def runTest(self):
nx, ny, nz, str_f, pt0, pt1 = self.args
slidx = common.slices_two_points(pt0, pt1)
str_fs = common.convert_to_tuple(str_f)
# instance
fields = Fields(0, nx, ny, nz, '', 'single')
getf = GetFields(fields, str_f, pt0, pt1)
# host allocations
ehs = common_random.generate_ehs(nx, ny, nz, fields.dtype)
eh_dict = dict( zip(['ex', 'ey', 'ez', 'hx', 'hy', 'hz'], ehs) )
fields.set_eh_bufs(*ehs)
# verify
getf.wait()
for str_f in str_fs:
original = eh_dict[str_f][slidx]
copy = getf.get_fields(str_f)
norm = np.linalg.norm(original - copy)
self.assertEqual(norm, 0, '%s, %g' % (self.args, norm))
fields.context_pop()
def runTest(self):
nx, ny, nz, str_f, pt0, pt1 = self.args
slidx = common.slices_two_points(pt0, pt1)
str_fs = common.convert_to_tuple(str_f)
# instance
gpu_devices = common_gpu.gpu_device_list(print_info=False)
context = cl.Context(gpu_devices)
device = gpu_devices[0]
fields = Fields(context, device, nx, ny, nz, '', 'single')
getf = GetFields(fields, str_f, pt0, pt1)
# host allocations
ehs = common_random.generate_ehs(nx, ny, nz, fields.dtype)
eh_dict = dict( zip(['ex', 'ey', 'ez', 'hx', 'hy', 'hz'], ehs) )
fields.set_eh_bufs(*ehs)
# verify
getf.get_event().wait()
for str_f in str_fs:
original = eh_dict[str_f][slidx]
copy = getf.get_fields(str_f)
norm = np.linalg.norm(original - copy)
self.assertEqual(norm, 0, '%s, %g' % (self.args, norm))
def runTest(self):
nx, ny, nz, str_f, pt0, pt1 = self.args
slidx = common.slice_index_two_points(pt0, pt1)
str_fs = common.convert_to_tuple(str_f)
# instance
gpu_devices = common_gpu.gpu_device_list(print_info=False)
context = cl.Context(gpu_devices)
device = gpu_devices[0]
fields = Fields(context, device, nx, ny, nz, '', 'single')
getf = GetFields(fields, str_f, pt0, pt1)
# host allocations
ehs = common_update.generate_random_ehs(nx, ny, nz, fields.dtype)
eh_dict = dict( zip(['ex', 'ey', 'ez', 'hx', 'hy', 'hz'], ehs) )
fields.set_eh_bufs(*ehs)
# verify
getf.get_event().wait()
for str_f in str_fs:
original = eh_dict[str_f][slidx]
copy = getf.get_fields(str_f)
norm = np.linalg.norm(original - copy)
self.assertEqual(norm, 0, '%s, %g' % (self.args, norm))
def verify(s, pt0, pt1):
print('pt0 = %s, pt1 = %s' % (pt0, pt1))
slidx = s.get_slidx(pt0, pt1)
for strf in s.strf_list:
# non-spatial
fset = SetFields(s.fdtd, strf, pt0, pt1)
value = np.random.rand()
fset.set_fields(value)
fget = GetFields(s.fdtd, strf, pt0, pt1)
fget.get_event().wait()
copy = fget.get_fields(strf)
assert np.linalg.norm(value - copy) == 0
if pt0 != pt1:
for strf in s.strf_list:
# spatial
fset = SetFields(s.fdtd, strf, pt0, pt1, np.ndarray)
values = np.random.rand(*fset.rplist['shape']).astype(
s.fdtd.dtype)
fset.set_fields(values)
fget = GetFields(s.fdtd, strf, pt0, pt1)
fget.get_event().wait()
copy = fget.get_fields(strf)
assert np.linalg.norm(values - copy) == 0
def __init__(self, gpuf, core, direction):
common.check_type('gpuf', gpuf, Fields)
common.check_value('direction', direction, ('+', '-', '+-'))
if '+' in direction:
self.gf_h = gf_h = GetFields(gpuf, ['hy', 'hz'], (-1, 0, 0),
(-1, -1, -1))
self.sf_e = SetFields(gpuf, ['ey', 'ez'], (-1, 0, 0), (-1, -1, -1),
True)
target = 0 if rank == size - 1 else rank + 1
self.req_send_h = comm.Send_init(gf_h.host_array, target, tag=0)
self.tmp_recv_e = np.zeros(gf_h.host_array.shape, gpuf.dtype)
self.req_recv_e = comm.Recv_init(self.tmp_recv_e, target, tag=1)
if '-' in direction:
self.gf_e = gf_e = GetFields(gpuf, ['ey', 'ez'], (0, 0, 0),
(0, -1, -1))
self.sf_h = SetFields(gpuf, ['hy', 'hz'], (0, 0, 0), (0, -1, -1),
True)
target = size - 1 if rank == 0 else rank - 1
self.req_send_e = comm.Send_init(gf_e.host_array, target, tag=1)
self.tmp_recv_h = np.zeros(gf_e.host_array.shape, gpuf.dtype)
self.req_recv_h = comm.Recv_init(self.tmp_recv_h, target, tag=0)
# global variables
self.core = core
