-
Notifications
You must be signed in to change notification settings - Fork 24
/
Copy pathswap.c
437 lines (379 loc) · 19.3 KB
/
swap.c
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
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
/**
@section LICENSE
Copyright (c) 2013-2016, Regents of the University of California
All rights reserved.
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#include <stdio.h>
#include"pmcl3d.h"
#define MPIRANKX 100000
#define MPIRANKY 50000
void update_bound_y_H(float* u1, float* v1, float* w1, float* f_u1, float* f_v1, float* f_w1, float* b_u1, float* b_v1,
float* b_w1, int nxt, int nzt, cudaStream_t St1, cudaStream_t St2, int rank_f, int rank_b);
void mediaswap(Grid3D d1, Grid3D mu, Grid3D lam, Grid3D qp, Grid3D qs,
int rank, int x_rank_L, int x_rank_R, int y_rank_F, int y_rank_B,
int nxt, int nyt, int nzt, MPI_Comm MCW)
{
int i, j, k, idx, idy, idz;
int media_count_x, media_count_y;
int media_size_x, media_size_y;
MPI_Request request_x[4], request_y[4];
MPI_Status status_x[4], status_y[4];
Grid1D mediaL_S=NULL, mediaR_S=NULL, mediaF_S=NULL, mediaB_S=NULL;
Grid1D mediaL_R=NULL, mediaR_R=NULL, mediaF_R=NULL, mediaB_R=NULL;
if(x_rank_L<0 && x_rank_R<0 && y_rank_F<0 && y_rank_B<0)
return;
if(y_rank_F>=0 || y_rank_B>=0)
{
mediaF_S = Alloc1D(5*4*loop*(nxt+2)*(nzt+2));
mediaB_S = Alloc1D(5*4*loop*(nxt+2)*(nzt+2));
mediaF_R = Alloc1D(5*4*loop*(nxt+2)*(nzt+2));
mediaB_R = Alloc1D(5*4*loop*(nxt+2)*(nzt+2));
media_size_y = 5*(4*loop)*(nxt+2)*(nzt+2);
media_count_y = 0;
PostRecvMsg_Y(mediaF_R, mediaB_R, MCW, request_y, &media_count_y, media_size_y, y_rank_F, y_rank_B);
if(y_rank_F>=0)
{
for(i=1+4*loop;i<nxt+3+4*loop;i++)
for(j=2+4*loop;j<2+8*loop;j++)
for(k=align-1;k<nzt+align+1;k++)
{
idx = i-1-4*loop;
idy = (j-2-4*loop)*5;
idz = k-align+1;
mediaF_S[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz] = d1[i][j][k];
idy++;
mediaF_S[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz] = mu[i][j][k];
idy++;
mediaF_S[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz] = lam[i][j][k];
idy++;
mediaF_S[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz] = qp[i][j][k];
idy++;
mediaF_S[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz] = qs[i][j][k];
}
}
if(y_rank_B>=0)
{
for(i=1+4*loop;i<nxt+3+4*loop;i++)
for(j=nyt+2;j<nyt+2+4*loop;j++)
for(k=align-1;k<nzt+align+1;k++)
{
idx = i-1-4*loop;
idy = (j-nyt-2)*5;
idz = k-align+1;
mediaB_S[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz] = d1[i][j][k];
idy++;
mediaB_S[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz] = mu[i][j][k];
idy++;
mediaB_S[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz] = lam[i][j][k];
idy++;
mediaB_S[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz] = qp[i][j][k];
idy++;
mediaB_S[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz] = qs[i][j][k];
}
}
PostSendMsg_Y(mediaF_S, mediaB_S, MCW, request_y, &media_count_y, media_size_y, y_rank_F, y_rank_B, rank, Both);
MPI_Waitall(media_count_y, request_y, status_y);
if(y_rank_F>=0)
{
for(i=1+4*loop;i<nxt+3+4*loop;i++)
for(j=2;j<2+4*loop;j++)
for(k=align-1;k<nzt+align+1;k++)
{
idx = i-1-4*loop;
idy = (j-2)*5;
idz = k-align+1;
d1[i][j][k] = mediaF_R[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz];
idy++;
mu[i][j][k] = mediaF_R[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz];
idy++;
lam[i][j][k] = mediaF_R[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz];
idy++;
