Actual source code: baij.c
petsc-3.5.4 2015-05-23
2: /*
3: Defines the basic matrix operations for the BAIJ (compressed row)
4: matrix storage format.
5: */
6: #include <../src/mat/impls/baij/seq/baij.h> /*I "petscmat.h" I*/
7: #include <petscblaslapack.h>
8: #include <petsc-private/kernels/blockinvert.h>
9: #include <petsc-private/kernels/blockmatmult.h>
13: PetscErrorCode MatInvertBlockDiagonal_SeqBAIJ(Mat A,const PetscScalar **values)
14: {
15: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*) A->data;
17: PetscInt *diag_offset,i,bs = A->rmap->bs,mbs = a->mbs,ipvt[5],bs2 = bs*bs,*v_pivots;
18: MatScalar *v = a->a,*odiag,*diag,*mdiag,work[25],*v_work;
19: PetscReal shift = 0.0;
22: if (a->idiagvalid) {
23: if (values) *values = a->idiag;
24: return(0);
25: }
26: MatMarkDiagonal_SeqBAIJ(A);
27: diag_offset = a->diag;
28: if (!a->idiag) {
29: PetscMalloc1(2*bs2*mbs,&a->idiag);
30: PetscLogObjectMemory((PetscObject)A,2*bs2*mbs*sizeof(PetscScalar));
31: }
32: diag = a->idiag;
33: mdiag = a->idiag+bs2*mbs;
34: if (values) *values = a->idiag;
35: /* factor and invert each block */
36: switch (bs) {
37: case 1:
38: for (i=0; i<mbs; i++) {
39: odiag = v + 1*diag_offset[i];
40: diag[0] = odiag[0];
41: mdiag[0] = odiag[0];
42: diag[0] = (PetscScalar)1.0 / (diag[0] + shift);
43: diag += 1;
44: mdiag += 1;
45: }
46: break;
47: case 2:
48: for (i=0; i<mbs; i++) {
49: odiag = v + 4*diag_offset[i];
50: diag[0] = odiag[0]; diag[1] = odiag[1]; diag[2] = odiag[2]; diag[3] = odiag[3];
51: mdiag[0] = odiag[0]; mdiag[1] = odiag[1]; mdiag[2] = odiag[2]; mdiag[3] = odiag[3];
52: PetscKernel_A_gets_inverse_A_2(diag,shift);
53: diag += 4;
54: mdiag += 4;
55: }
56: break;
57: case 3:
58: for (i=0; i<mbs; i++) {
59: odiag = v + 9*diag_offset[i];
60: diag[0] = odiag[0]; diag[1] = odiag[1]; diag[2] = odiag[2]; diag[3] = odiag[3];
61: diag[4] = odiag[4]; diag[5] = odiag[5]; diag[6] = odiag[6]; diag[7] = odiag[7];
62: diag[8] = odiag[8];
63: mdiag[0] = odiag[0]; mdiag[1] = odiag[1]; mdiag[2] = odiag[2]; mdiag[3] = odiag[3];
64: mdiag[4] = odiag[4]; mdiag[5] = odiag[5]; mdiag[6] = odiag[6]; mdiag[7] = odiag[7];
65: mdiag[8] = odiag[8];
66: PetscKernel_A_gets_inverse_A_3(diag,shift);
67: diag += 9;
68: mdiag += 9;
69: }
70: break;
71: case 4:
72: for (i=0; i<mbs; i++) {
73: odiag = v + 16*diag_offset[i];
74: PetscMemcpy(diag,odiag,16*sizeof(PetscScalar));
75: PetscMemcpy(mdiag,odiag,16*sizeof(PetscScalar));
76: PetscKernel_A_gets_inverse_A_4(diag,shift);
77: diag += 16;
78: mdiag += 16;
79: }
80: break;
81: case 5:
82: for (i=0; i<mbs; i++) {
83: odiag = v + 25*diag_offset[i];
84: PetscMemcpy(diag,odiag,25*sizeof(PetscScalar));
85: PetscMemcpy(mdiag,odiag,25*sizeof(PetscScalar));
86: PetscKernel_A_gets_inverse_A_5(diag,ipvt,work,shift);
87: diag += 25;
88: mdiag += 25;
89: }
90: break;
91: case 6:
92: for (i=0; i<mbs; i++) {
93: odiag = v + 36*diag_offset[i];
94: PetscMemcpy(diag,odiag,36*sizeof(PetscScalar));
95: PetscMemcpy(mdiag,odiag,36*sizeof(PetscScalar));
96: PetscKernel_A_gets_inverse_A_6(diag,shift);
97: diag += 36;
98: mdiag += 36;
99: }
100: break;
101: case 7:
102: for (i=0; i<mbs; i++) {
103: odiag = v + 49*diag_offset[i];
104: PetscMemcpy(diag,odiag,49*sizeof(PetscScalar));
105: PetscMemcpy(mdiag,odiag,49*sizeof(PetscScalar));
106: PetscKernel_A_gets_inverse_A_7(diag,shift);
107: diag += 49;
108: mdiag += 49;
109: }
110: break;
111: default:
112: PetscMalloc2(bs,&v_work,bs,&v_pivots);
113: for (i=0; i<mbs; i++) {
114: odiag = v + bs2*diag_offset[i];
115: PetscMemcpy(diag,odiag,bs2*sizeof(PetscScalar));
116: PetscMemcpy(mdiag,odiag,bs2*sizeof(PetscScalar));
117: PetscKernel_A_gets_inverse_A(bs,diag,v_pivots,v_work);
118: diag += bs2;
119: mdiag += bs2;
120: }
121: PetscFree2(v_work,v_pivots);
122: }
123: a->idiagvalid = PETSC_TRUE;
124: return(0);
125: }
129: PetscErrorCode MatSOR_SeqBAIJ(Mat A,Vec bb,PetscReal omega,MatSORType flag,PetscReal fshift,PetscInt its,PetscInt lits,Vec xx)
130: {
131: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
132: PetscScalar *x,*work,*w,*workt,*t;
133: const MatScalar *v,*aa = a->a, *idiag;
134: const PetscScalar *b,*xb;
135: PetscScalar s[7], xw[7]={0}; /* avoid some compilers thinking xw is uninitialized */
136: PetscErrorCode ierr;
137: PetscInt m = a->mbs,i,i2,nz,bs = A->rmap->bs,bs2 = bs*bs,k,j,idx,it;
138: const PetscInt *diag,*ai = a->i,*aj = a->j,*vi;
141: its = its*lits;
142: if (flag & SOR_EISENSTAT) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"No support yet for Eisenstat");
143: if (its <= 0) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Relaxation requires global its %D and local its %D both positive",its,lits);
144: if (fshift) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Sorry, no support for diagonal shift");
145: if (omega != 1.0) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Sorry, no support for non-trivial relaxation factor");
146: if ((flag & SOR_APPLY_UPPER) || (flag & SOR_APPLY_LOWER)) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Sorry, no support for applying upper or lower triangular parts");
148: if (!a->idiagvalid) {MatInvertBlockDiagonal(A,NULL);}
150: if (!m) return(0);
151: diag = a->diag;
152: idiag = a->idiag;
153: k = PetscMax(A->rmap->n,A->cmap->n);
154: if (!a->mult_work) {
155: PetscMalloc1(k+1,&a->mult_work);
156: }
157: if (!a->sor_workt) {
158: PetscMalloc1(k,&a->sor_workt);
159: }
160: if (!a->sor_work) {
161: PetscMalloc1(bs,&a->sor_work);
162: }
163: work = a->mult_work;
164: t = a->sor_workt;
165: w = a->sor_work;
167: VecGetArray(xx,&x);
168: VecGetArrayRead(bb,&b);
170: if (flag & SOR_ZERO_INITIAL_GUESS) {
171: if (flag & SOR_FORWARD_SWEEP || flag & SOR_LOCAL_FORWARD_SWEEP) {
172: switch (bs) {
173: case 1:
174: PetscKernel_v_gets_A_times_w_1(x,idiag,b);
175: t[0] = b[0];
176: i2 = 1;
177: idiag += 1;
178: for (i=1; i<m; i++) {
179: v = aa + ai[i];
180: vi = aj + ai[i];
181: nz = diag[i] - ai[i];
182: s[0] = b[i2];
183: for (j=0; j<nz; j++) {
184: xw[0] = x[vi[j]];
185: PetscKernel_v_gets_v_minus_A_times_w_1(s,(v+j),xw);
186: }
187: t[i2] = s[0];
188: PetscKernel_v_gets_A_times_w_1(xw,idiag,s);
189: x[i2] = xw[0];
190: idiag += 1;
191: i2 += 1;
192: }
193: break;
194: case 2:
195: PetscKernel_v_gets_A_times_w_2(x,idiag,b);
196: t[0] = b[0]; t[1] = b[1];
197: i2 = 2;
198: idiag += 4;
199: for (i=1; i<m; i++) {
200: v = aa + 4*ai[i];
201: vi = aj + ai[i];
202: nz = diag[i] - ai[i];
203: s[0] = b[i2]; s[1] = b[i2+1];
204: for (j=0; j<nz; j++) {
205: idx = 2*vi[j];
206: it = 4*j;
207: xw[0] = x[idx]; xw[1] = x[1+idx];
208: PetscKernel_v_gets_v_minus_A_times_w_2(s,(v+it),xw);
209: }
210: t[i2] = s[0]; t[i2+1] = s[1];
211: PetscKernel_v_gets_A_times_w_2(xw,idiag,s);
212: x[i2] = xw[0]; x[i2+1] = xw[1];
213: idiag += 4;
214: i2 += 2;
215: }
216: break;
217: case 3:
218: PetscKernel_v_gets_A_times_w_3(x,idiag,b);
219: t[0] = b[0]; t[1] = b[1]; t[2] = b[2];
220: i2 = 3;
221: idiag += 9;
222: for (i=1; i<m; i++) {
223: v = aa + 9*ai[i];
224: vi = aj + ai[i];
225: nz = diag[i] - ai[i];
226: s[0] = b[i2]; s[1] = b[i2+1]; s[2] = b[i2+2];
227: while (nz--) {
228: idx = 3*(*vi++);
229: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx];
230: PetscKernel_v_gets_v_minus_A_times_w_3(s,v,xw);
231: v += 9;
232: }
233: t[i2] = s[0]; t[i2+1] = s[1]; t[i2+2] = s[2];
234: PetscKernel_v_gets_A_times_w_3(xw,idiag,s);
235: x[i2] = xw[0]; x[i2+1] = xw[1]; x[i2+2] = xw[2];
236: idiag += 9;
237: i2 += 3;
238: }
239: break;
240: case 4:
241: PetscKernel_v_gets_A_times_w_4(x,idiag,b);
242: t[0] = b[0]; t[1] = b[1]; t[2] = b[2]; t[3] = b[3];
243: i2 = 4;
244: idiag += 16;
245: for (i=1; i<m; i++) {
246: v = aa + 16*ai[i];
247: vi = aj + ai[i];
248: nz = diag[i] - ai[i];
249: s[0] = b[i2]; s[1] = b[i2+1]; s[2] = b[i2+2]; s[3] = b[i2+3];
250: while (nz--) {
251: idx = 4*(*vi++);
252: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx]; xw[3] = x[3+idx];
253: PetscKernel_v_gets_v_minus_A_times_w_4(s,v,xw);
254: v += 16;
255: }
256: t[i2] = s[0]; t[i2+1] = s[1]; t[i2+2] = s[2]; t[i2 + 3] = s[3];
257: PetscKernel_v_gets_A_times_w_4(xw,idiag,s);
258: x[i2] = xw[0]; x[i2+1] = xw[1]; x[i2+2] = xw[2]; x[i2+3] = xw[3];
259: idiag += 16;
260: i2 += 4;
261: }
262: break;
263: case 5:
264: PetscKernel_v_gets_A_times_w_5(x,idiag,b);
265: t[0] = b[0]; t[1] = b[1]; t[2] = b[2]; t[3] = b[3]; t[4] = b[4];
266: i2 = 5;
267: idiag += 25;
268: for (i=1; i<m; i++) {
269: v = aa + 25*ai[i];
270: vi = aj + ai[i];
271: nz = diag[i] - ai[i];
272: s[0] = b[i2]; s[1] = b[i2+1]; s[2] = b[i2+2]; s[3] = b[i2+3]; s[4] = b[i2+4];
273: while (nz--) {
274: idx = 5*(*vi++);
275: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx]; xw[3] = x[3+idx]; xw[4] = x[4+idx];
276: PetscKernel_v_gets_v_minus_A_times_w_5(s,v,xw);
277: v += 25;
278: }
279: t[i2] = s[0]; t[i2+1] = s[1]; t[i2+2] = s[2]; t[i2+3] = s[3]; t[i2+4] = s[4];
280: PetscKernel_v_gets_A_times_w_5(xw,idiag,s);
281: x[i2] = xw[0]; x[i2+1] = xw[1]; x[i2+2] = xw[2]; x[i2+3] = xw[3]; x[i2+4] = xw[4];
282: idiag += 25;
283: i2 += 5;
284: }
285: break;
286: case 6:
287: PetscKernel_v_gets_A_times_w_6(x,idiag,b);
288: t[0] = b[0]; t[1] = b[1]; t[2] = b[2]; t[3] = b[3]; t[4] = b[4]; t[5] = b[5];
289: i2 = 6;
290: idiag += 36;
291: for (i=1; i<m; i++) {
292: v = aa + 36*ai[i];
293: vi = aj + ai[i];
294: nz = diag[i] - ai[i];
295: s[0] = b[i2]; s[1] = b[i2+1]; s[2] = b[i2+2]; s[3] = b[i2+3]; s[4] = b[i2+4]; s[5] = b[i2+5];
296: while (nz--) {
297: idx = 6*(*vi++);
298: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx];
299: xw[3] = x[3+idx]; xw[4] = x[4+idx]; xw[5] = x[5+idx];
300: PetscKernel_v_gets_v_minus_A_times_w_6(s,v,xw);
301: v += 36;
302: }
303: t[i2] = s[0]; t[i2+1] = s[1]; t[i2+2] = s[2];
304: t[i2+3] = s[3]; t[i2+4] = s[4]; t[i2+5] = s[5];
305: PetscKernel_v_gets_A_times_w_6(xw,idiag,s);
306: x[i2] = xw[0]; x[i2+1] = xw[1]; x[i2+2] = xw[2]; x[i2+3] = xw[3]; x[i2+4] = xw[4]; x[i2+5] = xw[5];
307: idiag += 36;
308: i2 += 6;
309: }
310: break;
311: case 7:
312: PetscKernel_v_gets_A_times_w_7(x,idiag,b);
313: t[0] = b[0]; t[1] = b[1]; t[2] = b[2];
314: t[3] = b[3]; t[4] = b[4]; t[5] = b[5]; t[6] = b[6];
315: i2 = 7;
316: idiag += 49;
317: for (i=1; i<m; i++) {
318: v = aa + 49*ai[i];
319: vi = aj + ai[i];
320: nz = diag[i] - ai[i];
321: s[0] = b[i2]; s[1] = b[i2+1]; s[2] = b[i2+2];
322: s[3] = b[i2+3]; s[4] = b[i2+4]; s[5] = b[i2+5]; s[6] = b[i2+6];
323: while (nz--) {
324: idx = 7*(*vi++);
325: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx];
326: xw[3] = x[3+idx]; xw[4] = x[4+idx]; xw[5] = x[5+idx]; xw[6] = x[6+idx];
327: PetscKernel_v_gets_v_minus_A_times_w_7(s,v,xw);
328: v += 49;
329: }
330: t[i2] = s[0]; t[i2+1] = s[1]; t[i2+2] = s[2];
331: t[i2+3] = s[3]; t[i2+4] = s[4]; t[i2+5] = s[5]; t[i2+6] = s[6];
332: PetscKernel_v_gets_A_times_w_7(xw,idiag,s);
333: x[i2] = xw[0]; x[i2+1] = xw[1]; x[i2+2] = xw[2];
334: x[i2+3] = xw[3]; x[i2+4] = xw[4]; x[i2+5] = xw[5]; x[i2+6] = xw[6];
335: idiag += 49;
336: i2 += 7;
337: }
338: break;
339: default:
340: PetscKernel_w_gets_Ar_times_v(bs,bs,b,idiag,x);
341: PetscMemcpy(t,b,bs*sizeof(PetscScalar));
342: i2 = bs;
343: idiag += bs2;
344: for (i=1; i<m; i++) {
345: v = aa + bs2*ai[i];
346: vi = aj + ai[i];
347: nz = diag[i] - ai[i];
349: PetscMemcpy(w,b+i2,bs*sizeof(PetscScalar));
350: /* copy all rows of x that are needed into contiguous space */
351: workt = work;
352: for (j=0; j<nz; j++) {
353: PetscMemcpy(workt,x + bs*(*vi++),bs*sizeof(PetscScalar));
354: workt += bs;
355: }
356: PetscKernel_w_gets_w_minus_Ar_times_v(bs,bs*nz,w,v,work);
357: PetscMemcpy(t+i2,w,bs*sizeof(PetscScalar));
358: PetscKernel_w_gets_Ar_times_v(bs,bs,w,idiag,x+i2);
360: idiag += bs2;
361: i2 += bs;
362: }
363: break;
364: }
365: /* for logging purposes assume number of nonzero in lower half is 1/2 of total */
366: PetscLogFlops(1.0*bs2*a->nz);
367: xb = t;
368: }
369: else xb = b;
370: if (flag & SOR_BACKWARD_SWEEP || flag & SOR_LOCAL_BACKWARD_SWEEP) {
371: idiag = a->idiag+bs2*(a->mbs-1);
372: i2 = bs * (m-1);
373: switch (bs) {
374: case 1:
375: s[0] = xb[i2];
376: PetscKernel_v_gets_A_times_w_1(xw,idiag,s);
377: x[i2] = xw[0];
378: i2 -= 1;
379: for (i=m-2; i>=0; i--) {
380: v = aa + (diag[i]+1);
381: vi = aj + diag[i] + 1;
382: nz = ai[i+1] - diag[i] - 1;
383: s[0] = xb[i2];
384: for (j=0; j<nz; j++) {
385: xw[0] = x[vi[j]];
386: PetscKernel_v_gets_v_minus_A_times_w_1(s,(v+j),xw);
387: }
388: PetscKernel_v_gets_A_times_w_1(xw,idiag,s);
389: x[i2] = xw[0];
390: idiag -= 1;
391: i2 -= 1;
392: }
393: break;
394: case 2:
395: s[0] = xb[i2]; s[1] = xb[i2+1];
396: PetscKernel_v_gets_A_times_w_2(xw,idiag,s);
397: x[i2] = xw[0]; x[i2+1] = xw[1];
398: i2 -= 2;
399: idiag -= 4;
400: for (i=m-2; i>=0; i--) {
401: v = aa + 4*(diag[i] + 1);
402: vi = aj + diag[i] + 1;
403: nz = ai[i+1] - diag[i] - 1;
404: s[0] = xb[i2]; s[1] = xb[i2+1];
405: for (j=0; j<nz; j++) {
406: idx = 2*vi[j];
407: it = 4*j;
408: xw[0] = x[idx]; xw[1] = x[1+idx];
409: PetscKernel_v_gets_v_minus_A_times_w_2(s,(v+it),xw);
410: }
411: PetscKernel_v_gets_A_times_w_2(xw,idiag,s);
412: x[i2] = xw[0]; x[i2+1] = xw[1];
413: idiag -= 4;
414: i2 -= 2;
415: }
416: break;
417: case 3:
418: s[0] = xb[i2]; s[1] = xb[i2+1]; s[2] = xb[i2+2];
419: PetscKernel_v_gets_A_times_w_3(xw,idiag,s);
420: x[i2] = xw[0]; x[i2+1] = xw[1]; x[i2+2] = xw[2];
421: i2 -= 3;
422: idiag -= 9;
423: for (i=m-2; i>=0; i--) {
424: v = aa + 9*(diag[i]+1);
425: vi = aj + diag[i] + 1;
426: nz = ai[i+1] - diag[i] - 1;
427: s[0] = xb[i2]; s[1] = xb[i2+1]; s[2] = xb[i2+2];
428: while (nz--) {
429: idx = 3*(*vi++);
430: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx];
431: PetscKernel_v_gets_v_minus_A_times_w_3(s,v,xw);
432: v += 9;
433: }
434: PetscKernel_v_gets_A_times_w_3(xw,idiag,s);
435: x[i2] = xw[0]; x[i2+1] = xw[1]; x[i2+2] = xw[2];
436: idiag -= 9;
437: i2 -= 3;
438: }
439: break;
440: case 4:
