Actual source code: csrperm.c

petsc-3.7.5 2017-01-01
Report Typos and Errors
  2: /*
  3:   Defines basic operations for the MATSEQAIJPERM matrix class.
  4:   This class is derived from the MATSEQAIJ class and retains the
  5:   compressed row storage (aka Yale sparse matrix format) but augments
  6:   it with some permutation information that enables some operations
  7:   to be more vectorizable.  A physically rearranged copy of the matrix
  8:   may be stored if the user desires.

 10:   Eventually a variety of permutations may be supported.
 11: */

 13: #include <../src/mat/impls/aij/seq/aij.h>

 15: #define NDIM 512
 16: /* NDIM specifies how many rows at a time we should work with when
 17:  * performing the vectorized mat-vec.  This depends on various factors
 18:  * such as vector register length, etc., and I really need to add a
 19:  * way for the user (or the library) to tune this.  I'm setting it to
 20:  * 512 for now since that is what Ed D'Azevedo was using in his Fortran
 21:  * routines. */

 23: typedef struct {
 24:   PetscInt ngroup;
 25:   PetscInt *xgroup;
 26:   /* Denotes where groups of rows with same number of nonzeros
 27:    * begin and end, i.e., xgroup[i] gives us the position in iperm[]
 28:    * where the ith group begins. */
 29:   PetscInt *nzgroup; /*  how many nonzeros each row that is a member of group i has. */
 30:   PetscInt *iperm;  /* The permutation vector. */

 32:   /* Flag that indicates whether we need to clean up permutation
 33:    * information during the MatDestroy. */
 34:   PetscBool CleanUpAIJPERM;

 36:   /* Some of this stuff is for Ed's recursive triangular solve.
 37:    * I'm not sure what I need yet. */
 38:   PetscInt blocksize;
 39:   PetscInt nstep;
 40:   PetscInt *jstart_list;
 41:   PetscInt *jend_list;
 42:   PetscInt *action_list;
 43:   PetscInt *ngroup_list;
 44:   PetscInt **ipointer_list;
 45:   PetscInt **xgroup_list;
 46:   PetscInt **nzgroup_list;
 47:   PetscInt **iperm_list;
 48: } Mat_SeqAIJPERM;

 50: extern PetscErrorCode MatAssemblyEnd_SeqAIJ(Mat,MatAssemblyType);

 54: PETSC_INTERN PetscErrorCode MatConvert_SeqAIJPERM_SeqAIJ(Mat A,MatType type,MatReuse reuse,Mat *newmat)
 55: {
 56:   /* This routine is only called to convert a MATAIJPERM to its base PETSc type, */
 57:   /* so we will ignore 'MatType type'. */
 59:   Mat            B       = *newmat;
 60:   Mat_SeqAIJPERM *aijperm=(Mat_SeqAIJPERM*)A->spptr;

 63:   if (reuse == MAT_INITIAL_MATRIX) {
 64:     MatDuplicate(A,MAT_COPY_VALUES,&B);
 65:   }

 67:   /* Reset the original function pointers. */
 68:   B->ops->assemblyend = MatAssemblyEnd_SeqAIJ;
 69:   B->ops->destroy     = MatDestroy_SeqAIJ;
 70:   B->ops->duplicate   = MatDuplicate_SeqAIJ;

 72:   /* Free everything in the Mat_SeqAIJPERM data structure.
 73:    * We don't free the Mat_SeqAIJPERM struct itself, as this will
 74:    * cause problems later when MatDestroy() tries to free it. */
 75:   if (aijperm->CleanUpAIJPERM) {
 76:     PetscFree(aijperm->xgroup);
 77:     PetscFree(aijperm->nzgroup);
 78:     PetscFree(aijperm->iperm);
 79:   }

 81:   /* Change the type of B to MATSEQAIJ. */
 82:   PetscObjectChangeTypeName((PetscObject)B, MATSEQAIJ);

 84:   *newmat = B;
 85:   return(0);
 86: }

 90: PetscErrorCode MatDestroy_SeqAIJPERM(Mat A)
 91: {
 93:   Mat_SeqAIJPERM *aijperm = (Mat_SeqAIJPERM*) A->spptr;

