/[escript]/branches/domexper/dudley/src/Mesh_optimizeDOFDistribution.c
ViewVC logotype

Contents of /branches/domexper/dudley/src/Mesh_optimizeDOFDistribution.c

Parent Directory Parent Directory | Revision Log Revision Log


Revision 3126 - (show annotations)
Wed Sep 1 00:37:53 2010 UTC (8 years, 9 months ago) by jfenwick
File MIME type: text/plain
File size: 11511 byte(s)
Updated referenceElements and enums
1
2 /*******************************************************
3 *
4 * Copyright (c) 2003-2010 by University of Queensland
5 * Earth Systems Science Computational Center (ESSCC)
6 * http://www.uq.edu.au/esscc
7 *
8 * Primary Business: Queensland, Australia
9 * Licensed under the Open Software License version 3.0
10 * http://www.opensource.org/licenses/osl-3.0.php
11 *
12 *******************************************************/
13
14
15 /**************************************************************/
16
17 /* Dudley: Mesh: optimizes the distribution of DOFs across processors */
18 /* using ParMETIS. On return a new distribution is given and the globalDOF are relabled */
19 /* accordingly but the mesh has not been redesitributed yet */
20
21 /**************************************************************/
22
23 #include "Mesh.h"
24 #include "IndexList.h"
25 #ifdef _OPENMP
26 #include <omp.h>
27 #endif
28 #ifdef USE_PARMETIS
29 #include "parmetis.h"
30 #endif
31
32 /**************************************************************
33 Check whether there is any node which has no vertex. In case
34 such node exists, we don't use parmetis since parmetis requires
35 that every node has at least 1 vertex (at line 129 of file
36 "xyzpart.c" in parmetis 3.1.1, variable "nvtxs" would be 0 if
37 any node has no vertex).
38 **************************************************************/
39 #ifdef USE_PARMETIS
40 int Check_Inputs_For_Parmetis(dim_t mpiSize, dim_t rank, dim_t *distribution, MPI_Comm *comm)
41 {
42 dim_t i, len;
43 int ret_val = 1;
44
45 if (rank == 0){
46 for (i=0; i<mpiSize; i++){
47 len = distribution[i+1] - distribution[i];
48 if (len == 0){
49 ret_val = 0;
50 break;
51 }
52 }
53 }
54 MPI_Bcast(&ret_val, 1, MPI_INTEGER, 0, *comm);
55 if (ret_val == 0)
56 printf("INFO: Parmetis is not used since some nodes have no vertex!\n");
57 return ret_val;
58 }
59 #endif
60
61
62
63 /**************************************************************/
64
65 void Dudley_Mesh_optimizeDOFDistribution(Dudley_Mesh* in,dim_t *distribution) {
66
67 dim_t dim, i,j,k, myNumVertices,p, mpiSize, len, globalNumVertices,*partition_count=NULL, *new_distribution=NULL, *loc_partition_count=NULL;
68 bool_t *setNewDOFId=NULL;
69 index_t myFirstVertex, myLastVertex, firstVertex, lastVertex, *newGlobalDOFID=NULL;
70 size_t mpiSize_size;
71 index_t* partition=NULL;
72 Paso_Pattern *pattern=NULL;
73 Paso_MPI_rank myRank,dest,source,current_rank, rank;
74 Dudley_IndexList* index_list=NULL;
75 float *xyz=NULL;
76 int c;
77
78 #ifdef PASO_MPI
79 MPI_Status status;
80 #endif
81
82 if (in==NULL) return;
83 if (in->Nodes == NULL) return;
84
85 myRank=in->MPIInfo->rank;
86 mpiSize=in->MPIInfo->size;
87 mpiSize_size=mpiSize*sizeof(dim_t);
88 dim=in->Nodes->numDim;
89 /* first step is to distribute the elements according to a global X of DOF */
90
91 myFirstVertex=distribution[myRank];
