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
|
/*****************************************************************************
*
* Copyright (c) 2003-2020 by The University of Queensland
* http://www.uq.edu.au
*
* Primary Business: Queensland, Australia
* Licensed under the Apache License, version 2.0
* http://www.apache.org/licenses/LICENSE-2.0
*
* Development until 2012 by Earth Systems Science Computational Center (ESSCC)
* Development 2012-2013 by School of Earth Sciences
* Development from 2014-2017 by Centre for Geoscience Computing (GeoComp)
* Development from 2019 by School of Earth and Environmental Sciences
**
*****************************************************************************/
#include "Assemble.h"
#ifdef ESYS_HAVE_PASO
#include <paso/SystemMatrix.h>
#endif
#ifdef ESYS_HAVE_TRILINOS
#include <trilinoswrap/TrilinosMatrixAdapter.h>
using esys_trilinos::TrilinosMatrixAdapter;
#endif
namespace finley {
using escript::DataTypes::real_t;
using escript::DataTypes::cplx_t;
#ifdef ESYS_HAVE_PASO
static void addToSystemMatrixPasoCSC(paso::SystemMatrix<double>* S, int NN_Equa,
const index_t* Nodes_Equa, int num_Equa,
int NN_Sol, const index_t* Nodes_Sol,
int num_Sol, const real_t* array);
template <typename T>
static void addToSystemMatrixPasoCSR(paso::SystemMatrix<T>* S, int NN_Equa,
const index_t* Nodes_Equa, int num_Equa,
int NN_Sol, const index_t* Nodes_Sol,
int num_Sol, const T* array);
#endif
template<>
void Assemble_addToSystemMatrix<real_t>(escript::ASM_ptr S, int NN_Equa,
const index_t* Nodes_Equa, int num_Equa,
int NN_Sol, const index_t* Nodes_Sol,
int num_Sol, const real_t* array)
{
#ifdef ESYS_HAVE_PASO
paso::SystemMatrix<real_t>* pmat = dynamic_cast<paso::SystemMatrix<real_t>*>(S.get());
if (pmat) {
// call the right function depending on storage type
if (pmat->type & MATRIX_FORMAT_CSC) {
addToSystemMatrixPasoCSC(pmat, NN_Equa, Nodes_Equa,
num_Equa, NN_Sol, Nodes_Sol,
num_Sol, array);
} else { // type == CSR
addToSystemMatrixPasoCSR(pmat, NN_Equa, Nodes_Equa,
num_Equa, NN_Sol, Nodes_Sol,
num_Sol, array);
}
return;
}
#endif
#ifdef ESYS_HAVE_TRILINOS
TrilinosMatrixAdapter* tmat(dynamic_cast<TrilinosMatrixAdapter*>(S.get()));
if (tmat) {
IndexVector rowIdx(Nodes_Equa, Nodes_Equa+NN_Equa);
//IndexVector colIdx(Nodes_Sol, Nodes_Sol+NN_Sol);
std::vector<real_t> arr(array, array+(NN_Equa*NN_Sol*num_Sol*num_Equa));
tmat->add(rowIdx, arr);
return;
}
#endif
throw FinleyException("Assemble_addToSystemMatrix: unknown system "
"matrix type.");
}
template<>
void Assemble_addToSystemMatrix<cplx_t>(escript::ASM_ptr S, int NN_Equa,
const index_t* Nodes_Equa, int num_Equa,
int NN_Sol, const index_t* Nodes_Sol,
int num_Sol, const cplx_t* array)
{
#ifdef ESYS_HAVE_MUMPS
paso::SystemMatrix<cplx_t>* pmat = dynamic_cast<paso::SystemMatrix<cplx_t>*>(S.get());
if (pmat) {
if (pmat->type & (MATRIX_FORMAT_OFFSET1 + MATRIX_FORMAT_BLK1)) {
addToSystemMatrixPasoCSR(pmat, NN_Equa, Nodes_Equa,
num_Equa, NN_Sol, Nodes_Sol,
num_Sol, array);
} else {
throw FinleyException("addToSystemMatrix: MUMPS requires CSR format with "
"index offset 1 and block size 1.");
}
return;
}
#endif
#ifdef ESYS_HAVE_TRILINOS
TrilinosMatrixAdapter* tmat = dynamic_cast<TrilinosMatrixAdapter*>(S.get());
if (tmat) {
IndexVector rowIdx(Nodes_Equa, Nodes_Equa+NN_Equa);
//IndexVector colIdx(Nodes_Sol, Nodes_Sol+NN_Sol);
std::vector<cplx_t> arr(array, array+(NN_Equa*NN_Sol*num_Sol*num_Equa));
tmat->add(rowIdx, arr);
return;
}
#endif
throw FinleyException("addToSystemMatrix: only Trilinos matrices support "
"complex-valued assembly!");
}
#ifdef ESYS_HAVE_PASO
void addToSystemMatrixPasoCSC(paso::SystemMatrix<double>* in, int NN_Equa,
const index_t* Nodes_Equa, int num_Equa,
int NN_Sol, const index_t* Nodes_Sol,