qp[i][j][k] = mediaF_R[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz];
idy++;
qs[i][j][k] = mediaF_R[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz];
}
}
if(y_rank_B>=0)
{
for(i=1+4*loop;i<nxt+3+4*loop;i++)
for(j=nyt+2+4*loop;j<nyt+2+8*loop;j++)
for(k=align-1;k<nzt+align+1;k++)
{
idx = i-1-4*loop;
idy = (j-nyt-2-4*loop)*5;
idz = k-align+1;
d1[i][j][k] = mediaB_R[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz];
idy++;
mu[i][j][k] = mediaB_R[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz];
idy++;
lam[i][j][k] = mediaB_R[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz];
idy++;
qp[i][j][k] = mediaB_R[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz];
idy++;
qs[i][j][k] = mediaB_R[idx*5*(4*loop)*(nzt+2)+idy*(nzt+2)+idz];
}
}
Delloc1D(mediaF_S);
Delloc1D(mediaB_S);
Delloc1D(mediaF_R);
Delloc1D(mediaB_R);
}
if(x_rank_L>=0 || x_rank_R>=0)
{
mediaL_S = Alloc1D(5*4*loop*(nyt+8*loop)*(nzt+2));
mediaR_S = Alloc1D(5*4*loop*(nyt+8*loop)*(nzt+2));
mediaL_R = Alloc1D(5*4*loop*(nyt+8*loop)*(nzt+2));
mediaR_R = Alloc1D(5*4*loop*(nyt+8*loop)*(nzt+2));
media_size_x = 5*(4*loop)*(nyt+8*loop)*(nzt+2);
media_count_x = 0;
PostRecvMsg_X(mediaL_R, mediaR_R, MCW, request_x, &media_count_x, media_size_x, x_rank_L, x_rank_R);
if(x_rank_L>=0)
{
for(i=2+4*loop;i<2+8*loop;i++)
for(j=2;j<nyt+2+8*loop;j++)
for(k=align-1;k<nzt+align+1;k++)
{
idx = (i-2-4*loop)*5;
idy = j-2;
idz = k-align+1;
mediaL_S[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz] = d1[i][j][k];
idx++;
mediaL_S[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz] = mu[i][j][k];
idx++;
mediaL_S[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz] = lam[i][j][k];
idx++;
mediaL_S[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz] = qp[i][j][k];
idx++;
mediaL_S[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz] = qs[i][j][k];
}
}
if(x_rank_R>=0)
{
for(i=nxt+2;i<nxt+2+4*loop;i++)
for(j=2;j<nyt+2+8*loop;j++)
for(k=align-1;k<nzt+align+1;k++)
{
idx = (i-nxt-2)*5;
idy = j-2;
idz = k-align+1;
mediaR_S[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz] = d1[i][j][k];
idx++;
mediaR_S[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz] = mu[i][j][k];
idx++;
mediaR_S[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz] = lam[i][j][k];
idx++;
mediaR_S[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz] = qp[i][j][k];
idx++;
mediaR_S[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz] = qs[i][j][k];
}
}
PostSendMsg_X(mediaL_S, mediaR_S, MCW, request_x, &media_count_x, media_size_x, x_rank_L, x_rank_R, rank, Both);
MPI_Waitall(media_count_x, request_x, status_x);
if(x_rank_L>=0)
{
for(i=2;i<2+4*loop;i++)
for(j=2;j<nyt+2+8*loop;j++)
for(k=align-1;k<nzt+align+1;k++)
{
idx = (i-2)*5;
idy = j-2;
idz = k-align+1;
d1[i][j][k] = mediaL_R[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz];
idx++;
mu[i][j][k] = mediaL_R[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz];
idx++;
lam[i][j][k] = mediaL_R[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz];
idx++;
qp[i][j][k] = mediaL_R[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz];
idx++;
qs[i][j][k] = mediaL_R[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz];
}
}
if(x_rank_R>=0)
{
for(i=nxt+2+4*loop;i<nxt+2+8*loop;i++)
for(j=2;j<nyt+2+8*loop;j++)
for(k=align-1;k<nzt+align+1;k++)
{
idx = (i-nxt-2-4*loop)*5;
idy = j-2;
idz = k-align+1;
d1[i][j][k] = mediaR_R[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz];
idx++;
