projects / LbmBenchmarkKernelsPublic.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
bulk commit
[LbmBenchmarkKernelsPublic.git] / src / BenchKernelD3Q19Aa.c
CommitLineData
e3f82424
MW		 1// --------------------------------------------------------------------------
2//
3// Copyright
4// Markus Wittmann, 2016-2017
5// RRZE, University of Erlangen-Nuremberg, Germany
6// markus.wittmann -at- fau.de or hpc -at- rrze.fau.de
7//
8// Viktor Haag, 2016
9// LSS, University of Erlangen-Nuremberg, Germany
10//
11// This file is part of the Lattice Boltzmann Benchmark Kernels (LbmBenchKernels).
12//
13// LbmBenchKernels is free software: you can redistribute it and/or modify
14// it under the terms of the GNU General Public License as published by
15// the Free Software Foundation, either version 3 of the License, or
16// (at your option) any later version.
17//
18// LbmBenchKernels is distributed in the hope that it will be useful,
19// but WITHOUT ANY WARRANTY; without even the implied warranty of
20// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
21// GNU General Public License for more details.
22//
23// You should have received a copy of the GNU General Public License
24// along with LbmBenchKernels. If not, see <http://www.gnu.org/licenses/>.
25//
26// --------------------------------------------------------------------------
27#include "BenchKernelD3Q19AaCommon.h"
28
29#include "Memory.h"
30#include "Vtk.h"
31#include "LikwidIf.h"
32
33#include <inttypes.h>
34#include <math.h>
35
36#ifdef _OPENMP
37 #include <omp.h>
38#endif
39
40void FNAME(D3Q19AaKernel)(LatticeDesc * ld, KernelData * kernelData, CaseData * cd)
41{
42 Assert(ld != NULL);
43 Assert(kernelData != NULL);
44 Assert(cd != NULL);
45
46 Assert(cd->Omega > 0.0);
47 Assert(cd->Omega < 2.0);
48
49 KernelData * kd = (KernelData *)kernelData;
50
51
52 int nX = ld->Dims[0];
53 int nY = ld->Dims[1];
54 int nZ = ld->Dims[2];
55
56 int * gDims = kd->GlobalDims;
57
58 int oX = kd->Offsets[0];
59 int oY = kd->Offsets[1];
60 int oZ = kd->Offsets[2];
61
62 KernelDataAa * kda = KDA(kd);
63
64 int blk[3];
65 blk[0] = kda->Blk[0];
66 blk[1] = kda->Blk[1];
67 blk[2] = kda->Blk[2];
68
69 PdfT omega = cd->Omega;
70 PdfT omegaEven = omega;
71 PdfT magicParam = 1.0 / 12.0;
72 // 1/4: best stability;
73 // 1/12: removes third-order advection error (best advection);
74 // 1/6: removes fourth-order diffusion error (best diffusion);
75 // 3/16: exact location of bounce back for poiseuille flow
76
77 PdfT omegaOdd = 1.0/( 0.5 + magicParam/(1.0/omega - 0.5) );
78
79 PdfT evenPart = 0.0;
80 PdfT oddPart = 0.0;
81 PdfT dir_indep_trm = 0.0;
82
83 PdfT w_0 = 1.0 / 3.0;
84 PdfT w_1 = 1.0 / 18.0;
85 PdfT w_2 = 1.0 / 36.0;
86
87 PdfT w_1_x3 = w_1 * 3.0; PdfT w_1_nine_half = w_1 * 9.0/2.0; PdfT w_1_indep = 0.0;
88 PdfT w_2_x3 = w_2 * 3.0; PdfT w_2_nine_half = w_2 * 9.0/2.0; PdfT w_2_indep = 0.0;
89
90 PdfT ux, uy, uz, ui;
91 PdfT dens;
92
93 // Declare pdf_N, pdf_E, pdf_S, pdf_W, ...