 96:   /* Free everything in the Mat_SeqAIJPERM data structure.
 97:    * Note that we don't need to free the Mat_SeqAIJPERM struct
 98:    * itself, as MatDestroy() will do so. */
 99:   if (aijperm && aijperm->CleanUpAIJPERM) {
100:     PetscFree(aijperm->xgroup);
101:     PetscFree(aijperm->nzgroup);
102:     PetscFree(aijperm->iperm);
103:   }
104:   PetscFree(A->spptr);

106:   /* Change the type of A back to SEQAIJ and use MatDestroy_SeqAIJ()
107:    * to destroy everything that remains. */
108:   PetscObjectChangeTypeName((PetscObject)A, MATSEQAIJ);
109:   /* Note that I don't call MatSetType().  I believe this is because that
110:    * is only to be called when *building* a matrix.  I could be wrong, but
111:    * that is how things work for the SuperLU matrix class. */
112:   MatDestroy_SeqAIJ(A);
113:   return(0);
114: }

116: PetscErrorCode MatDuplicate_SeqAIJPERM(Mat A, MatDuplicateOption op, Mat *M)
117: {
119:   Mat_SeqAIJPERM *aijperm      = (Mat_SeqAIJPERM*) A->spptr;
120:   Mat_SeqAIJPERM *aijperm_dest = (Mat_SeqAIJPERM*) (*M)->spptr;

123:   MatDuplicate_SeqAIJ(A,op,M);
124:   PetscMemcpy((*M)->spptr,aijperm,sizeof(Mat_SeqAIJPERM));
125:   /* Allocate space for, and copy the grouping and permutation info.
126:    * I note that when the groups are initially determined in
127:    * MatSeqAIJPERM_create_perm, xgroup and nzgroup may be sized larger than
128:    * necessary.  But at this point, we know how large they need to be, and
129:    * allocate only the necessary amount of memory.  So the duplicated matrix
130:    * may actually use slightly less storage than the original! */
131:   PetscMalloc1(A->rmap->n, &aijperm_dest->iperm);
132:   PetscMalloc1(aijperm->ngroup+1, &aijperm_dest->xgroup);
133:   PetscMalloc1(aijperm->ngroup, &aijperm_dest->nzgroup);
134:   PetscMemcpy(aijperm_dest->iperm,aijperm->iperm,sizeof(PetscInt)*A->rmap->n);
135:   PetscMemcpy(aijperm_dest->xgroup,aijperm->xgroup,sizeof(PetscInt)*(aijperm->ngroup+1));
136:   PetscMemcpy(aijperm_dest->nzgroup,aijperm->nzgroup,sizeof(PetscInt)*aijperm->ngroup);
137:   return(0);
138: }

142: PetscErrorCode MatSeqAIJPERM_create_perm(Mat A)
143: {
144:   PetscInt   m;       /* Number of rows in the matrix. */
145:   Mat_SeqAIJ *a = (Mat_SeqAIJ*)(A)->data;
146:   PetscInt   *ia;       /* From the CSR representation; points to the beginning  of each row. */
147:   PetscInt   maxnz;      /* Maximum number of nonzeros in any row. */
148:   PetscInt   *rows_in_bucket;
149:   /* To construct the permutation, we sort each row into one of maxnz
150:    * buckets based on how many nonzeros are in the row. */
151:   PetscInt nz;
152:   PetscInt *nz_in_row;         /* the number of nonzero elements in row k. */
153:   PetscInt *ipnz;
154:   /* When constructing the iperm permutation vector,
155:    * ipnz[nz] is used to point to the next place in the permutation vector
156:    * that a row with nz nonzero elements should be placed.*/
157:   Mat_SeqAIJPERM *aijperm = (Mat_SeqAIJPERM*) A->spptr;
158:   /* Points to the MATSEQAIJPERM-specific data in the matrix B. */
160:   PetscInt       i, ngroup, istart, ipos;

162:   /* I really ought to put something in here to check if B is of
163:    * type MATSEQAIJPERM and return an error code if it is not.
164:    * Come back and do this! */
166:   m  = A->rmap->n;
167:   ia = a->i;

169:   /* Allocate the arrays that will hold the permutation vector. */
170:   PetscMalloc1(m, &aijperm->iperm);