mu[i][j][k] = mediaR_R[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz];
idx++;
lam[i][j][k] = mediaR_R[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz];
idx++;
qp[i][j][k] = mediaR_R[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz];
idx++;
qs[i][j][k] = mediaR_R[idx*(nyt+8*loop)*(nzt+2)+idy*(nzt+2)+idz];
}
}
Delloc1D(mediaL_S);
Delloc1D(mediaR_S);
Delloc1D(mediaL_R);
Delloc1D(mediaR_R);
}
return;
}
void PostRecvMsg_X(float* RL_M, float* RR_M, MPI_Comm MCW, MPI_Request* request, int* count, int msg_size, int rank_L, int rank_R)
{
int temp_count = 0;
if(rank_L>=0){
MPI_Irecv(RL_M, msg_size, MPI_FLOAT, rank_L, MPIRANKX+rank_L, MCW, &request[temp_count]);
++temp_count;
}
if(rank_R>=0){
MPI_Irecv(RR_M, msg_size, MPI_FLOAT, rank_R, MPIRANKX+rank_R, MCW, &request[temp_count]);
++temp_count;
}
*count = temp_count;
return;
}
void PostSendMsg_X(float* SL_M, float* SR_M, MPI_Comm MCW, MPI_Request* request, int* count, int msg_size,
int rank_L, int rank_R, int rank, int flag)
{
int temp_count, flag_L=-1, flag_R=-1;
temp_count = *count;
if(rank<0)
return;
if(flag==Both || flag==Left) flag_L=1;
if(flag==Both || flag==Right) flag_R=1;
if(rank_L>=0 && flag_L==1){
MPI_Isend(SL_M, msg_size, MPI_FLOAT, rank_L, MPIRANKX+rank, MCW, &request[temp_count]);
++temp_count;
}
if(rank_R>=0 && flag_R==1){
MPI_Isend(SR_M, msg_size, MPI_FLOAT, rank_R, MPIRANKX+rank, MCW, &request[temp_count]);
++temp_count;
}
*count = temp_count;
return;
}
void PostRecvMsg_Y(float* RF_M, float* RB_M, MPI_Comm MCW, MPI_Request* request, int* count, int msg_size, int rank_F, int rank_B)
{
int temp_count = 0;
if(rank_F>=0){
MPI_Irecv(RF_M, msg_size, MPI_FLOAT, rank_F, MPIRANKY+rank_F, MCW, &request[temp_count]);
++temp_count;
}
if(rank_B>=0){
MPI_Irecv(RB_M, msg_size, MPI_FLOAT, rank_B, MPIRANKY+rank_B, MCW, &request[temp_count]);
++temp_count;
}
*count = temp_count;
return;
}
void PostSendMsg_Y(float* SF_M, float* SB_M, MPI_Comm MCW, MPI_Request* request, int* count, int msg_size,
int rank_F, int rank_B, int rank, int flag)
{
int temp_count, flag_F=-1, flag_B=-1;
temp_count = *count;
if(rank<0)
return;
if(flag==Both || flag==Front) flag_F=1;
if(flag==Both || flag==Back) flag_B=1;
if(rank_F>=0 && flag_F==1){
MPI_Isend(SF_M, msg_size, MPI_FLOAT, rank_F, MPIRANKY+rank, MCW, &request[temp_count]);
++temp_count;
}
if(rank_B>=0 && flag_B==1){
MPI_Isend(SB_M, msg_size, MPI_FLOAT, rank_B, MPIRANKY+rank, MCW, &request[temp_count]);
++temp_count;
}
*count = temp_count;
return;
}
void Cpy2Device_source(int npsrc, int READ_STEP,
int index_offset,
Grid1D taxx, Grid1D tayy, Grid1D tazz,
Grid1D taxz, Grid1D tayz, Grid1D taxy,
float *d_taxx, float *d_tayy, float *d_tazz,
float *d_taxz, float *d_tayz, float *d_taxy){
long int num_bytes;
cudaError_t cerr;
num_bytes = sizeof(float)*npsrc*READ_STEP;
cerr=cudaMemcpy(d_taxx,taxx+index_offset,num_bytes,cudaMemcpyHostToDevice);
if(cerr!=cudaSuccess) printf("CUDA ERROR: Cpy2Device: %s\n",cudaGetErrorString(cerr));
cerr=cudaMemcpy(d_tayy,tayy+index_offset,num_bytes,cudaMemcpyHostToDevice);
if(cerr!=cudaSuccess) printf("CUDA ERROR: Cpy2Device: %s\n",cudaGetErrorString(cerr));
cerr=cudaMemcpy(d_tazz,tazz+index_offset,num_bytes,cudaMemcpyHostToDevice);
if(cerr!=cudaSuccess) printf("CUDA ERROR: Cpy2Device: %s\n",cudaGetErrorString(cerr));
cerr=cudaMemcpy(d_taxz,taxz+index_offset,num_bytes,cudaMemcpyHostToDevice);
if(cerr!=cudaSuccess) printf("CUDA ERROR: Cpy2Device: %s\n",cudaGetErrorString(cerr));
