Actual source code: crl.c

petsc-3.4.5 2014-06-29
  2: /*
  3:   Defines a matrix-vector product for the MATSEQAIJCRL 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 a column oriented storage that is more efficient for
  7:   matrix vector products on Vector machines.

  9:   CRL stands for constant row length (that is the same number of columns
 10:   is kept (padded with zeros) for each row of the sparse matrix.
 11: */
 12: #include <../src/mat/impls/aij/seq/crl/crl.h>

 16: PetscErrorCode MatDestroy_SeqAIJCRL(Mat A)
 17: {
 19:   Mat_AIJCRL     *aijcrl = (Mat_AIJCRL*) A->spptr;

 21:   /* Free everything in the Mat_AIJCRL data structure. */
 22:   if (aijcrl) {
 23:     PetscFree2(aijcrl->acols,aijcrl->icols);
 24:   }
 25:   PetscFree(A->spptr);
 26:   PetscObjectChangeTypeName((PetscObject)A, MATSEQAIJ);
 27:   MatDestroy_SeqAIJ(A);
 28:   return(0);
 29: }

 31: PetscErrorCode MatDuplicate_AIJCRL(Mat A, MatDuplicateOption op, Mat *M)
 32: {
 34:   SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Cannot duplicate AIJCRL matrices yet");
 35:   return(0);
 36: }

 40: PetscErrorCode MatSeqAIJCRL_create_aijcrl(Mat A)
 41: {
 42:   Mat_SeqAIJ     *a      = (Mat_SeqAIJ*)(A)->data;
 43:   Mat_AIJCRL     *aijcrl = (Mat_AIJCRL*) A->spptr;
 44:   PetscInt       m       = A->rmap->n; /* Number of rows in the matrix. */
 45:   PetscInt       *aj     = a->j; /* From the CSR representation; points to the beginning  of each row. */
 46:   PetscInt       i, j,rmax = a->rmax,*icols, *ilen = a->ilen;
 47:   MatScalar      *aa = a->a;
 48:   PetscScalar    *acols;

 52:   aijcrl->nz   = a->nz;
 53:   aijcrl->m    = A->rmap->n;
 54:   aijcrl->rmax = rmax;

 56:   PetscFree2(aijcrl->acols,aijcrl->icols);
 57:   PetscMalloc2(rmax*m,PetscScalar,&aijcrl->acols,rmax*m,PetscInt,&aijcrl->icols);
 58:   acols = aijcrl->acols;
 59:   icols = aijcrl->icols;
 60:   for (i=0; i<m; i++) {
 61:     for (j=0; j<ilen[i]; j++) {
 62:       acols[j*m+i] = *aa++;
 63:       icols[j*m+i] = *aj++;
 64:     }
 65:     for (; j<rmax; j++) { /* empty column entries */
 66:       acols[j*m+i] = 0.0;
 67:       icols[j*m+i] = (j) ? icols[(j-1)*m+i] : 0;  /* handle case where row is EMPTY */
 68:     }
 69:   }
 70:   PetscInfo2(A,"Percentage of 0's introduced for vectorized multiply %g. Rmax= %D\n",1.0-((double)a->nz)/((double)(rmax*m)),rmax);
 71:   return(0);
 72: }

 74: extern PetscErrorCode MatAssemblyEnd_SeqAIJ(Mat,MatAssemblyType);

 78: PetscErrorCode MatAssemblyEnd_SeqAIJCRL(Mat A, MatAssemblyType mode)
 79: {
 81:   Mat_SeqAIJ     *a = (Mat_SeqAIJ*)A->data;

 84:   a->inode.use = PETSC_FALSE;

 86:   MatAssemblyEnd_SeqAIJ(A,mode);
 87:   if (mode == MAT_FLUSH_ASSEMBLY) return(0);

 89:   /* Now calculate the permutation and grouping information. */
 90:   MatSeqAIJCRL_create_aijcrl(A);
 91:   return(0);
 92: }

 94: #include <../src/mat/impls/aij/seq/crl/ftn-kernels/fmultcrl.h>

 98: /*
 99:     Shared by both sequential and parallel versions of CRL matrix: MATMPIAIJCRL and MATSEQAIJCRL
100:     - the scatter is used only in the parallel version

102: */
103: PetscErrorCode MatMult_AIJCRL(Mat A,Vec xx,Vec yy)
104: {
105:   Mat_AIJCRL     *aijcrl = (Mat_AIJCRL*) A->spptr;
106:   PetscInt       m       = aijcrl->m; /* Number of rows in the matrix. */
107:   PetscInt       rmax    = aijcrl->rmax,*icols = aijcrl->icols;
108:   PetscScalar    *acols  = aijcrl->acols;
110:   PetscScalar    *x,*y;
111: #if !defined(PETSC_USE_FORTRAN_KERNEL_MULTCRL)
112:   PetscInt i,j,ii;
113: #endif


116: #if defined(PETSC_HAVE_PRAGMA_DISJOINT)
117: #pragma disjoint(*x,*y,*aa)
118: #endif

