Actual source code: matproduct.c
1: /*
2: Routines for matrix products. Calling procedure:
4: MatProductCreate(A,B,C,&D); or MatProductCreateWithMat(A,B,C,D)
5: MatProductSetType(D, MATPRODUCT_AB/AtB/ABt/PtAP/RARt/ABC)
6: MatProductSetAlgorithm(D, alg)
7: MatProductSetFill(D,fill)
8: MatProductSetFromOptions(D)
9: -> MatProductSetFromOptions_Private(D)
10: # Check matrix global sizes
11: if the matrices have the same setfromoptions routine, use it
12: if not, try:
13: -> Query MatProductSetFromOptions_Atype_Btype_Ctype_C(D) from A, B and C (in order)
14: if found -> run the specific setup that must set the symbolic operation (these callbacks should never fail)
15: if callback not found or no symbolic operation set
16: -> Query MatProductSetFromOptions_anytype_C(D) from A, B and C (in order) (e.g, matrices may have inner matrices like MATTRANSPOSEVIRTUAL)
17: if dispatch found but combination still not present do
18: -> check if B is dense and product type AtB or AB -> if true, basic looping of dense columns
19: -> check if triple product (PtAP, RARt or ABC) -> if true, set the Basic routines
21: # The setfromoptions calls MatProductSetFromOptions_Atype_Btype_Ctype should
22: # Check matrix local sizes for mpi matrices
23: # Set default algorithm
24: # Get runtime option
25: # Set D->ops->productsymbolic = MatProductSymbolic_productype_Atype_Btype_Ctype if found
27: MatProductSymbolic(D)
28: # Call MatProductSymbolic_productype_Atype_Btype_Ctype()
29: the callback must set the numeric phase D->ops->productnumeric = MatProductNumeric_productype_Atype_Btype_Ctype
31: MatProductNumeric(D)
32: # Call the numeric phase
34: # The symbolic phases are allowed to set extra data structures and attach those to the product
35: # this additional data can be reused between multiple numeric phases with the same matrices
36: # if not needed, call
37: MatProductClear(D)
38: */
40: #include <petsc/private/matimpl.h>
42: const char *const MatProductTypes[] = {"UNSPECIFIED", "AB", "AtB", "ABt", "PtAP", "RARt", "ABC"};
44: /* these are basic implementations relying on the old function pointers
45: * they are dangerous and should be removed in the future */
46: static PetscErrorCode MatProductNumeric_PtAP_Unsafe(Mat C)
47: {
48: Mat_Product *product = C->product;
49: Mat P = product->B, AP = product->Dwork;
51: PetscFunctionBegin;
52: /* AP = A*P */
53: PetscCall(MatProductNumeric(AP));
54: /* C = P^T*AP */
55: PetscCall((*C->ops->transposematmultnumeric)(P, AP, C));
56: PetscFunctionReturn(PETSC_SUCCESS);
57: }
59: static PetscErrorCode MatProductSymbolic_PtAP_Unsafe(Mat C)
60: {
61: Mat_Product *product = C->product;
62: Mat A = product->A, P = product->B, AP;
63: PetscReal fill = product->fill;
65: PetscFunctionBegin;
66: PetscCall(PetscInfo((PetscObject)C, "for A %s, P %s is used\n", ((PetscObject)product->A)->type_name, ((PetscObject)product->B)->type_name));
67: /* AP = A*P */
68: PetscCall(MatProductCreate(A, P, NULL, &AP));
69: PetscCall(MatProductSetType(AP, MATPRODUCT_AB));
70: PetscCall(MatProductSetAlgorithm(AP, MATPRODUCTALGORITHMDEFAULT));
71: PetscCall(MatProductSetFill(AP, fill));
72: PetscCall(MatProductSetFromOptions(AP));
73: PetscCall(MatProductSymbolic(AP));
75: /* C = P^T*AP */
76: PetscCall(MatProductSetType(C, MATPRODUCT_AtB));
77: PetscCall(MatProductSetAlgorithm(C, MATPRODUCTALGORITHMDEFAULT));
78: product->A = P;
79: product->B = AP;
80: PetscCall(MatProductSetFromOptions(C));
81: PetscCall(MatProductSymbolic(C));
83: /* resume user's original input matrix setting for A and B */
84: product->A = A;
85: product->B = P;
86: product->Dwork = AP;
88: C->ops->productnumeric = MatProductNumeric_PtAP_Unsafe;
89: PetscFunctionReturn(PETSC_SUCCESS);
90: }
92: static PetscErrorCode MatProductNumeric_RARt_Unsafe(Mat C)
93: {
94: Mat_Product *product = C->product;
95: Mat R = product->B, RA = product->Dwork;
97: PetscFunctionBegin;
98: /* RA = R*A */
99: PetscCall(MatProductNumeric(RA));
100: /* C = RA*R^T */
101: PetscCall((*C->ops->mattransposemultnumeric)(RA, R, C));
102: PetscFunctionReturn(PETSC_SUCCESS);
103: }
105: static PetscErrorCode MatProductSymbolic_RARt_Unsafe(Mat C)
106: {
107: Mat_Product *product = C->product;
108: Mat A = product->A, R = product->B, RA;
109: PetscReal fill = product->fill;
111: PetscFunctionBegin;
112: PetscCall(PetscInfo((PetscObject)C, "for A %s, R %s is used\n", ((PetscObject)product->A)->type_name, ((PetscObject)product->B)->type_name));
113: /* RA = R*A */
114: PetscCall(MatProductCreate(R, A, NULL, &RA));
115: PetscCall(MatProductSetType(RA, MATPRODUCT_AB));
116: PetscCall(MatProductSetAlgorithm(RA, MATPRODUCTALGORITHMDEFAULT));
117: PetscCall(MatProductSetFill(RA, fill));
118: PetscCall(MatProductSetFromOptions(RA));
119: PetscCall(MatProductSymbolic(RA));
121: /* C = RA*R^T */
122: PetscCall(MatProductSetType(C, MATPRODUCT_ABt));
123: PetscCall(MatProductSetAlgorithm(C, MATPRODUCTALGORITHMDEFAULT));
124: product->A = RA;
125: PetscCall(MatProductSetFromOptions(C));
126: PetscCall(MatProductSymbolic(C));
128: /* resume user's original input matrix setting for A */
129: product->A = A;
130: product->Dwork = RA; /* save here so it will be destroyed with product C */