172:   /* Allocate some temporary work arrays that will be used in
173:    * calculating the permuation vector and groupings. */
174:   PetscMalloc1(m+1, &rows_in_bucket);
175:   PetscMalloc1(m+1, &ipnz);
176:   PetscMalloc1(m, &nz_in_row);

178:   /* Now actually figure out the permutation and grouping. */

180:   /* First pass: Determine number of nonzeros in each row, maximum
181:    * number of nonzeros in any row, and how many rows fall into each
182:    * "bucket" of rows with same number of nonzeros. */
183:   maxnz = 0;
184:   for (i=0; i<m; i++) {
185:     nz_in_row[i] = ia[i+1]-ia[i];
186:     if (nz_in_row[i] > maxnz) maxnz = nz_in_row[i];
187:   }

189:   for (i=0; i<=maxnz; i++) {
190:     rows_in_bucket[i] = 0;
191:   }
192:   for (i=0; i<m; i++) {
193:     nz = nz_in_row[i];
194:     rows_in_bucket[nz]++;
195:   }

197:   /* Allocate space for the grouping info.  There will be at most (maxnz + 1)
198:    * groups.  (It is maxnz + 1 instead of simply maxnz because there may be
199:    * rows with no nonzero elements.)  If there are (maxnz + 1) groups,
200:    * then xgroup[] must consist of (maxnz + 2) elements, since the last
201:    * element of xgroup will tell us where the (maxnz + 1)th group ends.
202:    * We allocate space for the maximum number of groups;
203:    * that is potentially a little wasteful, but not too much so.
204:    * Perhaps I should fix it later. */
205:   PetscMalloc1(maxnz+2, &aijperm->xgroup);
206:   PetscMalloc1(maxnz+1, &aijperm->nzgroup);

208:   /* Second pass.  Look at what is in the buckets and create the groupings.
209:    * Note that it is OK to have a group of rows with no non-zero values. */
210:   ngroup = 0;
211:   istart = 0;
212:   for (i=0; i<=maxnz; i++) {
213:     if (rows_in_bucket[i] > 0) {
214:       aijperm->nzgroup[ngroup] = i;
215:       aijperm->xgroup[ngroup]  = istart;
216:       ngroup++;
217:       istart += rows_in_bucket[i];
218:     }
219:   }

221:   aijperm->xgroup[ngroup] = istart;
222:   aijperm->ngroup         = ngroup;

224:   /* Now fill in the permutation vector iperm. */
225:   ipnz[0] = 0;
226:   for (i=0; i<maxnz; i++) {
227:     ipnz[i+1] = ipnz[i] + rows_in_bucket[i];
228:   }

230:   for (i=0; i<m; i++) {
231:     nz                   = nz_in_row[i];
232:     ipos                 = ipnz[nz];
233:     aijperm->iperm[ipos] = i;
234:     ipnz[nz]++;
235:   }

237:   /* Clean up temporary work arrays. */
238:   PetscFree(rows_in_bucket);
239:   PetscFree(ipnz);
240:   PetscFree(nz_in_row);

242:   /* Since we've allocated some memory to hold permutation info,
243:    * flip the CleanUpAIJPERM flag to true. */
244:   aijperm->CleanUpAIJPERM = PETSC_TRUE;
245:   return(0);
246: }


251: PetscErrorCode MatAssemblyEnd_SeqAIJPERM(Mat A, MatAssemblyType mode)
252: {
254:   Mat_SeqAIJ     *a = (Mat_SeqAIJ*)A->data;

257:   if (mode == MAT_FLUSH_ASSEMBLY) return(0);

259:   /* Since a MATSEQAIJPERM matrix is really just a MATSEQAIJ with some
260:    * extra information, call the AssemblyEnd routine for a MATSEQAIJ.
261:    * I'm not sure if this is the best way to do this, but it avoids
262:    * a lot of code duplication.
263:    * I also note that currently MATSEQAIJPERM doesn't know anything about
264:    * the Mat_CompressedRow data structure that SeqAIJ now uses when there
265:    * are many zero rows.  If the SeqAIJ assembly end routine decides to use
266:    * this, this may break things.  (Don't know... haven't looked at it.) */
267:   a->inode.use = PETSC_FALSE;
268:   MatAssemblyEnd_SeqAIJ(A, mode);