441: s[0] = xb[i2]; s[1] = xb[i2+1]; s[2] = xb[i2+2]; s[3] = xb[i2+3];
442: PetscKernel_v_gets_A_times_w_4(xw,idiag,s);
443: x[i2] = xw[0]; x[i2+1] = xw[1]; x[i2+2] = xw[2]; x[i2+3] = xw[3];
444: i2 -= 4;
445: idiag -= 16;
446: for (i=m-2; i>=0; i--) {
447: v = aa + 16*(diag[i]+1);
448: vi = aj + diag[i] + 1;
449: nz = ai[i+1] - diag[i] - 1;
450: s[0] = xb[i2]; s[1] = xb[i2+1]; s[2] = xb[i2+2]; s[3] = xb[i2+3];
451: while (nz--) {
452: idx = 4*(*vi++);
453: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx]; xw[3] = x[3+idx];
454: PetscKernel_v_gets_v_minus_A_times_w_4(s,v,xw);
455: v += 16;
456: }
457: PetscKernel_v_gets_A_times_w_4(xw,idiag,s);
458: x[i2] = xw[0]; x[i2+1] = xw[1]; x[i2+2] = xw[2]; x[i2+3] = xw[3];
459: idiag -= 16;
460: i2 -= 4;
461: }
462: break;
463: case 5:
464: s[0] = xb[i2]; s[1] = xb[i2+1]; s[2] = xb[i2+2]; s[3] = xb[i2+3]; s[4] = xb[i2+4];
465: PetscKernel_v_gets_A_times_w_5(xw,idiag,s);
466: x[i2] = xw[0]; x[i2+1] = xw[1]; x[i2+2] = xw[2]; x[i2+3] = xw[3]; x[i2+4] = xw[4];
467: i2 -= 5;
468: idiag -= 25;
469: for (i=m-2; i>=0; i--) {
470: v = aa + 25*(diag[i]+1);
471: vi = aj + diag[i] + 1;
472: nz = ai[i+1] - diag[i] - 1;
473: s[0] = xb[i2]; s[1] = xb[i2+1]; s[2] = xb[i2+2]; s[3] = xb[i2+3]; s[4] = xb[i2+4];
474: while (nz--) {
475: idx = 5*(*vi++);
476: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx]; xw[3] = x[3+idx]; xw[4] = x[4+idx];
477: PetscKernel_v_gets_v_minus_A_times_w_5(s,v,xw);
478: v += 25;
479: }
480: PetscKernel_v_gets_A_times_w_5(xw,idiag,s);
481: x[i2] = xw[0]; x[i2+1] = xw[1]; x[i2+2] = xw[2]; x[i2+3] = xw[3]; x[i2+4] = xw[4];
482: idiag -= 25;
483: i2 -= 5;
484: }
485: break;
486: case 6:
487: s[0] = xb[i2]; s[1] = xb[i2+1]; s[2] = xb[i2+2]; s[3] = xb[i2+3]; s[4] = xb[i2+4]; s[5] = xb[i2+5];
488: PetscKernel_v_gets_A_times_w_6(xw,idiag,s);
489: x[i2] = xw[0]; x[i2+1] = xw[1]; x[i2+2] = xw[2]; x[i2+3] = xw[3]; x[i2+4] = xw[4]; x[i2+5] = xw[5];
490: i2 -= 6;
491: idiag -= 36;
492: for (i=m-2; i>=0; i--) {
493: v = aa + 36*(diag[i]+1);
494: vi = aj + diag[i] + 1;
495: nz = ai[i+1] - diag[i] - 1;
496: s[0] = xb[i2]; s[1] = xb[i2+1]; s[2] = xb[i2+2]; s[3] = xb[i2+3]; s[4] = xb[i2+4]; s[5] = xb[i2+5];
497: while (nz--) {
498: idx = 6*(*vi++);
499: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx];
500: xw[3] = x[3+idx]; xw[4] = x[4+idx]; xw[5] = x[5+idx];
501: PetscKernel_v_gets_v_minus_A_times_w_6(s,v,xw);
502: v += 36;
503: }
504: PetscKernel_v_gets_A_times_w_6(xw,idiag,s);
505: x[i2] = xw[0]; x[i2+1] = xw[1]; x[i2+2] = xw[2]; x[i2+3] = xw[3]; x[i2+4] = xw[4]; x[i2+5] = xw[5];
506: idiag -= 36;
507: i2 -= 6;
508: }
509: break;
510: case 7:
511: s[0] = xb[i2]; s[1] = xb[i2+1]; s[2] = xb[i2+2];
512: s[3] = xb[i2+3]; s[4] = xb[i2+4]; s[5] = xb[i2+5]; s[6] = xb[i2+6];
513: PetscKernel_v_gets_A_times_w_7(x,idiag,b);
514: x[i2] = xw[0]; x[i2+1] = xw[1]; x[i2+2] = xw[2];
515: x[i2+3] = xw[3]; x[i2+4] = xw[4]; x[i2+5] = xw[5]; x[i2+6] = xw[6];
516: i2 -= 7;
517: idiag -= 49;
518: for (i=m-2; i>=0; i--) {
519: v = aa + 49*(diag[i]+1);
520: vi = aj + diag[i] + 1;
521: nz = ai[i+1] - diag[i] - 1;
522: s[0] = xb[i2]; s[1] = xb[i2+1]; s[2] = xb[i2+2];
523: s[3] = xb[i2+3]; s[4] = xb[i2+4]; s[5] = xb[i2+5]; s[6] = xb[i2+6];
524: while (nz--) {
525: idx = 7*(*vi++);
526: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx];
527: xw[3] = x[3+idx]; xw[4] = x[4+idx]; xw[5] = x[5+idx]; xw[6] = x[6+idx];
528: PetscKernel_v_gets_v_minus_A_times_w_7(s,v,xw);
529: v += 49;
530: }
531: PetscKernel_v_gets_A_times_w_7(xw,idiag,s);
532: x[i2] = xw[0]; x[i2+1] = xw[1]; x[i2+2] = xw[2];
533: x[i2+3] = xw[3]; x[i2+4] = xw[4]; x[i2+5] = xw[5]; x[i2+6] = xw[6];
534: idiag -= 49;
535: i2 -= 7;
536: }
537: break;
538: default:
539: PetscMemcpy(w,xb+i2,bs*sizeof(PetscScalar));
540: PetscKernel_w_gets_Ar_times_v(bs,bs,w,idiag,x+i2);
541: i2 -= bs;
542: idiag -= bs2;
543: for (i=m-2; i>=0; i--) {
544: v = aa + bs2*(diag[i]+1);
545: vi = aj + diag[i] + 1;
546: nz = ai[i+1] - diag[i] - 1;
548: PetscMemcpy(w,xb+i2,bs*sizeof(PetscScalar));
549: /* copy all rows of x that are needed into contiguous space */
550: workt = work;
551: for (j=0; j<nz; j++) {
552: PetscMemcpy(workt,x + bs*(*vi++),bs*sizeof(PetscScalar));
553: workt += bs;
554: }
555: PetscKernel_w_gets_w_minus_Ar_times_v(bs,bs*nz,w,v,work);
556: PetscKernel_w_gets_Ar_times_v(bs,bs,w,idiag,x+i2);
558: idiag -= bs2;
559: i2 -= bs;
560: }
561: break;
562: }
563: PetscLogFlops(1.0*bs2*(a->nz));
564: }
565: its--;
566: }
567: while (its--) {
568: if (flag & SOR_FORWARD_SWEEP || flag & SOR_LOCAL_FORWARD_SWEEP) {
569: idiag = a->idiag;
570: i2 = 0;
571: switch (bs) {
572: case 1:
573: for (i=0; i<m; i++) {
574: v = aa + ai[i];
575: vi = aj + ai[i];
576: nz = ai[i+1] - ai[i];
577: s[0] = b[i2];
578: for (j=0; j<nz; j++) {
579: xw[0] = x[vi[j]];
580: PetscKernel_v_gets_v_minus_A_times_w_1(s,(v+j),xw);
581: }
582: PetscKernel_v_gets_A_times_w_1(xw,idiag,s);
583: x[i2] += xw[0];
584: idiag += 1;
585: i2 += 1;
586: }
587: break;
588: case 2:
589: for (i=0; i<m; i++) {
590: v = aa + 4*ai[i];
591: vi = aj + ai[i];
592: nz = ai[i+1] - ai[i];
593: s[0] = b[i2]; s[1] = b[i2+1];
594: for (j=0; j<nz; j++) {
595: idx = 2*vi[j];
596: it = 4*j;
597: xw[0] = x[idx]; xw[1] = x[1+idx];
598: PetscKernel_v_gets_v_minus_A_times_w_2(s,(v+it),xw);
599: }
600: PetscKernel_v_gets_A_times_w_2(xw,idiag,s);
601: x[i2] += xw[0]; x[i2+1] += xw[1];
602: idiag += 4;
603: i2 += 2;
604: }
605: break;
606: case 3:
607: for (i=0; i<m; i++) {
608: v = aa + 9*ai[i];
609: vi = aj + ai[i];
610: nz = ai[i+1] - ai[i];
611: s[0] = b[i2]; s[1] = b[i2+1]; s[2] = b[i2+2];
612: while (nz--) {
613: idx = 3*(*vi++);
614: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx];
615: PetscKernel_v_gets_v_minus_A_times_w_3(s,v,xw);
616: v += 9;
617: }
618: PetscKernel_v_gets_A_times_w_3(xw,idiag,s);
619: x[i2] += xw[0]; x[i2+1] += xw[1]; x[i2+2] += xw[2];
620: idiag += 9;
621: i2 += 3;
622: }
623: break;
624: case 4:
625: for (i=0; i<m; i++) {
626: v = aa + 16*ai[i];
627: vi = aj + ai[i];
628: nz = ai[i+1] - ai[i];
629: s[0] = b[i2]; s[1] = b[i2+1]; s[2] = b[i2+2]; s[3] = b[i2+3];
630: while (nz--) {
631: idx = 4*(*vi++);
632: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx]; xw[3] = x[3+idx];
633: PetscKernel_v_gets_v_minus_A_times_w_4(s,v,xw);
634: v += 16;
635: }
636: PetscKernel_v_gets_A_times_w_4(xw,idiag,s);
637: x[i2] += xw[0]; x[i2+1] += xw[1]; x[i2+2] += xw[2]; x[i2+3] += xw[3];
638: idiag += 16;
639: i2 += 4;
640: }
641: break;
642: case 5:
643: for (i=0; i<m; i++) {
644: v = aa + 25*ai[i];
645: vi = aj + ai[i];
646: nz = ai[i+1] - ai[i];
647: s[0] = b[i2]; s[1] = b[i2+1]; s[2] = b[i2+2]; s[3] = b[i2+3]; s[4] = b[i2+4];
648: while (nz--) {
649: idx = 5*(*vi++);
650: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx]; xw[3] = x[3+idx]; xw[4] = x[4+idx];
651: PetscKernel_v_gets_v_minus_A_times_w_5(s,v,xw);
652: v += 25;
653: }
654: PetscKernel_v_gets_A_times_w_5(xw,idiag,s);
655: x[i2] += xw[0]; x[i2+1] += xw[1]; x[i2+2] += xw[2]; x[i2+3] += xw[3]; x[i2+4] += xw[4];
656: idiag += 25;
657: i2 += 5;
658: }
659: break;
660: case 6:
661: for (i=0; i<m; i++) {
662: v = aa + 36*ai[i];
663: vi = aj + ai[i];
664: nz = ai[i+1] - ai[i];
665: s[0] = b[i2]; s[1] = b[i2+1]; s[2] = b[i2+2]; s[3] = b[i2+3]; s[4] = b[i2+4]; s[5] = b[i2+5];
666: while (nz--) {
667: idx = 6*(*vi++);
668: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx];
669: xw[3] = x[3+idx]; xw[4] = x[4+idx]; xw[5] = x[5+idx];
670: PetscKernel_v_gets_v_minus_A_times_w_6(s,v,xw);
671: v += 36;
672: }
673: PetscKernel_v_gets_A_times_w_6(xw,idiag,s);
674: x[i2] += xw[0]; x[i2+1] += xw[1]; x[i2+2] += xw[2];
675: x[i2+3] += xw[3]; x[i2+4] += xw[4]; x[i2+5] += xw[5];
676: idiag += 36;
677: i2 += 6;
678: }
679: break;
680: case 7:
681: for (i=0; i<m; i++) {
682: v = aa + 49*ai[i];
683: vi = aj + ai[i];
684: nz = ai[i+1] - ai[i];
685: s[0] = b[i2]; s[1] = b[i2+1]; s[2] = b[i2+2];
686: s[3] = b[i2+3]; s[4] = b[i2+4]; s[5] = b[i2+5]; s[6] = b[i2+6];
687: while (nz--) {
688: idx = 7*(*vi++);
689: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx];
690: xw[3] = x[3+idx]; xw[4] = x[4+idx]; xw[5] = x[5+idx]; xw[6] = x[6+idx];
691: PetscKernel_v_gets_v_minus_A_times_w_7(s,v,xw);
692: v += 49;
693: }
694: PetscKernel_v_gets_A_times_w_7(xw,idiag,s);
695: x[i2] += xw[0]; x[i2+1] += xw[1]; x[i2+2] += xw[2];
696: x[i2+3] += xw[3]; x[i2+4] += xw[4]; x[i2+5] += xw[5]; x[i2+6] += xw[6];
697: idiag += 49;
698: i2 += 7;
699: }
700: break;
701: default:
702: for (i=0; i<m; i++) {
703: v = aa + bs2*ai[i];
704: vi = aj + ai[i];
705: nz = ai[i+1] - ai[i];
707: PetscMemcpy(w,b+i2,bs*sizeof(PetscScalar));
708: /* copy all rows of x that are needed into contiguous space */
709: workt = work;
710: for (j=0; j<nz; j++) {
711: PetscMemcpy(workt,x + bs*(*vi++),bs*sizeof(PetscScalar));
712: workt += bs;
713: }
714: PetscKernel_w_gets_w_minus_Ar_times_v(bs,bs*nz,w,v,work);
715: PetscKernel_w_gets_w_plus_Ar_times_v(bs,bs,w,idiag,x+i2);
717: idiag += bs2;
718: i2 += bs;
719: }
720: break;
721: }
722: PetscLogFlops(2.0*bs2*a->nz);
723: }
724: if (flag & SOR_BACKWARD_SWEEP || flag & SOR_LOCAL_BACKWARD_SWEEP) {
725: idiag = a->idiag+bs2*(a->mbs-1);
726: i2 = bs * (m-1);
727: switch (bs) {
728: case 1:
729: for (i=m-1; i>=0; i--) {
730: v = aa + ai[i];
731: vi = aj + ai[i];
732: nz = ai[i+1] - ai[i];
733: s[0] = b[i2];
734: for (j=0; j<nz; j++) {
735: xw[0] = x[vi[j]];
736: PetscKernel_v_gets_v_minus_A_times_w_1(s,(v+j),xw);
737: }
738: PetscKernel_v_gets_A_times_w_1(xw,idiag,s);
739: x[i2] += xw[0];
740: idiag -= 1;
741: i2 -= 1;
742: }
743: break;
744: case 2:
745: for (i=m-1; i>=0; i--) {
746: v = aa + 4*ai[i];
747: vi = aj + ai[i];
748: nz = ai[i+1] - ai[i];
749: s[0] = b[i2]; s[1] = b[i2+1];
750: for (j=0; j<nz; j++) {
751: idx = 2*vi[j];
752: it = 4*j;
753: xw[0] = x[idx]; xw[1] = x[1+idx];
754: PetscKernel_v_gets_v_minus_A_times_w_2(s,(v+it),xw);
755: }
756: PetscKernel_v_gets_A_times_w_2(xw,idiag,s);
757: x[i2] += xw[0]; x[i2+1] += xw[1];
758: idiag -= 4;
759: i2 -= 2;
760: }
761: break;
762: case 3:
763: for (i=m-1; i>=0; i--) {
764: v = aa + 9*ai[i];
765: vi = aj + ai[i];
766: nz = ai[i+1] - ai[i];
767: s[0] = b[i2]; s[1] = b[i2+1]; s[2] = b[i2+2];
768: while (nz--) {
769: idx = 3*(*vi++);
770: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx];
771: PetscKernel_v_gets_v_minus_A_times_w_3(s,v,xw);
772: v += 9;
773: }
774: PetscKernel_v_gets_A_times_w_3(xw,idiag,s);
775: x[i2] += xw[0]; x[i2+1] += xw[1]; x[i2+2] += xw[2];
776: idiag -= 9;
777: i2 -= 3;
778: }
779: break;
780: case 4:
781: for (i=m-1; i>=0; i--) {
782: v = aa + 16*ai[i];
783: vi = aj + ai[i];
784: nz = ai[i+1] - ai[i];
785: s[0] = b[i2]; s[1] = b[i2+1]; s[2] = b[i2+2]; s[3] = b[i2+3];
786: while (nz--) {
787: idx = 4*(*vi++);
788: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx]; xw[3] = x[3+idx];
789: PetscKernel_v_gets_v_minus_A_times_w_4(s,v,xw);
790: v += 16;
791: }
792: PetscKernel_v_gets_A_times_w_4(xw,idiag,s);
793: x[i2] += xw[0]; x[i2+1] += xw[1]; x[i2+2] += xw[2]; x[i2+3] += xw[3];
794: idiag -= 16;
795: i2 -= 4;
796: }
797: break;
798: case 5:
799: for (i=m-1; i>=0; i--) {
800: v = aa + 25*ai[i];
801: vi = aj + ai[i];
802: nz = ai[i+1] - ai[i];
803: s[0] = b[i2]; s[1] = b[i2+1]; s[2] = b[i2+2]; s[3] = b[i2+3]; s[4] = b[i2+4];
804: while (nz--) {
805: idx = 5*(*vi++);
806: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx]; xw[3] = x[3+idx]; xw[4] = x[4+idx];
807: PetscKernel_v_gets_v_minus_A_times_w_5(s,v,xw);
808: v += 25;
809: }
810: PetscKernel_v_gets_A_times_w_5(xw,idiag,s);
811: x[i2] += xw[0]; x[i2+1] += xw[1]; x[i2+2] += xw[2]; x[i2+3] += xw[3]; x[i2+4] += xw[4];
812: idiag -= 25;
813: i2 -= 5;
814: }
815: break;
816: case 6:
817: for (i=m-1; i>=0; i--) {
818: v = aa + 36*ai[i];
819: vi = aj + ai[i];
820: nz = ai[i+1] - ai[i];
821: s[0] = b[i2]; s[1] = b[i2+1]; s[2] = b[i2+2]; s[3] = b[i2+3]; s[4] = b[i2+4]; s[5] = b[i2+5];
822: while (nz--) {
823: idx = 6*(*vi++);
824: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx];
825: xw[3] = x[3+idx]; xw[4] = x[4+idx]; xw[5] = x[5+idx];
826: PetscKernel_v_gets_v_minus_A_times_w_6(s,v,xw);
827: v += 36;
828: }
829: PetscKernel_v_gets_A_times_w_6(xw,idiag,s);
830: x[i2] += xw[0]; x[i2+1] += xw[1]; x[i2+2] += xw[2];
831: x[i2+3] += xw[3]; x[i2+4] += xw[4]; x[i2+5] += xw[5];
832: idiag -= 36;
833: i2 -= 6;
834: }
835: break;
836: case 7:
837: for (i=m-1; i>=0; i--) {
838: v = aa + 49*ai[i];
839: vi = aj + ai[i];
840: nz = ai[i+1] - ai[i];
841: s[0] = b[i2]; s[1] = b[i2+1]; s[2] = b[i2+2];
842: s[3] = b[i2+3]; s[4] = b[i2+4]; s[5] = b[i2+5]; s[6] = b[i2+6];
843: while (nz--) {
844: idx = 7*(*vi++);
845: xw[0] = x[idx]; xw[1] = x[1+idx]; xw[2] = x[2+idx];
846: xw[3] = x[3+idx]; xw[4] = x[4+idx]; xw[5] = x[5+idx]; xw[6] = x[6+idx];
847: PetscKernel_v_gets_v_minus_A_times_w_7(s,v,xw);
848: v += 49;
849: }
850: PetscKernel_v_gets_A_times_w_7(xw,idiag,s);
851: x[i2] += xw[0]; x[i2+1] += xw[1]; x[i2+2] += xw[2];
852: x[i2+3] += xw[3]; x[i2+4] += xw[4]; x[i2+5] += xw[5]; x[i2+6] += xw[6];
853: idiag -= 49;
854: i2 -= 7;
855: }
856: break;
857: default:
858: for (i=m-1; i>=0; i--) {
859: v = aa + bs2*ai[i];
860: vi = aj + ai[i];
861: nz = ai[i+1] - ai[i];
863: PetscMemcpy(w,b+i2,bs*sizeof(PetscScalar));
864: /* copy all rows of x that are needed into contiguous space */
865: workt = work;
866: for (j=0; j<nz; j++) {
867: PetscMemcpy(workt,x + bs*(*vi++),bs*sizeof(PetscScalar));
868: workt += bs;
869: }
870: PetscKernel_w_gets_w_minus_Ar_times_v(bs,bs*nz,w,v,work);
871: PetscKernel_w_gets_w_plus_Ar_times_v(bs,bs,w,idiag,x+i2);
873: idiag -= bs2;
874: i2 -= bs;
875: }
876: break;
877: }
878: PetscLogFlops(2.0*bs2*(a->nz));
879: }
880: }
881: VecRestoreArray(xx,&x);
882: VecRestoreArrayRead(bb,&b);
883: return(0);
884: }
887: /*
888: Special version for direct calls from Fortran (Used in PETSc-fun3d)
889: */
890: #if defined(PETSC_HAVE_FORTRAN_CAPS)