131: C->ops->productnumeric = MatProductNumeric_RARt_Unsafe;
132: PetscFunctionReturn(PETSC_SUCCESS);
133: }
135: static PetscErrorCode MatProductNumeric_ABC_Unsafe(Mat mat)
136: {
137: Mat_Product *product = mat->product;
138: Mat A = product->A, BC = product->Dwork;
140: PetscFunctionBegin;
141: /* Numeric BC = B*C */
142: PetscCall(MatProductNumeric(BC));
143: /* Numeric mat = A*BC */
144: PetscCall((*mat->ops->matmultnumeric)(A, BC, mat));
145: PetscFunctionReturn(PETSC_SUCCESS);
146: }
148: static PetscErrorCode MatProductSymbolic_ABC_Unsafe(Mat mat)
149: {
150: Mat_Product *product = mat->product;
151: Mat B = product->B, C = product->C, BC;
152: PetscReal fill = product->fill;
154: PetscFunctionBegin;
155: PetscCall(PetscInfo((PetscObject)mat, "for A %s, B %s, C %s is used\n", ((PetscObject)product->A)->type_name, ((PetscObject)product->B)->type_name, ((PetscObject)product->C)->type_name));
156: /* Symbolic BC = B*C */
157: PetscCall(MatProductCreate(B, C, NULL, &BC));
158: PetscCall(MatProductSetType(BC, MATPRODUCT_AB));
159: PetscCall(MatProductSetAlgorithm(BC, MATPRODUCTALGORITHMDEFAULT));
160: PetscCall(MatProductSetFill(BC, fill));
161: PetscCall(MatProductSetFromOptions(BC));
162: PetscCall(MatProductSymbolic(BC));
164: /* Symbolic mat = A*BC */
165: PetscCall(MatProductSetType(mat, MATPRODUCT_AB));
166: PetscCall(MatProductSetAlgorithm(mat, MATPRODUCTALGORITHMDEFAULT));
167: product->B = BC;
168: product->Dwork = BC;
169: PetscCall(MatProductSetFromOptions(mat));
170: PetscCall(MatProductSymbolic(mat));
172: /* resume user's original input matrix setting for B */
173: product->B = B;
174: mat->ops->productnumeric = MatProductNumeric_ABC_Unsafe;
175: PetscFunctionReturn(PETSC_SUCCESS);
176: }
178: static PetscErrorCode MatProductSymbolic_Unsafe(Mat mat)
179: {
180: Mat_Product *product = mat->product;
182: PetscFunctionBegin;
183: switch (product->type) {
184: case MATPRODUCT_PtAP:
185: PetscCall(MatProductSymbolic_PtAP_Unsafe(mat));
186: break;
187: case MATPRODUCT_RARt:
188: PetscCall(MatProductSymbolic_RARt_Unsafe(mat));
189: break;
190: case MATPRODUCT_ABC:
191: PetscCall(MatProductSymbolic_ABC_Unsafe(mat));
192: break;
193: default:
194: SETERRQ(PetscObjectComm((PetscObject)mat), PETSC_ERR_SUP, "ProductType %s is not supported", MatProductTypes[product->type]);
195: }
196: PetscFunctionReturn(PETSC_SUCCESS);
197: }
199: /*@
200: MatProductReplaceMats - Replace the input matrices for the matrix-matrix product operation inside the computed matrix
202: Collective
204: Input Parameters:
205: + A - the matrix or `NULL` if not being replaced
206: . B - the matrix or `NULL` if not being replaced
207: . C - the matrix or `NULL` if not being replaced
208: - D - the matrix whose values are computed via a matrix-matrix product operation
210: Level: intermediate
212: Note:
213: To reuse the symbolic phase, the input matrices must have exactly the same data structure as the replaced one.
214: If the type of any of the input matrices is different than what was previously used, or their symmetry flag changed but
215: the symbolic phase took advantage of their symmetry, the product is cleared and `MatProductSetFromOptions()`
216: and `MatProductSymbolic()` are invoked again.
218: .seealso: [](ch_matrices), `MatProduct`, `Mat`, `MatProductCreate()`, `MatProductSetFromOptions()`, `MatProductSymbolic().` `MatProductClear()`
219: @*/
220: PetscErrorCode MatProductReplaceMats(Mat A, Mat B, Mat C, Mat D)
221: {
222: Mat_Product *product;
223: PetscBool flgA = PETSC_TRUE, flgB = PETSC_TRUE, flgC = PETSC_TRUE, isset, issym;
225: PetscFunctionBegin;
227: MatCheckProduct(D, 4);
228: product = D->product;
229: if (A) {
231: PetscCall(PetscObjectReference((PetscObject)A));
232: PetscCall(PetscObjectTypeCompare((PetscObject)product->A, ((PetscObject)A)->type_name, &flgA));
233: PetscCall(MatIsSymmetricKnown(A, &isset, &issym));
234: if (product->symbolic_used_the_fact_A_is_symmetric && isset && !issym) { /* symbolic was built around a symmetric A, but the new A is not anymore */
235: flgA = PETSC_FALSE;
236: product->symbolic_used_the_fact_A_is_symmetric = PETSC_FALSE; /* reinit */
237: }
238: PetscCall(MatDestroy(&product->A));
239: product->A = A;
240: }
241: if (B) {
243: PetscCall(PetscObjectReference((PetscObject)B));
244: PetscCall(PetscObjectTypeCompare((PetscObject)product->B, ((PetscObject)B)->type_name, &flgB));
245: PetscCall(MatIsSymmetricKnown(B, &isset, &issym));
246: if (product->symbolic_used_the_fact_B_is_symmetric && isset && !issym) {
247: flgB = PETSC_FALSE;
248: product->symbolic_used_the_fact_B_is_symmetric = PETSC_FALSE; /* reinit */
249: }
250: PetscCall(MatDestroy(&product->B));
251: product->B = B;
252: }
253: if (C) {
255: PetscCall(PetscObjectReference((PetscObject)C));
256: PetscCall(PetscObjectTypeCompare((PetscObject)product->C, ((PetscObject)C)->type_name, &flgC));
257: PetscCall(MatIsSymmetricKnown(C, &isset, &issym));
258: if (product->symbolic_used_the_fact_C_is_symmetric && isset && !issym) {
259: flgC = PETSC_FALSE;
260: product->symbolic_used_the_fact_C_is_symmetric = PETSC_FALSE; /* reinit */
261: }