self.direction = direction
def test_boundary(s):
print('\n-- test boundary (two fields) --')
shape_dict = {
'x': (s.ny * 2, s.nz),
'y': (s.nx * 2, s.nz),
'z': (s.nx * 2, s.ny)
}
print('E fields')
str_fs_dict = {'x': ['ey', 'ez'], 'y': ['ex', 'ez'], 'z': ['ex', 'ey']}
pt0_dict = {
'x': (s.nx - 1, 0, 0),
'y': (0, s.ny - 1, 0),
'z': (0, 0, s.nz - 1)
}
pt1 = (s.nx - 1, s.ny - 1, s.nz - 1)
for axis in str_fs_dict.keys():
print('direction : %s' % axis)
str_fs = str_fs_dict[axis]
pt0 = pt0_dict[axis]
slidx = s.get_slidx(pt0, pt1)
fset = SetFields(s.fdtd, str_fs, pt0, pt1, np.ndarray)
values = np.random.rand(*shape_dict[axis]).astype(s.fdtd.dtype)
fset.set_fields(values)
fget = GetFields(s.fdtd, str_fs, pt0, pt1)
fget.get_event().wait()
copy = fget.get_fields()
assert np.linalg.norm(values - copy) == 0
print('H fields')
str_fs_dict = {'x': ['hy', 'hz'], 'y': ['hx', 'hz'], 'z': ['hx', 'hy']}
pt0 = (0, 0, 0)
pt1_dict = {
'x': (0, s.ny - 1, s.nz - 1),
'y': (s.nx - 1, 0, s.nz - 1),
'z': (s.nx - 1, s.ny - 1, 0)
}
for axis in str_fs_dict.keys():
print('direction : %s' % axis)
str_fs = str_fs_dict[axis]
pt1 = pt1_dict[axis]
slidx = s.get_slidx(pt0, pt1)
fset = SetFields(s.fdtd, str_fs, pt0, pt1, np.ndarray)
values = np.random.rand(*shape_dict[axis]).astype(s.fdtd.dtype)
fset.set_fields(values)
fget = GetFields(s.fdtd, str_fs, pt0, pt1)
fget.get_event().wait()
copy = fget.get_fields()
assert np.linalg.norm(values - copy) == 0
def verify(s, pt0, pt1):
print('pt0 = %s, pt1 = %s' % (pt0, pt1))
slidx = s.get_slidx(pt0, pt1)
for strf in s.strf_list:
fget = GetFields(s.fdtd, strf, pt0, pt1)
fget.get_event().wait()
original = s.fhosts[strf][slidx]
copy = fget.get_fields(strf)
#print original, copy
assert np.linalg.norm(original - copy) == 0
def verify(s, pt0, pt1):
print('pt0 = %s, pt1 = %s' % (pt0, pt1))
slidx = s.get_slidx(pt0, pt1)
for strf in s.strf_list:
fget = GetFields(s.fdtd, strf, pt0, pt1)
fget.get_event().wait()
original = s.fhosts[strf][slidx]
copy = fget.get_fields(strf)
#print original, copy
assert np.linalg.norm(original - copy) == 0
def __init__(self, node_fields, target_rank, direction):
common.check_type('node_fields', node_fields, Fields)
common.check_type('target_rank', target_rank, int)
# local variables
nodef = node_fields
dtype = nodef.dtype
assert rank != target_rank, 'The target_rank %d is same as the my_rank %d.' % (
target_rank, rank)
# create instances (getf, setf and mpi requests)
if '+' in direction:
mainf = nodef.mainf_list[-1]
nx, ny, nz = mainf.ns
getf = GetFields(mainf, ['hy', 'hz'], (nx - 1, 0, 0),
(nx - 1, ny - 1, nz - 1))
setf = SetFields(mainf, ['ey', 'ez'], (nx - 1, 0, 0),
(nx - 1, ny - 1, nz - 1), True)
if rank < target_rank:
tag_send, tag_recv = 0, 1
elif rank > target_rank:
tag_send, tag_recv = 2, 3
elif '-' in direction:
mainf = nodef.mainf_list[0]
nx, ny, nz = mainf.ns
getf = GetFields(mainf, ['ey', 'ez'], (0, 0, 0),
(0, ny - 1, nz - 1))
setf = SetFields(mainf, ['hy', 'hz'], (0, 0, 0),
(0, ny - 1, nz - 1), True)
if rank > target_rank:
tag_send, tag_recv = 1, 0
elif rank < target_rank: # pbc
tag_send, tag_recv = 3, 2
# global variables
self.target_rank = target_rank
self.getf = getf
self.setf = setf
self.tag_send = tag_send
self.tag_recv = tag_recv
self.tmp_recv = np.zeros(getf.host_array.shape, dtype)
# global functions
self.update_e = self.recv if '+' in direction else self.send
self.update_h = self.send if '+' in direction else self.recv
# append instance
self.priority_type = 'mpi'
nodef.append_instance(self)
def __init__(s, fields_list, axis):
emf_list = fields_list
e_strfs, h_strfs, pt0, pt1 = common_gpu.get_strfs_pts(
axis, *emf_list[0].ns)
s.e_getf = GetFields(emf_list[0], e_strfs, pt0['-'], pt1['-'])
s.h_setf = SetFields(emf_list[0], h_strfs, pt0['-'], pt1['-'],
np.ndarray, True)