270:   /* Now calculate the permutation and grouping information. */
271:   MatSeqAIJPERM_create_perm(A);
272:   return(0);
273: }

277: PetscErrorCode MatMult_SeqAIJPERM(Mat A,Vec xx,Vec yy)
278: {
279:   Mat_SeqAIJ        *a = (Mat_SeqAIJ*)A->data;
280:   const PetscScalar *x;
281:   PetscScalar       *y;
282:   const MatScalar   *aa;
283:   PetscErrorCode    ierr;
284:   const PetscInt    *aj,*ai;
285: #if !(defined(PETSC_USE_FORTRAN_KERNEL_MULTAIJPERM) && defined(notworking))
286:   PetscInt i,j;
287: #endif

289:   /* Variables that don't appear in MatMult_SeqAIJ. */
290:   Mat_SeqAIJPERM *aijperm = (Mat_SeqAIJPERM*) A->spptr;
291:   PetscInt       *iperm;  /* Points to the permutation vector. */
292:   PetscInt       *xgroup;
293:   /* Denotes where groups of rows with same number of nonzeros
294:    * begin and end in iperm. */
295:   PetscInt *nzgroup;
296:   PetscInt ngroup;
297:   PetscInt igroup;
298:   PetscInt jstart,jend;
299:   /* jstart is used in loops to denote the position in iperm where a
300:    * group starts; jend denotes the position where it ends.
301:    * (jend + 1 is where the next group starts.) */
302:   PetscInt    iold,nz;
303:   PetscInt    istart,iend,isize;
304:   PetscInt    ipos;
305:   PetscScalar yp[NDIM];
306:   PetscInt    ip[NDIM];    /* yp[] and ip[] are treated as vector "registers" for performing the mat-vec. */

308: #if defined(PETSC_HAVE_PRAGMA_DISJOINT)
309: #pragma disjoint(*x,*y,*aa)
310: #endif

313:   VecGetArrayRead(xx,&x);
314:   VecGetArray(yy,&y);
315:   aj   = a->j;   /* aj[k] gives column index for element aa[k]. */
316:   aa   = a->a; /* Nonzero elements stored row-by-row. */
317:   ai   = a->i;  /* ai[k] is the position in aa and aj where row k starts. */

319:   /* Get the info we need about the permutations and groupings. */
320:   iperm   = aijperm->iperm;
321:   ngroup  = aijperm->ngroup;
322:   xgroup  = aijperm->xgroup;
323:   nzgroup = aijperm->nzgroup;

325: #if defined(PETSC_USE_FORTRAN_KERNEL_MULTAIJPERM) && defined(notworking)
326:   fortranmultaijperm_(&m,x,ii,aj,aa,y);
327: #else

329:   for (igroup=0; igroup<ngroup; igroup++) {
330:     jstart = xgroup[igroup];
331:     jend   = xgroup[igroup+1] - 1;
332:     nz     = nzgroup[igroup];

334:     /* Handle the special cases where the number of nonzeros per row
335:      * in the group is either 0 or 1. */
336:     if (nz == 0) {
337:       for (i=jstart; i<=jend; i++) {
338:         y[iperm[i]] = 0.0;
339:       }
340:     } else if (nz == 1) {
341:       for (i=jstart; i<=jend; i++) {
342:         iold    = iperm[i];
343:         ipos    = ai[iold];
344:         y[iold] = aa[ipos] * x[aj[ipos]];
345:       }
346:     } else {

348:       /* We work our way through the current group in chunks of NDIM rows
349:        * at a time. */

351:       for (istart=jstart; istart<=jend; istart+=NDIM) {
352:         /* Figure out where the chunk of 'isize' rows ends in iperm.
353:          * 'isize may of course be less than NDIM for the last chunk. */
354:         iend = istart + (NDIM - 1);

356:         if (iend > jend) iend = jend;

358:         isize = iend - istart + 1;

360:         /* Initialize the yp[] array that will be used to hold part of
361:          * the permuted results vector, and figure out where in aa each
362:          * row of the chunk will begin. */
363:         for (i=0; i<isize; i++) {
364:           iold = iperm[istart + i];
365:           /* iold is a row number from the matrix A *before* reordering. */
366:           ip[i] = ai[iold];
367:           /* ip[i] tells us where the ith row of the chunk begins in aa. */
368:           yp[i] = (PetscScalar) 0.0;
369:         }