121:   if (aijcrl->xscat) {
122:     VecCopy(xx,aijcrl->xwork);
123:     /* get remote values needed for local part of multiply */
124:     VecScatterBegin(aijcrl->xscat,xx,aijcrl->fwork,INSERT_VALUES,SCATTER_FORWARD);
125:     VecScatterEnd(aijcrl->xscat,xx,aijcrl->fwork,INSERT_VALUES,SCATTER_FORWARD);
126:     xx   = aijcrl->xwork;
127:   }

129:   VecGetArray(xx,&x);
130:   VecGetArray(yy,&y);

132: #if defined(PETSC_USE_FORTRAN_KERNEL_MULTCRL)
133:   fortranmultcrl_(&m,&rmax,x,y,icols,acols);
134: #else

136:   /* first column */
137:   for (j=0; j<m; j++) y[j] = acols[j]*x[icols[j]];

139:   /* other columns */
140: #if defined(PETSC_HAVE_CRAY_VECTOR)
141: #pragma _CRI preferstream
142: #endif
143:   for (i=1; i<rmax; i++) {
144:     ii = i*m;
145: #if defined(PETSC_HAVE_CRAY_VECTOR)
146: #pragma _CRI prefervector
147: #endif
148:     for (j=0; j<m; j++) y[j] = y[j] + acols[ii+j]*x[icols[ii+j]];
149:   }
150: #if defined(PETSC_HAVE_CRAY_VECTOR)
151: #pragma _CRI ivdep
152: #endif

154: #endif
155:   PetscLogFlops(2.0*aijcrl->nz - m);
156:   VecRestoreArray(xx,&x);
157:   VecRestoreArray(yy,&y);
158:   return(0);
159: }


162: /* MatConvert_SeqAIJ_SeqAIJCRL converts a SeqAIJ matrix into a
163:  * SeqAIJCRL matrix.  This routine is called by the MatCreate_SeqAIJCRL()
164:  * routine, but can also be used to convert an assembled SeqAIJ matrix
165:  * into a SeqAIJCRL one. */
168: PETSC_EXTERN PetscErrorCode MatConvert_SeqAIJ_SeqAIJCRL(Mat A,MatType type,MatReuse reuse,Mat *newmat)
169: {
171:   Mat            B = *newmat;
172:   Mat_AIJCRL     *aijcrl;

175:   if (reuse == MAT_INITIAL_MATRIX) {
176:     MatDuplicate(A,MAT_COPY_VALUES,&B);
177:   }

179:   PetscNewLog(B,Mat_AIJCRL,&aijcrl);
180:   B->spptr = (void*) aijcrl;

182:   /* Set function pointers for methods that we inherit from AIJ but override. */
183:   B->ops->duplicate   = MatDuplicate_AIJCRL;
184:   B->ops->assemblyend = MatAssemblyEnd_SeqAIJCRL;
185:   B->ops->destroy     = MatDestroy_SeqAIJCRL;
186:   B->ops->mult        = MatMult_AIJCRL;

188:   /* If A has already been assembled, compute the permutation. */
189:   if (A->assembled) {
190:     MatSeqAIJCRL_create_aijcrl(B);
191:   }
192:   PetscObjectChangeTypeName((PetscObject)B,MATSEQAIJCRL);
193:   *newmat = B;
194:   return(0);
195: }

199: /*@C
200:    MatCreateSeqAIJCRL - Creates a sparse matrix of type SEQAIJCRL.
201:    This type inherits from AIJ, but stores some additional
202:    information that is used to allow better vectorization of
203:    the matrix-vector product. At the cost of increased storage, the AIJ formatted
204:    matrix can be copied to a format in which pieces of the matrix are
205:    stored in ELLPACK format, allowing the vectorized matrix multiply
206:    routine to use stride-1 memory accesses.  As with the AIJ type, it is
207:    important to preallocate matrix storage in order to get good assembly
208:    performance.

210:    Collective on MPI_Comm

212:    Input Parameters:
213: +  comm - MPI communicator, set to PETSC_COMM_SELF
214: .  m - number of rows
215: .  n - number of columns
216: .  nz - number of nonzeros per row (same for all rows)
217: -  nnz - array containing the number of nonzeros in the various rows
218:          (possibly different for each row) or NULL

220:    Output Parameter:
221: .  A - the matrix

223:    Notes:
224:    If nnz is given then nz is ignored

226:    Level: intermediate

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

230: .seealso: MatCreate(), MatCreateMPIAIJPERM(), MatSetValues()
231: @*/
232: PetscErrorCode  MatCreateSeqAIJCRL(MPI_Comm comm,PetscInt m,PetscInt n,PetscInt nz,const PetscInt nnz[],Mat *A)
233: {

237:   MatCreate(comm,A);
238:   MatSetSizes(*A,m,n,m,n);
239:   MatSetType(*A,MATSEQAIJCRL);
240:   MatSeqAIJSetPreallocation_SeqAIJ(*A,nz,nnz);
241:   return(0);
242: }

246: PETSC_EXTERN PetscErrorCode MatCreate_SeqAIJCRL(Mat A)
247: {

251:   MatSetType(A,MATSEQAIJ);
252:   MatConvert_SeqAIJ_SeqAIJCRL(A,MATSEQAIJCRL,MAT_REUSE_MATRIX,&A);
253:   return(0);
254: }