File: test_custom_assembler.py

package info (click to toggle)
fenics-dolfinx 1%3A0.9.0-10
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid
  • size: 5,376 kB
  • sloc: cpp: 33,701; python: 22,338; makefile: 230; sh: 170; xml: 55
file content (368 lines) | stat: -rw-r--r-- 12,838 bytes parent folder | download | duplicates (2)
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
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
# Copyright (C) 2019-2024 Garth N. Wells
#
# This file is part of DOLFINx (https://www.fenicsproject.org)
#
# SPDX-License-Identifier:    LGPL-3.0-or-later
"""Tests for custom Python assemblers."""

import importlib
import math
import os
import pathlib
import time

from mpi4py import MPI

try:
    from petsc4py import PETSc

    from dolfinx.fem.petsc import assemble_matrix
except ImportError:
    pass

import numpy as np
import pytest

import dolfinx
import dolfinx.pkgconfig
import ufl
from dolfinx.fem import Function, form, functionspace
from dolfinx.mesh import create_unit_square
from dolfinx.utils import cffi_utils as petsc_cffi
from dolfinx.utils import ctypes_utils as petsc_ctypes
from dolfinx.utils import numba_utils as petsc_numba

cffi = pytest.importorskip("cffi")
cffi_support = pytest.importorskip("numba.core.typing.cffi_utils")
numba = pytest.importorskip("numba")

# Get PETSc MatSetValuesLocal interfaces
try:
    MatSetValuesLocal = petsc_numba.MatSetValuesLocal
    MatSetValuesLocal_ctypes = petsc_ctypes.MatSetValuesLocal
    MatSetValuesLocal_abi = petsc_cffi.MatSetValuesLocal
except AttributeError:
    MatSetValuesLocal_abi = None


@numba.njit
def set_vals_numba(A, m, rows, n, cols, data, mode):
    MatSetValuesLocal(A, 3, rows.ctypes, 3, cols.ctypes, data.ctypes, mode)


@numba.njit
def set_vals_cffi(A, m, rows, n, cols, data, mode):
    MatSetValuesLocal_abi(
        A, m, ffi.from_buffer(rows), n, ffi.from_buffer(cols), ffi.from_buffer(data), mode
    )


@numba.njit
def set_vals_ctypes(A, m, rows, n, cols, data, mode):
    MatSetValuesLocal_ctypes(A, m, rows.ctypes, n, cols.ctypes, data.ctypes, mode)


ffi = cffi.FFI()


def get_matsetvalues_cffi_api():
    """Make MatSetValuesLocal from PETSc available via cffi in API mode.

    This function is not (yet) in the DOLFINx module because it is complicated
    by needing to compile code.
    """
    has_petsc_complex = np.issubdtype(PETSc.ScalarType, np.complexfloating)
    dolfinx_pc_name = "dolfinx_complex" if has_petsc_complex else "dolfinx_real"
    if dolfinx.pkgconfig.exists(dolfinx_pc_name):
        dolfinx_pc = dolfinx.pkgconfig.parse(dolfinx_pc_name)
    else:
        raise RuntimeError("Could not find DOLFINx pkg-config file")

    import petsc4py.lib
    from petsc4py import get_config as PETSc_get_config

    cffi_support.register_type(ffi.typeof("float _Complex"), numba.types.complex64)
    cffi_support.register_type(ffi.typeof("double _Complex"), numba.types.complex128)

    petsc_dir = PETSc_get_config()["PETSC_DIR"]
    petsc_arch = petsc4py.lib.getPathArchPETSc()[1]