371:         /* If the number of zeros per row exceeds the number of rows in
372:          * the chunk, we should vectorize along nz, that is, perform the
373:          * mat-vec one row at a time as in the usual CSR case. */
374:         if (nz > isize) {
375: #if defined(PETSC_HAVE_CRAY_VECTOR)
376: #pragma _CRI preferstream
377: #endif
378:           for (i=0; i<isize; i++) {
379: #if defined(PETSC_HAVE_CRAY_VECTOR)
380: #pragma _CRI prefervector
381: #endif
382:             for (j=0; j<nz; j++) {
383:               ipos   = ip[i] + j;
384:               yp[i] += aa[ipos] * x[aj[ipos]];
385:             }
386:           }
387:         } else {
388:           /* Otherwise, there are enough rows in the chunk to make it
389:            * worthwhile to vectorize across the rows, that is, to do the
390:            * matvec by operating with "columns" of the chunk. */
391:           for (j=0; j<nz; j++) {
392:             for (i=0; i<isize; i++) {
393:               ipos   = ip[i] + j;
394:               yp[i] += aa[ipos] * x[aj[ipos]];
395:             }
396:           }
397:         }

399: #if defined(PETSC_HAVE_CRAY_VECTOR)
400: #pragma _CRI ivdep
401: #endif
402:         /* Put results from yp[] into non-permuted result vector y. */
403:         for (i=0; i<isize; i++) {
404:           y[iperm[istart+i]] = yp[i];
405:         }
406:       } /* End processing chunk of isize rows of a group. */
407:     } /* End handling matvec for chunk with nz > 1. */
408:   } /* End loop over igroup. */
409: #endif
410:   PetscLogFlops(2.0*a->nz - A->rmap->n);
411:   VecRestoreArrayRead(xx,&x);
412:   VecRestoreArray(yy,&y);
413:   return(0);
414: }


417: /* MatMultAdd_SeqAIJPERM() calculates yy = ww + A * xx.
418:  * Note that the names I used to designate the vectors differs from that
419:  * used in MatMultAdd_SeqAIJ().  I did this to keep my notation consistent
420:  * with the MatMult_SeqAIJPERM() routine, which is very similar to this one. */
421: /*
422:     I hate having virtually identical code for the mult and the multadd!!!
423: */
426: PetscErrorCode MatMultAdd_SeqAIJPERM(Mat A,Vec xx,Vec ww,Vec yy)
427: {
428:   Mat_SeqAIJ        *a = (Mat_SeqAIJ*)A->data;
429:   const PetscScalar *x;
430:   PetscScalar       *y,*w;
431:   const MatScalar   *aa;
432:   PetscErrorCode    ierr;
433:   const PetscInt    *aj,*ai;
434: #if !defined(PETSC_USE_FORTRAN_KERNEL_MULTADDAIJPERM)
435:   PetscInt i,j;
436: #endif

438:   /* Variables that don't appear in MatMultAdd_SeqAIJ. */
439:   Mat_SeqAIJPERM * aijperm;
440:   PetscInt       *iperm;    /* Points to the permutation vector. */
441:   PetscInt       *xgroup;
442:   /* Denotes where groups of rows with same number of nonzeros
443:    * begin and end in iperm. */
444:   PetscInt *nzgroup;
445:   PetscInt ngroup;
446:   PetscInt igroup;
447:   PetscInt jstart,jend;
448:   /* jstart is used in loops to denote the position in iperm where a
449:    * group starts; jend denotes the position where it ends.
450:    * (jend + 1 is where the next group starts.) */
451:   PetscInt    iold,nz;
452:   PetscInt    istart,iend,isize;
453:   PetscInt    ipos;
454:   PetscScalar yp[NDIM];
455:   PetscInt    ip[NDIM];
456:   /* yp[] and ip[] are treated as vector "registers" for performing
457:    * the mat-vec. */

459: #if defined(PETSC_HAVE_PRAGMA_DISJOINT)
460: #pragma disjoint(*x,*y,*aa)
461: #endif