891: #define matsetvaluesblocked4_ MATSETVALUESBLOCKED4
892: #elif !defined(PETSC_HAVE_FORTRAN_UNDERSCORE)
893: #define matsetvaluesblocked4_ matsetvaluesblocked4
894: #endif
898: PETSC_EXTERN void matsetvaluesblocked4_(Mat *AA,PetscInt *mm,const PetscInt im[],PetscInt *nn,const PetscInt in[],const PetscScalar v[])
899: {
900: Mat A = *AA;
901: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
902: PetscInt *rp,k,low,high,t,ii,jj,row,nrow,i,col,l,N,m = *mm,n = *nn;
903: PetscInt *ai =a->i,*ailen=a->ilen;
904: PetscInt *aj =a->j,stepval,lastcol = -1;
905: const PetscScalar *value = v;
906: MatScalar *ap,*aa = a->a,*bap;
909: if (A->rmap->bs != 4) SETERRABORT(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_WRONG,"Can only be called with a block size of 4");
910: stepval = (n-1)*4;
911: for (k=0; k<m; k++) { /* loop over added rows */
912: row = im[k];
913: rp = aj + ai[row];
914: ap = aa + 16*ai[row];
915: nrow = ailen[row];
916: low = 0;
917: high = nrow;
918: for (l=0; l<n; l++) { /* loop over added columns */
919: col = in[l];
920: if (col <= lastcol) low = 0;
921: else high = nrow;
922: lastcol = col;
923: value = v + k*(stepval+4 + l)*4;
924: while (high-low > 7) {
925: t = (low+high)/2;
926: if (rp[t] > col) high = t;
927: else low = t;
928: }
929: for (i=low; i<high; i++) {
930: if (rp[i] > col) break;
931: if (rp[i] == col) {
932: bap = ap + 16*i;
933: for (ii=0; ii<4; ii++,value+=stepval) {
934: for (jj=ii; jj<16; jj+=4) {
935: bap[jj] += *value++;
936: }
937: }
938: goto noinsert2;
939: }
940: }
941: N = nrow++ - 1;
942: high++; /* added new column index thus must search to one higher than before */
943: /* shift up all the later entries in this row */
944: for (ii=N; ii>=i; ii--) {
945: rp[ii+1] = rp[ii];
946: PetscMemcpy(ap+16*(ii+1),ap+16*(ii),16*sizeof(MatScalar));
947: }
948: if (N >= i) {
949: PetscMemzero(ap+16*i,16*sizeof(MatScalar));
950: }
951: rp[i] = col;
952: bap = ap + 16*i;
953: for (ii=0; ii<4; ii++,value+=stepval) {
954: for (jj=ii; jj<16; jj+=4) {
955: bap[jj] = *value++;
956: }
957: }
958: noinsert2:;
959: low = i;
960: }
961: ailen[row] = nrow;
962: }
963: PetscFunctionReturnVoid();
964: }
966: #if defined(PETSC_HAVE_FORTRAN_CAPS)
967: #define matsetvalues4_ MATSETVALUES4
968: #elif !defined(PETSC_HAVE_FORTRAN_UNDERSCORE)
969: #define matsetvalues4_ matsetvalues4
970: #endif
974: PETSC_EXTERN void matsetvalues4_(Mat *AA,PetscInt *mm,PetscInt *im,PetscInt *nn,PetscInt *in,PetscScalar *v)
975: {
976: Mat A = *AA;
977: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
978: PetscInt *rp,k,low,high,t,ii,row,nrow,i,col,l,N,n = *nn,m = *mm;
979: PetscInt *ai=a->i,*ailen=a->ilen;
980: PetscInt *aj=a->j,brow,bcol;
981: PetscInt ridx,cidx,lastcol = -1;
982: MatScalar *ap,value,*aa=a->a,*bap;
985: for (k=0; k<m; k++) { /* loop over added rows */
986: row = im[k]; brow = row/4;
987: rp = aj + ai[brow];
988: ap = aa + 16*ai[brow];
989: nrow = ailen[brow];
990: low = 0;
991: high = nrow;
992: for (l=0; l<n; l++) { /* loop over added columns */
993: col = in[l]; bcol = col/4;
994: ridx = row % 4; cidx = col % 4;
995: value = v[l + k*n];
996: if (col <= lastcol) low = 0;
997: else high = nrow;
998: lastcol = col;
999: while (high-low > 7) {
1000: t = (low+high)/2;
1001: if (rp[t] > bcol) high = t;
1002: else low = t;
1003: }
1004: for (i=low; i<high; i++) {
1005: if (rp[i] > bcol) break;
1006: if (rp[i] == bcol) {
1007: bap = ap + 16*i + 4*cidx + ridx;
1008: *bap += value;
1009: goto noinsert1;
1010: }
1011: }
1012: N = nrow++ - 1;
1013: high++; /* added new column thus must search to one higher than before */
1014: /* shift up all the later entries in this row */
1015: for (ii=N; ii>=i; ii--) {
1016: rp[ii+1] = rp[ii];
1017: PetscMemcpy(ap+16*(ii+1),ap+16*(ii),16*sizeof(MatScalar));
1018: }
1019: if (N>=i) {
1020: PetscMemzero(ap+16*i,16*sizeof(MatScalar));
1021: }
1022: rp[i] = bcol;
1023: ap[16*i + 4*cidx + ridx] = value;
1024: noinsert1:;
1025: low = i;
1026: }
1027: ailen[brow] = nrow;
1028: }
1029: PetscFunctionReturnVoid();
1030: }
1032: /*
1033: Checks for missing diagonals
1034: */
1037: PetscErrorCode MatMissingDiagonal_SeqBAIJ(Mat A,PetscBool *missing,PetscInt *d)
1038: {
1039: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
1041: PetscInt *diag,*ii = a->i,i;
1044: MatMarkDiagonal_SeqBAIJ(A);
1045: *missing = PETSC_FALSE;
1046: if (A->rmap->n > 0 && !ii) {
1047: *missing = PETSC_TRUE;
1048: if (d) *d = 0;
1049: PetscInfo(A,"Matrix has no entries therefore is missing diagonal");
1050: } else {
1051: diag = a->diag;
1052: for (i=0; i<a->mbs; i++) {
1053: if (diag[i] >= ii[i+1]) {
1054: *missing = PETSC_TRUE;
1055: if (d) *d = i;
1056: PetscInfo1(A,"Matrix is missing block diagonal number %D",i);
1057: break;
1058: }
1059: }
1060: }
1061: return(0);
1062: }
1066: PetscErrorCode MatMarkDiagonal_SeqBAIJ(Mat A)
1067: {
1068: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
1070: PetscInt i,j,m = a->mbs;
1073: if (!a->diag) {
1074: PetscMalloc1(m,&a->diag);
1075: PetscLogObjectMemory((PetscObject)A,m*sizeof(PetscInt));
1076: a->free_diag = PETSC_TRUE;
1077: }
1078: for (i=0; i<m; i++) {
1079: a->diag[i] = a->i[i+1];
1080: for (j=a->i[i]; j<a->i[i+1]; j++) {
1081: if (a->j[j] == i) {
1082: a->diag[i] = j;
1083: break;
1084: }
1085: }
1086: }
1087: return(0);
1088: }
1093: static PetscErrorCode MatGetRowIJ_SeqBAIJ(Mat A,PetscInt oshift,PetscBool symmetric,PetscBool blockcompressed,PetscInt *nn,const PetscInt *inia[],const PetscInt *inja[],PetscBool *done)
1094: {
1095: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
1097: PetscInt i,j,n = a->mbs,nz = a->i[n],*tia,*tja,bs = A->rmap->bs,k,l,cnt;
1098: PetscInt **ia = (PetscInt**)inia,**ja = (PetscInt**)inja;
1101: *nn = n;
1102: if (!ia) return(0);
1103: if (symmetric) {
1104: MatToSymmetricIJ_SeqAIJ(n,a->i,a->j,0,0,&tia,&tja);
1105: nz = tia[n];
1106: } else {
1107: tia = a->i; tja = a->j;
1108: }
1110: if (!blockcompressed && bs > 1) {
1111: (*nn) *= bs;
1112: /* malloc & create the natural set of indices */
1113: PetscMalloc1((n+1)*bs,ia);
1114: if (n) {
1115: (*ia)[0] = 0;
1116: for (j=1; j<bs; j++) {
1117: (*ia)[j] = (tia[1]-tia[0])*bs+(*ia)[j-1];
1118: }
1119: }
1121: for (i=1; i<n; i++) {
1122: (*ia)[i*bs] = (tia[i]-tia[i-1])*bs + (*ia)[i*bs-1];
1123: for (j=1; j<bs; j++) {
1124: (*ia)[i*bs+j] = (tia[i+1]-tia[i])*bs + (*ia)[i*bs+j-1];
1125: }
1126: }
1127: if (n) {
1128: (*ia)[n*bs] = (tia[n]-tia[n-1])*bs + (*ia)[n*bs-1];
1129: }
1131: if (inja) {
1132: PetscMalloc1(nz*bs*bs,ja);
1133: cnt = 0;
1134: for (i=0; i<n; i++) {
1135: for (j=0; j<bs; j++) {
1136: for (k=tia[i]; k<tia[i+1]; k++) {
1137: for (l=0; l<bs; l++) {
1138: (*ja)[cnt++] = bs*tja[k] + l;
1139: }
1140: }
1141: }
1142: }
1143: }
1145: if (symmetric) { /* deallocate memory allocated in MatToSymmetricIJ_SeqAIJ() */
1146: PetscFree(tia);
1147: PetscFree(tja);
1148: }
1149: } else if (oshift == 1) {
1150: if (symmetric) {
1151: nz = tia[A->rmap->n/bs];
1152: /* add 1 to i and j indices */
1153: for (i=0; i<A->rmap->n/bs+1; i++) tia[i] = tia[i] + 1;
1154: *ia = tia;
1155: if (ja) {
1156: for (i=0; i<nz; i++) tja[i] = tja[i] + 1;
1157: *ja = tja;
1158: }
1159: } else {
1160: nz = a->i[A->rmap->n/bs];
1161: /* malloc space and add 1 to i and j indices */
1162: PetscMalloc1((A->rmap->n/bs+1),ia);
1163: for (i=0; i<A->rmap->n/bs+1; i++) (*ia)[i] = a->i[i] + 1;
1164: if (ja) {
1165: PetscMalloc1(nz,ja);
1166: for (i=0; i<nz; i++) (*ja)[i] = a->j[i] + 1;
1167: }
1168: }
1169: } else {
1170: *ia = tia;
1171: if (ja) *ja = tja;
1172: }
1173: return(0);
1174: }
1178: static PetscErrorCode MatRestoreRowIJ_SeqBAIJ(Mat A,PetscInt oshift,PetscBool symmetric,PetscBool blockcompressed,PetscInt *nn,const PetscInt *ia[],const PetscInt *ja[],PetscBool *done)
1179: {
1183: if (!ia) return(0);
1184: if ((!blockcompressed && A->rmap->bs > 1) || (symmetric || oshift == 1)) {
1185: PetscFree(*ia);
1186: if (ja) {PetscFree(*ja);}
1187: }
1188: return(0);
1189: }
1193: PetscErrorCode MatDestroy_SeqBAIJ(Mat A)
1194: {
1195: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
1199: #if defined(PETSC_USE_LOG)
1200: PetscLogObjectState((PetscObject)A,"Rows=%D, Cols=%D, NZ=%D",A->rmap->N,A->cmap->n,a->nz);
1201: #endif
1202: MatSeqXAIJFreeAIJ(A,&a->a,&a->j,&a->i);
1203: ISDestroy(&a->row);
1204: ISDestroy(&a->col);
1205: if (a->free_diag) {PetscFree(a->diag);}
1206: PetscFree(a->idiag);
1207: if (a->free_imax_ilen) {PetscFree2(a->imax,a->ilen);}
1208: PetscFree(a->solve_work);
1209: PetscFree(a->mult_work);
1210: PetscFree(a->sor_workt);
1211: PetscFree(a->sor_work);
1212: ISDestroy(&a->icol);
1213: PetscFree(a->saved_values);
1214: PetscFree(a->xtoy);
1215: PetscFree2(a->compressedrow.i,a->compressedrow.rindex);
1217: MatDestroy(&a->sbaijMat);
1218: MatDestroy(&a->parent);
1219: PetscFree(A->data);
1221: PetscObjectChangeTypeName((PetscObject)A,0);
1222: PetscObjectComposeFunction((PetscObject)A,"MatInvertBlockDiagonal_C",NULL);
1223: PetscObjectComposeFunction((PetscObject)A,"MatStoreValues_C",NULL);
1224: PetscObjectComposeFunction((PetscObject)A,"MatRetrieveValues_C",NULL);
1225: PetscObjectComposeFunction((PetscObject)A,"MatSeqBAIJSetColumnIndices_C",NULL);
1226: PetscObjectComposeFunction((PetscObject)A,"MatConvert_seqbaij_seqaij_C",NULL);
1227: PetscObjectComposeFunction((PetscObject)A,"MatConvert_seqbaij_seqsbaij_C",NULL);
1228: PetscObjectComposeFunction((PetscObject)A,"MatSeqBAIJSetPreallocation_C",NULL);
1229: PetscObjectComposeFunction((PetscObject)A,"MatSeqBAIJSetPreallocationCSR_C",NULL);
1230: PetscObjectComposeFunction((PetscObject)A,"MatConvert_seqbaij_seqbstrm_C",NULL);
1231: PetscObjectComposeFunction((PetscObject)A,"MatIsTranspose_C",NULL);
1232: return(0);
1233: }
1237: PetscErrorCode MatSetOption_SeqBAIJ(Mat A,MatOption op,PetscBool flg)
1238: {
1239: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
1243: switch (op) {
1244: case MAT_ROW_ORIENTED:
1245: a->roworiented = flg;
1246: break;
1247: case MAT_KEEP_NONZERO_PATTERN:
1248: a->keepnonzeropattern = flg;
1249: break;
1250: case MAT_NEW_NONZERO_LOCATIONS:
1251: a->nonew = (flg ? 0 : 1);
1252: break;
1253: case MAT_NEW_NONZERO_LOCATION_ERR:
1254: a->nonew = (flg ? -1 : 0);
1255: break;
1256: case MAT_NEW_NONZERO_ALLOCATION_ERR:
1257: a->nonew = (flg ? -2 : 0);
1258: break;
1259: case MAT_UNUSED_NONZERO_LOCATION_ERR:
1260: a->nounused = (flg ? -1 : 0);
1261: break;
1262: case MAT_NEW_DIAGONALS:
1263: case MAT_IGNORE_OFF_PROC_ENTRIES:
1264: case MAT_USE_HASH_TABLE:
1265: PetscInfo1(A,"Option %s ignored\n",MatOptions[op]);
1266: break;
1267: case MAT_SPD:
1268: case MAT_SYMMETRIC:
1269: case MAT_STRUCTURALLY_SYMMETRIC:
1270: case MAT_HERMITIAN:
1271: case MAT_SYMMETRY_ETERNAL:
1272: /* These options are handled directly by MatSetOption() */
1273: break;
1274: default:
1275: SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"unknown option %d",op);
1276: }
1277: return(0);
1278: }
1280: /* used for both SeqBAIJ and SeqSBAIJ matrices */
1283: PetscErrorCode MatGetRow_SeqBAIJ_private(Mat A,PetscInt row,PetscInt *nz,PetscInt **idx,PetscScalar **v,PetscInt *ai,PetscInt *aj,PetscScalar *aa)
1284: {
1286: PetscInt itmp,i,j,k,M,bn,bp,*idx_i,bs,bs2;
1287: MatScalar *aa_i;
1288: PetscScalar *v_i;
1291: bs = A->rmap->bs;
1292: bs2 = bs*bs;
1293: if (row < 0 || row >= A->rmap->N) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Row %D out of range", row);
1295: bn = row/bs; /* Block number */
1296: bp = row % bs; /* Block Position */
1297: M = ai[bn+1] - ai[bn];
1298: *nz = bs*M;
1300: if (v) {
1301: *v = 0;
1302: if (*nz) {
1303: PetscMalloc1((*nz),v);
1304: for (i=0; i<M; i++) { /* for each block in the block row */
1305: v_i = *v + i*bs;
1306: aa_i = aa + bs2*(ai[bn] + i);
1307: for (j=bp,k=0; j<bs2; j+=bs,k++) v_i[k] = aa_i[j];
1308: }
1309: }
1310: }
1312: if (idx) {
1313: *idx = 0;
1314: if (*nz) {
1315: PetscMalloc1((*nz),idx);
1316: for (i=0; i<M; i++) { /* for each block in the block row */
1317: idx_i = *idx + i*bs;
1318: itmp = bs*aj[ai[bn] + i];
1319: for (j=0; j<bs; j++) idx_i[j] = itmp++;
1320: }
1321: }
1322: }
1323: return(0);
1324: }
1328: PetscErrorCode MatGetRow_SeqBAIJ(Mat A,PetscInt row,PetscInt *nz,PetscInt **idx,PetscScalar **v)
1329: {
1330: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
1332:
1334: MatGetRow_SeqBAIJ_private(A,row,nz,idx,v,a->i,a->j,a->a);
1335: return(0);
1336: }
1340: PetscErrorCode MatRestoreRow_SeqBAIJ(Mat A,PetscInt row,PetscInt *nz,PetscInt **idx,PetscScalar **v)
1341: {
1345: if (idx) {PetscFree(*idx);}
1346: if (v) {PetscFree(*v);}
1347: return(0);
1348: }
1350: extern PetscErrorCode MatSetValues_SeqBAIJ(Mat,PetscInt,const PetscInt[],PetscInt,const PetscInt[],const PetscScalar[],InsertMode);
1354: PetscErrorCode MatTranspose_SeqBAIJ(Mat A,MatReuse reuse,Mat *B)
1355: {
1356: Mat_SeqBAIJ *a=(Mat_SeqBAIJ*)A->data;
1357: Mat C;
1359: PetscInt i,j,k,*aj=a->j,*ai=a->i,bs=A->rmap->bs,mbs=a->mbs,nbs=a->nbs,len,*col;
1360: PetscInt *rows,*cols,bs2=a->bs2;
1361: MatScalar *array;
1364: if (reuse == MAT_REUSE_MATRIX && A == *B && mbs != nbs) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Square matrix only for in-place");
1365: if (reuse == MAT_INITIAL_MATRIX || A == *B) {
1366: PetscCalloc1((1+nbs),&col);
1368: for (i=0; i<ai[mbs]; i++) col[aj[i]] += 1;
1369: MatCreate(PetscObjectComm((PetscObject)A),&C);
1370: MatSetSizes(C,A->cmap->n,A->rmap->N,A->cmap->n,A->rmap->N);
1371: MatSetType(C,((PetscObject)A)->type_name);
1372: MatSeqBAIJSetPreallocation_SeqBAIJ(C,bs,0,col);
1373: PetscFree(col);
1374: } else {
1375: C = *B;
1376: }
1378: array = a->a;
1379: PetscMalloc2(bs,&rows,bs,&cols);
1380: for (i=0; i<mbs; i++) {
1381: cols[0] = i*bs;
1382: for (k=1; k<bs; k++) cols[k] = cols[k-1] + 1;
1383: len = ai[i+1] - ai[i];
1384: for (j=0; j<len; j++) {
1385: rows[0] = (*aj++)*bs;
1386: for (k=1; k<bs; k++) rows[k] = rows[k-1] + 1;