s.h_getf = GetFields(emf_list[-1], h_strfs, pt0['+'], pt1['+'])
s.e_setf = SetFields(emf_list[-1], e_strfs, pt0['+'], pt1['+'],
np.ndarray, True)
def verify(s, pt0, pt1):
print('pt0 = %s, pt1 = %s' % (pt0, pt1))
slidx = s.get_slidx(pt0, pt1)
for strf in s.strf_list:
# non-spatial
fset = SetFields(s.fdtd, strf, pt0, pt1)
value = np.random.rand()
fset.set_fields(value)
fget = GetFields(s.fdtd, strf, pt0, pt1)
fget.get_event().wait()
copy = fget.get_fields(strf)
assert np.linalg.norm(value - copy) == 0
if pt0 != pt1:
for strf in s.strf_list:
# spatial
fset = SetFields(s.fdtd, strf, pt0, pt1, np.ndarray)
values = np.random.rand(*fset.rplist['shape']).astype(s.fdtd.dtype)
fset.set_fields(values)
fget = GetFields(s.fdtd, strf, pt0, pt1)
fget.get_event().wait()
copy = fget.get_fields(strf)
assert np.linalg.norm(values - copy) == 0
def test_boundary(s):
print('\n-- test boundary (two fields) --')
print('E fields')
str_fs_dict = {'x': ['ey', 'ez'], 'y': ['ex', 'ez'], 'z': ['ex', 'ey']}
pt0 = (0, 0, 0)
pt1_dict = {
'x': (0, s.ny - 1, s.nz - 1),
'y': (s.nx - 1, 0, s.nz - 1),
'z': (s.nx - 1, s.ny - 1, 0)
}
for axis in str_fs_dict.keys():
print('direction : %s' % axis)
str_fs = str_fs_dict[axis]
pt1 = pt1_dict[axis]
slidx = s.get_slidx(pt0, pt1)
fget = GetFields(s.fdtd, str_fs, pt0, pt1)
fget.get_event().wait()
for strf in str_fs:
original = s.fhosts[strf][slidx]
copy = fget.get_fields(strf)
assert np.linalg.norm(original - copy) == 0
print('H fields')
str_fs_dict = {'x': ['hy', 'hz'], 'y': ['hx', 'hz'], 'z': ['hx', 'hy']}
pt0_dict = {
'x': (s.nx - 1, 0, 0),
'y': (0, s.ny - 1, 0),
'z': (0, 0, s.nz - 1)
}
pt1 = (s.nx - 1, s.ny - 1, s.nz - 1)
for axis in str_fs_dict.keys():
print('direction : %s' % axis)
str_fs = str_fs_dict[axis]
pt0 = pt0_dict[axis]
slidx = s.get_slidx(pt0, pt1)
fget = GetFields(s.fdtd, str_fs, pt0, pt1)
fget.get_event().wait()
for strf in str_fs:
original = s.fhosts[strf][slidx]
copy = fget.get_fields(strf)
assert np.linalg.norm(original - copy) == 0
def __init__(s, fields_list, axis):
emf_list = fields_list
e_strfs, h_strfs, pt0, pt1 = common.get_strfs_pts_for_boundary(
axis, *emf_list[0].ns)
s.e_getfs, s.h_getfs = [], []
s.e_setfs, s.h_setfs = [], []
for emf in emf_list[1:]:
s.e_getfs.append(GetFields(emf, e_strfs, pt0['-'], pt1['-']))
s.h_setfs.append(
SetFields(emf, h_strfs, pt0['-'], pt1['-'], np.ndarray))
for emf in emf_list[:-1]:
s.h_getfs.append(GetFields(emf, h_strfs, pt0['+'], pt1['+']))
s.e_setfs.append(
SetFields(emf, e_strfs, pt0['+'], pt1['+'], np.ndarray))
def __init__(self, gpuf, core, direction, tmax):
common.check_type('gpuf', gpuf, Fields)
common.check_value('direction', direction, ('+', '-', '+-'))
if '+' in direction:
self.gf_h = gf_h = GetFields(gpuf, ['hy', 'hz'], (-1, 0, 0),
(-1, -1, -1))
self.sf_e = SetFields(gpuf, ['ey', 'ez'], (-1, 0, 0), (-1, -1, -1),
True)
self.req_send_h = comm.Send_init(gf_h.host_array, rank + 1, tag=0)
self.tmp_recv_e_list = [
np.zeros(gf_h.host_array.shape, gpuf.dtype) for i in range(2)
]
self.req_recv_e_list = [
comm.Recv_init(tmp_recv_e, rank + 1, tag=1)
for tmp_recv_e in self.tmp_recv_e_list
]
self.switch_e = 0
if '-' in direction:
self.gf_e = gf_e = GetFields(gpuf, ['ey', 'ez'], (0, 0, 0),
(0, -1, -1))
self.sf_h = SetFields(gpuf, ['hy', 'hz'], (0, 0, 0), (0, -1, -1),
True)
self.req_send_e = comm.Send_init(gf_e.host_array, rank - 1, tag=1)
self.tmp_recv_h_list = [
np.zeros(gf_e.host_array.shape, gpuf.dtype) for i in range(2)
]
self.req_recv_h_list = [
comm.Recv_init(tmp_recv_h, rank - 1, tag=0)
for tmp_recv_h in self.tmp_recv_h_list
]
self.switch_h = 0
# global variables
self.gpuf = gpuf
self.core = core
self.direction = direction
self.tmax = tmax
self.tstep = 1
def test_boundary(s):
print('\n-- test boundary (two fields) --')
shape_dict = {'x':(s.ny*2, s.nz), 'y':(s.nx*2, s.nz), 'z':(s.nx*2, s.ny)}
print('E fields')