464:   VecGetArrayRead(xx,&x);
465:   VecGetArrayPair(yy,ww,&y,&w);

467:   aj = a->j;   /* aj[k] gives column index for element aa[k]. */
468:   aa = a->a;   /* Nonzero elements stored row-by-row. */
469:   ai = a->i;   /* ai[k] is the position in aa and aj where row k starts. */

471:   /* Get the info we need about the permutations and groupings. */
472:   aijperm = (Mat_SeqAIJPERM*) A->spptr;
473:   iperm   = aijperm->iperm;
474:   ngroup  = aijperm->ngroup;
475:   xgroup  = aijperm->xgroup;
476:   nzgroup = aijperm->nzgroup;

478: #if defined(PETSC_USE_FORTRAN_KERNEL_MULTADDAIJPERM)
479:   fortranmultaddaijperm_(&m,x,ii,aj,aa,y,w);
480: #else

482:   for (igroup=0; igroup<ngroup; igroup++) {
483:     jstart = xgroup[igroup];
484:     jend   = xgroup[igroup+1] - 1;

486:     nz = nzgroup[igroup];

488:     /* Handle the special cases where the number of nonzeros per row
489:      * in the group is either 0 or 1. */
490:     if (nz == 0) {
491:       for (i=jstart; i<=jend; i++) {
492:         iold    = iperm[i];
493:         y[iold] = w[iold];
494:       }
495:     }
496:     else if (nz == 1) {
497:       for (i=jstart; i<=jend; i++) {
498:         iold    = iperm[i];
499:         ipos    = ai[iold];
500:         y[iold] = w[iold] + aa[ipos] * x[aj[ipos]];
501:       }
502:     }
503:     /* For the general case: */
504:     else {

506:       /* We work our way through the current group in chunks of NDIM rows
507:        * at a time. */

509:       for (istart=jstart; istart<=jend; istart+=NDIM) {
510:         /* Figure out where the chunk of 'isize' rows ends in iperm.
511:          * 'isize may of course be less than NDIM for the last chunk. */
512:         iend = istart + (NDIM - 1);
513:         if (iend > jend) iend = jend;
514:         isize = iend - istart + 1;

516:         /* Initialize the yp[] array that will be used to hold part of
517:          * the permuted results vector, and figure out where in aa each
518:          * row of the chunk will begin. */
519:         for (i=0; i<isize; i++) {
520:           iold = iperm[istart + i];
521:           /* iold is a row number from the matrix A *before* reordering. */
522:           ip[i] = ai[iold];
523:           /* ip[i] tells us where the ith row of the chunk begins in aa. */
524:           yp[i] = w[iold];
525:         }

527:         /* If the number of zeros per row exceeds the number of rows in
528:          * the chunk, we should vectorize along nz, that is, perform the
529:          * mat-vec one row at a time as in the usual CSR case. */
530:         if (nz > isize) {
531: #if defined(PETSC_HAVE_CRAY_VECTOR)
532: #pragma _CRI preferstream
533: #endif
534:           for (i=0; i<isize; i++) {
535: #if defined(PETSC_HAVE_CRAY_VECTOR)
536: #pragma _CRI prefervector
537: #endif
538:             for (j=0; j<nz; j++) {
539:               ipos   = ip[i] + j;
540:               yp[i] += aa[ipos] * x[aj[ipos]];
541:             }
542:           }
543:         }
544:         /* Otherwise, there are enough rows in the chunk to make it
545:          * worthwhile to vectorize across the rows, that is, to do the
546:          * matvec by operating with "columns" of the chunk. */
547:         else {
548:           for (j=0; j<nz; j++) {
549:             for (i=0; i<isize; i++) {
550:               ipos   = ip[i] + j;
551:               yp[i] += aa[ipos] * x[aj[ipos]];
552:             }
553:           }
554:         }

556: #if defined(PETSC_HAVE_CRAY_VECTOR)
557: #pragma _CRI ivdep
558: #endif
559:         /* Put results from yp[] into non-permuted result vector y. */
560:         for (i=0; i<isize; i++) {
561:           y[iperm[istart+i]] = yp[i];
562:         }
563:       } /* End processing chunk of isize rows of a group. */