262: PetscCall(MatDestroy(&product->C));
263: product->C = C;
264: }
265: /* Any of the replaced mats is of a different type, reset */
266: if (!flgA || !flgB || !flgC) {
267: if (D->product->destroy) PetscCall((*D->product->destroy)(D->product->data));
268: D->product->destroy = NULL;
269: D->product->data = NULL;
270: if (D->ops->productnumeric || D->ops->productsymbolic) {
271: PetscCall(MatProductSetFromOptions(D));
272: PetscCall(MatProductSymbolic(D));
273: }
274: }
275: PetscFunctionReturn(PETSC_SUCCESS);
276: }
278: static PetscErrorCode MatProductNumeric_X_Dense(Mat C)
279: {
280: Mat_Product *product = C->product;
281: Mat A = product->A, B = product->B;
282: PetscInt k, K = B->cmap->N;
283: PetscBool t = PETSC_TRUE, iscuda = PETSC_FALSE;
284: PetscBool Bcpu = PETSC_TRUE, Ccpu = PETSC_TRUE;
285: char *Btype = NULL, *Ctype = NULL;
287: PetscFunctionBegin;
288: switch (product->type) {
289: case MATPRODUCT_AB:
290: t = PETSC_FALSE;
291: case MATPRODUCT_AtB:
292: break;
293: default:
294: SETERRQ(PetscObjectComm((PetscObject)C), PETSC_ERR_SUP, "MatProductNumeric type %s not supported for %s and %s matrices", MatProductTypes[product->type], ((PetscObject)A)->type_name, ((PetscObject)B)->type_name);
295: }
296: if (PetscDefined(HAVE_CUDA)) {
297: VecType vtype;
299: PetscCall(MatGetVecType(A, &vtype));
300: PetscCall(PetscStrcmp(vtype, VECCUDA, &iscuda));
301: if (!iscuda) PetscCall(PetscStrcmp(vtype, VECSEQCUDA, &iscuda));
302: if (!iscuda) PetscCall(PetscStrcmp(vtype, VECMPICUDA, &iscuda));
303: if (iscuda) { /* Make sure we have up-to-date data on the GPU */
304: PetscCall(PetscStrallocpy(((PetscObject)B)->type_name, &Btype));
305: PetscCall(PetscStrallocpy(((PetscObject)C)->type_name, &Ctype));
306: PetscCall(MatConvert(B, MATDENSECUDA, MAT_INPLACE_MATRIX, &B));
307: if (!C->assembled) { /* need to flag the matrix as assembled, otherwise MatConvert will complain */
308: PetscCall(MatAssemblyBegin(C, MAT_FINAL_ASSEMBLY));
309: PetscCall(MatAssemblyEnd(C, MAT_FINAL_ASSEMBLY));
310: }
311: PetscCall(MatConvert(C, MATDENSECUDA, MAT_INPLACE_MATRIX, &C));
312: } else { /* Make sure we have up-to-date data on the CPU */
313: #if defined(PETSC_HAVE_CUDA) || defined(PETSC_HAVE_VIENNACL)
314: Bcpu = B->boundtocpu;
315: Ccpu = C->boundtocpu;
316: #endif
317: PetscCall(MatBindToCPU(B, PETSC_TRUE));
318: PetscCall(MatBindToCPU(C, PETSC_TRUE));
319: }
320: }
321: for (k = 0; k < K; k++) {
322: Vec x, y;
324: PetscCall(MatDenseGetColumnVecRead(B, k, &x));
325: PetscCall(MatDenseGetColumnVecWrite(C, k, &y));
326: if (t) {
327: PetscCall(MatMultTranspose(A, x, y));
328: } else {
329: PetscCall(MatMult(A, x, y));
330: }
331: PetscCall(MatDenseRestoreColumnVecRead(B, k, &x));
332: PetscCall(MatDenseRestoreColumnVecWrite(C, k, &y));
333: }
334: PetscCall(MatAssemblyBegin(C, MAT_FINAL_ASSEMBLY));
335: PetscCall(MatAssemblyEnd(C, MAT_FINAL_ASSEMBLY));
336: if (PetscDefined(HAVE_CUDA)) {
337: if (iscuda) {
338: PetscCall(MatConvert(B, Btype, MAT_INPLACE_MATRIX, &B));
339: PetscCall(MatConvert(C, Ctype, MAT_INPLACE_MATRIX, &C));
340: } else {
341: PetscCall(MatBindToCPU(B, Bcpu));
342: PetscCall(MatBindToCPU(C, Ccpu));
343: }
344: }
345: PetscCall(PetscFree(Btype));
346: PetscCall(PetscFree(Ctype));
347: PetscFunctionReturn(PETSC_SUCCESS);
348: }
350: static PetscErrorCode MatProductSymbolic_X_Dense(Mat C)
351: {
352: Mat_Product *product = C->product;
353: Mat A = product->A, B = product->B;
354: PetscBool isdense;
356: PetscFunctionBegin;
357: switch (product->type) {
358: case MATPRODUCT_AB:
359: PetscCall(MatSetSizes(C, A->rmap->n, B->cmap->n, A->rmap->N, B->cmap->N));
360: break;
361: case MATPRODUCT_AtB:
362: PetscCall(MatSetSizes(C, A->cmap->n, B->cmap->n, A->cmap->N, B->cmap->N));
363: break;
364: default:
365: SETERRQ(PetscObjectComm((PetscObject)C), PETSC_ERR_SUP, "MatProductSymbolic type %s not supported for %s and %s matrices", MatProductTypes[product->type], ((PetscObject)A)->type_name, ((PetscObject)B)->type_name);
366: }
367: PetscCall(PetscObjectBaseTypeCompareAny((PetscObject)C, &isdense, MATSEQDENSE, MATMPIDENSE, ""));
368: if (!isdense) {
369: PetscCall(MatSetType(C, ((PetscObject)B)->type_name));
370: /* If matrix type of C was not set or not dense, we need to reset the pointer */
371: C->ops->productsymbolic = MatProductSymbolic_X_Dense;
372: }
373: C->ops->productnumeric = MatProductNumeric_X_Dense;
374: PetscCall(MatSetUp(C));
375: PetscFunctionReturn(PETSC_SUCCESS);
376: }
378: /* a single driver to query the dispatching */
379: static PetscErrorCode MatProductSetFromOptions_Private(Mat mat)
380: {
381: Mat_Product *product = mat->product;
382: PetscInt Am, An, Bm, Bn, Cm, Cn;
383: Mat A = product->A, B = product->B, C = product->C;
384: const char *const Bnames[] = {"B", "R", "P"};
385: const char *bname;
386: PetscErrorCode (*fA)(Mat);
387: PetscErrorCode (*fB)(Mat);
388: PetscErrorCode (*fC)(Mat);
389: PetscErrorCode (*f)(Mat) = NULL;
391: PetscFunctionBegin;
392: mat->ops->productsymbolic = NULL;
393: mat->ops->productnumeric = NULL;
394: if (product->type == MATPRODUCT_UNSPECIFIED) PetscFunctionReturn(PETSC_SUCCESS);
395: PetscCheck(A, PetscObjectComm((PetscObject)mat), PETSC_ERR_PLIB, "Missing A mat");