str_fs_dict = {'x':['ey','ez'], 'y':['ex','ez'], 'z':['ex','ey']}
pt0_dict = {'x':(s.nx-1, 0, 0), 'y':(0, s.ny-1, 0), 'z':(0, 0, s.nz-1)}
pt1 = (s.nx-1, s.ny-1, s.nz-1)
for axis in str_fs_dict.keys():
print('direction : %s' % axis)
str_fs = str_fs_dict[axis]
pt0 = pt0_dict[axis]
slidx = s.get_slidx(pt0, pt1)
fset = SetFields(s.fdtd, str_fs, pt0, pt1, np.ndarray)
values = np.random.rand(*shape_dict[axis]).astype(s.fdtd.dtype)
fset.set_fields(values)
fget = GetFields(s.fdtd, str_fs, pt0, pt1)
fget.get_event().wait()
copy = fget.get_fields()
assert np.linalg.norm(values - copy) == 0
print('H fields')
str_fs_dict = {'x':['hy','hz'], 'y':['hx','hz'], 'z':['hx','hy']}
pt0 = (0, 0, 0)
pt1_dict = {'x':(0, s.ny-1, s.nz-1), 'y':(s.nx-1, 0, s.nz-1), 'z':(s.nx-1, s.ny-1, 0)}
for axis in str_fs_dict.keys():
print('direction : %s' % axis)
str_fs = str_fs_dict[axis]
pt1 = pt1_dict[axis]
slidx = s.get_slidx(pt0, pt1)
fset = SetFields(s.fdtd, str_fs, pt0, pt1, np.ndarray)
values = np.random.rand(*shape_dict[axis]).astype(s.fdtd.dtype)
fset.set_fields(values)
fget = GetFields(s.fdtd, str_fs, pt0, pt1)
fget.get_event().wait()
copy = fget.get_fields()
assert np.linalg.norm(values - copy) == 0
def test_boundary(s):
print('\n-- test boundary (two fields) --')
print('E fields')
str_fs_dict = {'x':['ey','ez'], 'y':['ex','ez'], 'z':['ex','ey']}
pt0 = (0, 0, 0)
pt1_dict = {'x':(0, s.ny-1, s.nz-1), 'y':(s.nx-1, 0, s.nz-1), 'z':(s.nx-1, s.ny-1, 0)}
for axis in str_fs_dict.keys():
print('direction : %s' % axis)
str_fs = str_fs_dict[axis]
pt1 = pt1_dict[axis]
slidx = s.get_slidx(pt0, pt1)
fget = GetFields(s.fdtd, str_fs, pt0, pt1)
fget.get_event().wait()
for strf in str_fs:
original = s.fhosts[strf][slidx]
copy = fget.get_fields(strf)
assert np.linalg.norm(original - copy) == 0
print('H fields')
str_fs_dict = {'x':['hy','hz'], 'y':['hx','hz'], 'z':['hx','hy']}
pt0_dict = {'x':(s.nx-1, 0, 0), 'y':(0, s.ny-1, 0), 'z':(0, 0, s.nz-1)}
pt1 = (s.nx-1, s.ny-1, s.nz-1)
for axis in str_fs_dict.keys():
print('direction : %s' % axis)
str_fs = str_fs_dict[axis]
pt0 = pt0_dict[axis]
slidx = s.get_slidx(pt0, pt1)
fget = GetFields(s.fdtd, str_fs, pt0, pt1)
fget.get_event().wait()
for strf in str_fs:
original = s.fhosts[strf][slidx]
copy = fget.get_fields(strf)
assert np.linalg.norm(original - copy) == 0
# plot
import matplotlib.pyplot as plt
plt.ion()
fig = plt.figure(figsize=(12,8))
imag = plt.imshow(np.zeros((nx, ny), fields.dtype).T, interpolation='nearest', origin='lower', vmin=-1.1, vmax=1.1)
plt.colorbar()
'''
# main loop
from datetime import datetime
from time import time
t0 = datetime.now()
t00 = time()
gtmp = GetFields(fields, 'ez', (0, 0, 0), (0, 0, 0))
for tstep in xrange(1, tmax + 1):
fields.update_e()
fields.update_h()
'''
if tstep % tgap == 0:
print('[%s] %d/%d (%d %%)\r' % (datetime.now() - t0, tstep, tmax, float(tstep)/tmax*100)),
sys.stdout.flush()
getf.get_event().wait()
imag.set_array( getf.get_fields().T )
#plt.savefig('./png/%.6d.png' % tstep)
plt.draw()
'''
gtmp.get_event().wait()
# plot
import matplotlib.pyplot as plt
plt.ion()
fig = plt.figure(figsize=(12,8))
imag = plt.imshow(np.zeros((nx, ny), fields.dtype).T, interpolation='nearest', origin='lower', vmin=-1.1, vmax=1.1)
plt.colorbar()
'''
# main loop
from datetime import datetime
from time import time
t0 = datetime.now()
t00 = time()
gtmp = GetFields(fields, 'ez', (0, 0, 0), (0, 0, 0))
for tstep in xrange(1, tmax+1):
fields.update_e()
fields.update_h()
'''
if tstep % tgap == 0:
print('[%s] %d/%d (%d %%)\r' % (datetime.now() - t0, tstep, tmax, float(tstep)/tmax*100)),
sys.stdout.flush()
getf.get_event().wait()
imag.set_array( getf.get_fields().T )
#plt.savefig('./png/%.6d.png' % tstep)
plt.draw()
'''
device = gpu_devices[0]
# z-axis
nx, ny, nz = 180, 160, 2
fields = Fields(context, device, nx, ny, nz)
Core(fields)