565:     } /* End handling matvec for chunk with nz > 1. */
566:   } /* End loop over igroup. */

568: #endif
569:   PetscLogFlops(2.0*a->nz);
570:   VecRestoreArrayRead(xx,&x);
571:   VecRestoreArrayPair(yy,ww,&y,&w);
572:   return(0);
573: }


576: /* MatConvert_SeqAIJ_SeqAIJPERM converts a SeqAIJ matrix into a
577:  * SeqAIJPERM matrix.  This routine is called by the MatCreate_SeqAIJPERM()
578:  * routine, but can also be used to convert an assembled SeqAIJ matrix
579:  * into a SeqAIJPERM one. */
582: PETSC_INTERN PetscErrorCode MatConvert_SeqAIJ_SeqAIJPERM(Mat A,MatType type,MatReuse reuse,Mat *newmat)
583: {
585:   Mat            B = *newmat;
586:   Mat_SeqAIJPERM *aijperm;

589:   if (reuse == MAT_INITIAL_MATRIX) {
590:     MatDuplicate(A,MAT_COPY_VALUES,&B);
591:   }

593:   PetscNewLog(B,&aijperm);
594:   B->spptr = (void*) aijperm;

596:   /* Set function pointers for methods that we inherit from AIJ but override. */
597:   B->ops->duplicate   = MatDuplicate_SeqAIJPERM;
598:   B->ops->assemblyend = MatAssemblyEnd_SeqAIJPERM;
599:   B->ops->destroy     = MatDestroy_SeqAIJPERM;
600:   B->ops->mult        = MatMult_SeqAIJPERM;
601:   B->ops->multadd     = MatMultAdd_SeqAIJPERM;

603:   /* If A has already been assembled, compute the permutation. */
604:   if (A->assembled) {
605:     MatSeqAIJPERM_create_perm(B);
606:   }

608:   PetscObjectComposeFunction((PetscObject)B,"MatConvert_seqaijperm_seqaij_C",MatConvert_SeqAIJPERM_SeqAIJ);

610:   PetscObjectChangeTypeName((PetscObject)B,MATSEQAIJPERM);
611:   *newmat = B;
612:   return(0);
613: }

617: /*@C
618:    MatCreateSeqAIJPERM - Creates a sparse matrix of type SEQAIJPERM.
619:    This type inherits from AIJ, but calculates some additional permutation
620:    information that is used to allow better vectorization of some
621:    operations.  At the cost of increased storage, the AIJ formatted
622:    matrix can be copied to a format in which pieces of the matrix are
623:    stored in ELLPACK format, allowing the vectorized matrix multiply
624:    routine to use stride-1 memory accesses.  As with the AIJ type, it is
625:    important to preallocate matrix storage in order to get good assembly
626:    performance.

628:    Collective on MPI_Comm

630:    Input Parameters:
631: +  comm - MPI communicator, set to PETSC_COMM_SELF
632: .  m - number of rows
633: .  n - number of columns
634: .  nz - number of nonzeros per row (same for all rows)
635: -  nnz - array containing the number of nonzeros in the various rows
636:          (possibly different for each row) or NULL

638:    Output Parameter:
639: .  A - the matrix

641:    Notes:
642:    If nnz is given then nz is ignored

644:    Level: intermediate

646: .keywords: matrix, cray, sparse, parallel

648: .seealso: MatCreate(), MatCreateMPIAIJPERM(), MatSetValues()
649: @*/
650: PetscErrorCode  MatCreateSeqAIJPERM(MPI_Comm comm,PetscInt m,PetscInt n,PetscInt nz,const PetscInt nnz[],Mat *A)
651: {

655:   MatCreate(comm,A);
656:   MatSetSizes(*A,m,n,m,n);
657:   MatSetType(*A,MATSEQAIJPERM);
658:   MatSeqAIJSetPreallocation_SeqAIJ(*A,nz,nnz);
659:   return(0);
660: }

664: PETSC_EXTERN PetscErrorCode MatCreate_SeqAIJPERM(Mat A)
665: {

669:   MatSetType(A,MATSEQAIJ);
670:   MatConvert_SeqAIJ_SeqAIJPERM(A,MATSEQAIJPERM,MAT_INPLACE_MATRIX,&A);
671:   return(0);
672: }