    worker = os.getenv("PYTEST_XDIST_WORKER", None)
    module_name = f"_petsc_cffi_{worker}"
    if MPI.COMM_WORLD.Get_rank() == 0:
        ffibuilder = cffi.FFI()
        ffibuilder.cdef(
            """typedef int... PetscInt;
                           typedef ... PetscScalar;
                           typedef int... InsertMode;
                           int MatSetValuesLocal(void* mat, PetscInt nrow, const PetscInt* irow,
                                PetscInt ncol, const PetscInt* icol,
                                const PetscScalar* y, InsertMode addv);"""
        )
        ffibuilder.set_source(
            module_name,
            '#include "petscmat.h"',
            libraries=["petsc"],
            include_dirs=[
                os.path.join(petsc_dir, petsc_arch, "include"),
                os.path.join(petsc_dir, "include"),
            ]
            + dolfinx_pc["include_dirs"],
            library_dirs=[os.path.join(petsc_dir, petsc_arch, "lib")],
            extra_compile_args=[],
        )

        # Build module in same directory as test file
        path = pathlib.Path(__file__).parent.absolute()
        ffibuilder.compile(tmpdir=path, verbose=True)

    MPI.COMM_WORLD.Barrier()
    spec = importlib.util.find_spec(module_name)
    if spec is None:
        raise ImportError("Failed to find CFFI generated module")
    module = importlib.util.module_from_spec(spec)
    cffi_support.register_module(module)
    cffi_support.register_type(module.ffi.typeof("PetscScalar"), petsc_numba._scalar)
    return module.lib.MatSetValuesLocal


# See https://github.com/numba/numba/issues/4036 for why we need 'sink'
@numba.njit
def sink(*args):
    pass


@numba.njit(fastmath=True)
def area(x0, x1, x2) -> float:
    """Compute the area of a triangle embedded in 2D from the three vertices"""
    a = (x1[0] - x2[0]) ** 2 + (x1[1] - x2[1]) ** 2
    b = (x0[0] - x2[0]) ** 2 + (x0[1] - x2[1]) ** 2
    c = (x0[0] - x1[0]) ** 2 + (x0[1] - x1[1]) ** 2
    return math.sqrt(2 * (a * b + a * c + b * c) - (a**2 + b**2 + c**2)) / 4.0


@numba.njit(fastmath=True)
def assemble_vector(b, mesh, dofmap, num_cells):
    """Assemble simple linear form over a mesh into the array b"""
    v, x = mesh
    q0, q1 = 1 / 3.0, 1 / 3.0
    for cell in range(num_cells):
        # FIXME: This assumes a particular geometry dof layout
        A = area(x[v[cell, 0]], x[v[cell, 1]], x[v[cell, 2]])
        b[dofmap[cell, 0]] += A * (1.0 - q0 - q1)
        b[dofmap[cell, 1]] += A * q0
        b[dofmap[cell, 2]] += A * q1


@numba.njit(parallel=(not numba.core.config.IS_32BITS), fastmath=True)
def assemble_vector_parallel(b, v, x, dofmap_t_data, dofmap_t_offsets, num_cells):
    """Assemble simple linear form over a mesh into the array b using a parallel loop"""
    q0 = 1 / 3.0
    q1 = 1 / 3.0
    b_unassembled = np.empty((num_cells, 3), dtype=b.dtype)
    for cell in numba.prange(num_cells):
        # FIXME: This assumes a particular geometry dof layout
        A = area(x[v[cell, 0]], x[v[cell, 1]], x[v[cell, 2]])
        b_unassembled[cell, 0] = A * (1.0 - q0 - q1)
        b_unassembled[cell, 1] = A * q0
        b_unassembled[cell, 2] = A * q1

    # Accumulate values in RHS
    _b_unassembled = b_unassembled.reshape(num_cells * 3)
    for index in numba.prange(dofmap_t_offsets.shape[0] - 1):
        for p in range(dofmap_t_offsets[index], dofmap_t_offsets[index + 1]):
            b[index] += _b_unassembled[dofmap_t_data[p]]


