Actual source code: baijsolvtran5.c

petsc-3.11.4 2019-09-28
Report Typos and Errors
  1:  #include <../src/mat/impls/baij/seq/baij.h>
  2:  #include <petsc/private/kernels/blockinvert.h>

  4: PetscErrorCode MatSolveTranspose_SeqBAIJ_5_inplace(Mat A,Vec bb,Vec xx)
  5: {
  6:   Mat_SeqBAIJ       *a   =(Mat_SeqBAIJ*)A->data;
  7:   IS                iscol=a->col,isrow=a->row;
  8:   PetscErrorCode    ierr;
  9:   const PetscInt    *r,*c,*rout,*cout;
 10:   const PetscInt    *diag=a->diag,n=a->mbs,*vi,*ai=a->i,*aj=a->j;
 11:   PetscInt          i,nz,idx,idt,ii,ic,ir,oidx;
 12:   const MatScalar   *aa=a->a,*v;
 13:   PetscScalar       s1,s2,s3,s4,s5,x1,x2,x3,x4,x5,*x,*t;
 14:   const PetscScalar *b;

 17:   VecGetArrayRead(bb,&b);
 18:   VecGetArray(xx,&x);
 19:   t    = a->solve_work;

 21:   ISGetIndices(isrow,&rout); r = rout;
 22:   ISGetIndices(iscol,&cout); c = cout;

 24:   /* copy the b into temp work space according to permutation */
 25:   ii = 0;
 26:   for (i=0; i<n; i++) {
 27:     ic      = 5*c[i];
 28:     t[ii]   = b[ic];
 29:     t[ii+1] = b[ic+1];
 30:     t[ii+2] = b[ic+2];
 31:     t[ii+3] = b[ic+3];
 32:     t[ii+4] = b[ic+4];
 33:     ii     += 5;
 34:   }

 36:   /* forward solve the U^T */
 37:   idx = 0;
 38:   for (i=0; i<n; i++) {

 40:     v = aa + 25*diag[i];
 41:     /* multiply by the inverse of the block diagonal */
 42:     x1 = t[idx];   x2 = t[1+idx]; x3    = t[2+idx]; x4 = t[3+idx]; x5 = t[4+idx];
 43:     s1 = v[0]*x1  +  v[1]*x2 +  v[2]*x3 +  v[3]*x4 +  v[4]*x5;
 44:     s2 = v[5]*x1  +  v[6]*x2 +  v[7]*x3 +  v[8]*x4 +  v[9]*x5;
 45:     s3 = v[10]*x1 + v[11]*x2 + v[12]*x3 + v[13]*x4 + v[14]*x5;
 46:     s4 = v[15]*x1 + v[16]*x2 + v[17]*x3 + v[18]*x4 + v[19]*x5;
 47:     s5 = v[20]*x1 + v[21]*x2 + v[22]*x3 + v[23]*x4 + v[24]*x5;
 48:     v += 25;