396: PetscCheck(B, PetscObjectComm((PetscObject)mat), PETSC_ERR_PLIB, "Missing B mat");
397: PetscCheck(product->type != MATPRODUCT_ABC || C, PetscObjectComm((PetscObject)mat), PETSC_ERR_PLIB, "Missing C mat");
398: if (product->type != MATPRODUCT_ABC) C = NULL; /* do not use C if not needed */
399: if (product->type == MATPRODUCT_RARt) bname = Bnames[1];
400: else if (product->type == MATPRODUCT_PtAP) bname = Bnames[2];
401: else bname = Bnames[0];
403: /* Check matrices sizes */
404: Am = A->rmap->N;
405: An = A->cmap->N;
406: Bm = B->rmap->N;
407: Bn = B->cmap->N;
408: Cm = C ? C->rmap->N : 0;
409: Cn = C ? C->cmap->N : 0;
410: if (product->type == MATPRODUCT_RARt || product->type == MATPRODUCT_ABt) {
411: PetscInt t = Bn;
412: Bn = Bm;
413: Bm = t;
414: }
415: if (product->type == MATPRODUCT_AtB) {
416: PetscInt t = An;
417: An = Am;
418: Am = t;
419: }
420: PetscCheck(An == Bm, PetscObjectComm((PetscObject)mat), PETSC_ERR_ARG_SIZ, "Matrix dimensions of A and %s are incompatible for MatProductType %s: A %" PetscInt_FMT "x%" PetscInt_FMT ", %s %" PetscInt_FMT "x%" PetscInt_FMT, bname,
421: MatProductTypes[product->type], A->rmap->N, A->cmap->N, bname, B->rmap->N, B->cmap->N);
422: PetscCheck(!Cm || Cm == Bn, PetscObjectComm((PetscObject)mat), PETSC_ERR_ARG_SIZ, "Matrix dimensions of B and C are incompatible for MatProductType %s: B %" PetscInt_FMT "x%" PetscInt_FMT ", C %" PetscInt_FMT "x%" PetscInt_FMT,
423: MatProductTypes[product->type], B->rmap->N, B->cmap->N, Cm, Cn);
425: fA = A->ops->productsetfromoptions;
426: fB = B->ops->productsetfromoptions;
427: fC = C ? C->ops->productsetfromoptions : fA;
428: if (C) {
429: PetscCall(PetscInfo(mat, "MatProductType %s for A %s, %s %s, C %s\n", MatProductTypes[product->type], ((PetscObject)A)->type_name, bname, ((PetscObject)B)->type_name, ((PetscObject)C)->type_name));
430: } else {
431: PetscCall(PetscInfo(mat, "MatProductType %s for A %s, %s %s\n", MatProductTypes[product->type], ((PetscObject)A)->type_name, bname, ((PetscObject)B)->type_name));
432: }
433: if (fA == fB && fA == fC && fA) {
434: PetscCall(PetscInfo(mat, " matching op\n"));
435: PetscCall((*fA)(mat));
436: }
437: /* We may have found f but it did not succeed */
438: if (!mat->ops->productsymbolic) { /* query MatProductSetFromOptions_Atype_Btype_Ctype */
439: char mtypes[256];
440: PetscCall(PetscStrncpy(mtypes, "MatProductSetFromOptions_", sizeof(mtypes)));
441: PetscCall(PetscStrlcat(mtypes, ((PetscObject)A)->type_name, sizeof(mtypes)));
442: PetscCall(PetscStrlcat(mtypes, "_", sizeof(mtypes)));
443: PetscCall(PetscStrlcat(mtypes, ((PetscObject)B)->type_name, sizeof(mtypes)));
444: if (C) {
445: PetscCall(PetscStrlcat(mtypes, "_", sizeof(mtypes)));
446: PetscCall(PetscStrlcat(mtypes, ((PetscObject)C)->type_name, sizeof(mtypes)));
447: }
448: PetscCall(PetscStrlcat(mtypes, "_C", sizeof(mtypes)));
449: #if defined(__clang__)
450: PETSC_PRAGMA_DIAGNOSTIC_IGNORED_BEGIN("-Wformat-pedantic")
451: #elif defined(__GNUC__) || defined(__GNUG__)
452: PETSC_PRAGMA_DIAGNOSTIC_IGNORED_BEGIN("-Wformat")
453: #endif
454: PetscCall(PetscObjectQueryFunction((PetscObject)A, mtypes, &f));
455: PetscCall(PetscInfo(mat, " querying %s from A? %p\n", mtypes, f));
456: if (!f) {
457: PetscCall(PetscObjectQueryFunction((PetscObject)B, mtypes, &f));
458: PetscCall(PetscInfo(mat, " querying %s from %s? %p\n", mtypes, bname, f));
459: }
460: if (!f && C) {
461: PetscCall(PetscObjectQueryFunction((PetscObject)C, mtypes, &f));
462: PetscCall(PetscInfo(mat, " querying %s from C? %p\n", mtypes, f));
463: }
464: if (f) PetscCall((*f)(mat));
466: /* We may have found f but it did not succeed */
467: /* some matrices (i.e. MATTRANSPOSEVIRTUAL, MATSHELL constructed from MatConvert), knows what to do with their inner matrices */
468: if (!mat->ops->productsymbolic) {
469: PetscCall(PetscStrncpy(mtypes, "MatProductSetFromOptions_anytype_C", sizeof(mtypes)));
470: PetscCall(PetscObjectQueryFunction((PetscObject)A, mtypes, &f));
471: PetscCall(PetscInfo(mat, " querying %s from A? %p\n", mtypes, f));
472: if (!f) {
473: PetscCall(PetscObjectQueryFunction((PetscObject)B, mtypes, &f));
474: PetscCall(PetscInfo(mat, " querying %s from %s? %p\n", mtypes, bname, f));
475: }
476: if (!f && C) {
477: PetscCall(PetscObjectQueryFunction((PetscObject)C, mtypes, &f));
478: PetscCall(PetscInfo(mat, " querying %s from C? %p\n", mtypes, f));
479: }
480: }
481: if (f) PetscCall((*f)(mat));
482: }
483: PETSC_PRAGMA_DIAGNOSTIC_IGNORED_END()
484: /* We may have found f but it did not succeed */
485: if (!mat->ops->productsymbolic) {
486: /* we can still compute the product if B is of type dense */
487: if (product->type == MATPRODUCT_AB || product->type == MATPRODUCT_AtB) {
488: PetscBool isdense;
490: PetscCall(PetscObjectBaseTypeCompareAny((PetscObject)B, &isdense, MATSEQDENSE, MATMPIDENSE, ""));
491: if (isdense) {
492: mat->ops->productsymbolic = MatProductSymbolic_X_Dense;
493: PetscCall(PetscInfo(mat, " using basic looping over columns of a dense matrix\n"));
494: }
495: } else if (product->type != MATPRODUCT_ABt) { /* use MatProductSymbolic/Numeric_Unsafe() for triple products only */