92 myLastVertex=distribution[myRank+1];
93 myNumVertices=myLastVertex-myFirstVertex;
94 globalNumVertices=distribution[mpiSize];
95 len=0;
96 for (p=0;p<mpiSize;++p) len=MAX(len,distribution[p+1]-distribution[p]);
97 partition=TMPMEMALLOC(len,index_t); /* len is used for the sending around of partition later on */
98 xyz=TMPMEMALLOC(myNumVertices*dim,float);
99 partition_count=TMPMEMALLOC(mpiSize+1,dim_t);
100 new_distribution=TMPMEMALLOC(mpiSize+1,dim_t);
101 newGlobalDOFID=TMPMEMALLOC(len,index_t);
102 setNewDOFId=TMPMEMALLOC(in->Nodes->numNodes,bool_t);
103 if (!(Dudley_checkPtr(partition) || Dudley_checkPtr(xyz) || Dudley_checkPtr(partition_count) || Dudley_checkPtr(partition_count) || Dudley_checkPtr(newGlobalDOFID) || Dudley_checkPtr(setNewDOFId))) {
104 dim_t *recvbuf=TMPMEMALLOC(mpiSize*mpiSize,dim_t);
105
106 /* set the coordinates: */
107 /* it is assumed that at least one node on this processor provides a coordinate */
108 #pragma omp parallel for private(i,j,k)
109 for (i=0;i<in->Nodes->numNodes;++i) {
110 k=in->Nodes->globalDegreesOfFreedom[i]-myFirstVertex;
111 if ((k>=0) && (k<myNumVertices)) {
112 for (j=0;j<dim;++j) xyz[k*dim+j]=(float)(in->Nodes->Coordinates[INDEX2(j,i,dim)]);
113 }
114 }
115
116 index_list=TMPMEMALLOC(myNumVertices,Dudley_IndexList);
117 /* ksteube CSR of DOF IDs */
118 /* create the adjacency structure xadj and adjncy */
119 if (! Dudley_checkPtr(index_list)) {
120 #pragma omp parallel private(i)
121 {
122 #pragma omp for schedule(static)
123 for(i=0;i<myNumVertices;++i) {
124 index_list[i].extension=NULL;
125 index_list[i].n=0;
126 }
127 /* ksteube build CSR format */
128 /* insert contributions from element matrices into colums index index_list: */
129 Dudley_IndexList_insertElementsWithRowRangeNoMainDiagonal(index_list, myFirstVertex, myLastVertex,
130 in->Elements,in->Nodes->globalDegreesOfFreedom,
131 in->Nodes->globalDegreesOfFreedom);
132 Dudley_IndexList_insertElementsWithRowRangeNoMainDiagonal(index_list, myFirstVertex, myLastVertex,
133 in->FaceElements,in->Nodes->globalDegreesOfFreedom,
134 in->Nodes->globalDegreesOfFreedom);
135 Dudley_IndexList_insertElementsWithRowRangeNoMainDiagonal(index_list, myFirstVertex, myLastVertex,
136 in->Points,in->Nodes->globalDegreesOfFreedom,
137 in->Nodes->globalDegreesOfFreedom);
138 }
139
140 /* create the local matrix pattern */
141 pattern=Dudley_IndexList_createPattern(0,myNumVertices,index_list,0,globalNumVertices,0);
142
143 /* clean up index list */
144 if (index_list!=NULL) {
145 #pragma omp parallel for private(i)
146 for(i=0;i<myNumVertices;++i) Dudley_IndexList_free(index_list[i].extension);
147 }
148
149 if (Dudley_noError()) {
150
151 #ifdef USE_PARMETIS
152
153 if (mpiSize>1 &&
154 Check_Inputs_For_Parmetis(mpiSize, myRank, distribution, &(in->MPIInfo->comm))>0 ) {
155 int i;
156 int wgtflag = 0;
157 int numflag = 0; /* pattern->ptr is C style: starting from 0 instead of 1 */
158 int ncon = 1;
159 int edgecut;
160 int options[2];
161 float *tpwgts = TMPMEMALLOC(ncon*mpiSize,float);
162 float *ubvec = TMPMEMALLOC(ncon,float);
163 for (i=0; i<ncon*mpiSize; i++) tpwgts[i] = 1.0/(float)mpiSize;
164 for (i=0; i<ncon; i++) ubvec[i] = 1.05;
165 options[0] = 3;
166 options[1] = 15;
167 ParMETIS_V3_PartGeomKway(distribution,
168 pattern->ptr,
169 pattern->index,