1387: MatSetValues_SeqBAIJ(C,bs,rows,bs,cols,array,INSERT_VALUES);
1388: array += bs2;
1389: }
1390: }
1391: PetscFree2(rows,cols);
1393: MatAssemblyBegin(C,MAT_FINAL_ASSEMBLY);
1394: MatAssemblyEnd(C,MAT_FINAL_ASSEMBLY);
1396: if (reuse == MAT_INITIAL_MATRIX || *B != A) {
1397: *B = C;
1398: } else {
1399: MatHeaderMerge(A,C);
1400: }
1401: return(0);
1402: }
1406: PetscErrorCode MatIsTranspose_SeqBAIJ(Mat A,Mat B,PetscReal tol,PetscBool *f)
1407: {
1409: Mat Btrans;
1412: *f = PETSC_FALSE;
1413: MatTranspose_SeqBAIJ(A,MAT_INITIAL_MATRIX,&Btrans);
1414: MatEqual_SeqBAIJ(B,Btrans,f);
1415: MatDestroy(&Btrans);
1416: return(0);
1417: }
1421: static PetscErrorCode MatView_SeqBAIJ_Binary(Mat A,PetscViewer viewer)
1422: {
1423: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
1425: PetscInt i,*col_lens,bs = A->rmap->bs,count,*jj,j,k,l,bs2=a->bs2;
1426: int fd;
1427: PetscScalar *aa;
1428: FILE *file;
1431: PetscViewerBinaryGetDescriptor(viewer,&fd);
1432: PetscMalloc1((4+A->rmap->N),&col_lens);
1433: col_lens[0] = MAT_FILE_CLASSID;
1435: col_lens[1] = A->rmap->N;
1436: col_lens[2] = A->cmap->n;
1437: col_lens[3] = a->nz*bs2;
1439: /* store lengths of each row and write (including header) to file */
1440: count = 0;
1441: for (i=0; i<a->mbs; i++) {
1442: for (j=0; j<bs; j++) {
1443: col_lens[4+count++] = bs*(a->i[i+1] - a->i[i]);
1444: }
1445: }
1446: PetscBinaryWrite(fd,col_lens,4+A->rmap->N,PETSC_INT,PETSC_TRUE);
1447: PetscFree(col_lens);
1449: /* store column indices (zero start index) */
1450: PetscMalloc1((a->nz+1)*bs2,&jj);
1451: count = 0;
1452: for (i=0; i<a->mbs; i++) {
1453: for (j=0; j<bs; j++) {
1454: for (k=a->i[i]; k<a->i[i+1]; k++) {
1455: for (l=0; l<bs; l++) {
1456: jj[count++] = bs*a->j[k] + l;
1457: }
1458: }
1459: }
1460: }
1461: PetscBinaryWrite(fd,jj,bs2*a->nz,PETSC_INT,PETSC_FALSE);
1462: PetscFree(jj);
1464: /* store nonzero values */
1465: PetscMalloc1((a->nz+1)*bs2,&aa);
1466: count = 0;
1467: for (i=0; i<a->mbs; i++) {
1468: for (j=0; j<bs; j++) {
1469: for (k=a->i[i]; k<a->i[i+1]; k++) {
1470: for (l=0; l<bs; l++) {
1471: aa[count++] = a->a[bs2*k + l*bs + j];
1472: }
1473: }
1474: }
1475: }
1476: PetscBinaryWrite(fd,aa,bs2*a->nz,PETSC_SCALAR,PETSC_FALSE);
1477: PetscFree(aa);
1479: PetscViewerBinaryGetInfoPointer(viewer,&file);
1480: if (file) {
1481: fprintf(file,"-matload_block_size %d\n",(int)A->rmap->bs);
1482: }
1483: return(0);
1484: }
1488: static PetscErrorCode MatView_SeqBAIJ_ASCII(Mat A,PetscViewer viewer)
1489: {
1490: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
1491: PetscErrorCode ierr;
1492: PetscInt i,j,bs = A->rmap->bs,k,l,bs2=a->bs2;
1493: PetscViewerFormat format;
1496: PetscViewerGetFormat(viewer,&format);
1497: if (format == PETSC_VIEWER_ASCII_INFO || format == PETSC_VIEWER_ASCII_INFO_DETAIL) {
1498: PetscViewerASCIIPrintf(viewer," block size is %D\n",bs);
1499: } else if (format == PETSC_VIEWER_ASCII_MATLAB) {
1500: const char *matname;
1501: Mat aij;
1502: MatConvert(A,MATSEQAIJ,MAT_INITIAL_MATRIX,&aij);
1503: PetscObjectGetName((PetscObject)A,&matname);
1504: PetscObjectSetName((PetscObject)aij,matname);
1505: MatView(aij,viewer);
1506: MatDestroy(&aij);
1507: } else if (format == PETSC_VIEWER_ASCII_FACTOR_INFO) {
1508: return(0);
1509: } else if (format == PETSC_VIEWER_ASCII_COMMON) {
1510: PetscViewerASCIIUseTabs(viewer,PETSC_FALSE);
1511: for (i=0; i<a->mbs; i++) {
1512: for (j=0; j<bs; j++) {
1513: PetscViewerASCIIPrintf(viewer,"row %D:",i*bs+j);
1514: for (k=a->i[i]; k<a->i[i+1]; k++) {
1515: for (l=0; l<bs; l++) {
1516: #if defined(PETSC_USE_COMPLEX)
1517: if (PetscImaginaryPart(a->a[bs2*k + l*bs + j]) > 0.0 && PetscRealPart(a->a[bs2*k + l*bs + j]) != 0.0) {
1518: PetscViewerASCIIPrintf(viewer," (%D, %g + %gi) ",bs*a->j[k]+l,
1519: (double)PetscRealPart(a->a[bs2*k + l*bs + j]),(double)PetscImaginaryPart(a->a[bs2*k + l*bs + j]));
1520: } else if (PetscImaginaryPart(a->a[bs2*k + l*bs + j]) < 0.0 && PetscRealPart(a->a[bs2*k + l*bs + j]) != 0.0) {
1521: PetscViewerASCIIPrintf(viewer," (%D, %g - %gi) ",bs*a->j[k]+l,
1522: (double)PetscRealPart(a->a[bs2*k + l*bs + j]),-(double)PetscImaginaryPart(a->a[bs2*k + l*bs + j]));
1523: } else if (PetscRealPart(a->a[bs2*k + l*bs + j]) != 0.0) {
1524: PetscViewerASCIIPrintf(viewer," (%D, %g) ",bs*a->j[k]+l,(double)PetscRealPart(a->a[bs2*k + l*bs + j]));
1525: }
1526: #else
1527: if (a->a[bs2*k + l*bs + j] != 0.0) {
1528: PetscViewerASCIIPrintf(viewer," (%D, %g) ",bs*a->j[k]+l,(double)a->a[bs2*k + l*bs + j]);
1529: }
1530: #endif
1531: }
1532: }
1533: PetscViewerASCIIPrintf(viewer,"\n");
1534: }
1535: }
1536: PetscViewerASCIIUseTabs(viewer,PETSC_TRUE);
1537: } else {
1538: PetscViewerASCIIUseTabs(viewer,PETSC_FALSE);
1539: for (i=0; i<a->mbs; i++) {
1540: for (j=0; j<bs; j++) {
1541: PetscViewerASCIIPrintf(viewer,"row %D:",i*bs+j);
1542: for (k=a->i[i]; k<a->i[i+1]; k++) {
1543: for (l=0; l<bs; l++) {
1544: #if defined(PETSC_USE_COMPLEX)
1545: if (PetscImaginaryPart(a->a[bs2*k + l*bs + j]) > 0.0) {
1546: PetscViewerASCIIPrintf(viewer," (%D, %g + %g i) ",bs*a->j[k]+l,
1547: (double)PetscRealPart(a->a[bs2*k + l*bs + j]),(double)PetscImaginaryPart(a->a[bs2*k + l*bs + j]));
1548: } else if (PetscImaginaryPart(a->a[bs2*k + l*bs + j]) < 0.0) {
1549: PetscViewerASCIIPrintf(viewer," (%D, %g - %g i) ",bs*a->j[k]+l,
1550: (double)PetscRealPart(a->a[bs2*k + l*bs + j]),-(double)PetscImaginaryPart(a->a[bs2*k + l*bs + j]));
1551: } else {
1552: PetscViewerASCIIPrintf(viewer," (%D, %g) ",bs*a->j[k]+l,(double)PetscRealPart(a->a[bs2*k + l*bs + j]));
1553: }
1554: #else
1555: PetscViewerASCIIPrintf(viewer," (%D, %g) ",bs*a->j[k]+l,(double)a->a[bs2*k + l*bs + j]);
1556: #endif
1557: }
1558: }
1559: PetscViewerASCIIPrintf(viewer,"\n");
1560: }
1561: }
1562: PetscViewerASCIIUseTabs(viewer,PETSC_TRUE);
1563: }
1564: PetscViewerFlush(viewer);
1565: return(0);
1566: }
1568: #include <petscdraw.h>
1571: static PetscErrorCode MatView_SeqBAIJ_Draw_Zoom(PetscDraw draw,void *Aa)
1572: {
1573: Mat A = (Mat) Aa;
1574: Mat_SeqBAIJ *a=(Mat_SeqBAIJ*)A->data;
1575: PetscErrorCode ierr;
1576: PetscInt row,i,j,k,l,mbs=a->mbs,color,bs=A->rmap->bs,bs2=a->bs2;
1577: PetscReal xl,yl,xr,yr,x_l,x_r,y_l,y_r;
1578: MatScalar *aa;
1579: PetscViewer viewer;
1580: PetscViewerFormat format;
1583: PetscObjectQuery((PetscObject)A,"Zoomviewer",(PetscObject*)&viewer);
1584: PetscViewerGetFormat(viewer,&format);
1586: PetscDrawGetCoordinates(draw,&xl,&yl,&xr,&yr);
1588: /* loop over matrix elements drawing boxes */
1590: if (format != PETSC_VIEWER_DRAW_CONTOUR) {
1591: color = PETSC_DRAW_BLUE;
1592: for (i=0,row=0; i<mbs; i++,row+=bs) {
1593: for (j=a->i[i]; j<a->i[i+1]; j++) {
1594: y_l = A->rmap->N - row - 1.0; y_r = y_l + 1.0;
1595: x_l = a->j[j]*bs; x_r = x_l + 1.0;
1596: aa = a->a + j*bs2;
1597: for (k=0; k<bs; k++) {
1598: for (l=0; l<bs; l++) {
1599: if (PetscRealPart(*aa++) >= 0.) continue;
1600: PetscDrawRectangle(draw,x_l+k,y_l-l,x_r+k,y_r-l,color,color,color,color);
1601: }
1602: }
1603: }
1604: }
1605: color = PETSC_DRAW_CYAN;
1606: for (i=0,row=0; i<mbs; i++,row+=bs) {
1607: for (j=a->i[i]; j<a->i[i+1]; j++) {
1608: y_l = A->rmap->N - row - 1.0; y_r = y_l + 1.0;
1609: x_l = a->j[j]*bs; x_r = x_l + 1.0;
1610: aa = a->a + j*bs2;
1611: for (k=0; k<bs; k++) {
1612: for (l=0; l<bs; l++) {
1613: if (PetscRealPart(*aa++) != 0.) continue;
1614: PetscDrawRectangle(draw,x_l+k,y_l-l,x_r+k,y_r-l,color,color,color,color);
1615: }
1616: }
1617: }
1618: }
1619: color = PETSC_DRAW_RED;
1620: for (i=0,row=0; i<mbs; i++,row+=bs) {
1621: for (j=a->i[i]; j<a->i[i+1]; j++) {
1622: y_l = A->rmap->N - row - 1.0; y_r = y_l + 1.0;
1623: x_l = a->j[j]*bs; x_r = x_l + 1.0;
1624: aa = a->a + j*bs2;
1625: for (k=0; k<bs; k++) {
1626: for (l=0; l<bs; l++) {
1627: if (PetscRealPart(*aa++) <= 0.) continue;
1628: PetscDrawRectangle(draw,x_l+k,y_l-l,x_r+k,y_r-l,color,color,color,color);
1629: }
1630: }
1631: }
1632: }
1633: } else {
1634: /* use contour shading to indicate magnitude of values */
1635: /* first determine max of all nonzero values */
1636: PetscDraw popup;
1637: PetscReal scale,maxv = 0.0;
1639: for (i=0; i<a->nz*a->bs2; i++) {
1640: if (PetscAbsScalar(a->a[i]) > maxv) maxv = PetscAbsScalar(a->a[i]);
1641: }
1642: scale = (245.0 - PETSC_DRAW_BASIC_COLORS)/maxv;
1643: PetscDrawGetPopup(draw,&popup);
1644: if (popup) {
1645: PetscDrawScalePopup(popup,0.0,maxv);
1646: }
1647: for (i=0,row=0; i<mbs; i++,row+=bs) {
1648: for (j=a->i[i]; j<a->i[i+1]; j++) {
1649: y_l = A->rmap->N - row - 1.0; y_r = y_l + 1.0;
1650: x_l = a->j[j]*bs; x_r = x_l + 1.0;
1651: aa = a->a + j*bs2;
1652: for (k=0; k<bs; k++) {
1653: for (l=0; l<bs; l++) {
1654: color = PETSC_DRAW_BASIC_COLORS + (PetscInt)(scale*PetscAbsScalar(*aa++));
1655: PetscDrawRectangle(draw,x_l+k,y_l-l,x_r+k,y_r-l,color,color,color,color);
1656: }
1657: }
1658: }
1659: }
1660: }
1661: return(0);
1662: }
1666: static PetscErrorCode MatView_SeqBAIJ_Draw(Mat A,PetscViewer viewer)
1667: {
1669: PetscReal xl,yl,xr,yr,w,h;
1670: PetscDraw draw;
1671: PetscBool isnull;
1674: PetscViewerDrawGetDraw(viewer,0,&draw);
1675: PetscDrawIsNull(draw,&isnull); if (isnull) return(0);
1677: PetscObjectCompose((PetscObject)A,"Zoomviewer",(PetscObject)viewer);
1678: xr = A->cmap->n; yr = A->rmap->N; h = yr/10.0; w = xr/10.0;
1679: xr += w; yr += h; xl = -w; yl = -h;
1680: PetscDrawSetCoordinates(draw,xl,yl,xr,yr);
1681: PetscDrawZoom(draw,MatView_SeqBAIJ_Draw_Zoom,A);
1682: PetscObjectCompose((PetscObject)A,"Zoomviewer",NULL);
1683: return(0);
1684: }
1688: PetscErrorCode MatView_SeqBAIJ(Mat A,PetscViewer viewer)
1689: {
1691: PetscBool iascii,isbinary,isdraw;
1694: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERASCII,&iascii);
1695: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERBINARY,&isbinary);
1696: PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERDRAW,&isdraw);
1697: if (iascii) {
1698: MatView_SeqBAIJ_ASCII(A,viewer);
1699: } else if (isbinary) {
1700: MatView_SeqBAIJ_Binary(A,viewer);
1701: } else if (isdraw) {
1702: MatView_SeqBAIJ_Draw(A,viewer);
1703: } else {
1704: Mat B;
1705: MatConvert(A,MATSEQAIJ,MAT_INITIAL_MATRIX,&B);
1706: MatView(B,viewer);
1707: MatDestroy(&B);
1708: }
1709: return(0);
1710: }
1715: PetscErrorCode MatGetValues_SeqBAIJ(Mat A,PetscInt m,const PetscInt im[],PetscInt n,const PetscInt in[],PetscScalar v[])
1716: {
1717: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
1718: PetscInt *rp,k,low,high,t,row,nrow,i,col,l,*aj = a->j;
1719: PetscInt *ai = a->i,*ailen = a->ilen;
1720: PetscInt brow,bcol,ridx,cidx,bs=A->rmap->bs,bs2=a->bs2;
1721: MatScalar *ap,*aa = a->a;
1724: for (k=0; k<m; k++) { /* loop over rows */
1725: row = im[k]; brow = row/bs;
1726: if (row < 0) {v += n; continue;} /* SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Negative row"); */
1727: if (row >= A->rmap->N) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Row %D too large", row);
1728: rp = aj + ai[brow]; ap = aa + bs2*ai[brow];
1729: nrow = ailen[brow];
1730: for (l=0; l<n; l++) { /* loop over columns */
1731: if (in[l] < 0) {v++; continue;} /* SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Negative column"); */
1732: if (in[l] >= A->cmap->n) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Column %D too large", in[l]);
1733: col = in[l];
1734: bcol = col/bs;
1735: cidx = col%bs;
1736: ridx = row%bs;
1737: high = nrow;
1738: low = 0; /* assume unsorted */
1739: while (high-low > 5) {
1740: t = (low+high)/2;
1741: if (rp[t] > bcol) high = t;
1742: else low = t;
1743: }
1744: for (i=low; i<high; i++) {
1745: if (rp[i] > bcol) break;
1746: if (rp[i] == bcol) {
1747: *v++ = ap[bs2*i+bs*cidx+ridx];
1748: goto finished;
1749: }
1750: }
1751: *v++ = 0.0;
1752: finished:;
1753: }
1754: }
1755: return(0);
1756: }
1760: PetscErrorCode MatSetValuesBlocked_SeqBAIJ(Mat A,PetscInt m,const PetscInt im[],PetscInt n,const PetscInt in[],const PetscScalar v[],InsertMode is)
1761: {
1762: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
1763: PetscInt *rp,k,low,high,t,ii,jj,row,nrow,i,col,l,rmax,N,lastcol = -1;
1764: PetscInt *imax=a->imax,*ai=a->i,*ailen=a->ilen;
1765: PetscErrorCode ierr;
1766: PetscInt *aj =a->j,nonew=a->nonew,bs2=a->bs2,bs=A->rmap->bs,stepval;
1767: PetscBool roworiented=a->roworiented;
1768: const PetscScalar *value = v;
1769: MatScalar *ap,*aa = a->a,*bap;
1772: if (roworiented) {
1773: stepval = (n-1)*bs;
1774: } else {
1775: stepval = (m-1)*bs;
1776: }
1777: for (k=0; k<m; k++) { /* loop over added rows */
1778: row = im[k];
1779: if (row < 0) continue;
1780: #if defined(PETSC_USE_DEBUG)
1781: if (row >= a->mbs) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Row too large: row %D max %D",row,a->mbs-1);
1782: #endif
1783: rp = aj + ai[row];
1784: ap = aa + bs2*ai[row];
1785: rmax = imax[row];
1786: nrow = ailen[row];
1787: low = 0;
1788: high = nrow;
1789: for (l=0; l<n; l++) { /* loop over added columns */
1790: if (in[l] < 0) continue;
1791: #if defined(PETSC_USE_DEBUG)
1792: if (in[l] >= a->nbs) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Column too large: col %D max %D",in[l],a->nbs-1);
1793: #endif
1794: col = in[l];
1795: if (roworiented) {
1796: value = v + (k*(stepval+bs) + l)*bs;
1797: } else {
1798: value = v + (l*(stepval+bs) + k)*bs;
1799: }
1800: if (col <= lastcol) low = 0;
1801: else high = nrow;
1802: lastcol = col;
1803: while (high-low > 7) {
1804: t = (low+high)/2;
1805: if (rp[t] > col) high = t;
1806: else low = t;
1807: }
1808: for (i=low; i<high; i++) {
1809: if (rp[i] > col) break;
1810: if (rp[i] == col) {
1811: bap = ap + bs2*i;
1812: if (roworiented) {
1813: if (is == ADD_VALUES) {
1814: for (ii=0; ii<bs; ii++,value+=stepval) {
1815: for (jj=ii; jj<bs2; jj+=bs) {
1816: bap[jj] += *value++;
1817: }
1818: }
1819: } else {
1820: for (ii=0; ii<bs; ii++,value+=stepval) {
1821: for (jj=ii; jj<bs2; jj+=bs) {
1822: bap[jj] = *value++;
1823: }
1824: }
1825: }