 50:     vi = aj + diag[i] + 1;
 51:     nz = ai[i+1] - diag[i] - 1;
 52:     while (nz--) {
 53:       oidx       = 5*(*vi++);
 54:       t[oidx]   -= v[0]*s1  +  v[1]*s2 +  v[2]*s3 +  v[3]*s4 +  v[4]*s5;
 55:       t[oidx+1] -= v[5]*s1  +  v[6]*s2 +  v[7]*s3 +  v[8]*s4 +  v[9]*s5;
 56:       t[oidx+2] -= v[10]*s1 + v[11]*s2 + v[12]*s3 + v[13]*s4 + v[14]*s5;
 57:       t[oidx+3] -= v[15]*s1 + v[16]*s2 + v[17]*s3 + v[18]*s4 + v[19]*s5;
 58:       t[oidx+4] -= v[20]*s1 + v[21]*s2 + v[22]*s3 + v[23]*s4 + v[24]*s5;
 59:       v         += 25;
 60:     }
 61:     t[idx] = s1;t[1+idx] = s2; t[2+idx] = s3;t[3+idx] = s4; t[4+idx] = s5;
 62:     idx   += 5;
 63:   }
 64:   /* backward solve the L^T */
 65:   for (i=n-1; i>=0; i--) {
 66:     v   = aa + 25*diag[i] - 25;
 67:     vi  = aj + diag[i] - 1;
 68:     nz  = diag[i] - ai[i];
 69:     idt = 5*i;
 70:     s1  = t[idt];  s2 = t[1+idt]; s3 = t[2+idt];s4 = t[3+idt]; s5 = t[4+idt];
 71:     while (nz--) {
 72:       idx       = 5*(*vi--);
 73:       t[idx]   -=  v[0]*s1 +  v[1]*s2 +  v[2]*s3 +  v[3]*s4 +  v[4]*s5;
 74:       t[idx+1] -=  v[5]*s1 +  v[6]*s2 +  v[7]*s3 +  v[8]*s4 +  v[9]*s5;
 75:       t[idx+2] -= v[10]*s1 + v[11]*s2 + v[12]*s3 + v[13]*s4 + v[14]*s5;
 76:       t[idx+3] -= v[15]*s1 + v[16]*s2 + v[17]*s3 + v[18]*s4 + v[19]*s5;
 77:       t[idx+4] -= v[20]*s1 + v[21]*s2 + v[22]*s3 + v[23]*s4 + v[24]*s5;
 78:       v        -= 25;
 79:     }
 80:   }

 82:   /* copy t into x according to permutation */
 83:   ii = 0;
 84:   for (i=0; i<n; i++) {
 85:     ir      = 5*r[i];
 86:     x[ir]   = t[ii];
 87:     x[ir+1] = t[ii+1];
 88:     x[ir+2] = t[ii+2];
 89:     x[ir+3] = t[ii+3];
 90:     x[ir+4] = t[ii+4];
 91:     ii     += 5;
 92:   }

 94:   ISRestoreIndices(isrow,&rout);
 95:   ISRestoreIndices(iscol,&cout);
 96:   VecRestoreArrayRead(bb,&b);
 97:   VecRestoreArray(xx,&x);
 98:   PetscLogFlops(2.0*25*(a->nz) - 5.0*A->cmap->n);
 99:   return(0);
100: }

102: PetscErrorCode MatSolveTranspose_SeqBAIJ_5(Mat A,Vec bb,Vec xx)
103: {
104:   Mat_SeqBAIJ       *a=(Mat_SeqBAIJ*)A->data;
105:   PetscErrorCode    ierr;
106:   IS                iscol=a->col,isrow=a->row;
107:   const PetscInt    n    =a->mbs,*vi,*ai=a->i,*aj=a->j,*diag=a->diag;
108:   const PetscInt    *r,*c,*rout,*cout;
109:   PetscInt          nz,idx,idt,j,i,oidx,ii,ic,ir;
110:   const PetscInt    bs =A->rmap->bs,bs2=a->bs2;
111:   const MatScalar   *aa=a->a,*v;
112:   PetscScalar       s1,s2,s3,s4,s5,x1,x2,x3,x4,x5,*x,*t;
113:   const PetscScalar *b;

116:   VecGetArrayRead(bb,&b);
117:   VecGetArray(xx,&x);
118:   t    = a->solve_work;

120:   ISGetIndices(isrow,&rout); r = rout;
121:   ISGetIndices(iscol,&cout); c = cout;

123:   /* copy b into temp work space according to permutation */
124:   for (i=0; i<n; i++) {
125:     ii      = bs*i; ic = bs*c[i];
126:     t[ii]   = b[ic]; t[ii+1] = b[ic+1]; t[ii+2] = b[ic+2]; t[ii+3] = b[ic+3];
127:     t[ii+4] = b[ic+4];
128:   }

