1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
|
import os
import glob
import sys
import catalyst
import catalyst_conduit
from mpi4py import MPI
def initialize():
# initialize ParaView Catalyst V2.
comm = MPI.COMM_WORLD
rank = comm.Get_rank()
node = catalyst_conduit.Node()
count = 0
for i in sys.argv[1:]:
node['catalyst/scripts/script'+str(count)] = i
count = count + 1
if rank == 0:
print('Using Catalyst script', i)
node['catalyst_load/implementation'] = 'paraview'
# if we know the location of libcatalyst-paraview.so and want to pass that along via
# Python we would do that like:
# node['catalyst_load/search_paths/paraview'] = </full/path/to/libcatalyst-paraview.so>
catalyst.initialize(node)
def finalize():
# finalize ParaView Catalyst V2.
node = catalyst_conduit.Node()
catalyst.finalize(node)
def coprocess(time, timeStep, grid, attributes):
# do the actual in situ analysis and visualization.
node = catalyst_conduit.Node()
node['catalyst/state/timestep'] = timeStep
node['catalyst/state/time'] = time
node['catalyst/state/pipelines/0'] = 'script0'
# the Catalyst channel is "input"
node['catalyst/channels/input/type'] = 'mesh'
mesh = node['catalyst/channels/input/data']
mesh['coordsets/coords/type'] = 'uniform'
mesh['coordsets/coords/dims/i'] = grid.XEndPoint - grid.XStartPoint + 1
mesh['coordsets/coords/dims/j'] = grid.NumberOfYPoints
mesh['coordsets/coords/dims/k'] = grid.NumberOfZPoints
mesh['coordsets/coords/origin/x'] = grid.XStartPoint * grid.Spacing[0]
mesh['coordsets/coords/origin/y'] = 0.0
mesh['coordsets/coords/origin/z'] = 0.0
mesh['coordsets/coords/spacing/dx'] = grid.Spacing[0]
mesh['coordsets/coords/spacing/dy'] = grid.Spacing[1]
mesh['coordsets/coords/spacing/dz'] = grid.Spacing[2]
mesh['topologies/mesh/type'] = 'uniform'
mesh['topologies/mesh/coordset'] = 'coords'
fields = mesh['fields']
# velocity is point-data.
# velocity is stored in interlaced form.
fields['velocitydeepcopy/association'] = 'vertex'
fields['velocitydeepcopy/topology'] = 'mesh'
fields['velocitydeepcopy/volume_dependent'] = 'false'
# flatten creates a deep copy of the array so we need to deep copy
# this data into the Conduit Node.
fields['velocitydeepcopy/values/x'] = attributes.Velocity.flatten()[0::3]
fields['velocitydeepcopy/values/y'] = attributes.Velocity.flatten()[1::3]
fields['velocitydeepcopy/values/z'] = attributes.Velocity.flatten()[2::3]
# pressure is cell-data.
fields['pressure/association'] = 'element'
fields['pressure/topology'] = 'mesh'
fields['pressure/volume_dependent'] = 'false'
# pressure is zero-copied into Catalyst because the simulation
# stores the data the same way that Conduit expects it and it is scalar.
fields['pressure/values'].set_external(attributes.Pressure)
# zero-copied numpy arrays don't seem to work yet
# for non-flat, non-scalar arrays. it should look something like
# the following though.
fields['velocityzerocopy/association'] = 'vertex'
fields['velocityzerocopy/topology'] = 'mesh'
fields['velocityzerocopy/volume_dependent'] = 'false'
# we can also use reshape to get do a zero-copy into Conduit.
fields['velocityzerocopy/values/x'].set_external(attributes.Velocity.reshape(-1)[0::3])
fields['velocityzerocopy/values/y'].set_external(attributes.Velocity.reshape(-1)[1::3])
fields['velocityzerocopy/values/z'].set_external(attributes.Velocity.reshape(-1)[2::3])
# can look at
# https://github.com/LLNL/conduit/blob/50a5dbe8e975c1914187cf4ef3279b61f7753085/src/tests/conduit/python/t_python_conduit_node.py#L671
# for other examples on how to do zero-copy using set_external.
catalyst.execute(node)
return
|