94 #define X(name, idx, idxinv, x, y, z) PdfT JOIN(pdf_,name);
95 D3Q19_LIST
96 #undef X
97
98 PdfT * src = kd->Pdfs[0];
99
100 int maxIterations = cd->MaxIterations;
101
102 #ifdef VTK_OUTPUT
103 if (cd->VtkOutput) {
104 kd->PdfsActive = src;
105 VtkWrite(ld, kd, cd, -1);
106 }
107 #endif
108
109 #ifdef STATISTICS
110 kd->PdfsActive = src;
111 KernelStatistics(kd, ld, cd, 0);
112 #endif
113
114
115 int nThreads = 1;
116
117 #ifdef _OPENMP
118 nThreads = omp_get_max_threads();
119 #endif
120
121 for (int iter = 0; iter < maxIterations; iter += 2) {
122
123 // --------------------------------------------------------------------
124 // even time step
125
126 X_LIKWID_START("aa-even");
127
128 // {{{
129
130 #ifdef _OPENMP
131 #pragma omp parallel for default(none) \
132 shared(gDims,src, w_0, w_1, w_2, omegaEven, omegaOdd, \
133 w_1_x3, w_2_x3, w_1_nine_half, w_2_nine_half, cd, \
134 oX, oY, oZ, nX, nY, nZ, blk, nThreads, ld) \
135 private(ux, uy, uz, ui, dens, dir_indep_trm, \
136 pdf_C, \
137 pdf_N, pdf_E, pdf_S, pdf_W, \
138 pdf_NE, pdf_SE, pdf_SW, pdf_NW, \
139 pdf_T, pdf_TN, pdf_TE, pdf_TS, pdf_TW, \
140 pdf_B, pdf_BN, pdf_BE, pdf_BS, pdf_BW, \
141 evenPart, oddPart, w_1_indep, w_2_indep)
142 #endif
143
144 for (int i = 0; i < nThreads; ++i) {
145
146 int threadStartX = nX / nThreads * i;
147 int threadEndX = nX / nThreads * (i + 1);
148
149 if (nX % nThreads > 0) {
150 if (nX % nThreads > i) {
151 threadStartX += i;
152 threadEndX += i + 1;
153 }
154 else {
155 threadStartX += nX % nThreads;
156 threadEndX += nX % nThreads;
157 }
158 }
159
160 for (int bX = oX + threadStartX; bX < threadEndX + oX; bX += blk[0]) {
161 for (int bY = oY; bY < nY + oY; bY += blk[1]) {
162 for (int bZ = oZ; bZ < nZ + oZ; bZ += blk[2]) {
163
164 int eX = MIN(bX + blk[0], threadEndX + oX);
165 int eY = MIN(bY + blk[1], nY + oY);
166 int eZ = MIN(bZ + blk[2], nZ + oZ);
167
168 // printf("%d: %d-%d %d-%d %d-%d %d - %d\n", omp_get_thread_num(), bZ, eZ, bY, eY, bX, eX, threadStartX, threadEndX);
169
170 for (int x = bX; x < eX; ++x) {
171 for (int y = bY; y < eY; ++y) {
172 for (int z = bZ; z < eZ; ++z) {
173
174
175 if (ld->Lattice[L_INDEX_4(ld->Dims, x - oX, y - oY, z - oZ)] == LAT_CELL_OBSTACLE) {
176 continue;
177 }
178
179 #define I(x, y, z, dir) P_INDEX_5(gDims, (x), (y), (z), (dir))
180
181
182 // Load PDFs of local cell: pdf_N = src[I(x, y, z, D3Q19_N)]; ...
183 #define X(name, idx, idxinv, _x, _y, _z) JOIN(pdf_,name) = src[I(x, y, z, idx)];
184 D3Q19_LIST
185 #undef X
186
187// #define LID_DRIVEN_CAVITY
188
189#ifdef LID_DRIVEN_CAVITY
190
191 if (z == nZ - 4 + oZ && x > 3 + oX && x < (nX - 4 + oX) && y > 3 + oY && y < (nY - 4 + oY)) {
192 ux = 0.1 * 0.577;
193 uy = 0.0;
194 uz = 0.0;
195
196 } else {
197 #endif
198 ux = pdf_E + pdf_NE + pdf_SE + pdf_TE + pdf_BE -
199 pdf_W - pdf_NW - pdf_SW - pdf_TW - pdf_BW;
200 uy = pdf_N + pdf_NE + pdf_NW + pdf_TN + pdf_BN -
201 pdf_S - pdf_SE - pdf_SW - pdf_TS - pdf_BS;
202 uz = pdf_T + pdf_TE + pdf_TW + pdf_TN + pdf_TS -
203 pdf_B - pdf_BE - pdf_BW - pdf_BN - pdf_BS;
204 #ifdef LID_DRIVEN_CAVITY
205 }
206 #endif
207
208 dens = pdf_C +
209 pdf_N + pdf_E + pdf_S + pdf_W +
210 pdf_NE + pdf_SE + pdf_SW + pdf_NW +
211 pdf_T + pdf_TN + pdf_TE + pdf_TS + pdf_TW +