1826: } else {
1827: if (is == ADD_VALUES) {
1828: for (ii=0; ii<bs; ii++,value+=bs+stepval) {
1829: for (jj=0; jj<bs; jj++) {
1830: bap[jj] += value[jj];
1831: }
1832: bap += bs;
1833: }
1834: } else {
1835: for (ii=0; ii<bs; ii++,value+=bs+stepval) {
1836: for (jj=0; jj<bs; jj++) {
1837: bap[jj] = value[jj];
1838: }
1839: bap += bs;
1840: }
1841: }
1842: }
1843: goto noinsert2;
1844: }
1845: }
1846: if (nonew == 1) goto noinsert2;
1847: if (nonew == -1) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Inserting a new nonzero (%D, %D) in the matrix", row, col);
1848: MatSeqXAIJReallocateAIJ(A,a->mbs,bs2,nrow,row,col,rmax,aa,ai,aj,rp,ap,imax,nonew,MatScalar);
1849: N = nrow++ - 1; high++;
1850: /* shift up all the later entries in this row */
1851: for (ii=N; ii>=i; ii--) {
1852: rp[ii+1] = rp[ii];
1853: PetscMemcpy(ap+bs2*(ii+1),ap+bs2*(ii),bs2*sizeof(MatScalar));
1854: }
1855: if (N >= i) {
1856: PetscMemzero(ap+bs2*i,bs2*sizeof(MatScalar));
1857: }
1858: rp[i] = col;
1859: bap = ap + bs2*i;
1860: if (roworiented) {
1861: for (ii=0; ii<bs; ii++,value+=stepval) {
1862: for (jj=ii; jj<bs2; jj+=bs) {
1863: bap[jj] = *value++;
1864: }
1865: }
1866: } else {
1867: for (ii=0; ii<bs; ii++,value+=stepval) {
1868: for (jj=0; jj<bs; jj++) {
1869: *bap++ = *value++;
1870: }
1871: }
1872: }
1873: noinsert2:;
1874: low = i;
1875: }
1876: ailen[row] = nrow;
1877: }
1878: return(0);
1879: }
1883: PetscErrorCode MatAssemblyEnd_SeqBAIJ(Mat A,MatAssemblyType mode)
1884: {
1885: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
1886: PetscInt fshift = 0,i,j,*ai = a->i,*aj = a->j,*imax = a->imax;
1887: PetscInt m = A->rmap->N,*ip,N,*ailen = a->ilen;
1889: PetscInt mbs = a->mbs,bs2 = a->bs2,rmax = 0;
1890: MatScalar *aa = a->a,*ap;
1891: PetscReal ratio=0.6;
1894: if (mode == MAT_FLUSH_ASSEMBLY) return(0);
1896: if (m) rmax = ailen[0];
1897: for (i=1; i<mbs; i++) {
1898: /* move each row back by the amount of empty slots (fshift) before it*/
1899: fshift += imax[i-1] - ailen[i-1];
1900: rmax = PetscMax(rmax,ailen[i]);
1901: if (fshift) {
1902: ip = aj + ai[i]; ap = aa + bs2*ai[i];
1903: N = ailen[i];
1904: for (j=0; j<N; j++) {
1905: ip[j-fshift] = ip[j];
1907: PetscMemcpy(ap+(j-fshift)*bs2,ap+j*bs2,bs2*sizeof(MatScalar));
1908: }
1909: }
1910: ai[i] = ai[i-1] + ailen[i-1];
1911: }
1912: if (mbs) {
1913: fshift += imax[mbs-1] - ailen[mbs-1];
1914: ai[mbs] = ai[mbs-1] + ailen[mbs-1];
1915: }
1917: /* reset ilen and imax for each row */
1918: a->nonzerorowcnt = 0;
1919: for (i=0; i<mbs; i++) {
1920: ailen[i] = imax[i] = ai[i+1] - ai[i];
1921: a->nonzerorowcnt += ((ai[i+1] - ai[i]) > 0);
1922: }
1923: a->nz = ai[mbs];
1925: /* diagonals may have moved, so kill the diagonal pointers */
1926: a->idiagvalid = PETSC_FALSE;
1927: if (fshift && a->diag) {
1928: PetscFree(a->diag);
1929: PetscLogObjectMemory((PetscObject)A,-(mbs+1)*sizeof(PetscInt));
1930: a->diag = 0;
1931: }
1932: if (fshift && a->nounused == -1) SETERRQ4(PETSC_COMM_SELF,PETSC_ERR_PLIB, "Unused space detected in matrix: %D X %D block size %D, %D unneeded", m, A->cmap->n, A->rmap->bs, fshift*bs2);
1933: PetscInfo5(A,"Matrix size: %D X %D, block size %D; storage space: %D unneeded, %D used\n",m,A->cmap->n,A->rmap->bs,fshift*bs2,a->nz*bs2);
1934: PetscInfo1(A,"Number of mallocs during MatSetValues is %D\n",a->reallocs);
1935: PetscInfo1(A,"Most nonzeros blocks in any row is %D\n",rmax);
1937: A->info.mallocs += a->reallocs;
1938: a->reallocs = 0;
1939: A->info.nz_unneeded = (PetscReal)fshift*bs2;
1941: MatCheckCompressedRow(A,a->nonzerorowcnt,&a->compressedrow,a->i,mbs,ratio);
1942: return(0);
1943: }
1945: /*
1946: This function returns an array of flags which indicate the locations of contiguous
1947: blocks that should be zeroed. for eg: if bs = 3 and is = [0,1,2,3,5,6,7,8,9]
1948: then the resulting sizes = [3,1,1,3,1] correspondig to sets [(0,1,2),(3),(5),(6,7,8),(9)]
1949: Assume: sizes should be long enough to hold all the values.
1950: */
1953: static PetscErrorCode MatZeroRows_SeqBAIJ_Check_Blocks(PetscInt idx[],PetscInt n,PetscInt bs,PetscInt sizes[], PetscInt *bs_max)
1954: {
1955: PetscInt i,j,k,row;
1956: PetscBool flg;
1959: for (i=0,j=0; i<n; j++) {
1960: row = idx[i];
1961: if (row%bs!=0) { /* Not the begining of a block */
1962: sizes[j] = 1;
1963: i++;
1964: } else if (i+bs > n) { /* complete block doesn't exist (at idx end) */
1965: sizes[j] = 1; /* Also makes sure atleast 'bs' values exist for next else */
1966: i++;
1967: } else { /* Begining of the block, so check if the complete block exists */
1968: flg = PETSC_TRUE;
1969: for (k=1; k<bs; k++) {
1970: if (row+k != idx[i+k]) { /* break in the block */
1971: flg = PETSC_FALSE;
1972: break;
1973: }
1974: }
1975: if (flg) { /* No break in the bs */
1976: sizes[j] = bs;
1977: i += bs;
1978: } else {
1979: sizes[j] = 1;
1980: i++;
1981: }
1982: }
1983: }
1984: *bs_max = j;
1985: return(0);
1986: }
1990: PetscErrorCode MatZeroRows_SeqBAIJ(Mat A,PetscInt is_n,const PetscInt is_idx[],PetscScalar diag,Vec x, Vec b)
1991: {
1992: Mat_SeqBAIJ *baij=(Mat_SeqBAIJ*)A->data;
1993: PetscErrorCode ierr;
1994: PetscInt i,j,k,count,*rows;
1995: PetscInt bs=A->rmap->bs,bs2=baij->bs2,*sizes,row,bs_max;
1996: PetscScalar zero = 0.0;
1997: MatScalar *aa;
1998: const PetscScalar *xx;
1999: PetscScalar *bb;
2002: /* fix right hand side if needed */
2003: if (x && b) {
2004: VecGetArrayRead(x,&xx);
2005: VecGetArray(b,&bb);
2006: for (i=0; i<is_n; i++) {
2007: bb[is_idx[i]] = diag*xx[is_idx[i]];
2008: }
2009: VecRestoreArrayRead(x,&xx);
2010: VecRestoreArray(b,&bb);
2011: }
2013: /* Make a copy of the IS and sort it */
2014: /* allocate memory for rows,sizes */
2015: PetscMalloc2(is_n,&rows,2*is_n,&sizes);
2017: /* copy IS values to rows, and sort them */
2018: for (i=0; i<is_n; i++) rows[i] = is_idx[i];
2019: PetscSortInt(is_n,rows);
2021: if (baij->keepnonzeropattern) {
2022: for (i=0; i<is_n; i++) sizes[i] = 1;
2023: bs_max = is_n;
2024: } else {
2025: MatZeroRows_SeqBAIJ_Check_Blocks(rows,is_n,bs,sizes,&bs_max);
2026: A->nonzerostate++;
2027: }
2029: for (i=0,j=0; i<bs_max; j+=sizes[i],i++) {
2030: row = rows[j];
2031: if (row < 0 || row > A->rmap->N) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"row %D out of range",row);
2032: count = (baij->i[row/bs +1] - baij->i[row/bs])*bs;
2033: aa = ((MatScalar*)(baij->a)) + baij->i[row/bs]*bs2 + (row%bs);
2034: if (sizes[i] == bs && !baij->keepnonzeropattern) {
2035: if (diag != (PetscScalar)0.0) {
2036: if (baij->ilen[row/bs] > 0) {
2037: baij->ilen[row/bs] = 1;
2038: baij->j[baij->i[row/bs]] = row/bs;
2040: PetscMemzero(aa,count*bs*sizeof(MatScalar));
2041: }
2042: /* Now insert all the diagonal values for this bs */
2043: for (k=0; k<bs; k++) {
2044: (*A->ops->setvalues)(A,1,rows+j+k,1,rows+j+k,&diag,INSERT_VALUES);
2045: }
2046: } else { /* (diag == 0.0) */
2047: baij->ilen[row/bs] = 0;
2048: } /* end (diag == 0.0) */
2049: } else { /* (sizes[i] != bs) */
2050: #if defined(PETSC_USE_DEBUG)
2051: if (sizes[i] != 1) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Internal Error. Value should be 1");
2052: #endif
2053: for (k=0; k<count; k++) {
2054: aa[0] = zero;
2055: aa += bs;
2056: }
2057: if (diag != (PetscScalar)0.0) {
2058: (*A->ops->setvalues)(A,1,rows+j,1,rows+j,&diag,INSERT_VALUES);
2059: }
2060: }
2061: }
2063: PetscFree2(rows,sizes);
2064: MatAssemblyEnd_SeqBAIJ(A,MAT_FINAL_ASSEMBLY);
2065: return(0);
2066: }
2070: PetscErrorCode MatZeroRowsColumns_SeqBAIJ(Mat A,PetscInt is_n,const PetscInt is_idx[],PetscScalar diag,Vec x, Vec b)
2071: {
2072: Mat_SeqBAIJ *baij=(Mat_SeqBAIJ*)A->data;
2073: PetscErrorCode ierr;
2074: PetscInt i,j,k,count;
2075: PetscInt bs =A->rmap->bs,bs2=baij->bs2,row,col;
2076: PetscScalar zero = 0.0;
2077: MatScalar *aa;
2078: const PetscScalar *xx;
2079: PetscScalar *bb;
2080: PetscBool *zeroed,vecs = PETSC_FALSE;
2083: /* fix right hand side if needed */
2084: if (x && b) {
2085: VecGetArrayRead(x,&xx);
2086: VecGetArray(b,&bb);
2087: vecs = PETSC_TRUE;
2088: }
2090: /* zero the columns */
2091: PetscCalloc1(A->rmap->n,&zeroed);
2092: for (i=0; i<is_n; i++) {
2093: if (is_idx[i] < 0 || is_idx[i] >= A->rmap->N) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"row %D out of range",is_idx[i]);
2094: zeroed[is_idx[i]] = PETSC_TRUE;
2095: }
2096: for (i=0; i<A->rmap->N; i++) {
2097: if (!zeroed[i]) {
2098: row = i/bs;
2099: for (j=baij->i[row]; j<baij->i[row+1]; j++) {
2100: for (k=0; k<bs; k++) {
2101: col = bs*baij->j[j] + k;
2102: if (zeroed[col]) {
2103: aa = ((MatScalar*)(baij->a)) + j*bs2 + (i%bs) + bs*k;
2104: if (vecs) bb[i] -= aa[0]*xx[col];
2105: aa[0] = 0.0;
2106: }
2107: }
2108: }
2109: } else if (vecs) bb[i] = diag*xx[i];
2110: }
2111: PetscFree(zeroed);
2112: if (vecs) {
2113: VecRestoreArrayRead(x,&xx);
2114: VecRestoreArray(b,&bb);
2115: }
2117: /* zero the rows */
2118: for (i=0; i<is_n; i++) {
2119: row = is_idx[i];
2120: count = (baij->i[row/bs +1] - baij->i[row/bs])*bs;
2121: aa = ((MatScalar*)(baij->a)) + baij->i[row/bs]*bs2 + (row%bs);
2122: for (k=0; k<count; k++) {
2123: aa[0] = zero;
2124: aa += bs;
2125: }
2126: if (diag != (PetscScalar)0.0) {
2127: (*A->ops->setvalues)(A,1,&row,1,&row,&diag,INSERT_VALUES);
2128: }
2129: }
2130: MatAssemblyEnd_SeqBAIJ(A,MAT_FINAL_ASSEMBLY);
2131: return(0);
2132: }
2136: PetscErrorCode MatSetValues_SeqBAIJ(Mat A,PetscInt m,const PetscInt im[],PetscInt n,const PetscInt in[],const PetscScalar v[],InsertMode is)
2137: {
2138: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
2139: PetscInt *rp,k,low,high,t,ii,row,nrow,i,col,l,rmax,N,lastcol = -1;
2140: PetscInt *imax=a->imax,*ai=a->i,*ailen=a->ilen;
2141: PetscInt *aj =a->j,nonew=a->nonew,bs=A->rmap->bs,brow,bcol;
2143: PetscInt ridx,cidx,bs2=a->bs2;
2144: PetscBool roworiented=a->roworiented;
2145: MatScalar *ap,value,*aa=a->a,*bap;
2148: for (k=0; k<m; k++) { /* loop over added rows */
2149: row = im[k];
2150: brow = row/bs;
2151: if (row < 0) continue;
2152: #if defined(PETSC_USE_DEBUG)
2153: if (row >= A->rmap->N) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Row too large: row %D max %D",row,A->rmap->N-1);
2154: #endif
2155: rp = aj + ai[brow];
2156: ap = aa + bs2*ai[brow];
2157: rmax = imax[brow];
2158: nrow = ailen[brow];
2159: low = 0;
2160: high = nrow;
2161: for (l=0; l<n; l++) { /* loop over added columns */
2162: if (in[l] < 0) continue;
2163: #if defined(PETSC_USE_DEBUG)
2164: if (in[l] >= A->cmap->n) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Column too large: col %D max %D",in[l],A->cmap->n-1);
2165: #endif
2166: col = in[l]; bcol = col/bs;
2167: ridx = row % bs; cidx = col % bs;
2168: if (roworiented) {
2169: value = v[l + k*n];
2170: } else {
2171: value = v[k + l*m];
2172: }
2173: if (col <= lastcol) low = 0; else high = nrow;
2174: lastcol = col;
2175: while (high-low > 7) {
2176: t = (low+high)/2;
2177: if (rp[t] > bcol) high = t;
2178: else low = t;
2179: }
2180: for (i=low; i<high; i++) {
2181: if (rp[i] > bcol) break;
2182: if (rp[i] == bcol) {
2183: bap = ap + bs2*i + bs*cidx + ridx;
2184: if (is == ADD_VALUES) *bap += value;
2185: else *bap = value;
2186: goto noinsert1;
2187: }
2188: }
2189: if (nonew == 1) goto noinsert1;
2190: if (nonew == -1) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Inserting a new nonzero (%D, %D) in the matrix", row, col);
2191: MatSeqXAIJReallocateAIJ(A,a->mbs,bs2,nrow,brow,bcol,rmax,aa,ai,aj,rp,ap,imax,nonew,MatScalar);
2192: N = nrow++ - 1; high++;
2193: /* shift up all the later entries in this row */
2194: for (ii=N; ii>=i; ii--) {
2195: rp[ii+1] = rp[ii];
2196: PetscMemcpy(ap+bs2*(ii+1),ap+bs2*(ii),bs2*sizeof(MatScalar));
2197: }
2198: if (N>=i) {
2199: PetscMemzero(ap+bs2*i,bs2*sizeof(MatScalar));
2200: }
2201: rp[i] = bcol;
2202: ap[bs2*i + bs*cidx + ridx] = value;
2203: a->nz++;
2204: A->nonzerostate++;
2205: noinsert1:;
2206: low = i;
2207: }
2208: ailen[brow] = nrow;
2209: }
2210: return(0);
2211: }
2215: PetscErrorCode MatILUFactor_SeqBAIJ(Mat inA,IS row,IS col,const MatFactorInfo *info)
2216: {
2217: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)inA->data;
2218: Mat outA;
2220: PetscBool row_identity,col_identity;
2223: if (info->levels != 0) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Only levels = 0 supported for in-place ILU");
2224: ISIdentity(row,&row_identity);
2225: ISIdentity(col,&col_identity);
2226: if (!row_identity || !col_identity) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Row and column permutations must be identity for in-place ILU");
2228: outA = inA;
2229: inA->factortype = MAT_FACTOR_LU;
2231: MatMarkDiagonal_SeqBAIJ(inA);
2233: PetscObjectReference((PetscObject)row);
2234: ISDestroy(&a->row);
2235: a->row = row;
2236: PetscObjectReference((PetscObject)col);
2237: ISDestroy(&a->col);
2238: a->col = col;
2240: /* Create the invert permutation so that it can be used in MatLUFactorNumeric() */
2241: ISDestroy(&a->icol);
2242: ISInvertPermutation(col,PETSC_DECIDE,&a->icol);
2243: PetscLogObjectParent((PetscObject)inA,(PetscObject)a->icol);
2245: MatSeqBAIJSetNumericFactorization_inplace(inA,(PetscBool)(row_identity && col_identity));
2246: if (!a->solve_work) {
2247: PetscMalloc1((inA->rmap->N+inA->rmap->bs),&a->solve_work);
2248: PetscLogObjectMemory((PetscObject)inA,(inA->rmap->N+inA->rmap->bs)*sizeof(PetscScalar));
2249: }
2250: MatLUFactorNumeric(outA,inA,info);
2251: return(0);
2252: }
2256: PetscErrorCode MatSeqBAIJSetColumnIndices_SeqBAIJ(Mat mat,PetscInt *indices)
2257: {
2258: Mat_SeqBAIJ *baij = (Mat_SeqBAIJ*)mat->data;
2259: PetscInt i,nz,mbs;
2262: nz = baij->maxnz;
2263: mbs = baij->mbs;
2264: for (i=0; i<nz; i++) {
2265: baij->j[i] = indices[i];
2266: }
2267: baij->nz = nz;
2268: for (i=0; i<mbs; i++) {
2269: baij->ilen[i] = baij->imax[i];
2270: }
2271: return(0);
2272: }
2276: /*@
2277: MatSeqBAIJSetColumnIndices - Set the column indices for all the rows
2278: in the matrix.