Pbc(fields, 'xyz')
IncidentDirect(fields, 'ey', (20, 0, 0), (20, ny-1, nz-1), tfunc)
IncidentDirect(fields, 'ex', (0, 20, 0), (nx-1, 20, nz-1), tfunc)
for tstep in xrange(1, tmax+1):
fields.update_e()
fields.update_h()
ax1 = fig.add_subplot(2, 3, 1)
getf = GetFields(fields, 'ey', (0, 0, nz/2), (nx-1, ny-1, nz/2))
getf.get_event().wait()
ax1.imshow(getf.get_fields().T, vmin=-1.1, vmax=1.1)
ax1.set_title('%s, ey[20,:,:]' % repr(fields.ns))
ax1.set_xlabel('x')
ax1.set_ylabel('y')
ax2 = fig.add_subplot(2, 3, 4)
getf = GetFields(fields, 'ex', (0, 0, nz/2), (nx-1, ny-1, nz/2))
getf.get_event().wait()
ax2.imshow(getf.get_fields().T, vmin=-1.1, vmax=1.1)
ax2.set_title('%s, ex[:,20,:]' % repr(fields.ns))
ax2.set_xlabel('x')
ax2.set_ylabel('y')
# y-axis
import pyopencl as cl
from kemp.fdtd3d import common_gpu
from kemp.fdtd3d.gpu import Fields, GetFields, Npml
nx, ny, nz = 100, 110, 128
gpu_id = 0
gpu_devices = common_gpu.get_gpu_devices()
context = cl.Context(gpu_devices)
device = gpu_devices[gpu_id]
fdtd = Fields(context, device, nx, ny, nz, coeff_use='')
fhosts = {}
pml = Npml(fdtd, -0.428571428571, 0.714285714286, 0.6, 0.2, 0.6, 0.2)
fhost = np.random.rand(nx, nz).astype(fdtd.dtype)
cl.enqueue_write_buffer(fdtd.queue, pml.pex, fhost)
pml.update_h()
fget = GetFields(fdtd, ['hz', 'hx'], (0, ny - 1, 0), (nx - 1, ny - 1, nz - 1))
fget.get_event().wait()
hz = fget.get_fields('hz')
hx = fget.get_fields('hx')
print fhost
print hz
print fhost.shape
print hz.shape
assert np.linalg.norm(fhost - hz) == 0
import numpy as np
import pyopencl as cl
nx, ny, nz = 2, 640, 640
tmax, tgap = 1000, 10
gpu_id = 0
gpu_devices = common_gpu.get_gpu_devices()
context = cl.Context(gpu_devices)
device = gpu_devices[gpu_id]
fdtd = Fields(context, device, nx, ny, nz, coeff_use='')
src = DirectSrc(fdtd, 'ex', (1, ny / 5 * 4, nz / 5 * 3),
(1, ny / 5 * 4, nz / 5 * 3), lambda tstep: np.sin(0.1 * tstep))
pbc = PbcInt(fdtd, 'x')
output = GetFields(fdtd, 'ex', (1, 0, 0), (1, ny - 1, nz - 1))
# Plot
import matplotlib.pyplot as plt
plt.ion()
imag = plt.imshow(output.get_fields().T,
cmap=plt.cm.hot,
origin='lower',
vmin=0,
vmax=0.05)
plt.colorbar()
# Main loop
from datetime import datetime
t0 = datetime.now()
import pyopencl as cl
from kemp.fdtd3d import common_gpu
from kemp.fdtd3d.gpu import Fields, GetFields, Npml
nx, ny, nz = 100, 110, 128
gpu_id = 0
gpu_devices = common_gpu.get_gpu_devices()
context = cl.Context(gpu_devices)
device = gpu_devices[gpu_id]
fdtd = Fields(context, device, nx, ny, nz, coeff_use='')
fhosts = {}
pml = Npml(fdtd, -0.428571428571, 0.714285714286, 0.6, 0.2, 0.6, 0.2)
fhost = np.random.rand(nx, nz).astype(fdtd.dtype)
cl.enqueue_write_buffer(fdtd.queue, pml.pex, fhost)
pml.update_h()
fget = GetFields(fdtd, ['hz','hx'], (0, ny-1, 0), (nx-1, ny-1, nz-1))
fget.get_event().wait()
hz = fget.get_fields('hz')
hx = fget.get_fields('hx')
print fhost
print hz
print fhost.shape
print hz.shape
assert np.linalg.norm(fhost - hz) == 0
import numpy as np
import pyopencl as cl
nx, ny, nz = 120, 320, 320
tmax, tgap = 300, 5
divide_axes = 'x'
gpu_devices = common_gpu.get_gpu_devices()
context = cl.Context(gpu_devices)
ngpu = len(gpu_devices)
fdtds = [
Fields(context, device, nx, ny, nz, coeff_use='') for device in gpu_devices
]
outputs = [
GetFields(fdtd, 'ez', (0, 0, nz / 2), (nx - 2, ny - 1, nz / 2))
for fdtd in fdtds
]
src_e = DirectSrc(fdtds[1], 'ez', (nx / 3 * 2, ny / 2, 0),
(nx / 3 * 2, ny / 2, nz - 1),
lambda tstep: np.sin(0.1 * tstep))
exch = ExchangeInternal(fdtds, 'x')
# Plot
import matplotlib.pyplot as plt
plt.ion()
idxs = [0] + [i * nx - i for i in range(1, ngpu + 1)] # [0, 239, 478, 717]
for idx in idxs[1:]:
plt.plot((idx, idx), (0, ny), color='w', linewidth=0.2)
global_ez = np.ones((idxs[-1], ny), dtype=fdtds[0].dtype)
#nx, ny, nz = 240, 256, 256 # 540 MB