cerr=cudaMemcpy(d_tayz,tayz+index_offset,num_bytes,cudaMemcpyHostToDevice);
if(cerr!=cudaSuccess) printf("CUDA ERROR: Cpy2Device: %s\n",cudaGetErrorString(cerr));
cerr=cudaMemcpy(d_taxy,taxy+index_offset,num_bytes,cudaMemcpyHostToDevice);
if(cerr!=cudaSuccess) printf("CUDA ERROR: Cpy2Device: %s\n",cudaGetErrorString(cerr));
return;
}
void Cpy2Host_VX(float* u1, float* v1, float* w1, float* h_m, int nxt, int nyt, int nzt, cudaStream_t St, int rank, int flag)
{
int d_offset=0, h_offset=0, msg_size=0;
if(rank<0 || flag<1 || flag>2)
return;
if(flag==Left) d_offset = (2+4*loop)*(nyt+4+8*loop)*(nzt+2*align);
if(flag==Right) d_offset = (nxt+2)*(nyt+4+8*loop)*(nzt+2*align);
h_offset = (4*loop)*(nyt+4+8*loop)*(nzt+2*align);
msg_size = sizeof(float)*(4*loop)*(nyt+4+8*loop)*(nzt+2*align);
cudaMemcpyAsync(h_m, u1+d_offset, msg_size, cudaMemcpyDeviceToHost, St);
cudaMemcpyAsync(h_m+h_offset, v1+d_offset, msg_size, cudaMemcpyDeviceToHost, St);
cudaMemcpyAsync(h_m+h_offset*2, w1+d_offset, msg_size, cudaMemcpyDeviceToHost, St);
return;
}
void Cpy2Host_VY(float* s_u1, float* s_v1, float* s_w1, float* h_m, int nxt, int nzt, cudaStream_t St, int rank)
{
int h_offset, msg_size;
if(rank<0)
return;
h_offset = (4*loop)*(nxt+4+8*loop)*(nzt+2*align);
msg_size = sizeof(float)*(4*loop)*(nxt+4+8*loop)*(nzt+2*align);
cudaMemcpyAsync(h_m, s_u1, msg_size, cudaMemcpyDeviceToHost, St);
cudaMemcpyAsync(h_m+h_offset, s_v1, msg_size, cudaMemcpyDeviceToHost, St);
cudaMemcpyAsync(h_m+h_offset*2, s_w1, msg_size, cudaMemcpyDeviceToHost, St);
return;
}
void Cpy2Device_VX(float* u1, float* v1, float* w1, float* L_m, float* R_m, int nxt,
int nyt, int nzt, cudaStream_t St1, cudaStream_t St2, int rank_L, int rank_R)
{
int d_offset, h_offset, msg_size;
h_offset = (4*loop)*(nyt+4+8*loop)*(nzt+2*align);
msg_size = sizeof(float)*(4*loop)*(nyt+4+8*loop)*(nzt+2*align);
if(rank_L>=0){
d_offset = 2*(nyt+4+8*loop)*(nzt+2*align);
cudaMemcpyAsync(u1+d_offset, L_m, msg_size, cudaMemcpyHostToDevice, St1);
cudaMemcpyAsync(v1+d_offset, L_m+h_offset, msg_size, cudaMemcpyHostToDevice, St1);
cudaMemcpyAsync(w1+d_offset, L_m+h_offset*2, msg_size, cudaMemcpyHostToDevice, St1);
}
if(rank_R>=0){
d_offset = (nxt+4*loop+2)*(nyt+4+8*loop)*(nzt+2*align);
cudaMemcpyAsync(u1+d_offset, R_m, msg_size, cudaMemcpyHostToDevice, St2);
cudaMemcpyAsync(v1+d_offset, R_m+h_offset, msg_size, cudaMemcpyHostToDevice, St2);
cudaMemcpyAsync(w1+d_offset, R_m+h_offset*2, msg_size, cudaMemcpyHostToDevice, St2);
}
return;
}
void Cpy2Device_VY(float* u1, float *v1, float *w1, float* f_u1, float* f_v1, float* f_w1, float* b_u1, float* b_v1,
float* b_w1, float* F_m, float* B_m, int nxt, int nyt, int nzt, cudaStream_t St1, cudaStream_t St2,
int rank_F, int rank_B)
{
int h_offset, msg_size;
h_offset = (4*loop)*(nxt+4+8*loop)*(nzt+2*align);
msg_size = sizeof(float)*(4*loop)*(nxt+4+8*loop)*(nzt+2*align);
if(rank_F>=0){
cudaMemcpyAsync(f_u1, F_m, msg_size, cudaMemcpyHostToDevice, St1);
cudaMemcpyAsync(f_v1, F_m+h_offset, msg_size, cudaMemcpyHostToDevice, St1);
cudaMemcpyAsync(f_w1, F_m+h_offset*2, msg_size, cudaMemcpyHostToDevice, St1);
}
if(rank_B>=0){
cudaMemcpyAsync(b_u1, B_m, msg_size, cudaMemcpyHostToDevice, St2);
cudaMemcpyAsync(b_v1, B_m+h_offset, msg_size, cudaMemcpyHostToDevice, St2);
cudaMemcpyAsync(b_w1, B_m+h_offset*2, msg_size, cudaMemcpyHostToDevice, St2);
}
update_bound_y_H(u1, v1, w1, f_u1, f_v1, f_w1, b_u1, b_v1, b_w1, nxt, nzt, St1, St2, rank_F, rank_B);
return;
}