496: /*
497: TODO: this should be changed to a proper setfromoptions, not setting the symbolic pointer here, because we do not know if
498: the combination will succeed. In order to be sure, we need MatProductGetProductType to return the type of the result
499: before computing the symbolic phase
500: */
501: PetscCall(PetscInfo(mat, " symbolic product not supported, using MatProductSymbolic_Unsafe() implementation\n"));
502: mat->ops->productsymbolic = MatProductSymbolic_Unsafe;
503: }
504: }
505: if (!mat->ops->productsymbolic) PetscCall(PetscInfo(mat, " symbolic product is not supported\n"));
506: PetscFunctionReturn(PETSC_SUCCESS);
507: }
509: /*@
510: MatProductSetFromOptions - Sets the options for the computation of a matrix-matrix product operation where the type,
511: the algorithm etc are determined from the options database.
513: Logically Collective
515: Input Parameter:
516: . mat - the matrix whose values are computed via a matrix-matrix product operation
518: Options Database Keys:
519: + -mat_product_clear - Clear intermediate data structures after `MatProductNumeric()` has been called
520: . -mat_product_algorithm <algorithm> - Sets the algorithm, see `MatProductAlgorithm` for possible values
521: - -mat_product_algorithm_backend_cpu - Use the CPU to perform the computation even if the matrix is a GPU matrix
523: Level: intermediate
525: Note:
526: The `-mat_product_clear` option reduces memory usage but means that the matrix cannot be re-used for a matrix-matrix product operation
528: .seealso: [](ch_matrices), `MatProduct`, `Mat`, `MatSetFromOptions()`, `MatProductCreate()`, `MatProductCreateWithMat()`, `MatProductNumeric()`,
529: `MatProductSetType()`, `MatProductSetAlgorithm()`, `MatProductAlgorithm`
530: @*/
531: PetscErrorCode MatProductSetFromOptions(Mat mat)
532: {
533: PetscFunctionBegin;
535: MatCheckProduct(mat, 1);
536: PetscCheck(!mat->product->data, PetscObjectComm((PetscObject)mat), PETSC_ERR_ORDER, "Cannot call MatProductSetFromOptions with already present data");
537: PetscObjectOptionsBegin((PetscObject)mat);
538: PetscCall(PetscOptionsBool("-mat_product_clear", "Clear intermediate data structures after MatProductNumeric() has been called", "MatProductClear", mat->product->clear, &mat->product->clear, NULL));
539: PetscCall(PetscOptionsDeprecated("-mat_freeintermediatedatastructures", "-mat_product_clear", "3.13", "Or call MatProductClear() after MatProductNumeric()"));
540: PetscOptionsEnd();
541: PetscCall(MatProductSetFromOptions_Private(mat));
542: PetscCheck(mat->product, PetscObjectComm((PetscObject)mat), PETSC_ERR_PLIB, "Missing product after setup phase");
543: PetscFunctionReturn(PETSC_SUCCESS);
544: }
546: /*@C
547: MatProductView - View the private matrix-matrix algorithm object within a matrix
549: Logically Collective
551: Input Parameters:
552: + mat - the matrix obtained with `MatProductCreate()` or `MatProductCreateWithMat()`
553: - viewer - where the information on the matrix-matrix algorithm of `mat` should be reviewed
555: Level: intermediate
557: .seealso: [](ch_matrices), `MatProductType`, `Mat`, `MatProductSetFromOptions()`, `MatView()`, `MatProductCreate()`, `MatProductCreateWithMat()`
558: @*/
559: PetscErrorCode MatProductView(Mat mat, PetscViewer viewer)
560: {
561: PetscFunctionBegin;
563: if (!mat->product) PetscFunctionReturn(PETSC_SUCCESS);
564: if (!viewer) PetscCall(PetscViewerASCIIGetStdout(PetscObjectComm((PetscObject)mat), &viewer));
566: PetscCheckSameComm(mat, 1, viewer, 2);
567: if (mat->product->view) PetscCall((*mat->product->view)(mat, viewer));
568: PetscFunctionReturn(PETSC_SUCCESS);
569: }
571: /* these are basic implementations relying on the old function pointers
572: * they are dangerous and should be removed in the future */
573: PetscErrorCode MatProductNumeric_AB(Mat mat)
574: {
575: Mat_Product *product = mat->product;
576: Mat A = product->A, B = product->B;
578: PetscFunctionBegin;
579: PetscCall((*mat->ops->matmultnumeric)(A, B, mat));
580: PetscFunctionReturn(PETSC_SUCCESS);
581: }
583: PetscErrorCode MatProductNumeric_AtB(Mat mat)
584: {
585: Mat_Product *product = mat->product;
586: Mat A = product->A, B = product->B;
588: PetscFunctionBegin;
589: PetscCall((*mat->ops->transposematmultnumeric)(A, B, mat));
590: PetscFunctionReturn(PETSC_SUCCESS);
591: }
593: PetscErrorCode MatProductNumeric_ABt(Mat mat)
594: {
595: Mat_Product *product = mat->product;
596: Mat A = product->A, B = product->B;
598: PetscFunctionBegin;
599: PetscCall((*mat->ops->mattransposemultnumeric)(A, B, mat));
600: PetscFunctionReturn(PETSC_SUCCESS);
601: }
603: PetscErrorCode MatProductNumeric_PtAP(Mat mat)
604: {
605: Mat_Product *product = mat->product;
606: Mat A = product->A, B = product->B;
608: PetscFunctionBegin;
609: PetscCall((*mat->ops->ptapnumeric)(A, B, mat));
610: PetscFunctionReturn(PETSC_SUCCESS);
611: }
613: PetscErrorCode MatProductNumeric_RARt(Mat mat)
614: {
615: Mat_Product *product = mat->product;
616: Mat A = product->A, B = product->B;
618: PetscFunctionBegin;
619: PetscCall((*mat->ops->rartnumeric)(A, B, mat));
620: PetscFunctionReturn(PETSC_SUCCESS);
621: }