#nx, ny, nz = 512, 480, 480 # 3.96 GB
#nx, ny, nz = 480, 480, 480 # 3.71 GB
nx, ny, nz = 240, 640, 640
tmax, tgap = 200, 10
gpu_id = 0
gpu_devices = common_gpu.get_gpu_devices()
context = cl.Context(gpu_devices)
device = gpu_devices[gpu_id]
fdtd = Fields(context, device, nx, ny, nz, coeff_use='')
src = DirectSrc(fdtd, 'ez', (nx/5*4, ny/2, 0), (nx/5*4, ny/2, nz-1), lambda tstep: np.sin(0.1 * tstep))
pbc = PbcInt(fdtd, 'x')
output = GetFields(fdtd, 'ez', (0, 0, nz/2), (nx-1, ny-1, nz/2))
# Plot
import matplotlib.pyplot as plt
plt.ion()
imag = plt.imshow(output.get_fields('ez').T, cmap=plt.cm.hot, origin='lower', vmin=0, vmax=0.05)
plt.colorbar()
# Main loop
from datetime import datetime
t0 = datetime.now()
for tstep in xrange(1, tmax+1):
fdtd.update_e()
elif rank == size - 1: direction = '-'
else: direction = '+-'
#exch = node.ExchangeMpiNonBlock(fields, direction)
#exch = node.ExchangeMpiBufferBlock(fields, direction)
#exch = node.ExchangeMpiBufferBlockSplit(fields, direction)
exch = node.ExchangeMpiBufferNonBlockSplitEnqueue(fields, direction, tmax)
if '+' in direction: cpu.Core(exch.cpuf_p)
if '-' in direction: cpu.Core(exch.cpuf_m)
is_master = True if rank == 0 else False
if is_plot:
Pbc(fields, 'yz')
getf = GetFields(fields, 'ez', (0, 0, 0.5), (-1, -1, 0.5))
tfunc = lambda tstep: 40 * np.sin(0.05 * tstep)
if rank < size - 1:
IncidentDirect(fields, 'ez', (220, 0.5, 0), (220, 0.5, -1), tfunc)
#if rank > 0:
# IncidentDirect(fields, 'ez', (20, 0.5, 0), (20, 0.5, -1), tfunc)
if is_master:
# plot
import matplotlib.pyplot as plt
plt.ion()
fig = plt.figure(figsize=(12, 8))
arr = np.zeros((size * nx, ny), fields.dtype)
for i in range(1, size):
plt.plot((i * nx, i * nx), (0, ny), color='k', linewidth=1.2)
from kemp.fdtd3d.gpu import Fields, DirectSrc, GetFields, ExchangeFields
import numpy as np
import pyopencl as cl
nx, ny, nz = 240, 640, 640
tmax, tgap = 200, 10
divide_axes = 'x'
gpu_devices = common_gpu.get_gpu_devices()
context = cl.Context(gpu_devices)
ngpu = len(gpu_devices)
fdtds = [
Fields(context, device, nx, ny, nz, coeff_use='') for device in gpu_devices
]
outputs = [GetFields(fdtds[0], 'ez', (0, 0, nz / 2), (nx - 1, ny - 1, nz / 2))]
outputs += [
GetFields(fdtd, 'ez', (1, 0, nz / 2), (nx - 1, ny - 1, nz / 2))
for fdtd in fdtds[1:]
]
src = DirectSrc(fdtds[1], 'ez', (nx / 5 * 4, ny / 2, 0),
(nx / 5 * 4, ny / 2, nz - 1),
lambda tstep: np.sin(0.1 * tstep))
exch = ExchangeFields(fdtds, 'x')
# Plot
import matplotlib.pyplot as plt
plt.ion()
global_ez = np.ones((ngpu * (nx - 1) + 1, ny), dtype=fdtds[0].dtype)
imag = plt.imshow(global_ez.T,
nx, ny, nz = 2, 250, 300
tmax, tgap = 1000, 10
npml = 10
# instances
gpu_devices = common_gpu.gpu_device_list(print_info=False)
context = cl.Context(gpu_devices)
device = gpu_devices[0]
fields = Fields(context, device, nx, ny, nz)
Pbc(fields, 'x')
Pml(fields, ('', '+-', '+-'), npml)
Core(fields)
tfunc = lambda tstep: 50 * np.sin(0.05 * tstep)
IncidentDirect(fields, 'ex', (0, 0.4, 0.3), (-1, 0.4, 0.3), tfunc)
getf = GetFields(fields, 'ex', (0.5, 0, 0), (0.5, -1, -1))
print fields.instance_list
# plot
import matplotlib.pyplot as plt
plt.ion()
fig = plt.figure(figsize=(12, 8))
imag = plt.imshow(np.zeros((ny, nz), fields.dtype).T,
interpolation='nearest',
origin='lower',
vmin=-1.1,
vmax=1.1)
plt.xlabel('y')
plt.ylabel('z')
plt.colorbar()
nx, ny, nz = 160, 140, 32 tmax, tgap = 150, 10 # instances gpu_devices = common_gpu.gpu_device_list(print_info=False) context = cl.Context(gpu_devices) device = gpu_devices[0] fields = Fields(context, device, nx, ny, nz) Core(fields) Pbc(fields, 'xyz') tfunc = lambda tstep: np.sin(0.05 * tstep) IncidentDirect(fields, 'ez', (20, 0, 0), (20, ny-1, nz-1), tfunc) #IncidentDirect(fields, 'ez', (0, 20, 0), (nx-1, 20, nz-1), tfunc) getf = GetFields(fields, 'ez', (0, 0, nz/2), (nx-1, ny-1, nz/2)) print fields.instance_list # plot import matplotlib.pyplot as plt plt.ion() fig = plt.figure(figsize=(12,8)) imag = plt.imshow(np.zeros((nx, ny), fields.dtype).T, interpolation='nearest', origin='lower', vmin=-1.1, vmax=1.1) plt.colorbar() # main loop from datetime import datetime t0 = datetime.now() for tstep in xrange(1, tmax+1):