212 pdf_B + pdf_BN + pdf_BE + pdf_BS + pdf_BW;
213
214 dir_indep_trm = dens - (ux * ux + uy * uy + uz * uz)*3.0/2.0;
215
216 // direction: w_0
217 src[I(x, y, z, D3Q19_C)] = pdf_C - omegaEven*(pdf_C - w_0*dir_indep_trm);
218
219 // direction: w_1
220 w_1_indep = w_1*dir_indep_trm;
221
222 ui = uy;
223 evenPart = omegaEven*( 0.5*(pdf_N + pdf_S) - ui*ui*w_1_nine_half - w_1_indep );
224 oddPart = omegaOdd*(0.5*(pdf_N - pdf_S) - ui*w_1_x3 );
225 src[I(x, y, z, D3Q19_S)] = pdf_N - evenPart - oddPart;
226 src[I(x, y, z, D3Q19_N)] = pdf_S - evenPart + oddPart;
227
228 ui = ux;
229 evenPart = omegaEven*( 0.5*(pdf_E + pdf_W) - ui*ui*w_1_nine_half - w_1_indep );
230 oddPart = omegaOdd*(0.5*(pdf_E - pdf_W) - ui*w_1_x3 );
231 src[I(x, y, z, D3Q19_W)] = pdf_E - evenPart - oddPart;
232 src[I(x, y, z, D3Q19_E)] = pdf_W - evenPart + oddPart;
233
234 ui = uz;
235 evenPart = omegaEven*( 0.5*(pdf_T + pdf_B) - ui*ui*w_1_nine_half - w_1_indep );
236 oddPart = omegaOdd*(0.5*(pdf_T - pdf_B) - ui*w_1_x3 );
237 src[I(x, y, z, D3Q19_B)] = pdf_T - evenPart - oddPart;
238 src[I(x, y, z, D3Q19_T)] = pdf_B - evenPart + oddPart;
239
240 // direction: w_2
241 w_2_indep = w_2*dir_indep_trm;
242
243 ui = -ux + uy;
244 evenPart = omegaEven*( 0.5*(pdf_NW + pdf_SE) - ui*ui*w_2_nine_half - w_2_indep );
245 oddPart = omegaOdd*(0.5*(pdf_NW - pdf_SE) - ui*w_2_x3 );
246 src[I(x, y, z, D3Q19_SE)] = pdf_NW - evenPart - oddPart;
247 src[I(x, y, z, D3Q19_NW)] = pdf_SE - evenPart + oddPart;
248
249 ui = ux + uy;
250 evenPart = omegaEven*( 0.5*(pdf_NE + pdf_SW) - ui*ui*w_2_nine_half - w_2_indep );
251 oddPart = omegaOdd*(0.5*(pdf_NE - pdf_SW) - ui*w_2_x3 );
252 src[I(x, y, z, D3Q19_SW)] = pdf_NE - evenPart - oddPart;
253 src[I(x, y, z, D3Q19_NE)] = pdf_SW - evenPart + oddPart;
254
255 ui = -ux + uz;
256 evenPart = omegaEven*( 0.5*(pdf_TW + pdf_BE) - ui*ui*w_2_nine_half - w_2_indep );
257 oddPart = omegaOdd*(0.5*(pdf_TW - pdf_BE) - ui*w_2_x3 );
258 src[I(x, y, z, D3Q19_BE)] = pdf_TW - evenPart - oddPart;
259 src[I(x, y, z, D3Q19_TW)] = pdf_BE - evenPart + oddPart;
260
261 ui = ux + uz;
262 evenPart = omegaEven*( 0.5*(pdf_TE + pdf_BW) - ui*ui*w_2_nine_half - w_2_indep );
263 oddPart = omegaOdd*(0.5*(pdf_TE - pdf_BW) - ui*w_2_x3 );
264 src[I(x, y, z, D3Q19_BW)] = pdf_TE - evenPart - oddPart;
265 src[I(x, y, z, D3Q19_TE)] = pdf_BW - evenPart + oddPart;
266
267 ui = -uy + uz;
268 evenPart = omegaEven*( 0.5*(pdf_TS + pdf_BN) - ui*ui*w_2_nine_half - w_2_indep );
269 oddPart = omegaOdd*(0.5*(pdf_TS - pdf_BN) - ui*w_2_x3 );
270 src[I(x, y, z, D3Q19_BN)] = pdf_TS - evenPart - oddPart;
271 src[I(x, y, z, D3Q19_TS)] = pdf_BN - evenPart + oddPart;
272
273 ui = uy + uz;
274 evenPart = omegaEven*( 0.5*(pdf_TN + pdf_BS) - ui*ui*w_2_nine_half - w_2_indep );
275 oddPart = omegaOdd*(0.5*(pdf_TN - pdf_BS) - ui*w_2_x3 );
276 src[I(x, y, z, D3Q19_BS)] = pdf_TN - evenPart - oddPart;
277 src[I(x, y, z, D3Q19_TN)] = pdf_BS - evenPart + oddPart;
278
279 #undef I
280 } } } // z, y, x (from inner to outer)
281 } } } // z, y, x (from inner to outer)
282
283 } // loop over threads
284
285 // }}}
286
287 X_LIKWID_STOP("aa-even");
288
289 #ifdef STATISTICS
290 kd->PdfsActive = src;
291 KernelStatistics(kd, ld, cd, iter);
292 #endif
293
294 // Fixup bounce back PDFs.