623: PetscErrorCode MatProductNumeric_ABC(Mat mat)
624: {
625: Mat_Product *product = mat->product;
626: Mat A = product->A, B = product->B, C = product->C;
628: PetscFunctionBegin;
629: PetscCall((*mat->ops->matmatmultnumeric)(A, B, C, mat));
630: PetscFunctionReturn(PETSC_SUCCESS);
631: }
633: /*@
634: MatProductNumeric - Compute a matrix-matrix product operation with the numerical values
636: Collective
638: Input/Output Parameter:
639: . mat - the matrix whose values are computed via a matrix-matrix product operation
641: Level: intermediate
643: Note:
644: `MatProductSymbolic()` must have been called on `mat` before calling this function
646: .seealso: [](ch_matrices), `MatProduct`, `Mat`, `MatProductSetAlgorithm()`, `MatProductSetType()`, `MatProductCreate()`, `MatSetType()`, `MatProductSymbolic()`
647: @*/
648: PetscErrorCode MatProductNumeric(Mat mat)
649: {
650: PetscLogEvent eventtype = -1;
652: PetscFunctionBegin;
654: MatCheckProduct(mat, 1);
655: switch (mat->product->type) {
656: case MATPRODUCT_AB:
657: eventtype = MAT_MatMultNumeric;
658: break;
659: case MATPRODUCT_AtB:
660: eventtype = MAT_TransposeMatMultNumeric;
661: break;
662: case MATPRODUCT_ABt:
663: eventtype = MAT_MatTransposeMultNumeric;
664: break;
665: case MATPRODUCT_PtAP:
666: eventtype = MAT_PtAPNumeric;
667: break;
668: case MATPRODUCT_RARt:
669: eventtype = MAT_RARtNumeric;
670: break;
671: case MATPRODUCT_ABC:
672: eventtype = MAT_MatMatMultNumeric;
673: break;
674: default:
675: SETERRQ(PetscObjectComm((PetscObject)mat), PETSC_ERR_SUP, "ProductType %s is not supported", MatProductTypes[mat->product->type]);
676: }
678: if (mat->ops->productnumeric) {
679: PetscCall(PetscLogEventBegin(eventtype, mat, 0, 0, 0));
680: PetscUseTypeMethod(mat, productnumeric);
681: PetscCall(PetscLogEventEnd(eventtype, mat, 0, 0, 0));
682: } else if (mat->product) {
683: char errstr[256];
685: if (mat->product->type == MATPRODUCT_ABC) {
686: PetscCall(PetscSNPrintf(errstr, 256, "%s with A %s, B %s, C %s", MatProductTypes[mat->product->type], ((PetscObject)mat->product->A)->type_name, ((PetscObject)mat->product->B)->type_name, ((PetscObject)mat->product->C)->type_name));
687: } else {
688: PetscCall(PetscSNPrintf(errstr, 256, "%s with A %s, B %s", MatProductTypes[mat->product->type], ((PetscObject)mat->product->A)->type_name, ((PetscObject)mat->product->B)->type_name));
689: }
690: SETERRQ(PetscObjectComm((PetscObject)mat), PETSC_ERR_PLIB, "Unspecified numeric phase for product %s", errstr);
691: }
692: PetscCheck(mat->product, PetscObjectComm((PetscObject)mat), PETSC_ERR_PLIB, "Missing struct after numeric phase for product");
694: if (mat->product->clear) PetscCall(MatProductClear(mat));
695: PetscCall(PetscObjectStateIncrease((PetscObject)mat));
696: PetscFunctionReturn(PETSC_SUCCESS);
697: }
699: /* these are basic implementations relying on the old function pointers
700: * they are dangerous and should be removed in the future */
701: PetscErrorCode MatProductSymbolic_AB(Mat mat)
702: {
703: Mat_Product *product = mat->product;
704: Mat A = product->A, B = product->B;
706: PetscFunctionBegin;
707: PetscCall((*mat->ops->matmultsymbolic)(A, B, product->fill, mat));
708: mat->ops->productnumeric = MatProductNumeric_AB;
709: PetscFunctionReturn(PETSC_SUCCESS);
710: }
712: PetscErrorCode MatProductSymbolic_AtB(Mat mat)
713: {
714: Mat_Product *product = mat->product;
715: Mat A = product->A, B = product->B;
717: PetscFunctionBegin;
718: PetscCall((*mat->ops->transposematmultsymbolic)(A, B, product->fill, mat));
719: mat->ops->productnumeric = MatProductNumeric_AtB;
720: PetscFunctionReturn(PETSC_SUCCESS);
721: }
723: PetscErrorCode MatProductSymbolic_ABt(Mat mat)
724: {
725: Mat_Product *product = mat->product;
726: Mat A = product->A, B = product->B;
728: PetscFunctionBegin;
729: PetscCall((*mat->ops->mattransposemultsymbolic)(A, B, product->fill, mat));
730: mat->ops->productnumeric = MatProductNumeric_ABt;
731: PetscFunctionReturn(PETSC_SUCCESS);
732: }
734: PetscErrorCode MatProductSymbolic_ABC(Mat mat)
735: {
736: Mat_Product *product = mat->product;
737: Mat A = product->A, B = product->B, C = product->C;
739: PetscFunctionBegin;
740: PetscCall((*mat->ops->matmatmultsymbolic)(A, B, C, product->fill, mat));
741: mat->ops->productnumeric = MatProductNumeric_ABC;
742: PetscFunctionReturn(PETSC_SUCCESS);
743: }
745: /*@
746: MatProductSymbolic - Perform the symbolic portion of a matrix-matrix product operation, this creates a data structure for use with the numerical
747: product to be done with `MatProductNumeric()`
749: Collective
751: Input/Output Parameter:
752: . mat - the matrix whose values are to be computed via a matrix-matrix product operation
754: Level: intermediate
756: Note:
757: `MatProductSetFromOptions()` must have been called on `mat` before calling this function
759: .seealso: [](ch_matrices), `MatProduct`, `Mat`, `MatProductCreate()`, `MatProductCreateWithMat()`, `MatProductSetFromOptions()`, `MatProductNumeric()`, `MatProductSetType()`, `MatProductSetAlgorithm()`
760: @*/
761: PetscErrorCode MatProductSymbolic(Mat mat)
762: {
763: PetscLogEvent eventtype = -1;
764: PetscBool missing = PETSC_FALSE;