nx, ny, nz = 240, 256, 256 # 540 MB
#nx, ny, nz = 512, 480, 480 # 3.96 GB
#nx, ny, nz = 480, 480, 480 # 3.71 GB
tmax, tgap = 200, 10
gpu_id = 0
gpu_devices = common_gpu.get_gpu_devices()
context = cl.Context(gpu_devices)
device = gpu_devices[gpu_id]
fdtd = Fdtd(context, device, nx, ny, nz, coeff_use='')
src = DirectSrc(fdtd, 'ez', (nx / 3 * 2, ny / 2, 0),
(nx / 3 * 2, ny / 2, nz - 1),
lambda tstep: np.sin(0.1 * tstep))
output = GetFields(fdtd, 'ez', (0, 0, nz / 2), (nx - 1, ny - 1, nz / 2))
# Plot
import matplotlib.pyplot as plt
plt.ion()
imag = plt.imshow(output.get_fields('ez').T,
cmap=plt.cm.hot,
origin='lower',
vmin=0,
vmax=0.05)
plt.colorbar()
# Main loop
from datetime import datetime
t0 = datetime.now()
import pyopencl as cl nx, ny, nz = 2, 640, 640 tmax, tgap = 1000, 10 gpu_id = 0 gpu_devices = common_gpu.get_gpu_devices() context = cl.Context(gpu_devices) device = gpu_devices[gpu_id] fdtd = Fields(context, device, nx, ny, nz, coeff_use='') src_e = DirectSrc(fdtd, 'ex', (1, ny/5*4, nz/5*1), (1, ny/5*4, nz/5*1), lambda tstep: np.sin(0.1 * tstep)) pbc = PbcInt(fdtd, 'x') pml = Npml(fdtd, -0.428571428571, 0.714285714286, 0.6, 0.2, 0.6, 0.2) output = GetFields(fdtd, 'ex', (1, 0, 0), (1, ny-1, nz-1)) # Plot import matplotlib.pyplot as plt plt.ion() imag = plt.imshow(output.get_fields().T, cmap=plt.cm.hot, origin='lower', vmin=0, vmax=0.01) plt.colorbar() # Main loop from datetime import datetime t0 = datetime.now() for tstep in xrange(1, tmax+1): fdtd.update_e()
import pyopencl as cl
nx, ny, nz = 240, 256, 256 # 540 MB
#nx, ny, nz = 512, 480, 480 # 3.96 GB
#nx, ny, nz = 480, 480, 480 # 3.71 GB
tmax, tgap = 200, 10
gpu_id = 0
gpu_devices = common_gpu.get_gpu_devices()
context = cl.Context(gpu_devices)
device = gpu_devices[gpu_id]
fdtd = Fdtd(context, device, nx, ny, nz, coeff_use='')
src = DirectSrc(fdtd, 'ez', (nx/3*2, ny/2, 0), (nx/3*2, ny/2, nz-1), lambda tstep: np.sin(0.1 * tstep))
output = GetFields(fdtd, 'ez', (0, 0, nz/2), (nx-1, ny-1, nz/2))
# Plot
import matplotlib.pyplot as plt
plt.ion()
imag = plt.imshow(output.get_fields('ez').T, cmap=plt.cm.hot, origin='lower', vmin=0, vmax=0.05)
plt.colorbar()
# Main loop
from datetime import datetime
t0 = datetime.now()
for tstep in xrange(1, tmax+1):
fdtd.update_h()
nx, ny, nz = 160, 140, 32 tmax, tgap = 150, 10 # instances gpu_devices = common_gpu.gpu_device_list(print_info=False) context = cl.Context(gpu_devices) device = gpu_devices[0] qtask = QueueTask() fields = Fields(context, device, qtask, nx, ny, nz) Core(fields) Pbc(fields, 'xyz') tfunc = lambda tstep: np.sin(0.05 * tstep) IncidentDirect(fields, 'ez', (20, 0, 0), (20, ny-1, nz-1), tfunc) #IncidentDirect(fields, 'ez', (0, 20, 0), (nx-1, 20, nz-1), tfunc) getf = GetFields(fields, 'ez', (0, 0, nz/2), (nx-1, ny-1, nz/2)) print fields.instance_list # plot import matplotlib.pyplot as plt plt.ion() fig = plt.figure(figsize=(12,8)) imag = plt.imshow(np.zeros((nx, ny), fields.dtype).T, interpolation='nearest', origin='lower', vmin=-1.1, vmax=1.1) plt.colorbar() # main loop from datetime import datetime t0 = datetime.now() for tstep in xrange(1, tmax+1):
device = gpu_devices[0]
# z-axis
nx, ny, nz = 180, 160, 2
fields = Fields(context, device, nx, ny, nz)
Pbc(fields, 'xyz')
Core(fields)
IncidentDirect(fields, 'ey', (20, 0, 0), (20, ny - 1, nz - 1), tfunc)
IncidentDirect(fields, 'ex', (0, 20, 0), (nx - 1, 20, nz - 1), tfunc)
for tstep in xrange(1, tmax + 1):
fields.update_e()
fields.update_h()
ax1 = fig.add_subplot(2, 3, 1)