295 #ifdef _OPENMP
296 #pragma omp parallel for default(none) \
297 shared(kd, src)
298 #endif
299 for (int i = 0; i < kd->nBounceBackPdfs; ++i) {
300 src[kd->BounceBackPdfsSrc[i]] = src[kd->BounceBackPdfsDst[i]];
301 }
302
303 // save current iteration
304 kda->Iteration = iter;
305
306 #ifdef VERIFICATION
307 kd->PdfsActive = src;
308 KernelAddBodyForce(kd, ld, cd);
309 #endif
310
311 #ifdef VTK_OUTPUT
312 if (cd->VtkOutput && (iter % cd->VtkModulus) == 0) {
313 kd->PdfsActive = src;
314 VtkWrite(ld, kd, cd, iter);
315 }
316 #endif
317
318 #ifdef STATISTICS
319 kd->PdfsActive = src;
320 KernelStatistics(kd, ld, cd, iter);
321 #endif
322
323 // --------------------------------------------------------------------
324 // odd time step
325
326
327 X_LIKWID_START("aa-odd");
328
329 // {{{
330
331 #ifdef _OPENMP
332 #pragma omp parallel for default(none) \
333 shared(gDims,src, w_0, w_1, w_2, omegaEven, omegaOdd, \
334 w_1_x3, w_2_x3, w_1_nine_half, w_2_nine_half, cd, \
335 oX, oY, oZ, nX, nY, nZ, blk, nThreads) \
336 private(ux, uy, uz, ui, dens, dir_indep_trm, \
337 pdf_C, \
338 pdf_N, pdf_E, pdf_S, pdf_W, \
339 pdf_NE, pdf_SE, pdf_SW, pdf_NW, \
340 pdf_T, pdf_TN, pdf_TE, pdf_TS, pdf_TW, \
341 pdf_B, pdf_BN, pdf_BE, pdf_BS, pdf_BW, \
342 evenPart, oddPart, w_1_indep, w_2_indep)
343 #endif
344
345 for (int i = 0; i < nThreads; ++i) {
346
347 int threadStartX = nX / nThreads * i;
348 int threadEndX = nX / nThreads * (i + 1);
349
350 if (nX % nThreads > 0) {
351 if (nX % nThreads > i) {
352 threadStartX += i;
353 threadEndX += i + 1;
354 }
355 else {
356 threadStartX += nX % nThreads;
357 threadEndX += nX % nThreads;
358 }
359 }
360
361 for (int bX = oX + threadStartX; bX < threadEndX + oX; bX += blk[0]) {
362 for (int bY = oY; bY < nY + oY; bY += blk[1]) {
363 for (int bZ = oZ; bZ < nZ + oZ; bZ += blk[2]) {
364
365 // Must do everything here, else it would break collapse.
366 int eZ = MIN(bZ + blk[2], nZ + oZ);
367 int eY = MIN(bY + blk[1], nY + oY);
368 int eX = MIN(bX + blk[0], threadEndX + oX);
369
370 for (int x = bX; x < eX; ++x) {
371 for (int y = bY; y < eY; ++y) {
372 for (int z = bZ; z < eZ; ++z) {
373
374 #define I(x, y, z, dir) P_INDEX_5(gDims, (x), (y), (z), (dir))
375
376 // Load PDFs of local cell: pdf_N = src[I(x, y, z, D3Q19_N)]; ...