766: PetscFunctionBegin;
768: MatCheckProduct(mat, 1);
769: PetscCheck(!mat->product->data, PetscObjectComm((PetscObject)mat), PETSC_ERR_ORDER, "Cannot run symbolic phase. Product data not empty");
770: switch (mat->product->type) {
771: case MATPRODUCT_AB:
772: eventtype = MAT_MatMultSymbolic;
773: break;
774: case MATPRODUCT_AtB:
775: eventtype = MAT_TransposeMatMultSymbolic;
776: break;
777: case MATPRODUCT_ABt:
778: eventtype = MAT_MatTransposeMultSymbolic;
779: break;
780: case MATPRODUCT_PtAP:
781: eventtype = MAT_PtAPSymbolic;
782: break;
783: case MATPRODUCT_RARt:
784: eventtype = MAT_RARtSymbolic;
785: break;
786: case MATPRODUCT_ABC:
787: eventtype = MAT_MatMatMultSymbolic;
788: break;
789: default:
790: SETERRQ(PetscObjectComm((PetscObject)mat), PETSC_ERR_SUP, "ProductType %s is not supported", MatProductTypes[mat->product->type]);
791: }
792: mat->ops->productnumeric = NULL;
793: if (mat->ops->productsymbolic) {
794: PetscCall(PetscLogEventBegin(eventtype, mat, 0, 0, 0));
795: PetscUseTypeMethod(mat, productsymbolic);
796: PetscCall(PetscLogEventEnd(eventtype, mat, 0, 0, 0));
797: } else missing = PETSC_TRUE;
799: if (missing || !mat->product || !mat->ops->productnumeric) {
800: char errstr[256];
802: if (mat->product->type == MATPRODUCT_ABC) {
803: PetscCall(PetscSNPrintf(errstr, 256, "%s with A %s, B %s, C %s", MatProductTypes[mat->product->type], ((PetscObject)mat->product->A)->type_name, ((PetscObject)mat->product->B)->type_name, ((PetscObject)mat->product->C)->type_name));
804: } else {
805: PetscCall(PetscSNPrintf(errstr, 256, "%s with A %s, B %s", MatProductTypes[mat->product->type], ((PetscObject)mat->product->A)->type_name, ((PetscObject)mat->product->B)->type_name));
806: }
807: PetscCheck(!missing, PetscObjectComm((PetscObject)mat), PETSC_ERR_PLIB, "Unspecified symbolic phase for product %s. Call MatProductSetFromOptions() first", errstr);
808: PetscCheck(mat->product, PetscObjectComm((PetscObject)mat), PETSC_ERR_PLIB, "Missing struct after symbolic phase for product %s", errstr);
809: }
811: #if defined(PETSC_HAVE_DEVICE)
812: Mat A = mat->product->A;
813: Mat B = mat->product->B;
814: Mat C = mat->product->C;
815: PetscBool bindingpropagates;
816: bindingpropagates = (PetscBool)((A->boundtocpu && A->bindingpropagates) || (B->boundtocpu && B->bindingpropagates));
817: if (C) bindingpropagates = (PetscBool)(bindingpropagates || (C->boundtocpu && C->bindingpropagates));
818: if (bindingpropagates) {
819: PetscCall(MatBindToCPU(mat, PETSC_TRUE));
820: PetscCall(MatSetBindingPropagates(mat, PETSC_TRUE));
821: }
822: #endif
823: PetscFunctionReturn(PETSC_SUCCESS);
824: }
826: /*@
827: MatProductSetFill - Set an expected fill of the matrix whose values are computed via a matrix-matrix product operation
829: Collective
831: Input Parameters:
832: + mat - the matrix whose values are to be computed via a matrix-matrix product operation
833: - fill - expected fill as ratio of nnz(mat)/(nnz(A) + nnz(B) + nnz(C)); use `PETSC_DEFAULT` if you do not have a good estimate.
834: If the product is a dense matrix, this value is not used.
836: Level: intermediate
838: .seealso: [](ch_matrices), `MatProduct`, `Mat`, `MatProductSetFromOptions()`, `MatProductSetType()`, `MatProductSetAlgorithm()`, `MatProductCreate()`
839: @*/
840: PetscErrorCode MatProductSetFill(Mat mat, PetscReal fill)
841: {
842: PetscFunctionBegin;
844: MatCheckProduct(mat, 1);
845: if (fill == (PetscReal)PETSC_DEFAULT || fill == (PetscReal)PETSC_DECIDE) mat->product->fill = 2.0;
846: else mat->product->fill = fill;
847: PetscFunctionReturn(PETSC_SUCCESS);
848: }
850: /*@C
851: MatProductSetAlgorithm - Requests a particular algorithm for a matrix-matrix product operation that will perform to compute the given matrix
853: Collective
855: Input Parameters:
856: + mat - the matrix whose values are computed via a matrix-matrix product operation
857: - alg - particular implementation algorithm of the matrix product, e.g., `MATPRODUCTALGORITHMDEFAULT`.
859: Options Database Key:
860: . -mat_product_algorithm <algorithm> - Sets the algorithm, see `MatProductAlgorithm`
862: Level: intermediate
864: .seealso: [](ch_matrices), `MatProduct`, `Mat`, `MatProductClear()`, `MatProductSetType()`, `MatProductSetFill()`, `MatProductCreate()`, `MatProductAlgorithm`, `MatProductType`
865: @*/
866: PetscErrorCode MatProductSetAlgorithm(Mat mat, MatProductAlgorithm alg)
867: {
868: PetscFunctionBegin;
870: MatCheckProduct(mat, 1);
871: PetscCall(PetscFree(mat->product->alg));
872: PetscCall(PetscStrallocpy(alg, &mat->product->alg));
873: PetscFunctionReturn(PETSC_SUCCESS);
874: }
876: /*@
877: MatProductSetType - Sets a particular matrix-matrix product operation to be used to compute the values of the given matrix
879: Collective
881: Input Parameters:
882: + mat - the matrix whose values are computed via a matrix-matrix product operation
883: - productype - matrix product type, e.g., `MATPRODUCT_AB`,`MATPRODUCT_AtB`,`MATPRODUCT_ABt`,`MATPRODUCT_PtAP`,`MATPRODUCT_RARt`,`MATPRODUCT_ABC`,
884: see `MatProductType`
886: Level: intermediate
888: Note:
889: The small t represents the transpose operation.