getf = GetFields(fields, 'ey', (0, 0, nz / 2), (nx - 1, ny - 1, nz / 2))
getf.get_event().wait()
ax1.imshow(getf.get_fields().T, vmin=-1.1, vmax=1.1)
ax1.set_title('%s, ey[20,:,:]' % repr(fields.ns))
ax1.set_xlabel('x')
ax1.set_ylabel('y')
ax2 = fig.add_subplot(2, 3, 4)
getf = GetFields(fields, 'ex', (0, 0, nz / 2), (nx - 1, ny - 1, nz / 2))
getf.get_event().wait()
ax2.imshow(getf.get_fields().T, vmin=-1.1, vmax=1.1)
ax2.set_title('%s, ex[:,20,:]' % repr(fields.ns))
ax2.set_xlabel('x')
ax2.set_ylabel('y')
# y-axis
ax2 = fig.add_subplot(1, 3, 2)
ax3 = fig.add_subplot(1, 3, 3)
# gpu device
gpu_devices = common_gpu.gpu_device_list(print_info=False)
context = cl.Context(gpu_devices)
device = gpu_devices[0]
# xy-plane
nx, ny, nz = 180, 160, 2
fields = Fields(context, device, nx, ny, nz)
Pbc(fields, 'z')
Pml(fields, ('+-', '+-', ''), npml)
Core(fields)
IncidentDirect(fields, 'ez', (0.4, 0.3, 0), (0.4, 0.3, -1), tfunc)
getf = GetFields(fields, 'ez', (0, 0, 0.5), (-1, -1, 0.5))
for tstep in xrange(1, tmax+1):
fields.update_e()
fields.update_h()
getf.get_event().wait()
ax1.imshow(getf.get_fields().T, vmin=-1.1, vmax=1.1)
ax1.set_title('xy-plane')
ax1.set_xlabel('x')
ax1.set_ylabel('y')
# yz-plane
nx, ny, nz = 2, 180, 160
fields = Fields(context, device, nx, ny, nz)
Pbc(fields, 'x')
elif rank == size - 1: direction = '-'
else: direction = '+-'
#exch = node.ExchangeMpiNonBlock(fields, direction)
#exch = node.ExchangeMpiBufferBlock(fields, direction)
#exch = node.ExchangeMpiBufferBlockSplit(fields, direction)
exch = node.ExchangeMpiBufferNonBlockSplitEnqueue(fields, direction, tmax)
if '+' in direction: cpu.Core(exch.cpuf_p)
if '-' in direction: cpu.Core(exch.cpuf_m)
is_master = True if rank == 0 else False
if is_plot:
Pbc(fields, 'yz')
getf = GetFields(fields, 'ez', (0, 0, 0.5), (-1, -1, 0.5))
tfunc = lambda tstep: 40 * np.sin(0.05 * tstep)
if rank < size - 1:
IncidentDirect(fields, 'ez', (220, 0.5, 0), (220, 0.5, -1), tfunc)
#if rank > 0:
# IncidentDirect(fields, 'ez', (20, 0.5, 0), (20, 0.5, -1), tfunc)
if is_master:
# plot
import matplotlib.pyplot as plt
plt.ion()
fig = plt.figure(figsize=(12,8))
arr = np.zeros((size * nx, ny), fields.dtype)
for i in range(1, size):
plt.plot((i*nx, i*nx), (0, ny), color='k', linewidth=1.2)
def runTest(self):
axis, nx, ny, nz, precision_float = self.args
gpu_devices = common_gpu.gpu_device_list(print_info=False)
context = cl.Context(gpu_devices)
device = gpu_devices[0]
fields = Fields(context, device, nx, ny, nz, '', precision_float)
pbc = Pbc(fields, axis)
# allocations
ehs = common_update.generate_random_ehs(nx, ny, nz, fields.dtype)
fields.set_eh_bufs(*ehs)
# update
fields.update_e()
fields.update_h()
# verify
getf0, getf1 = {}, {}
strfs_e = {
'x': ['ey', 'ez'],
'y': ['ex', 'ez'],
'z': ['ex', 'ey']
}[axis]
strfs_h = {
'x': ['hy', 'hz'],
'y': ['hx', 'hz'],
'z': ['hx', 'hy']
}[axis]
pt0 = (0, 0, 0)
pt1 = { 'x': (0, ny-2, nz-2), \
'y': (nx-2, 0, nz-2), \
'z': (nx-2, ny-2, 0) }[axis]
getf0['e'] = GetFields(fields, strfs_e, pt0, pt1)
pt0 = { 'x': (nx-1, 0, 0), \
'y': (0, ny-1, 0), \
'z': (0, 0, nz-1) }[axis]
pt1 = { 'x': (nx-1, ny-2, nz-2), \
'y': (nx-2, ny-1, nz-2), \
'z': (nx-2, ny-2, nz-1) }[axis]
getf1['e'] = GetFields(fields, strfs_e, pt0, pt1)
pt0 = { 'x': (0, 1, 1), \
'y': (1, 0, 1), \
'z': (1, 1, 0) }[axis]
pt1 = { 'x': (0, ny-1, nz-1), \
'y': (nx-1, 0, nz-1), \
'z': (nx-1, ny-1, 0) }[axis]
getf0['h'] = GetFields(fields, strfs_h, pt0, pt1)
pt0 = { 'x': (nx-1, 1, 1), \
'y': (1, ny-1, 1), \
'z': (1, 1, nz-1) }[axis]
pt1 = (nx - 1, ny - 1, nz - 1)
getf1['h'] = GetFields(fields, strfs_h, pt0, pt1)
for getf in getf0.values() + getf1.values():
getf.get_event().wait()
for eh in ['e', 'h']:
norm = np.linalg.norm( \
getf0[eh].get_fields() - getf1[eh].get_fields() )
self.assertEqual(norm, 0, '%g, %s, %s' % (norm, self.args, eh))