Actual source code: baijsolvtrannat5.c
1: #include <../src/mat/impls/baij/seq/baij.h>
3: PetscErrorCode MatSolveTranspose_SeqBAIJ_5_NaturalOrdering_inplace(Mat A, Vec bb, Vec xx)
4: {
5: Mat_SeqBAIJ *a = (Mat_SeqBAIJ *)A->data;
6: const PetscInt *diag = a->diag, n = a->mbs, *vi, *ai = a->i, *aj = a->j;
7: PetscInt i, nz, idx, idt, oidx;
8: const MatScalar *aa = a->a, *v;
9: PetscScalar s1, s2, s3, s4, s5, x1, x2, x3, x4, x5, *x;
11: VecCopy(bb, xx);
12: VecGetArray(xx, &x);
14: /* forward solve the U^T */
15: idx = 0;
16: for (i = 0; i < n; i++) {
17: v = aa + 25 * diag[i];
18: /* multiply by the inverse of the block diagonal */
19: x1 = x[idx];
20: x2 = x[1 + idx];
21: x3 = x[2 + idx];
22: x4 = x[3 + idx];
23: x5 = x[4 + idx];
24: s1 = v[0] * x1 + v[1] * x2 + v[2] * x3 + v[3] * x4 + v[4] * x5;
25: s2 = v[5] * x1 + v[6] * x2 + v[7] * x3 + v[8] * x4 + v[9] * x5;
26: s3 = v[10] * x1 + v[11] * x2 + v[12] * x3 + v[13] * x4 + v[14] * x5;
27: s4 = v[15] * x1 + v[16] * x2 + v[17] * x3 + v[18] * x4 + v[19] * x5;
28: s5 = v[20] * x1 + v[21] * x2 + v[22] * x3 + v[23] * x4 + v[24] * x5;
29: v += 25;
31: vi = aj + diag[i] + 1;
32: nz = ai[i + 1] - diag[i] - 1;
33: while (nz--) {
34: oidx = 5 * (*vi++);
35: x[oidx] -= v[0] * s1 + v[1] * s2 + v[2] * s3 + v[3] * s4 + v[4] * s5;
36: x[oidx + 1] -= v[5] * s1 + v[6] * s2 + v[7] * s3 + v[8] * s4 + v[9] * s5;
37: x[oidx + 2] -= v[10] * s1 + v[11] * s2 + v[12] * s3 + v[13] * s4 + v[14] * s5;
38: x[oidx + 3] -= v[15] * s1 + v[16] * s2 + v[17] * s3 + v[18] * s4 + v[19] * s5;
39: x[oidx + 4] -= v[20] * s1 + v[21] * s2 + v[22] * s3 + v[23] * s4 + v[24] * s5;
40: v += 25;
41: }
42: x[idx] = s1;
43: x[1 + idx] = s2;
44: x[2 + idx] = s3;
45: x[3 + idx] = s4;
46: x[4 + idx] = s5;
47: idx += 5;
48: }
49: /* backward solve the L^T */
50: for (i = n - 1; i >= 0; i--) {
51: v = aa + 25 * diag[i] - 25;
52: vi = aj + diag[i] - 1;
53: nz = diag[i] - ai[i];
54: idt = 5 * i;
55: s1 = x[idt];
56: s2 = x[1 + idt];
57: s3 = x[2 + idt];
58: s4 = x[3 + idt];
59: s5 = x[4 + idt];
60: while (nz--) {
61: idx = 5 * (*vi--);
62: x[idx] -= v[0] * s1 + v[1] * s2 + v[2] * s3 + v[3] * s4 + v[4] * s5;
63: x[idx + 1] -= v[5] * s1 + v[6] * s2 + v[7] * s3 + v[8] * s4 + v[9] * s5;
64: x[idx + 2] -= v[10] * s1 + v[11] * s2 + v[12] * s3 + v[13] * s4 + v[14] * s5;
65: x[idx + 3] -= v[15] * s1 + v[16] * s2 + v[17] * s3 + v[18] * s4 + v[19] * s5;
66: x[idx + 4] -= v[20] * s1 + v[21] * s2 + v[22] * s3 + v[23] * s4 + v[24] * s5;
67: v -= 25;
68: }
69: }
70: VecRestoreArray(xx, &x);
71: PetscLogFlops(2.0 * 25 * (a->nz) - 5.0 * A->cmap->n);
72: return 0;
73: }