377 #define X(name, idx, idxinv, _x, _y, _z) JOIN(pdf_,name) = src[I(x - _x, y - _y, z - _z, idxinv)];
378 D3Q19_LIST
379 #undef X
380
381
382// #define LID_DRIVEN_CAVITY
383
384#ifdef LID_DRIVEN_CAVITY
385
386 if (z == nZ - 4 + oZ && x > 3 + oX && x < (nX - 4 + oX) && y > 3 + oY && y < (nY - 4 + oY)) {
387 ux = 0.1 * 0.577;
388 uy = 0.0;
389 uz = 0.0;
390
391 } else {
392#endif
393 ux = pdf_E + pdf_NE + pdf_SE + pdf_TE + pdf_BE -
394 pdf_W - pdf_NW - pdf_SW - pdf_TW - pdf_BW;
395 uy = pdf_N + pdf_NE + pdf_NW + pdf_TN + pdf_BN -
396 pdf_S - pdf_SE - pdf_SW - pdf_TS - pdf_BS;
397 uz = pdf_T + pdf_TE + pdf_TW + pdf_TN + pdf_TS -
398 pdf_B - pdf_BE - pdf_BW - pdf_BN - pdf_BS;
399#ifdef LID_DRIVEN_CAVITY
400 }
401#endif
402
403 dens = pdf_C +
404 pdf_N + pdf_E + pdf_S + pdf_W +
405 pdf_NE + pdf_SE + pdf_SW + pdf_NW +
406 pdf_T + pdf_TN + pdf_TE + pdf_TS + pdf_TW +
407 pdf_B + pdf_BN + pdf_BE + pdf_BS + pdf_BW;
408
409 dir_indep_trm = dens - (ux * ux + uy * uy + uz * uz)*3.0/2.0;
410
411 // direction: w_0
412 src[I(x, y, z, D3Q19_C)] = pdf_C - omegaEven*(pdf_C - w_0*dir_indep_trm);
413
414 // direction: w_1
415 w_1_indep = w_1*dir_indep_trm;
416
417 ui = uy;
418 evenPart = omegaEven*( 0.5*(pdf_N + pdf_S) - ui*ui*w_1_nine_half - w_1_indep );
419 oddPart = omegaOdd*(0.5*(pdf_N - pdf_S) - ui*w_1_x3 );
420 src[I(x, y + 1, z, D3Q19_N)] = pdf_N - evenPart - oddPart;
421 src[I(x, y - 1, z, D3Q19_S)] = pdf_S - evenPart + oddPart;
422
423 ui = ux;
424 evenPart = omegaEven*( 0.5*(pdf_E + pdf_W) - ui*ui*w_1_nine_half - w_1_indep );
425 oddPart = omegaOdd*(0.5*(pdf_E - pdf_W) - ui*w_1_x3 );
426 src[I(x + 1, y, z, D3Q19_E)] = pdf_E - evenPart - oddPart;
427 src[I(x - 1, y, z, D3Q19_W)] = pdf_W - evenPart + oddPart;
428
429 ui = uz;
430 evenPart = omegaEven*( 0.5*(pdf_T + pdf_B) - ui*ui*w_1_nine_half - w_1_indep );
431 oddPart = omegaOdd*(0.5*(pdf_T - pdf_B) - ui*w_1_x3 );
432 src[I(x, y, z + 1, D3Q19_T)] = pdf_T - evenPart - oddPart;
433 src[I(x, y, z - 1, D3Q19_B)] = pdf_B - evenPart + oddPart;
434
435 // direction: w_2
436 w_2_indep = w_2*dir_indep_trm;
437
438 ui = -ux + uy;
439 evenPart = omegaEven*( 0.5*(pdf_NW + pdf_SE) - ui*ui*w_2_nine_half - w_2_indep );
440 oddPart = omegaOdd*(0.5*(pdf_NW - pdf_SE) - ui*w_2_x3 );
441 src[I(x - 1, y + 1, z, D3Q19_NW)] = pdf_NW - evenPart - oddPart;
442 src[I(x + 1, y - 1, z, D3Q19_SE)] = pdf_SE - evenPart + oddPart;
443
444 ui = ux + uy;
445 evenPart = omegaEven*( 0.5*(pdf_NE + pdf_SW) - ui*ui*w_2_nine_half - w_2_indep );
446 oddPart = omegaOdd*(0.5*(pdf_NE - pdf_SW) - ui*w_2_x3 );
447 src[I(x + 1, y + 1, z, D3Q19_NE)] = pdf_NE - evenPart - oddPart;