@numba.njit(fastmath=True)
def assemble_vector_ufc(b, kernel, mesh, dofmap, num_cells, dtype):
    """Assemble provided FFCx/UFC kernel over a mesh into the array b"""
    v, x = mesh
    entity_local_index = np.array([0], dtype=np.intc)
    perm = np.array([0], dtype=np.uint8)
    geometry = np.zeros((3, 3), dtype=x.dtype)
    coeffs = np.zeros(1, dtype=dtype)
    constants = np.zeros(1, dtype=dtype)

    b_local = np.zeros(3, dtype=dtype)
    for cell in range(num_cells):
        # FIXME: This assumes a particular geometry dof layout
        for j in range(3):
            geometry[j] = x[v[cell, j], :]
        b_local.fill(0.0)
        kernel(
            ffi.from_buffer(b_local),
            ffi.from_buffer(coeffs),
            ffi.from_buffer(constants),
            ffi.from_buffer(geometry),
            ffi.from_buffer(entity_local_index),
            ffi.from_buffer(perm),
        )
        for j in range(3):
            b[dofmap[cell, j]] += b_local[j]


@numba.njit(fastmath=True)
def assemble_petsc_matrix(A, mesh, dofmap, num_cells, set_vals, mode):
    """Assemble P1 mass matrix over a mesh into the PETSc matrix A"""
    # Mesh data
    v, x = mesh

    # Quadrature points and weights
    q = np.array([[0.5, 0.0], [0.5, 0.5], [0.0, 0.5]], dtype=np.double)
    weights = np.full(3, 1.0 / 3.0, dtype=np.double)

    # Loop over cells
    N = np.empty(3, dtype=np.double)
    A_local = np.empty((3, 3), dtype=PETSc.ScalarType)
    for cell in range(num_cells):
        cell_area = area(x[v[cell, 0]], x[v[cell, 1]], x[v[cell, 2]])

        # Loop over quadrature points
        A_local[:] = 0.0
        for j in range(q.shape[0]):
            N[0], N[1], N[2] = 1.0 - q[j, 0] - q[j, 1], q[j, 0], q[j, 1]
            for row in range(3):
                for col in range(3):
                    A_local[row, col] += weights[j] * cell_area * N[row] * N[col]

        # Add to global tensor
        pos = dofmap[cell, :]
        set_vals(A, 3, pos, 3, pos, A_local, mode)
    sink(A_local, dofmap)


@pytest.mark.parametrize(
    "dtype",
    [
        np.float32,
        np.float64,
        pytest.param(np.complex64, marks=pytest.mark.xfail_win32_complex),
        pytest.param(np.complex128, marks=pytest.mark.xfail_win32_complex),
    ],
)
def test_custom_mesh_loop_rank1(dtype):
    mesh = create_unit_square(MPI.COMM_WORLD, 64, 64, dtype=dtype(0).real.dtype)
    V = functionspace(mesh, ("Lagrange", 1))

    # Unpack mesh and dofmap data
    num_owned_cells = mesh.topology.index_map(mesh.topology.dim).size_local
    x_dofs = mesh.geometry.dofmap
    x = mesh.geometry.x
    dofmap = V.dofmap.list

    # Assemble with pure Numba function (two passes, first will include
    # JIT overhead)
    b0 = Function(V, dtype=dtype)
    for i in range(2):
        b = b0.x.array
        b[:] = 0.0
        start = time.time()
        assemble_vector(b, (x_dofs, x), dofmap, num_owned_cells)
        end = time.time()
        print(f"Time (numba, pass {i}): {end - start}")
    b0.x.scatter_reverse(dolfinx.la.InsertMode.add)
    b0sum = np.sum(b0.x.array[: b0.x.index_map.size_local * b0.x.block_size])
    assert mesh.comm.allreduce(b0sum, op=MPI.SUM) == pytest.approx(1.0)