2280: Input Parameters:
2281: + mat - the SeqBAIJ matrix
2282: - indices - the column indices
2284: Level: advanced
2286: Notes:
2287: This can be called if you have precomputed the nonzero structure of the
2288: matrix and want to provide it to the matrix object to improve the performance
2289: of the MatSetValues() operation.
2291: You MUST have set the correct numbers of nonzeros per row in the call to
2292: MatCreateSeqBAIJ(), and the columns indices MUST be sorted.
2294: MUST be called before any calls to MatSetValues();
2296: @*/
2297: PetscErrorCode MatSeqBAIJSetColumnIndices(Mat mat,PetscInt *indices)
2298: {
2304: PetscUseMethod(mat,"MatSeqBAIJSetColumnIndices_C",(Mat,PetscInt*),(mat,indices));
2305: return(0);
2306: }
2310: PetscErrorCode MatGetRowMaxAbs_SeqBAIJ(Mat A,Vec v,PetscInt idx[])
2311: {
2312: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
2314: PetscInt i,j,n,row,bs,*ai,*aj,mbs;
2315: PetscReal atmp;
2316: PetscScalar *x,zero = 0.0;
2317: MatScalar *aa;
2318: PetscInt ncols,brow,krow,kcol;
2321: if (A->factortype) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Not for factored matrix");
2322: bs = A->rmap->bs;
2323: aa = a->a;
2324: ai = a->i;
2325: aj = a->j;
2326: mbs = a->mbs;
2328: VecSet(v,zero);
2329: VecGetArray(v,&x);
2330: VecGetLocalSize(v,&n);
2331: if (n != A->rmap->N) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Nonconforming matrix and vector");
2332: for (i=0; i<mbs; i++) {
2333: ncols = ai[1] - ai[0]; ai++;
2334: brow = bs*i;
2335: for (j=0; j<ncols; j++) {
2336: for (kcol=0; kcol<bs; kcol++) {
2337: for (krow=0; krow<bs; krow++) {
2338: atmp = PetscAbsScalar(*aa);aa++;
2339: row = brow + krow; /* row index */
2340: if (PetscAbsScalar(x[row]) < atmp) {x[row] = atmp; if (idx) idx[row] = bs*(*aj) + kcol;}
2341: }
2342: }
2343: aj++;
2344: }
2345: }
2346: VecRestoreArray(v,&x);
2347: return(0);
2348: }
2352: PetscErrorCode MatCopy_SeqBAIJ(Mat A,Mat B,MatStructure str)
2353: {
2357: /* If the two matrices have the same copy implementation, use fast copy. */
2358: if (str == SAME_NONZERO_PATTERN && (A->ops->copy == B->ops->copy)) {
2359: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
2360: Mat_SeqBAIJ *b = (Mat_SeqBAIJ*)B->data;
2361: PetscInt ambs=a->mbs,bmbs=b->mbs,abs=A->rmap->bs,bbs=B->rmap->bs,bs2=abs*abs;
2363: if (a->i[ambs] != b->i[bmbs]) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_INCOMP,"Number of nonzero blocks in matrices A %D and B %D are different",a->i[ambs],b->i[bmbs]);
2364: if (abs != bbs) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_INCOMP,"Block size A %D and B %D are different",abs,bbs);
2365: PetscMemcpy(b->a,a->a,(bs2*a->i[ambs])*sizeof(PetscScalar));
2366: } else {
2367: MatCopy_Basic(A,B,str);
2368: }
2369: return(0);
2370: }
2374: PetscErrorCode MatSetUp_SeqBAIJ(Mat A)
2375: {
2379: MatSeqBAIJSetPreallocation_SeqBAIJ(A,A->rmap->bs,PETSC_DEFAULT,0);
2380: return(0);
2381: }
2385: PetscErrorCode MatSeqBAIJGetArray_SeqBAIJ(Mat A,PetscScalar *array[])
2386: {
2387: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
2390: *array = a->a;
2391: return(0);
2392: }
2396: PetscErrorCode MatSeqBAIJRestoreArray_SeqBAIJ(Mat A,PetscScalar *array[])
2397: {
2399: return(0);
2400: }
2404: PetscErrorCode MatAXPYGetPreallocation_SeqBAIJ(Mat Y,Mat X,PetscInt *nnz)
2405: {
2406: PetscInt bs = Y->rmap->bs,mbs = Y->rmap->N/bs;
2407: Mat_SeqBAIJ *x = (Mat_SeqBAIJ*)X->data;
2408: Mat_SeqBAIJ *y = (Mat_SeqBAIJ*)Y->data;
2412: /* Set the number of nonzeros in the new matrix */
2413: MatAXPYGetPreallocation_SeqX_private(mbs,x->i,x->j,y->i,y->j,nnz);
2414: return(0);
2415: }
2419: PetscErrorCode MatAXPY_SeqBAIJ(Mat Y,PetscScalar a,Mat X,MatStructure str)
2420: {
2421: Mat_SeqBAIJ *x = (Mat_SeqBAIJ*)X->data,*y = (Mat_SeqBAIJ*)Y->data;
2423: PetscInt bs=Y->rmap->bs,bs2=bs*bs;
2424: PetscBLASInt one=1;
2427: if (str == SAME_NONZERO_PATTERN) {
2428: PetscScalar alpha = a;
2429: PetscBLASInt bnz;
2430: PetscBLASIntCast(x->nz*bs2,&bnz);
2431: PetscStackCallBLAS("BLASaxpy",BLASaxpy_(&bnz,&alpha,x->a,&one,y->a,&one));
2432: PetscObjectStateIncrease((PetscObject)Y);
2433: } else if (str == SUBSET_NONZERO_PATTERN) { /* nonzeros of X is a subset of Y's */
2434: MatAXPY_Basic(Y,a,X,str);
2435: } else {
2436: Mat B;
2437: PetscInt *nnz;
2438: if (bs != X->rmap->bs) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Matrices must have same block size");
2439: PetscMalloc1(Y->rmap->N,&nnz);
2440: MatCreate(PetscObjectComm((PetscObject)Y),&B);
2441: PetscObjectSetName((PetscObject)B,((PetscObject)Y)->name);
2442: MatSetSizes(B,Y->rmap->n,Y->cmap->n,Y->rmap->N,Y->cmap->N);
2443: MatSetBlockSizesFromMats(B,Y,Y);
2444: MatSetType(B,(MatType) ((PetscObject)Y)->type_name);
2445: MatAXPYGetPreallocation_SeqBAIJ(Y,X,nnz);
2446: MatSeqBAIJSetPreallocation(B,bs,0,nnz);
2447: MatAXPY_BasicWithPreallocation(B,Y,a,X,str);
2448: MatHeaderReplace(Y,B);
2449: PetscFree(nnz);
2450: }
2451: return(0);
2452: }
2456: PetscErrorCode MatRealPart_SeqBAIJ(Mat A)
2457: {
2458: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
2459: PetscInt i,nz = a->bs2*a->i[a->mbs];
2460: MatScalar *aa = a->a;
2463: for (i=0; i<nz; i++) aa[i] = PetscRealPart(aa[i]);
2464: return(0);
2465: }
2469: PetscErrorCode MatImaginaryPart_SeqBAIJ(Mat A)
2470: {
2471: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
2472: PetscInt i,nz = a->bs2*a->i[a->mbs];
2473: MatScalar *aa = a->a;
2476: for (i=0; i<nz; i++) aa[i] = PetscImaginaryPart(aa[i]);
2477: return(0);
2478: }
2482: /*
2483: Code almost idential to MatGetColumnIJ_SeqAIJ() should share common code
2484: */
2485: PetscErrorCode MatGetColumnIJ_SeqBAIJ(Mat A,PetscInt oshift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *nn,const PetscInt *ia[],const PetscInt *ja[],PetscBool *done)
2486: {
2487: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
2489: PetscInt bs = A->rmap->bs,i,*collengths,*cia,*cja,n = A->cmap->n/bs,m = A->rmap->n/bs;
2490: PetscInt nz = a->i[m],row,*jj,mr,col;
2493: *nn = n;
2494: if (!ia) return(0);
2495: if (symmetric) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Not for BAIJ matrices");
2496: else {
2497: PetscCalloc1((n+1),&collengths);
2498: PetscMalloc1((n+1),&cia);
2499: PetscMalloc1((nz+1),&cja);
2500: jj = a->j;
2501: for (i=0; i<nz; i++) {
2502: collengths[jj[i]]++;
2503: }
2504: cia[0] = oshift;
2505: for (i=0; i<n; i++) {
2506: cia[i+1] = cia[i] + collengths[i];
2507: }
2508: PetscMemzero(collengths,n*sizeof(PetscInt));
2509: jj = a->j;
2510: for (row=0; row<m; row++) {
2511: mr = a->i[row+1] - a->i[row];
2512: for (i=0; i<mr; i++) {
2513: col = *jj++;
2515: cja[cia[col] + collengths[col]++ - oshift] = row + oshift;
2516: }
2517: }
2518: PetscFree(collengths);
2519: *ia = cia; *ja = cja;
2520: }
2521: return(0);
2522: }
2526: PetscErrorCode MatRestoreColumnIJ_SeqBAIJ(Mat A,PetscInt oshift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *n,const PetscInt *ia[],const PetscInt *ja[],PetscBool *done)
2527: {
2531: if (!ia) return(0);
2532: PetscFree(*ia);
2533: PetscFree(*ja);
2534: return(0);
2535: }
2537: /*
2538: MatGetColumnIJ_SeqBAIJ_Color() and MatRestoreColumnIJ_SeqBAIJ_Color() are customized from
2539: MatGetColumnIJ_SeqBAIJ() and MatRestoreColumnIJ_SeqBAIJ() by adding an output
2540: spidx[], index of a->a, to be used in MatTransposeColoringCreate() and MatFDColoringCreate()
2541: */
2544: PetscErrorCode MatGetColumnIJ_SeqBAIJ_Color(Mat A,PetscInt oshift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *nn,const PetscInt *ia[],const PetscInt *ja[],PetscInt *spidx[],PetscBool *done)
2545: {
2546: Mat_SeqBAIJ *a = (Mat_SeqBAIJ*)A->data;
2548: PetscInt i,*collengths,*cia,*cja,n=a->nbs,m=a->mbs;
2549: PetscInt nz = a->i[m],row,*jj,mr,col;
2550: PetscInt *cspidx;
2553: *nn = n;
2554: if (!ia) return(0);
2556: PetscCalloc1((n+1),&collengths);
2557: PetscMalloc1((n+1),&cia);
2558: PetscMalloc1((nz+1),&cja);
2559: PetscMalloc1((nz+1),&cspidx);
2560: jj = a->j;
2561: for (i=0; i<nz; i++) {
2562: collengths[jj[i]]++;
2563: }
2564: cia[0] = oshift;
2565: for (i=0; i<n; i++) {
2566: cia[i+1] = cia[i] + collengths[i];
2567: }
2568: PetscMemzero(collengths,n*sizeof(PetscInt));
2569: jj = a->j;
2570: for (row=0; row<m; row++) {
2571: mr = a->i[row+1] - a->i[row];
2572: for (i=0; i<mr; i++) {
2573: col = *jj++;
2574: cspidx[cia[col] + collengths[col] - oshift] = a->i[row] + i; /* index of a->j */
2575: cja[cia[col] + collengths[col]++ - oshift] = row + oshift;
2576: }
2577: }
2578: PetscFree(collengths);
2579: *ia = cia; *ja = cja;
2580: *spidx = cspidx;
2581: return(0);
2582: }
2586: PetscErrorCode MatRestoreColumnIJ_SeqBAIJ_Color(Mat A,PetscInt oshift,PetscBool symmetric,PetscBool inodecompressed,PetscInt *n,const PetscInt *ia[],const PetscInt *ja[],PetscInt *spidx[],PetscBool *done)
2587: {
2591: MatRestoreColumnIJ_SeqBAIJ(A,oshift,symmetric,inodecompressed,n,ia,ja,done);
2592: PetscFree(*spidx);
2593: return(0);
2594: }
2596: /* -------------------------------------------------------------------*/
2597: static struct _MatOps MatOps_Values = {MatSetValues_SeqBAIJ,
2598: MatGetRow_SeqBAIJ,
2599: MatRestoreRow_SeqBAIJ,
2600: MatMult_SeqBAIJ_N,
2601: /* 4*/ MatMultAdd_SeqBAIJ_N,
2602: MatMultTranspose_SeqBAIJ,
2603: MatMultTransposeAdd_SeqBAIJ,
2604: 0,
2605: 0,
2606: 0,
2607: /* 10*/ 0,
2608: MatLUFactor_SeqBAIJ,
2609: 0,
2610: 0,
2611: MatTranspose_SeqBAIJ,
2612: /* 15*/ MatGetInfo_SeqBAIJ,
2613: MatEqual_SeqBAIJ,
2614: MatGetDiagonal_SeqBAIJ,
2615: MatDiagonalScale_SeqBAIJ,
2616: MatNorm_SeqBAIJ,
2617: /* 20*/ 0,
2618: MatAssemblyEnd_SeqBAIJ,
2619: MatSetOption_SeqBAIJ,
2620: MatZeroEntries_SeqBAIJ,
2621: /* 24*/ MatZeroRows_SeqBAIJ,
2622: 0,
2623: 0,
2624: 0,
2625: 0,
2626: /* 29*/ MatSetUp_SeqBAIJ,
2627: 0,
2628: 0,
2629: 0,
2630: 0,
2631: /* 34*/ MatDuplicate_SeqBAIJ,
2632: 0,
2633: 0,
2634: MatILUFactor_SeqBAIJ,
2635: 0,
2636: /* 39*/ MatAXPY_SeqBAIJ,
2637: MatGetSubMatrices_SeqBAIJ,
2638: MatIncreaseOverlap_SeqBAIJ,
2639: MatGetValues_SeqBAIJ,
2640: MatCopy_SeqBAIJ,
2641: /* 44*/ 0,
2642: MatScale_SeqBAIJ,
2643: 0,
2644: 0,
2645: MatZeroRowsColumns_SeqBAIJ,
2646: /* 49*/ 0,
2647: MatGetRowIJ_SeqBAIJ,
2648: MatRestoreRowIJ_SeqBAIJ,
2649: MatGetColumnIJ_SeqBAIJ,
2650: MatRestoreColumnIJ_SeqBAIJ,
2651: /* 54*/ MatFDColoringCreate_SeqXAIJ,
2652: 0,
2653: 0,
2654: 0,
2655: MatSetValuesBlocked_SeqBAIJ,
2656: /* 59*/ MatGetSubMatrix_SeqBAIJ,
2657: MatDestroy_SeqBAIJ,
2658: MatView_SeqBAIJ,
2659: 0,
2660: 0,
2661: /* 64*/ 0,
2662: 0,
2663: 0,
2664: 0,
2665: 0,
2666: /* 69*/ MatGetRowMaxAbs_SeqBAIJ,
2667: 0,
2668: MatConvert_Basic,
2669: 0,
2670: 0,
2671: /* 74*/ 0,
2672: MatFDColoringApply_BAIJ,
2673: 0,
2674: 0,
2675: 0,
2676: /* 79*/ 0,
2677: 0,
2678: 0,
2679: 0,
2680: MatLoad_SeqBAIJ,
2681: /* 84*/ 0,
2682: 0,
2683: 0,
2684: 0,
2685: 0,
2686: /* 89*/ 0,
2687: 0,
2688: 0,
2689: 0,
2690: 0,
2691: /* 94*/ 0,
2692: 0,
2693: 0,
2694: 0,
2695: 0,
2696: /* 99*/ 0,
2697: 0,
2698: 0,
2699: 0,
2700: 0,
2701: /*104*/ 0,
2702: MatRealPart_SeqBAIJ,
2703: MatImaginaryPart_SeqBAIJ,
2704: 0,
2705: 0,
2706: /*109*/ 0,
2707: 0,
2708: 0,
2709: 0,
2710: MatMissingDiagonal_SeqBAIJ,
2711: /*114*/ 0,
2712: 0,
2713: 0,
2714: 0,
2715: 0,
2716: /*119*/ 0,
2717: 0,
2718: MatMultHermitianTranspose_SeqBAIJ,
2719: MatMultHermitianTransposeAdd_SeqBAIJ,
2720: 0,
2721: /*124*/ 0,
2722: 0,
2723: MatInvertBlockDiagonal_SeqBAIJ,
2724: 0,
2725: 0,
2726: /*129*/ 0,
2727: 0,
2728: 0,
2729: 0,
2730: 0,
2731: /*134*/ 0,
2732: 0,
2733: 0,
2734: 0,
2735: 0,
2736: /*139*/ 0,
2737: 0,
2738: 0,
2739: MatFDColoringSetUp_SeqXAIJ
2740: };
2744: PetscErrorCode MatStoreValues_SeqBAIJ(Mat mat)
2745: {
2746: Mat_SeqBAIJ *aij = (Mat_SeqBAIJ*)mat->data;
2747: PetscInt nz = aij->i[aij->mbs]*aij->bs2;
2751: if (aij->nonew != 1) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ORDER,"Must call MatSetOption(A,MAT_NEW_NONZERO_LOCATIONS,PETSC_FALSE);first");
2753: /* allocate space for values if not already there */
2754: if (!aij->saved_values) {
2755: PetscMalloc1((nz+1),&aij->saved_values);
2756: PetscLogObjectMemory((PetscObject)mat,(nz+1)*sizeof(PetscScalar));
2757: }
2759: /* copy values over */
2760: PetscMemcpy(aij->saved_values,aij->a,nz*sizeof(PetscScalar));
2761: return(0);
2762: }
2766: PetscErrorCode MatRetrieveValues_SeqBAIJ(Mat mat)
2767: {
2768: Mat_SeqBAIJ *aij = (Mat_SeqBAIJ*)mat->data;
2770: PetscInt nz = aij->i[aij->mbs]*aij->bs2;
2773: if (aij->nonew != 1) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ORDER,"Must call MatSetOption(A,MAT_NEW_NONZERO_LOCATIONS,PETSC_FALSE);first");
2774: if (!aij->saved_values) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ORDER,"Must call MatStoreValues(A);first");
2776: /* copy values over */
2777: PetscMemcpy(aij->a,aij->saved_values,nz*sizeof(PetscScalar));
2778: return(0);
2779: }
2781: PETSC_EXTERN PetscErrorCode MatConvert_SeqBAIJ_SeqAIJ(Mat, MatType,MatReuse,Mat*);
2782: PETSC_EXTERN PetscErrorCode MatConvert_SeqBAIJ_SeqSBAIJ(Mat, MatType,MatReuse,Mat*);
2786: PetscErrorCode MatSeqBAIJSetPreallocation_SeqBAIJ(Mat B,PetscInt bs,PetscInt nz,PetscInt *nnz)
2787: {
2788: Mat_SeqBAIJ *b;
2790: PetscInt i,mbs,nbs,bs2;
2791: PetscBool flg,skipallocation = PETSC_FALSE,realalloc = PETSC_FALSE;
2794: if (nz >= 0 || nnz) realalloc = PETSC_TRUE;
2795: if (nz == MAT_SKIP_ALLOCATION) {
2796: skipallocation = PETSC_TRUE;
2797: nz = 0;
2798: }
2800: MatSetBlockSize(B,PetscAbs(bs));
2801: PetscLayoutSetUp(B->rmap);
2802: PetscLayoutSetUp(B->cmap);
2803: PetscLayoutGetBlockSize(B->rmap,&bs);
2805: B->preallocated = PETSC_TRUE;
2807: mbs = B->rmap->n/bs;
2808: nbs = B->cmap->n/bs;
2809: bs2 = bs*bs;
2811: if (mbs*bs!=B->rmap->n || nbs*bs!=B->cmap->n) SETERRQ3(PETSC_COMM_SELF,PETSC_ERR_ARG_SIZ,"Number rows %D, cols %D must be divisible by blocksize %D",B->rmap->N,B->cmap->n,bs);