int num_Sol, const real_t* array)
{
const int index_offset = (in->type & MATRIX_FORMAT_OFFSET1 ? 1 : 0);
const int row_block_size = in->row_block_size;
const int col_block_size = in->col_block_size;
const int block_size = in->block_size;
const int num_subblocks_Equa = num_Equa/row_block_size;
const int num_subblocks_Sol = num_Sol/col_block_size;
const dim_t numMyCols = in->pattern->mainPattern->numInput;
const dim_t numMyRows = in->pattern->mainPattern->numOutput;
const index_t* mainBlock_ptr = in->mainBlock->pattern->ptr;
const index_t* mainBlock_index = in->mainBlock->pattern->index;
real_t* mainBlock_val = in->mainBlock->val;
const index_t* col_coupleBlock_ptr = in->col_coupleBlock->pattern->ptr;
const index_t* col_coupleBlock_index = in->col_coupleBlock->pattern->index;
real_t* col_coupleBlock_val = in->col_coupleBlock->val;
//const index_t* row_coupleBlock_ptr = in->row_coupleBlock->pattern->ptr;
const index_t* row_coupleBlock_index = in->row_coupleBlock->pattern->index;
real_t* row_coupleBlock_val = in->row_coupleBlock->val;
for (int k_Sol = 0; k_Sol < NN_Sol; ++k_Sol) {
// Down columns of array
const index_t j_Sol = Nodes_Sol[k_Sol];
for (int l_col = 0; l_col < num_subblocks_Sol; ++l_col) {
const index_t i_col = j_Sol * num_subblocks_Sol + l_col;
if (i_col < numMyCols) {
for (int k_Equa = 0; k_Equa < NN_Equa; ++k_Equa) {
// Across cols of array
const index_t j_Equa = Nodes_Equa[k_Equa];
for (int l_row = 0; l_row < num_subblocks_Equa; ++l_row) {
const index_t i_row = j_Equa*num_subblocks_Equa+index_offset+l_row;
if (i_row < numMyRows + index_offset ) {
for (index_t k = mainBlock_ptr[i_col]-index_offset;
k < mainBlock_ptr[i_col + 1]-index_offset; ++k) {
if (mainBlock_index[k] == i_row) {
// Entry array(k_Equa, j_Sol) is a block
// (col_block_size x col_block_size)
for (int ic = 0; ic < col_block_size; ++ic) {
const int i_Sol = ic + col_block_size * l_col;
for (int ir = 0; ir < row_block_size; ++ir) {
const int i_Eq = ir + row_block_size * l_row;
mainBlock_val[k*block_size + ir + row_block_size*ic] +=
array[INDEX4
(i_Eq, i_Sol, k_Equa, k_Sol, num_Equa, num_Sol, NN_Equa)];
}
}
break;
}
}
} else {
for (index_t k = col_coupleBlock_ptr[i_col]-index_offset;
k < col_coupleBlock_ptr[i_col + 1]-index_offset; ++k) {
if (row_coupleBlock_index[k] == i_row - numMyRows) {
for (int ic = 0; ic < col_block_size; ++ic) {
const int i_Sol = ic + col_block_size * l_col;
for (int ir = 0; ir < row_block_size; ++ir) {
const int i_Eq = ir + row_block_size * l_row;
row_coupleBlock_val[k*block_size + ir + row_block_size*ic] +=
array[INDEX4
(i_Eq, i_Sol, k_Equa, k_Sol, num_Equa, num_Sol, NN_Equa)];
}
}
break;
}
}
}
}
}
} else { // i_col >= numMyCols
for (int k_Equa = 0; k_Equa < NN_Equa; ++k_Equa) {
// Across rows of array
const index_t j_Equa = Nodes_Equa[k_Equa];
for (int l_row = 0; l_row < num_subblocks_Equa; ++l_row) {
const index_t i_row = j_Equa * num_subblocks_Equa + index_offset + l_row;
if (i_row < numMyRows + index_offset) {
for (index_t k = col_coupleBlock_ptr[i_col-numMyCols]-index_offset;
k < col_coupleBlock_ptr[i_col - numMyCols + 1] - index_offset; ++k) {
if (col_coupleBlock_index[k] == i_row) {
for (int ic = 0; ic < col_block_size; ++ic) {
const int i_Sol = ic + col_block_size * l_col;
for (int ir = 0; ir < row_block_size; ++ir) {
const int i_Eq = ir + row_block_size * l_row;
col_coupleBlock_val[k*block_size + ir + row_block_size*ic] +=
array[INDEX4
(i_Eq, i_Sol, k_Equa, k_Sol, num_Equa, num_Sol, NN_Equa)];
}
}
break;
}
}
}
}
}
}
}
}
}
template <typename T>
void addToSystemMatrixPasoCSR(paso::SystemMatrix<T>* in, int NN_Equa,
const index_t* Nodes_Equa, int num_Equa,
int NN_Sol, const index_t* Nodes_Sol,
int num_Sol, const T* array)
{