    # NOTE: Parallel (threaded) Numba can cause problems with MPI
    # Assemble with pure Numba function using parallel loop (two passes,
    # first will include JIT overhead)
    # from dolfinx.fem import transpose_dofmap
    # dofmap_t = transpose_dofmap(V.dofmap.list, num_owned_cells)
    # btmp = Function(V)
    # for i in range(2):
    #     b = btmp.x.array
    #     b[:] = 0.0
    #     start = time.time()
    #     assemble_vector_parallel(b, x_dofs, x, dofmap_t.array, dofmap_t.offsets, num_owned_cells)
    #     end = time.time()
    #     print("Time (numba parallel, pass {}): {}".format(i, end - start))
    # btmp.x.petsc_vec.ghostUpdate(addv=PETSc.InsertMode.ADD, mode=PETSc.ScatterMode.REVERSE)
    # assert (btmp.x.petsc_vec - b0.x.petsc_vec).norm() == pytest.approx(0.0)

    # Test against generated code and general assembler
    v = ufl.TestFunction(V)
    L = ufl.inner(1.0, v) * ufl.dx
    Lf = form(L, dtype=dtype)
    start = time.time()
    b1 = dolfinx.fem.assemble_vector(Lf)
    end = time.time()
    print("Time (C++, pass 0):", end - start)

    b1.array[:] = 0
    start = time.time()
    dolfinx.fem.assemble_vector(b1.array, Lf)
    end = time.time()
    print("Time (C++, pass 1):", end - start)
    b1.scatter_reverse(dolfinx.la.InsertMode.add)
    assert np.linalg.norm(b1.array - b0.x.array) == pytest.approx(0.0, abs=1.0e-8)

    # Assemble using generated tabulate_tensor kernel and Numba
    # assembler
    b3 = Function(V, dtype=dtype)
    ufcx_form, module, code = dolfinx.jit.ffcx_jit(
        mesh.comm, L, form_compiler_options={"scalar_type": dtype}
    )

    # Get the one and only kernel
    kernel = getattr(ufcx_form.form_integrals[0], f"tabulate_tensor_{np.dtype(dtype).name}")
    for i in range(2):
        b = b3.x.array
        b[:] = 0.0
        start = time.time()
        assemble_vector_ufc(b, kernel, (x_dofs, x), dofmap, num_owned_cells, dtype)
        end = time.time()
        print(f"Time (numba/cffi, pass {i}): {end - start}")
    b3.x.scatter_reverse(dolfinx.la.InsertMode.add)
    assert np.linalg.norm(b3.x.array - b0.x.array) == pytest.approx(0.0, abs=1e-8)


@pytest.mark.petsc4py
@pytest.mark.parametrize(
    "set_vals,backend",
    [
        (set_vals_numba, "numba"),
        (set_vals_ctypes, "ctypes"),
        (set_vals_cffi, "cffi_abi"),
    ],
)
def test_custom_mesh_loop_petsc_rank2(set_vals, backend):
    """Test numba assembler for a bilinear form."""

    mesh = create_unit_square(MPI.COMM_WORLD, 64, 64)
    V = functionspace(mesh, ("Lagrange", 1))

    # Test against generated code and general assembler
    u, v = ufl.TrialFunction(V), ufl.TestFunction(V)
    a = form(ufl.inner(u, v) * ufl.dx)
    A0 = assemble_matrix(a)
    A0.assemble()

    A0.zeroEntries()
    start = time.time()
    assemble_matrix(A0, a)
    end = time.time()
    print("Time (C++, pass 2):", end - start)
    A0.assemble()

    # Unpack mesh and dofmap data
    num_owned_cells = mesh.topology.index_map(mesh.topology.dim).size_local
    x_dofs = mesh.geometry.dofmap
    x = mesh.geometry.x
    dofmap = V.dofmap.list.astype(np.dtype(PETSc.IntType))

    A1 = A0.copy()
    for i in range(2):
        A1.zeroEntries()
        start = time.time()
        assemble_petsc_matrix(
            A1.handle, (x_dofs, x), dofmap, num_owned_cells, set_vals, PETSc.InsertMode.ADD_VALUES
        )
        end = time.time()
        print(f"Time (Numba/{backend}, pass {i}): {end - start}")
        A1.assemble()
    assert (A1 - A0).norm() == pytest.approx(0.0, abs=1.0e-9)

    A0.destroy()
    A1.destroy()