130:   /* forward solve the U^T */
131:   idx = 0;
132:   for (i=0; i<n; i++) {
133:     v = aa + bs2*diag[i];
134:     /* multiply by the inverse of the block diagonal */
135:     x1 = t[idx];   x2 = t[1+idx]; x3    = t[2+idx]; x4 = t[3+idx]; x5 = t[4+idx];
136:     s1 = v[0]*x1  +  v[1]*x2 +  v[2]*x3 +  v[3]*x4 +  v[4]*x5;
137:     s2 = v[5]*x1  +  v[6]*x2 +  v[7]*x3 +  v[8]*x4 +  v[9]*x5;
138:     s3 = v[10]*x1 + v[11]*x2 + v[12]*x3 + v[13]*x4 + v[14]*x5;
139:     s4 = v[15]*x1 + v[16]*x2 + v[17]*x3 + v[18]*x4 + v[19]*x5;
140:     s5 = v[20]*x1 + v[21]*x2 + v[22]*x3 + v[23]*x4 + v[24]*x5;
141:     v -= bs2;

143:     vi = aj + diag[i] - 1;
144:     nz = diag[i] - diag[i+1] - 1;
145:     for (j=0; j>-nz; j--) {
146:       oidx       = bs*vi[j];
147:       t[oidx]   -= v[0]*s1  +  v[1]*s2 +  v[2]*s3 +  v[3]*s4 +  v[4]*s5;
148:       t[oidx+1] -= v[5]*s1  +  v[6]*s2 +  v[7]*s3 +  v[8]*s4 +  v[9]*s5;
149:       t[oidx+2] -= v[10]*s1 + v[11]*s2 + v[12]*s3 + v[13]*s4 + v[14]*s5;
150:       t[oidx+3] -= v[15]*s1 + v[16]*s2 + v[17]*s3 + v[18]*s4 + v[19]*s5;
151:       t[oidx+4] -= v[20]*s1 + v[21]*s2 + v[22]*s3 + v[23]*s4 + v[24]*s5;
152:       v         -= bs2;
153:     }
154:     t[idx] = s1;t[1+idx] = s2;  t[2+idx] = s3;  t[3+idx] = s4; t[4+idx] =s5;
155:     idx   += bs;
156:   }
157:   /* backward solve the L^T */
158:   for (i=n-1; i>=0; i--) {
159:     v   = aa + bs2*ai[i];
160:     vi  = aj + ai[i];
161:     nz  = ai[i+1] - ai[i];
162:     idt = bs*i;
163:     s1  = t[idt];  s2 = t[1+idt];  s3 = t[2+idt];  s4 = t[3+idt]; s5 = t[4+idt];
164:     for (j=0; j<nz; j++) {
165:       idx       = bs*vi[j];
166:       t[idx]   -= v[0]*s1  +  v[1]*s2 +  v[2]*s3 +  v[3]*s4 +  v[4]*s5;
167:       t[idx+1] -= v[5]*s1  +  v[6]*s2 +  v[7]*s3 +  v[8]*s4 +  v[9]*s5;
168:       t[idx+2] -= v[10]*s1 + v[11]*s2 + v[12]*s3 + v[13]*s4 + v[14]*s5;
169:       t[idx+3] -= v[15]*s1 + v[16]*s2 + v[17]*s3 + v[18]*s4 + v[19]*s5;
170:       t[idx+4] -= v[20]*s1 + v[21]*s2 + v[22]*s3 + v[23]*s4 + v[24]*s5;
171:       v        += bs2;
172:     }
173:   }

175:   /* copy t into x according to permutation */
176:   for (i=0; i<n; i++) {
177:     ii      = bs*i;  ir = bs*r[i];
178:     x[ir]   = t[ii];  x[ir+1] = t[ii+1]; x[ir+2] = t[ii+2];  x[ir+3] = t[ii+3];
179:     x[ir+4] = t[ii+4];
180:   }

182:   ISRestoreIndices(isrow,&rout);
183:   ISRestoreIndices(iscol,&cout);
184:   VecRestoreArrayRead(bb,&b);
185:   VecRestoreArray(xx,&x);
186:   PetscLogFlops(2.0*bs2*(a->nz) - bs*A->cmap->n);
187:   return(0);
188: }