2813: if (nz == PETSC_DEFAULT || nz == PETSC_DECIDE) nz = 5;
2814: if (nz < 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"nz cannot be less than 0: value %D",nz);
2815: if (nnz) {
2816: for (i=0; i<mbs; i++) {
2817: if (nnz[i] < 0) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"nnz cannot be less than 0: local row %D value %D",i,nnz[i]);
2818: if (nnz[i] > nbs) SETERRQ3(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"nnz cannot be greater than block row length: local row %D value %D rowlength %D",i,nnz[i],nbs);
2819: }
2820: }
2822: b = (Mat_SeqBAIJ*)B->data;
2823: PetscOptionsBegin(PetscObjectComm((PetscObject)B),NULL,"Optimize options for SEQBAIJ matrix 2 ","Mat");
2824: PetscOptionsBool("-mat_no_unroll","Do not optimize for block size (slow)",NULL,PETSC_FALSE,&flg,NULL);
2825: PetscOptionsEnd();
2827: if (!flg) {
2828: switch (bs) {
2829: case 1:
2830: B->ops->mult = MatMult_SeqBAIJ_1;
2831: B->ops->multadd = MatMultAdd_SeqBAIJ_1;
2832: break;
2833: case 2:
2834: B->ops->mult = MatMult_SeqBAIJ_2;
2835: B->ops->multadd = MatMultAdd_SeqBAIJ_2;
2836: break;
2837: case 3:
2838: B->ops->mult = MatMult_SeqBAIJ_3;
2839: B->ops->multadd = MatMultAdd_SeqBAIJ_3;
2840: break;
2841: case 4:
2842: B->ops->mult = MatMult_SeqBAIJ_4;
2843: B->ops->multadd = MatMultAdd_SeqBAIJ_4;
2844: break;
2845: case 5:
2846: B->ops->mult = MatMult_SeqBAIJ_5;
2847: B->ops->multadd = MatMultAdd_SeqBAIJ_5;
2848: break;
2849: case 6:
2850: B->ops->mult = MatMult_SeqBAIJ_6;
2851: B->ops->multadd = MatMultAdd_SeqBAIJ_6;
2852: break;
2853: case 7:
2854: B->ops->mult = MatMult_SeqBAIJ_7;
2855: B->ops->multadd = MatMultAdd_SeqBAIJ_7;
2856: break;
2857: case 15:
2858: B->ops->mult = MatMult_SeqBAIJ_15_ver1;
2859: B->ops->multadd = MatMultAdd_SeqBAIJ_N;
2860: break;
2861: default:
2862: B->ops->mult = MatMult_SeqBAIJ_N;
2863: B->ops->multadd = MatMultAdd_SeqBAIJ_N;
2864: break;
2865: }
2866: }
2867: B->ops->sor = MatSOR_SeqBAIJ;
2868: b->mbs = mbs;
2869: b->nbs = nbs;
2870: if (!skipallocation) {
2871: if (!b->imax) {
2872: PetscMalloc2(mbs,&b->imax,mbs,&b->ilen);
2873: PetscLogObjectMemory((PetscObject)B,2*mbs*sizeof(PetscInt));
2875: b->free_imax_ilen = PETSC_TRUE;
2876: }
2877: /* b->ilen will count nonzeros in each block row so far. */
2878: for (i=0; i<mbs; i++) b->ilen[i] = 0;
2879: if (!nnz) {
2880: if (nz == PETSC_DEFAULT || nz == PETSC_DECIDE) nz = 5;
2881: else if (nz < 0) nz = 1;
2882: for (i=0; i<mbs; i++) b->imax[i] = nz;
2883: nz = nz*mbs;
2884: } else {
2885: nz = 0;
2886: for (i=0; i<mbs; i++) {b->imax[i] = nnz[i]; nz += nnz[i];}
2887: }
2889: /* allocate the matrix space */
2890: MatSeqXAIJFreeAIJ(B,&b->a,&b->j,&b->i);
2891: PetscMalloc3(bs2*nz,&b->a,nz,&b->j,B->rmap->N+1,&b->i);
2892: PetscLogObjectMemory((PetscObject)B,(B->rmap->N+1)*sizeof(PetscInt)+nz*(bs2*sizeof(PetscScalar)+sizeof(PetscInt)));
2893: PetscMemzero(b->a,nz*bs2*sizeof(MatScalar));
2894: PetscMemzero(b->j,nz*sizeof(PetscInt));
2896: b->singlemalloc = PETSC_TRUE;
2897: b->i[0] = 0;
2898: for (i=1; i<mbs+1; i++) {
2899: b->i[i] = b->i[i-1] + b->imax[i-1];
2900: }
2901: b->free_a = PETSC_TRUE;
2902: b->free_ij = PETSC_TRUE;
2903: #if defined(PETSC_THREADCOMM_ACTIVE)
2904: MatZeroEntries_SeqBAIJ(B);
2905: #endif
2906: } else {
2907: b->free_a = PETSC_FALSE;
2908: b->free_ij = PETSC_FALSE;
2909: }
2911: b->bs2 = bs2;
2912: b->mbs = mbs;
2913: b->nz = 0;
2914: b->maxnz = nz;
2915: B->info.nz_unneeded = (PetscReal)b->maxnz*bs2;
2916: if (realalloc) {MatSetOption(B,MAT_NEW_NONZERO_ALLOCATION_ERR,PETSC_TRUE);}
2917: return(0);
2918: }
2922: PetscErrorCode MatSeqBAIJSetPreallocationCSR_SeqBAIJ(Mat B,PetscInt bs,const PetscInt ii[],const PetscInt jj[],const PetscScalar V[])
2923: {
2924: PetscInt i,m,nz,nz_max=0,*nnz;
2925: PetscScalar *values=0;
2926: PetscBool roworiented = ((Mat_SeqBAIJ*)B->data)->roworiented;
2930: if (bs < 1) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Invalid block size specified, must be positive but it is %D",bs);
2931: PetscLayoutSetBlockSize(B->rmap,bs);
2932: PetscLayoutSetBlockSize(B->cmap,bs);
2933: PetscLayoutSetUp(B->rmap);
2934: PetscLayoutSetUp(B->cmap);
2935: PetscLayoutGetBlockSize(B->rmap,&bs);
2936: m = B->rmap->n/bs;
2938: if (ii[0] != 0) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE, "ii[0] must be 0 but it is %D",ii[0]);
2939: PetscMalloc1((m+1), &nnz);
2940: for (i=0; i<m; i++) {
2941: nz = ii[i+1]- ii[i];
2942: if (nz < 0) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE, "Local row %D has a negative number of columns %D",i,nz);
2943: nz_max = PetscMax(nz_max, nz);
2944: nnz[i] = nz;
2945: }
2946: MatSeqBAIJSetPreallocation(B,bs,0,nnz);
2947: PetscFree(nnz);
2949: values = (PetscScalar*)V;
2950: if (!values) {
2951: PetscCalloc1(bs*bs*(nz_max+1),&values);
2952: }
2953: for (i=0; i<m; i++) {
2954: PetscInt ncols = ii[i+1] - ii[i];
2955: const PetscInt *icols = jj + ii[i];
2956: const PetscScalar *svals = values + (V ? (bs*bs*ii[i]) : 0);
2957: if (!roworiented) {
2958: MatSetValuesBlocked_SeqBAIJ(B,1,&i,ncols,icols,svals,INSERT_VALUES);
2959: } else {
2960: PetscInt j;
2961: for (j=0; j<ncols; j++) {
2962: const PetscScalar *svals = values + (V ? (bs*bs*(ii[i]+j)) : 0);
2963: MatSetValuesBlocked_SeqBAIJ(B,1,&i,1,&icols[j],svals,INSERT_VALUES);
2964: }
2965: }
2966: }
2967: if (!V) { PetscFree(values); }
2968: MatAssemblyBegin(B,MAT_FINAL_ASSEMBLY);
2969: MatAssemblyEnd(B,MAT_FINAL_ASSEMBLY);
2970: MatSetOption(B,MAT_NEW_NONZERO_LOCATION_ERR,PETSC_TRUE);
2971: return(0);
2972: }
2974: PETSC_EXTERN PetscErrorCode MatGetFactor_seqbaij_petsc(Mat,MatFactorType,Mat*);
2975: PETSC_EXTERN PetscErrorCode MatGetFactor_seqbaij_bstrm(Mat,MatFactorType,Mat*);
2976: #if defined(PETSC_HAVE_MUMPS)
2977: PETSC_EXTERN PetscErrorCode MatGetFactor_baij_mumps(Mat,MatFactorType,Mat*);
2978: #endif
2979: extern PetscErrorCode MatGetFactorAvailable_seqbaij_petsc(Mat,MatFactorType,PetscBool*);
2981: /*MC
2982: MATSEQBAIJ - MATSEQBAIJ = "seqbaij" - A matrix type to be used for sequential block sparse matrices, based on
2983: block sparse compressed row format.
2985: Options Database Keys:
2986: . -mat_type seqbaij - sets the matrix type to "seqbaij" during a call to MatSetFromOptions()
2988: Level: beginner
2990: .seealso: MatCreateSeqBAIJ()
2991: M*/
2993: PETSC_EXTERN PetscErrorCode MatConvert_SeqBAIJ_SeqBSTRM(Mat, MatType,MatReuse,Mat*);
2997: PETSC_EXTERN PetscErrorCode MatCreate_SeqBAIJ(Mat B)
2998: {
3000: PetscMPIInt size;
3001: Mat_SeqBAIJ *b;
3004: MPI_Comm_size(PetscObjectComm((PetscObject)B),&size);
3005: if (size > 1) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"Comm must be of size 1");
3007: PetscNewLog(B,&b);
3008: B->data = (void*)b;
3009: PetscMemcpy(B->ops,&MatOps_Values,sizeof(struct _MatOps));
3011: b->row = 0;
3012: b->col = 0;
3013: b->icol = 0;
3014: b->reallocs = 0;
3015: b->saved_values = 0;
3017: b->roworiented = PETSC_TRUE;
3018: b->nonew = 0;
3019: b->diag = 0;
3020: B->spptr = 0;
3021: B->info.nz_unneeded = (PetscReal)b->maxnz*b->bs2;
3022: b->keepnonzeropattern = PETSC_FALSE;
3023: b->xtoy = 0;
3024: b->XtoY = 0;
3026: PetscObjectComposeFunction((PetscObject)B,"MatGetFactorAvailable_petsc_C",MatGetFactorAvailable_seqbaij_petsc);
3027: PetscObjectComposeFunction((PetscObject)B,"MatGetFactor_petsc_C",MatGetFactor_seqbaij_petsc);
3028: PetscObjectComposeFunction((PetscObject)B,"MatGetFactor_bstrm_C",MatGetFactor_seqbaij_bstrm);
3029: #if defined(PETSC_HAVE_MUMPS)
3030: PetscObjectComposeFunction((PetscObject)B,"MatGetFactor_mumps_C", MatGetFactor_baij_mumps);
3031: #endif
3032: PetscObjectComposeFunction((PetscObject)B,"MatInvertBlockDiagonal_C",MatInvertBlockDiagonal_SeqBAIJ);
3033: PetscObjectComposeFunction((PetscObject)B,"MatStoreValues_C",MatStoreValues_SeqBAIJ);
3034: PetscObjectComposeFunction((PetscObject)B,"MatRetrieveValues_C",MatRetrieveValues_SeqBAIJ);
3035: PetscObjectComposeFunction((PetscObject)B,"MatSeqBAIJSetColumnIndices_C",MatSeqBAIJSetColumnIndices_SeqBAIJ);
3036: PetscObjectComposeFunction((PetscObject)B,"MatConvert_seqbaij_seqaij_C",MatConvert_SeqBAIJ_SeqAIJ);
3037: PetscObjectComposeFunction((PetscObject)B,"MatConvert_seqbaij_seqsbaij_C",MatConvert_SeqBAIJ_SeqSBAIJ);
3038: PetscObjectComposeFunction((PetscObject)B,"MatSeqBAIJSetPreallocation_C",MatSeqBAIJSetPreallocation_SeqBAIJ);
3039: PetscObjectComposeFunction((PetscObject)B,"MatSeqBAIJSetPreallocationCSR_C",MatSeqBAIJSetPreallocationCSR_SeqBAIJ);
3040: PetscObjectComposeFunction((PetscObject)B,"MatConvert_seqbaij_seqbstrm_C",MatConvert_SeqBAIJ_SeqBSTRM);
3041: PetscObjectComposeFunction((PetscObject)B,"MatIsTranspose_C",MatIsTranspose_SeqBAIJ);
3042: PetscObjectChangeTypeName((PetscObject)B,MATSEQBAIJ);
3043: return(0);
3044: }
3048: PetscErrorCode MatDuplicateNoCreate_SeqBAIJ(Mat C,Mat A,MatDuplicateOption cpvalues,PetscBool mallocmatspace)
3049: {
3050: Mat_SeqBAIJ *c = (Mat_SeqBAIJ*)C->data,*a = (Mat_SeqBAIJ*)A->data;
3052: PetscInt i,mbs = a->mbs,nz = a->nz,bs2 = a->bs2;
3055: if (a->i[mbs] != nz) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_PLIB,"Corrupt matrix");
3057: if (cpvalues == MAT_SHARE_NONZERO_PATTERN) {
3058: c->imax = a->imax;
3059: c->ilen = a->ilen;
3060: c->free_imax_ilen = PETSC_FALSE;
3061: } else {
3062: PetscMalloc2(mbs,&c->imax,mbs,&c->ilen);
3063: PetscLogObjectMemory((PetscObject)C,2*mbs*sizeof(PetscInt));
3064: for (i=0; i<mbs; i++) {
3065: c->imax[i] = a->imax[i];
3066: c->ilen[i] = a->ilen[i];
3067: }
3068: c->free_imax_ilen = PETSC_TRUE;
3069: }
3071: /* allocate the matrix space */
3072: if (mallocmatspace) {
3073: if (cpvalues == MAT_SHARE_NONZERO_PATTERN) {
3074: PetscCalloc1(bs2*nz,&c->a);
3075: PetscLogObjectMemory((PetscObject)C,a->i[mbs]*bs2*sizeof(PetscScalar));
3077: c->i = a->i;
3078: c->j = a->j;
3079: c->singlemalloc = PETSC_FALSE;
3080: c->free_a = PETSC_TRUE;
3081: c->free_ij = PETSC_FALSE;
3082: c->parent = A;
3083: C->preallocated = PETSC_TRUE;
3084: C->assembled = PETSC_TRUE;
3086: PetscObjectReference((PetscObject)A);
3087: MatSetOption(A,MAT_NEW_NONZERO_LOCATION_ERR,PETSC_TRUE);
3088: MatSetOption(C,MAT_NEW_NONZERO_LOCATION_ERR,PETSC_TRUE);
3089: } else {
3090: PetscMalloc3(bs2*nz,&c->a,nz,&c->j,mbs+1,&c->i);
3091: PetscLogObjectMemory((PetscObject)C,a->i[mbs]*(bs2*sizeof(PetscScalar)+sizeof(PetscInt))+(mbs+1)*sizeof(PetscInt));
3093: c->singlemalloc = PETSC_TRUE;
3094: c->free_a = PETSC_TRUE;
3095: c->free_ij = PETSC_TRUE;
3097: PetscMemcpy(c->i,a->i,(mbs+1)*sizeof(PetscInt));
3098: if (mbs > 0) {
3099: PetscMemcpy(c->j,a->j,nz*sizeof(PetscInt));
3100: if (cpvalues == MAT_COPY_VALUES) {
3101: PetscMemcpy(c->a,a->a,bs2*nz*sizeof(MatScalar));
3102: } else {
3103: PetscMemzero(c->a,bs2*nz*sizeof(MatScalar));
3104: }
3105: }
3106: C->preallocated = PETSC_TRUE;
3107: C->assembled = PETSC_TRUE;
3108: }
3109: }
3111: c->roworiented = a->roworiented;
3112: c->nonew = a->nonew;
3114: PetscLayoutReference(A->rmap,&C->rmap);
3115: PetscLayoutReference(A->cmap,&C->cmap);
3117: c->bs2 = a->bs2;
3118: c->mbs = a->mbs;
3119: c->nbs = a->nbs;
3121: if (a->diag) {
3122: if (cpvalues == MAT_SHARE_NONZERO_PATTERN) {
3123: c->diag = a->diag;
3124: c->free_diag = PETSC_FALSE;
3125: } else {
3126: PetscMalloc1((mbs+1),&c->diag);
3127: PetscLogObjectMemory((PetscObject)C,(mbs+1)*sizeof(PetscInt));
3128: for (i=0; i<mbs; i++) c->diag[i] = a->diag[i];
3129: c->free_diag = PETSC_TRUE;
3130: }
3131: } else c->diag = 0;
3133: c->nz = a->nz;
3134: c->maxnz = a->nz; /* Since we allocate exactly the right amount */
3135: c->solve_work = NULL;
3136: c->mult_work = NULL;
3137: c->sor_workt = NULL;
3138: c->sor_work = NULL;
3140: c->compressedrow.use = a->compressedrow.use;
3141: c->compressedrow.nrows = a->compressedrow.nrows;
3142: if (a->compressedrow.use) {
3143: i = a->compressedrow.nrows;
3144: PetscMalloc2(i+1,&c->compressedrow.i,i+1,&c->compressedrow.rindex);
3145: PetscLogObjectMemory((PetscObject)C,(2*i+1)*sizeof(PetscInt));
3146: PetscMemcpy(c->compressedrow.i,a->compressedrow.i,(i+1)*sizeof(PetscInt));
3147: PetscMemcpy(c->compressedrow.rindex,a->compressedrow.rindex,i*sizeof(PetscInt));
3148: } else {
3149: c->compressedrow.use = PETSC_FALSE;
3150: c->compressedrow.i = NULL;
3151: c->compressedrow.rindex = NULL;
3152: }
3153: C->nonzerostate = A->nonzerostate;
3155: PetscFunctionListDuplicate(((PetscObject)A)->qlist,&((PetscObject)C)->qlist);
3156: PetscMemcpy(C->ops,A->ops,sizeof(struct _MatOps));
3157: return(0);
3158: }
3162: PetscErrorCode MatDuplicate_SeqBAIJ(Mat A,MatDuplicateOption cpvalues,Mat *B)
3163: {
3167: MatCreate(PetscObjectComm((PetscObject)A),B);
3168: MatSetSizes(*B,A->rmap->N,A->cmap->n,A->rmap->N,A->cmap->n);
3169: MatSetType(*B,MATSEQBAIJ);
3170: MatDuplicateNoCreate_SeqBAIJ(*B,A,cpvalues,PETSC_TRUE);
3171: return(0);
3172: }
3176: PetscErrorCode MatLoad_SeqBAIJ(Mat newmat,PetscViewer viewer)
3177: {
3178: Mat_SeqBAIJ *a;
3180: PetscInt i,nz,header[4],*rowlengths=0,M,N,bs = newmat->rmap->bs;
3181: PetscInt *mask,mbs,*jj,j,rowcount,nzcount,k,*browlengths,maskcount;
3182: PetscInt kmax,jcount,block,idx,point,nzcountb,extra_rows,rows,cols;
3183: PetscInt *masked,nmask,tmp,bs2,ishift;
3184: PetscMPIInt size;
3185: int fd;
3186: PetscScalar *aa;
3187: MPI_Comm comm;
3190: PetscObjectGetComm((PetscObject)viewer,&comm);
3191: PetscOptionsBegin(comm,NULL,"Options for loading SEQBAIJ matrix","Mat");
3192: PetscOptionsInt("-matload_block_size","Set the blocksize used to store the matrix","MatLoad",bs,&bs,NULL);
3193: PetscOptionsEnd();
3194: if (bs < 0) bs = 1;
3195: bs2 = bs*bs;
3197: MPI_Comm_size(comm,&size);
3198: if (size > 1) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONG,"view must have one processor");