75: PetscErrorCode MatSolveTranspose_SeqBAIJ_5_NaturalOrdering(Mat A, Vec bb, Vec xx)
76: {
77: Mat_SeqBAIJ *a = (Mat_SeqBAIJ *)A->data;
78: const PetscInt n = a->mbs, *vi, *ai = a->i, *aj = a->j, *diag = a->diag;
79: PetscInt nz, idx, idt, j, i, oidx;
80: const PetscInt bs = A->rmap->bs, bs2 = a->bs2;
81: const MatScalar *aa = a->a, *v;
82: PetscScalar s1, s2, s3, s4, s5, x1, x2, x3, x4, x5, *x;
84: VecCopy(bb, xx);
85: VecGetArray(xx, &x);
87: /* forward solve the U^T */
88: idx = 0;
89: for (i = 0; i < n; i++) {
90: v = aa + bs2 * diag[i];
91: /* multiply by the inverse of the block diagonal */
92: x1 = x[idx];
93: x2 = x[1 + idx];
94: x3 = x[2 + idx];
95: x4 = x[3 + idx];
96: x5 = x[4 + idx];
97: s1 = v[0] * x1 + v[1] * x2 + v[2] * x3 + v[3] * x4 + v[4] * x5;
98: s2 = v[5] * x1 + v[6] * x2 + v[7] * x3 + v[8] * x4 + v[9] * x5;
99: s3 = v[10] * x1 + v[11] * x2 + v[12] * x3 + v[13] * x4 + v[14] * x5;
100: s4 = v[15] * x1 + v[16] * x2 + v[17] * x3 + v[18] * x4 + v[19] * x5;
101: s5 = v[20] * x1 + v[21] * x2 + v[22] * x3 + v[23] * x4 + v[24] * x5;
102: v -= bs2;
104: vi = aj + diag[i] - 1;
105: nz = diag[i] - diag[i + 1] - 1;
106: for (j = 0; j > -nz; j--) {
107: oidx = bs * vi[j];
108: x[oidx] -= v[0] * s1 + v[1] * s2 + v[2] * s3 + v[3] * s4 + v[4] * s5;
109: x[oidx + 1] -= v[5] * s1 + v[6] * s2 + v[7] * s3 + v[8] * s4 + v[9] * s5;
110: x[oidx + 2] -= v[10] * s1 + v[11] * s2 + v[12] * s3 + v[13] * s4 + v[14] * s5;
111: x[oidx + 3] -= v[15] * s1 + v[16] * s2 + v[17] * s3 + v[18] * s4 + v[19] * s5;
112: x[oidx + 4] -= v[20] * s1 + v[21] * s2 + v[22] * s3 + v[23] * s4 + v[24] * s5;
113: v -= bs2;
114: }
115: x[idx] = s1;
116: x[1 + idx] = s2;
117: x[2 + idx] = s3;
118: x[3 + idx] = s4;
119: x[4 + idx] = s5;
120: idx += bs;
121: }
122: /* backward solve the L^T */
123: for (i = n - 1; i >= 0; i--) {
124: v = aa + bs2 * ai[i];
125: vi = aj + ai[i];
126: nz = ai[i + 1] - ai[i];
127: idt = bs * i;
128: s1 = x[idt];
129: s2 = x[1 + idt];
130: s3 = x[2 + idt];
131: s4 = x[3 + idt];
132: s5 = x[4 + idt];
133: for (j = 0; j < nz; j++) {
134: idx = bs * vi[j];
135: x[idx] -= v[0] * s1 + v[1] * s2 + v[2] * s3 + v[3] * s4 + v[4] * s5;
136: x[idx + 1] -= v[5] * s1 + v[6] * s2 + v[7] * s3 + v[8] * s4 + v[9] * s5;
137: x[idx + 2] -= v[10] * s1 + v[11] * s2 + v[12] * s3 + v[13] * s4 + v[14] * s5;
138: x[idx + 3] -= v[15] * s1 + v[16] * s2 + v[17] * s3 + v[18] * s4 + v[19] * s5;
139: x[idx + 4] -= v[20] * s1 + v[21] * s2 + v[22] * s3 + v[23] * s4 + v[24] * s5;
140: v += bs2;
141: }
142: }
143: VecRestoreArray(xx, &x);
144: PetscLogFlops(2.0 * bs2 * (a->nz) - bs * A->cmap->n);
145: return 0;
146: }