ex,
Ey,
length_domain_x,
length_domain_y,
ghost_cells,
Jx,
Jy,
Jz,
dt,
):
forward_row = af.Array([1, -1, 0])
forward_column = af.Array([1, -1, 0])
backward_row = af.Array([0, 1, -1])
backward_column = af.Array([0, 1, -1])
identity = af.Array([0, 1, 0])
""" Number of grid points in the field's domain """
(x_number_of_points, y_number_of_points) = Bz.dims()
""" number of grid zones calculated from the input fields """
Nx = x_number_of_points - 2 * ghost_cells - 1
Ny = y_number_of_points - 2 * ghost_cells - 1
""" local variables for storing the input fields """
Bz_local = Bz.copy()
Ex_local = Ex.copy()
Ey_local = Ey.copy()
"""Enforcing periodic BC's"""
Bz_local = periodic_ghost(Bz_local, ghost_cells)
Ex_local = periodic_ghost(Ex_local, ghost_cells)
Ey_local = periodic_ghost(Ey_local, ghost_cells)
""" Setting division size and time steps"""
dx = np.float(length_domain_x / (Nx))
dy = np.float(length_domain_y / (Ny))
""" defining variable for convenience """
dt_by_dx = dt / (dx)
dt_by_dy = dt / (dy)
""" Updating the Electric fields using the current too """
Ex_local += dt_by_dy * (af.signal.convolve2_separable(
backward_row, identity, Bz_local)) - (Jx) * dt
# dEx/dt = + dBz/dy
Ey_local += -dt_by_dx * (af.signal.convolve2_separable(
identity, backward_column, Bz_local)) - (Jy) * dt
# dEy/dt = - dBz/dx
""" Implementing periodic boundary conditions using ghost cells """
Ex_local = periodic_ghost(Ex_local, ghost_cells)
Ey_local = periodic_ghost(Ey_local, ghost_cells)
""" Updating the Magnetic field """
Bz_local += - dt_by_dx * (af.signal.convolve2_separable(identity, forward_column, Ey_local)) \
+ dt_by_dy * (af.signal.convolve2_separable(forward_row, identity, Ex_local))
# dBz/dt = - ( dEy/dx - dEx/dy )
#Implementing periodic boundary conditions using ghost cells
Bz_local = periodic_ghost(Bz_local, ghost_cells)
af.eval(Bz_local, Ex_local, Ey_local)
return Bz_local, Ex_local, Ey_local
def testArrayfire(self):
a = af.Array([1, 2, 3, 4])
a_data = a.to_list()
b = Array.from_arrayfire(a)
np.testing.assert_array_equal(np.asarray(a_data), np.asarray(b.to_list()))
def af_gaussian2D(s): csize = ct.c_int(int(3*s)*2+1) csigma = ct.c_double(s) d_k = af.Array() af.safe_call(af.backend.get().af_gaussian_kernel(ct.pointer(d_k.arr),csize,csize,csigma,csigma)) return d_k
af.set_backend('cuda')
v = af.random.randu(size_of_vector)
time_sort_device = af.timer.timeit(af.algorithm.sort, v)
print('Device time: {}'.format(time_sort_device))
print('Speedup: {0:.2f}x'.format(time_sort_host / time_sort_device))
# Test de convolution
n = 1000
print('')
print('=== 2d convolution on a {}x{} matrix ==='.format(n, n))
af.set_backend('cpu')
m = af.random.randu(n, n)
numpy_kernel = np.array([[0.1, 0.2, 0.1], [0.2, 0.3, 0.2], [0.1, 0.2, 0.1]])
af_kernel = af.Array(numpy_kernel.ctypes.data, numpy_kernel.shape,
numpy_kernel.dtype.char)
time_convolve_host = af.timer.timeit(af.convolve, m, af_kernel)
print('Host time: {}'.format(time_convolve_host))
af.set_backend('cuda')
af_kernel = af.Array(numpy_kernel.ctypes.data, numpy_kernel.shape,
numpy_kernel.dtype.char)
m = af.random.randu(n, n)
time_convolve_device = af.timer.timeit(af.convolve, m, af_kernel)
print('Device time: {}'.format(time_convolve_device))
print('Speedup: {0:.2f}x'.format(time_convolve_host / time_convolve_device))
# Interop numpy
print('')