const int index_offset = (in->type & MATRIX_FORMAT_OFFSET1 ? 1 : 0);
const int row_block_size = in->row_block_size;
const int col_block_size = in->col_block_size;
const int block_size = in->block_size;
const int num_subblocks_Equa = num_Equa / row_block_size;
const int num_subblocks_Sol = num_Sol / col_block_size;
const dim_t numMyCols = in->pattern->mainPattern->numInput;
const dim_t numMyRows = in->pattern->mainPattern->numOutput;
const index_t* mainBlock_ptr = in->mainBlock->pattern->ptr;
const index_t* mainBlock_index = in->mainBlock->pattern->index;
T* mainBlock_val = in->mainBlock->val;
const index_t* col_coupleBlock_ptr = in->col_coupleBlock->pattern->ptr;
const index_t* col_coupleBlock_index = in->col_coupleBlock->pattern->index;
T* col_coupleBlock_val = in->col_coupleBlock->val;
const index_t* row_coupleBlock_ptr = in->row_coupleBlock->pattern->ptr;
const index_t* row_coupleBlock_index = in->row_coupleBlock->pattern->index;
T* row_coupleBlock_val = in->row_coupleBlock->val;
for (int k_Equa = 0; k_Equa < NN_Equa; ++k_Equa) {
// Down columns of array
const index_t j_Equa = Nodes_Equa[k_Equa];
for (int l_row = 0; l_row<num_subblocks_Equa; ++l_row) {
const index_t i_row = j_Equa*num_subblocks_Equa+l_row;
// only look at the matrix rows stored on this processor
if (i_row < numMyRows) {
for (int k_Sol=0; k_Sol<NN_Sol; ++k_Sol) {
// Across rows of array
const index_t j_Sol=Nodes_Sol[k_Sol];
for (int l_col=0; l_col<num_subblocks_Sol; ++l_col) {
// only look at the matrix rows stored on this processor
const index_t i_col = j_Sol * num_subblocks_Sol + index_offset + l_col;
if (i_col < numMyCols + index_offset) {
for (index_t k = mainBlock_ptr[i_row] - index_offset;
k < mainBlock_ptr[i_row + 1] - index_offset; ++k) {
if (mainBlock_index[k] == i_col) {
// Entry array(k_Sol, j_Equa) is a block
// (row_block_size x col_block_size)
for (int ic = 0; ic < col_block_size; ++ic) {
const int i_Sol = ic + col_block_size * l_col;
for (int ir = 0; ir < row_block_size; ++ir) {
const int i_Eq = ir + row_block_size * l_row;
mainBlock_val[k*block_size + ir + row_block_size*ic]+=
array[INDEX4
(i_Eq, i_Sol, k_Equa, k_Sol, num_Equa, num_Sol, NN_Equa)];
}
}
break;
}
}
} else {
for (index_t k = col_coupleBlock_ptr[i_row] - index_offset;
k < col_coupleBlock_ptr[i_row + 1] - index_offset; ++k) {
if (col_coupleBlock_index[k] == i_col - numMyCols) {
// Entry array(k_Sol, j_Equa) is a block
// (row_block_size x col_block_size)
for (int ic = 0; ic < col_block_size; ++ic) {
const int i_Sol = ic + col_block_size * l_col;
for (int ir = 0; ir < row_block_size; ++ir) {
const int i_Eq = ir+row_block_size*l_row;
col_coupleBlock_val[k*block_size + ir + row_block_size*ic]+=
array[INDEX4
(i_Eq, i_Sol, k_Equa, k_Sol, num_Equa, num_Sol, NN_Equa)];
}
}
break;
}
}
}
}
}
} else { // i_row >= numMyRows
for (int k_Sol = 0; k_Sol < NN_Sol; ++k_Sol) {
// Across rows of array
const index_t j_Sol = Nodes_Sol[k_Sol];
for (int l_col = 0; l_col < num_subblocks_Sol; ++l_col) {
const index_t i_col = j_Sol * num_subblocks_Sol + index_offset + l_col;
if (i_col < numMyCols + index_offset) {
for (index_t k = row_coupleBlock_ptr[i_row - numMyRows] - index_offset;
k < row_coupleBlock_ptr[i_row - numMyRows + 1] - index_offset; ++k) {
if (row_coupleBlock_index[k] == i_col) {
// Entry array(k_Sol, j_Equa) is a block
// (row_block_size x col_block_size)
for (int ic = 0; ic < col_block_size; ++ic) {
const int i_Sol = ic + col_block_size * l_col;
for (int ir = 0; ir < row_block_size; ++ir) {
const int i_Eq = ir + row_block_size * l_row;
row_coupleBlock_val[k*block_size + ir + row_block_size*ic]+=
array[INDEX4
(i_Eq, i_Sol, k_Equa, k_Sol, num_Equa, num_Sol, NN_Equa)];
}
}
break;
}
}
}
}
}
}
}
}
}
#endif // ESYS_HAVE_PASO
} // namespace finley
|