3199: PetscViewerBinaryGetDescriptor(viewer,&fd);
3200: PetscBinaryRead(fd,header,4,PETSC_INT);
3201: if (header[0] != MAT_FILE_CLASSID) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_FILE_UNEXPECTED,"not Mat object");
3202: M = header[1]; N = header[2]; nz = header[3];
3204: if (header[3] < 0) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_FILE_UNEXPECTED,"Matrix stored in special format, cannot load as SeqBAIJ");
3205: if (M != N) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_SUP,"Can only do square matrices");
3207: /*
3208: This code adds extra rows to make sure the number of rows is
3209: divisible by the blocksize
3210: */
3211: mbs = M/bs;
3212: extra_rows = bs - M + bs*(mbs);
3213: if (extra_rows == bs) extra_rows = 0;
3214: else mbs++;
3215: if (extra_rows) {
3216: PetscInfo(viewer,"Padding loaded matrix to match blocksize\n");
3217: }
3219: /* Set global sizes if not already set */
3220: if (newmat->rmap->n < 0 && newmat->rmap->N < 0 && newmat->cmap->n < 0 && newmat->cmap->N < 0) {
3221: MatSetSizes(newmat,PETSC_DECIDE,PETSC_DECIDE,M+extra_rows,N+extra_rows);
3222: } else { /* Check if the matrix global sizes are correct */
3223: MatGetSize(newmat,&rows,&cols);
3224: if (rows < 0 && cols < 0) { /* user might provide local size instead of global size */
3225: MatGetLocalSize(newmat,&rows,&cols);
3226: }
3227: if (M != rows || N != cols) SETERRQ4(PETSC_COMM_SELF,PETSC_ERR_FILE_UNEXPECTED,"Matrix in file of different length (%d, %d) than the input matrix (%d, %d)",M,N,rows,cols);
3228: }
3230: /* read in row lengths */
3231: PetscMalloc1((M+extra_rows),&rowlengths);
3232: PetscBinaryRead(fd,rowlengths,M,PETSC_INT);
3233: for (i=0; i<extra_rows; i++) rowlengths[M+i] = 1;
3235: /* read in column indices */
3236: PetscMalloc1((nz+extra_rows),&jj);
3237: PetscBinaryRead(fd,jj,nz,PETSC_INT);
3238: for (i=0; i<extra_rows; i++) jj[nz+i] = M+i;
3240: /* loop over row lengths determining block row lengths */
3241: PetscCalloc1(mbs,&browlengths);
3242: PetscMalloc2(mbs,&mask,mbs,&masked);
3243: PetscMemzero(mask,mbs*sizeof(PetscInt));
3244: rowcount = 0;
3245: nzcount = 0;
3246: for (i=0; i<mbs; i++) {
3247: nmask = 0;
3248: for (j=0; j<bs; j++) {
3249: kmax = rowlengths[rowcount];
3250: for (k=0; k<kmax; k++) {
3251: tmp = jj[nzcount++]/bs;
3252: if (!mask[tmp]) {masked[nmask++] = tmp; mask[tmp] = 1;}
3253: }
3254: rowcount++;
3255: }
3256: browlengths[i] += nmask;
3257: /* zero out the mask elements we set */
3258: for (j=0; j<nmask; j++) mask[masked[j]] = 0;
3259: }
3261: /* Do preallocation */
3262: MatSeqBAIJSetPreallocation_SeqBAIJ(newmat,bs,0,browlengths);
3263: a = (Mat_SeqBAIJ*)newmat->data;
3265: /* set matrix "i" values */
3266: a->i[0] = 0;
3267: for (i=1; i<= mbs; i++) {
3268: a->i[i] = a->i[i-1] + browlengths[i-1];
3269: a->ilen[i-1] = browlengths[i-1];
3270: }
3271: a->nz = 0;
3272: for (i=0; i<mbs; i++) a->nz += browlengths[i];
3274: /* read in nonzero values */
3275: PetscMalloc1((nz+extra_rows),&aa);
3276: PetscBinaryRead(fd,aa,nz,PETSC_SCALAR);
3277: for (i=0; i<extra_rows; i++) aa[nz+i] = 1.0;
3279: /* set "a" and "j" values into matrix */
3280: nzcount = 0; jcount = 0;
3281: for (i=0; i<mbs; i++) {
3282: nzcountb = nzcount;
3283: nmask = 0;
3284: for (j=0; j<bs; j++) {
3285: kmax = rowlengths[i*bs+j];
3286: for (k=0; k<kmax; k++) {
3287: tmp = jj[nzcount++]/bs;
3288: if (!mask[tmp]) { masked[nmask++] = tmp; mask[tmp] = 1;}
3289: }
3290: }
3291: /* sort the masked values */
3292: PetscSortInt(nmask,masked);
3294: /* set "j" values into matrix */
3295: maskcount = 1;
3296: for (j=0; j<nmask; j++) {
3297: a->j[jcount++] = masked[j];
3298: mask[masked[j]] = maskcount++;
3299: }
3300: /* set "a" values into matrix */
3301: ishift = bs2*a->i[i];
3302: for (j=0; j<bs; j++) {
3303: kmax = rowlengths[i*bs+j];
3304: for (k=0; k<kmax; k++) {
3305: tmp = jj[nzcountb]/bs;
3306: block = mask[tmp] - 1;
3307: point = jj[nzcountb] - bs*tmp;
3308: idx = ishift + bs2*block + j + bs*point;
3309: a->a[idx] = (MatScalar)aa[nzcountb++];
3310: }
3311: }
3312: /* zero out the mask elements we set */
3313: for (j=0; j<nmask; j++) mask[masked[j]] = 0;
3314: }
3315: if (jcount != a->nz) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_FILE_UNEXPECTED,"Bad binary matrix");
3317: PetscFree(rowlengths);
3318: PetscFree(browlengths);
3319: PetscFree(aa);
3320: PetscFree(jj);
3321: PetscFree2(mask,masked);
3323: MatAssemblyBegin(newmat,MAT_FINAL_ASSEMBLY);
3324: MatAssemblyEnd(newmat,MAT_FINAL_ASSEMBLY);
3325: return(0);
3326: }
3330: /*@C
3331: MatCreateSeqBAIJ - Creates a sparse matrix in block AIJ (block
3332: compressed row) format. For good matrix assembly performance the
3333: user should preallocate the matrix storage by setting the parameter nz
3334: (or the array nnz). By setting these parameters accurately, performance
3335: during matrix assembly can be increased by more than a factor of 50.
3337: Collective on MPI_Comm
3339: Input Parameters:
3340: + comm - MPI communicator, set to PETSC_COMM_SELF
3341: . bs - size of block
3342: . m - number of rows
3343: . n - number of columns
3344: . nz - number of nonzero blocks per block row (same for all rows)
3345: - nnz - array containing the number of nonzero blocks in the various block rows
3346: (possibly different for each block row) or NULL
3348: Output Parameter:
3349: . A - the matrix
3351: It is recommended that one use the MatCreate(), MatSetType() and/or MatSetFromOptions(),
3352: MatXXXXSetPreallocation() paradgm instead of this routine directly.
3353: [MatXXXXSetPreallocation() is, for example, MatSeqAIJSetPreallocation]
3355: Options Database Keys:
3356: . -mat_no_unroll - uses code that does not unroll the loops in the
3357: block calculations (much slower)
3358: . -mat_block_size - size of the blocks to use
3360: Level: intermediate
3362: Notes:
3363: The number of rows and columns must be divisible by blocksize.
3365: If the nnz parameter is given then the nz parameter is ignored
3367: A nonzero block is any block that as 1 or more nonzeros in it
3369: The block AIJ format is fully compatible with standard Fortran 77
3370: storage. That is, the stored row and column indices can begin at
3371: either one (as in Fortran) or zero. See the users' manual for details.
3373: Specify the preallocated storage with either nz or nnz (not both).
3374: Set nz=PETSC_DEFAULT and nnz=NULL for PETSc to control dynamic memory
3375: allocation. See Users-Manual: ch_mat for details.
3376: matrices.
3378: .seealso: MatCreate(), MatCreateSeqAIJ(), MatSetValues(), MatCreateBAIJ()
3379: @*/
3380: PetscErrorCode MatCreateSeqBAIJ(MPI_Comm comm,PetscInt bs,PetscInt m,PetscInt n,PetscInt nz,const PetscInt nnz[],Mat *A)
3381: {
3385: MatCreate(comm,A);
3386: MatSetSizes(*A,m,n,m,n);
3387: MatSetType(*A,MATSEQBAIJ);
3388: MatSeqBAIJSetPreallocation_SeqBAIJ(*A,bs,nz,(PetscInt*)nnz);
3389: return(0);
3390: }
3394: /*@C
3395: MatSeqBAIJSetPreallocation - Sets the block size and expected nonzeros
3396: per row in the matrix. For good matrix assembly performance the
3397: user should preallocate the matrix storage by setting the parameter nz
3398: (or the array nnz). By setting these parameters accurately, performance
3399: during matrix assembly can be increased by more than a factor of 50.
3401: Collective on MPI_Comm
3403: Input Parameters:
3404: + B - the matrix
3405: . bs - size of block
3406: . nz - number of block nonzeros per block row (same for all rows)
3407: - nnz - array containing the number of block nonzeros in the various block rows
3408: (possibly different for each block row) or NULL
3410: Options Database Keys:
3411: . -mat_no_unroll - uses code that does not unroll the loops in the
3412: block calculations (much slower)
3413: . -mat_block_size - size of the blocks to use
3415: Level: intermediate
3417: Notes:
3418: If the nnz parameter is given then the nz parameter is ignored
3420: You can call MatGetInfo() to get information on how effective the preallocation was;
3421: for example the fields mallocs,nz_allocated,nz_used,nz_unneeded;
3422: You can also run with the option -info and look for messages with the string
3423: malloc in them to see if additional memory allocation was needed.
3425: The block AIJ format is fully compatible with standard Fortran 77
3426: storage. That is, the stored row and column indices can begin at
3427: either one (as in Fortran) or zero. See the users' manual for details.
3429: Specify the preallocated storage with either nz or nnz (not both).
3430: Set nz=PETSC_DEFAULT and nnz=NULL for PETSc to control dynamic memory
3431: allocation. See Users-Manual: ch_mat for details.
3433: .seealso: MatCreate(), MatCreateSeqAIJ(), MatSetValues(), MatCreateBAIJ(), MatGetInfo()
3434: @*/
3435: PetscErrorCode MatSeqBAIJSetPreallocation(Mat B,PetscInt bs,PetscInt nz,const PetscInt nnz[])
3436: {
3443: PetscTryMethod(B,"MatSeqBAIJSetPreallocation_C",(Mat,PetscInt,PetscInt,const PetscInt[]),(B,bs,nz,nnz));
3444: return(0);
3445: }
3449: /*@C
3450: MatSeqBAIJSetPreallocationCSR - Allocates memory for a sparse sequential matrix in AIJ format
3451: (the default sequential PETSc format).
3453: Collective on MPI_Comm
3455: Input Parameters:
3456: + B - the matrix
3457: . i - the indices into j for the start of each local row (starts with zero)
3458: . j - the column indices for each local row (starts with zero) these must be sorted for each row
3459: - v - optional values in the matrix
3461: Level: developer
3463: Notes:
3464: The order of the entries in values is specified by the MatOption MAT_ROW_ORIENTED. For example, C programs
3465: may want to use the default MAT_ROW_ORIENTED=PETSC_TRUE and use an array v[nnz][bs][bs] where the second index is
3466: over rows within a block and the last index is over columns within a block row. Fortran programs will likely set
3467: MAT_ROW_ORIENTED=PETSC_FALSE and use a Fortran array v(bs,bs,nnz) in which the first index is over rows within a
3468: block column and the second index is over columns within a block.
3470: .keywords: matrix, aij, compressed row, sparse
3472: .seealso: MatCreate(), MatCreateSeqBAIJ(), MatSetValues(), MatSeqBAIJSetPreallocation(), MATSEQBAIJ
3473: @*/
3474: PetscErrorCode MatSeqBAIJSetPreallocationCSR(Mat B,PetscInt bs,const PetscInt i[],const PetscInt j[], const PetscScalar v[])
3475: {
3482: PetscTryMethod(B,"MatSeqBAIJSetPreallocationCSR_C",(Mat,PetscInt,const PetscInt[],const PetscInt[],const PetscScalar[]),(B,bs,i,j,v));
3483: return(0);
3484: }
3489: /*@
3490: MatCreateSeqBAIJWithArrays - Creates an sequential BAIJ matrix using matrix elements provided by the user.
3492: Collective on MPI_Comm
3494: Input Parameters:
3495: + comm - must be an MPI communicator of size 1
3496: . bs - size of block
3497: . m - number of rows
3498: . n - number of columns
3499: . i - row indices
3500: . j - column indices
3501: - a - matrix values
3503: Output Parameter:
3504: . mat - the matrix
3506: Level: advanced
3508: Notes:
3509: The i, j, and a arrays are not copied by this routine, the user must free these arrays
3510: once the matrix is destroyed
3512: You cannot set new nonzero locations into this matrix, that will generate an error.
3514: The i and j indices are 0 based
3516: When block size is greater than 1 the matrix values must be stored using the BAIJ storage format (see the BAIJ code to determine this).
3518: The order of the entries in values is the same as the block compressed sparse row storage format; that is, it is
3519: the same as a three dimensional array in Fortran values(bs,bs,nnz) that contains the first column of the first
3520: block, followed by the second column of the first block etc etc. That is, the blocks are contiguous in memory
3521: with column-major ordering within blocks.
3523: .seealso: MatCreate(), MatCreateBAIJ(), MatCreateSeqBAIJ()
3525: @*/
3526: PetscErrorCode MatCreateSeqBAIJWithArrays(MPI_Comm comm,PetscInt bs,PetscInt m,PetscInt n,PetscInt *i,PetscInt *j,PetscScalar *a,Mat *mat)
3527: {
3529: PetscInt ii;
3530: Mat_SeqBAIJ *baij;
3533: if (bs != 1) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_SUP,"block size %D > 1 is not supported yet",bs);
3534: if (i[0]) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"i (row indices) must start with 0");
3536: MatCreate(comm,mat);
3537: MatSetSizes(*mat,m,n,m,n);
3538: MatSetType(*mat,MATSEQBAIJ);
3539: MatSeqBAIJSetPreallocation_SeqBAIJ(*mat,bs,MAT_SKIP_ALLOCATION,0);
3540: baij = (Mat_SeqBAIJ*)(*mat)->data;
3541: PetscMalloc2(m,&baij->imax,m,&baij->ilen);
3542: PetscLogObjectMemory((PetscObject)*mat,2*m*sizeof(PetscInt));
3544: baij->i = i;
3545: baij->j = j;
3546: baij->a = a;
3548: baij->singlemalloc = PETSC_FALSE;
3549: baij->nonew = -1; /*this indicates that inserting a new value in the matrix that generates a new nonzero is an error*/
3550: baij->free_a = PETSC_FALSE;
3551: baij->free_ij = PETSC_FALSE;
3553: for (ii=0; ii<m; ii++) {
3554: baij->ilen[ii] = baij->imax[ii] = i[ii+1] - i[ii];
3555: #if defined(PETSC_USE_DEBUG)
3556: if (i[ii+1] - i[ii] < 0) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Negative row length in i (row indices) row = %d length = %d",ii,i[ii+1] - i[ii]);
3557: #endif
3558: }
3559: #if defined(PETSC_USE_DEBUG)
3560: for (ii=0; ii<baij->i[m]; ii++) {
3561: if (j[ii] < 0) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Negative column index at location = %d index = %d",ii,j[ii]);
3562: if (j[ii] > n - 1) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Column index to large at location = %d index = %d",ii,j[ii]);
3563: }
3564: #endif
3566: MatAssemblyBegin(*mat,MAT_FINAL_ASSEMBLY);
3567: MatAssemblyEnd(*mat,MAT_FINAL_ASSEMBLY);
3568: return(0);
3569: }