448 src[I(x - 1, y - 1, z, D3Q19_SW)] = pdf_SW - evenPart + oddPart;
449
450 ui = -ux + uz;
451 evenPart = omegaEven*( 0.5*(pdf_TW + pdf_BE) - ui*ui*w_2_nine_half - w_2_indep );
452 oddPart = omegaOdd*(0.5*(pdf_TW - pdf_BE) - ui*w_2_x3 );
453 src[I(x - 1, y, z + 1, D3Q19_TW)] = pdf_TW - evenPart - oddPart;
454 src[I(x + 1, y, z - 1, D3Q19_BE)] = pdf_BE - evenPart + oddPart;
455
456 ui = ux + uz;
457 evenPart = omegaEven*( 0.5*(pdf_TE + pdf_BW) - ui*ui*w_2_nine_half - w_2_indep );
458 oddPart = omegaOdd*(0.5*(pdf_TE - pdf_BW) - ui*w_2_x3 );
459 src[I(x + 1, y, z + 1, D3Q19_TE)] = pdf_TE - evenPart - oddPart;
460 src[I(x - 1, y, z - 1, D3Q19_BW)] = pdf_BW - evenPart + oddPart;
461
462 ui = -uy + uz;
463 evenPart = omegaEven*( 0.5*(pdf_TS + pdf_BN) - ui*ui*w_2_nine_half - w_2_indep );
464 oddPart = omegaOdd*(0.5*(pdf_TS - pdf_BN) - ui*w_2_x3 );
465 src[I(x, y - 1, z + 1, D3Q19_TS)] = pdf_TS - evenPart - oddPart;
466 src[I(x, y + 1, z - 1, D3Q19_BN)] = pdf_BN - evenPart + oddPart;
467
468 ui = uy + uz;
469 evenPart = omegaEven*( 0.5*(pdf_TN + pdf_BS) - ui*ui*w_2_nine_half - w_2_indep );
470 oddPart = omegaOdd*(0.5*(pdf_TN - pdf_BS) - ui*w_2_x3 );
471 src[I(x, y + 1, z + 1, D3Q19_TN)] = pdf_TN - evenPart - oddPart;
472 src[I(x, y - 1, z - 1, D3Q19_BS)] = pdf_BS - evenPart + oddPart;
473
474
475 #undef I
476 } } } // z, y, x (from inner to outer)
477 } } } // z, y, x (from inner to outer)
478 } // loop over threads
479
480 // }}}
481
482 // Stop counters before bounce back. Else computing loop balance will be incorrect.
483
484 X_LIKWID_STOP("aa-odd");
485
486 // Fixup bounce back PDFs.
487 #ifdef _OPENMP
488 #pragma omp parallel for default(none) \
489 shared(kd, src)
490 #endif
491 for (int i = 0; i < kd->nBounceBackPdfs; ++i) {
492 src[kd->BounceBackPdfsDst[i]] = src[kd->BounceBackPdfsSrc[i]];
493 }
494
495 // save current iteration
496 kda->Iteration = iter + 1;
497
498 #ifdef VERIFICATION
499 kd->PdfsActive = src;
500 KernelAddBodyForce(kd, ld, cd);
501 #endif
502
503 #ifdef VTK_OUTPUT
504 if (cd->VtkOutput && (iter + 1 % cd->VtkModulus) == 0) {
505 kd->PdfsActive = src;
506 VtkWrite(ld, kd, cd, iter + 1);
507 }
508 #endif
509
510 #ifdef STATISTICS
511 kd->PdfsActive = src;
512 KernelStatistics(kd, ld, cd, iter + 1);
513 #endif // }}}
514
515
516 } // for (int iter = 0; ...
517
518 #ifdef VTK_OUTPUT
519
520 if (cd->VtkOutput) {
521 kd->PdfsActive = src;
522 VtkWrite(ld, kd, cd, maxIterations);
523 }
524
525 #endif
526
527 return;
528}
529
LBM benchmark kernels
This page took 0.126313 seconds and 5 git commands to generate.