170 NULL,
171 NULL,
172 &wgtflag,
173 &numflag,
174 &dim,
175 xyz,
176 &ncon,
177 &mpiSize,
178 tpwgts,
179 ubvec,
180 options,
181 &edgecut,
182 partition, /* new CPU ownership of elements */
183 &(in->MPIInfo->comm));
184 /* printf("ParMETIS number of edges cut by partitioning per processor: %d\n", edgecut/MAX(in->MPIInfo->size,1)); */
185 TMPMEMFREE(ubvec);
186 TMPMEMFREE(tpwgts);
187 } else {
188 for (i=0;i<myNumVertices;++i) partition[i]=0; /* CPU 0 owns it */
189 }
190 #else
191 for (i=0;i<myNumVertices;++i) partition[i]=myRank; /* CPU 0 owns it */
192 #endif
193
194 }
195
196 Paso_Pattern_free(pattern);
197
198 /* create a new distributioin and labeling of the DOF */
199 memset(new_distribution,0,mpiSize_size);
200 #pragma omp parallel private(loc_partition_count)
201 {
202 loc_partition_count=THREAD_MEMALLOC(mpiSize,dim_t);
203 memset(loc_partition_count,0,mpiSize_size);
204 #pragma omp for private(i)
205 for (i=0;i<myNumVertices;++i) loc_partition_count[partition[i]]++ ;
206 #pragma omp critical
207 {
208 for (i=0;i<mpiSize;++i) new_distribution[i]+=loc_partition_count[i];
209 }
210 THREAD_MEMFREE(loc_partition_count);
211 }
212 #ifdef PASO_MPI
213 /* recvbuf will be the concatenation of each CPU's contribution to new_distribution */
214 MPI_Allgather(new_distribution, mpiSize, MPI_INT, recvbuf, mpiSize, MPI_INT, in->MPIInfo->comm);
215 #else
216 for (i=0;i<mpiSize;++i) recvbuf[i]=new_distribution[i];
217 #endif
218 new_distribution[0]=0;
219 for (rank=0; rank<mpiSize;rank++) {
220 c=0;
221 for (i=0;i<myRank;++i) c+=recvbuf[rank+mpiSize*i];
222 for (i=0;i<myNumVertices;++i) {
223 if (rank==partition[i]) {
224 newGlobalDOFID[i]=new_distribution[rank]+c;
225 c++;
226 }
227 }
228 for (i=myRank+1;i<mpiSize;++i) c+=recvbuf[rank+mpiSize*i];
229 new_distribution[rank+1]=new_distribution[rank]+c;
230 }
231 TMPMEMFREE(recvbuf);
232
233 /* now the overlap needs to be created by sending the partition around*/
234
235 dest=Paso_MPIInfo_mod(mpiSize, myRank + 1);
236 source=Paso_MPIInfo_mod(mpiSize, myRank - 1);
237 current_rank=myRank;
238 #pragma omp parallel for private(i)
239 for (i=0;i<in->Nodes->numNodes;++i) setNewDOFId[i]=TRUE;
240
241 for (p=0; p< mpiSize; ++p) {
242
243 firstVertex=distribution[current_rank];
244 lastVertex=distribution[current_rank+1];
245 #pragma omp parallel for private(i,j,k)
246 for (i=0;i<in->Nodes->numNodes;++i) {
247 k=in->Nodes->globalDegreesOfFreedom[i];
248 if (setNewDOFId[i] && (firstVertex<=k) && (k<lastVertex)) {
249 in->Nodes->globalDegreesOfFreedom[i]=newGlobalDOFID[k-firstVertex];
250 setNewDOFId[i]=FALSE;
251 }
252 }
253
254 if (p<mpiSize-1) { /* the final send can be skipped */
255 #ifdef PASO_MPI
256 MPI_Sendrecv_replace(newGlobalDOFID,len, MPI_INT,
257 dest, in->MPIInfo->msg_tag_counter,
258 source, in->MPIInfo->msg_tag_counter,
259 in->MPIInfo->comm,&status);
260 #endif
261 in->MPIInfo->msg_tag_counter++;
262 current_rank=Paso_MPIInfo_mod(mpiSize, current_rank-1);
263 }
264 }
265 for (i=0;i<mpiSize+1;++i) distribution[i]=new_distribution[i];
266 }
267 TMPMEMFREE(index_list);
268 }
269 TMPMEMFREE(newGlobalDOFID);
270 TMPMEMFREE(setNewDOFId);
271 TMPMEMFREE(new_distribution);
272 TMPMEMFREE(partition_count);
273 TMPMEMFREE(partition);
274 TMPMEMFREE(xyz);
275 return;
276 }

  ViewVC Help
Powered by ViewVC 1.1.26