891: .seealso: [](ch_matrices), `MatProduct`, `Mat`, `MatProductCreate()`, `MatProductType`,
892: `MATPRODUCT_AB`, `MATPRODUCT_AtB`, `MATPRODUCT_ABt`, `MATPRODUCT_PtAP`, `MATPRODUCT_RARt`, `MATPRODUCT_ABC`
893: @*/
894: PetscErrorCode MatProductSetType(Mat mat, MatProductType productype)
895: {
896: PetscFunctionBegin;
898: MatCheckProduct(mat, 1);
900: if (productype != mat->product->type) {
901: if (mat->product->destroy) PetscCall((*mat->product->destroy)(mat->product->data));
902: mat->product->destroy = NULL;
903: mat->product->data = NULL;
904: mat->ops->productsymbolic = NULL;
905: mat->ops->productnumeric = NULL;
906: }
907: mat->product->type = productype;
908: PetscFunctionReturn(PETSC_SUCCESS);
909: }
911: /*@
912: MatProductClear - Clears from the matrix any internal data structures related to the computation of the values of the matrix from matrix-matrix product operations
914: Collective
916: Input Parameter:
917: . mat - the matrix whose values are to be computed via a matrix-matrix product operation
919: Options Database Key:
920: . -mat_product_clear - Clear intermediate data structures after `MatProductNumeric()` has been called
922: Level: intermediate
924: Notes:
925: This function should be called to remove any intermediate data used to compute the matrix to free up memory.
927: After having called this function, matrix-matrix product operations can no longer be used on `mat`
929: .seealso: [](ch_matrices), `MatProduct`, `Mat`, `MatProductCreate()`
930: @*/
931: PetscErrorCode MatProductClear(Mat mat)
932: {
933: Mat_Product *product = mat->product;
935: PetscFunctionBegin;
937: if (product) {
938: PetscCall(MatDestroy(&product->A));
939: PetscCall(MatDestroy(&product->B));
940: PetscCall(MatDestroy(&product->C));
941: PetscCall(PetscFree(product->alg));
942: PetscCall(MatDestroy(&product->Dwork));
943: if (product->destroy) PetscCall((*product->destroy)(product->data));
944: }
945: PetscCall(PetscFree(mat->product));
946: mat->ops->productsymbolic = NULL;
947: mat->ops->productnumeric = NULL;
948: PetscFunctionReturn(PETSC_SUCCESS);
949: }
951: /* Create a supporting struct and attach it to the matrix product */
952: PetscErrorCode MatProductCreate_Private(Mat A, Mat B, Mat C, Mat D)
953: {
954: Mat_Product *product = NULL;
956: PetscFunctionBegin;
958: PetscCheck(!D->product, PetscObjectComm((PetscObject)D), PETSC_ERR_PLIB, "Product already present");
959: PetscCall(PetscNew(&product));
960: product->A = A;
961: product->B = B;
962: product->C = C;
963: product->type = MATPRODUCT_UNSPECIFIED;
964: product->Dwork = NULL;
965: product->api_user = PETSC_FALSE;
966: product->clear = PETSC_FALSE;
967: D->product = product;
969: PetscCall(MatProductSetAlgorithm(D, MATPRODUCTALGORITHMDEFAULT));
970: PetscCall(MatProductSetFill(D, PETSC_DEFAULT));
972: PetscCall(PetscObjectReference((PetscObject)A));
973: PetscCall(PetscObjectReference((PetscObject)B));
974: PetscCall(PetscObjectReference((PetscObject)C));
975: PetscFunctionReturn(PETSC_SUCCESS);
976: }
978: /*@
979: MatProductCreateWithMat - Set a given matrix to have its values computed via matrix-matrix operations on other matrices.
981: Collective
983: Input Parameters:
984: + A - the first matrix
985: . B - the second matrix
986: . C - the third matrix (optional, use `NULL` if not needed)
987: - D - the matrix whose values are to be computed via a matrix-matrix product operation
989: Level: intermediate
991: Notes:
992: Use `MatProductCreate()` if the matrix you wish computed (the `D` matrix) does not already exist
994: See `MatProductCreate()` for details on the usage of the matrix-matrix product operations
996: Any product data currently attached to `D` will be cleared
998: .seealso: [](ch_matrices), `MatProduct`, `Mat`, `MatProductType`, `MatProductSetType()`, `MatProductAlgorithm`,
999: `MatProductSetAlgorithm`, `MatProductCreate()`, `MatProductClear()`
1000: @*/
1001: PetscErrorCode MatProductCreateWithMat(Mat A, Mat B, Mat C, Mat D)
1002: {
1003: PetscFunctionBegin;
1006: MatCheckPreallocated(A, 1);
1007: PetscCheck(A->assembled, PetscObjectComm((PetscObject)A), PETSC_ERR_ARG_WRONGSTATE, "Not for unassembled matrix");
1008: PetscCheck(!A->factortype, PetscObjectComm((PetscObject)A), PETSC_ERR_ARG_WRONGSTATE, "Not for factored matrix");
1012: MatCheckPreallocated(B, 2);
1013: PetscCheck(B->assembled, PetscObjectComm((PetscObject)B), PETSC_ERR_ARG_WRONGSTATE, "Not for unassembled matrix");
1014: PetscCheck(!B->factortype, PetscObjectComm((PetscObject)B), PETSC_ERR_ARG_WRONGSTATE, "Not for factored matrix");
1016: if (C) {
1019: MatCheckPreallocated(C, 3);
1020: PetscCheck(C->assembled, PetscObjectComm((PetscObject)C), PETSC_ERR_ARG_WRONGSTATE, "Not for unassembled matrix");
1021: PetscCheck(!C->factortype, PetscObjectComm((PetscObject)C), PETSC_ERR_ARG_WRONGSTATE, "Not for factored matrix");
1022: }
1026: MatCheckPreallocated(D, 4);
1027: PetscCheck(D->assembled, PetscObjectComm((PetscObject)D), PETSC_ERR_ARG_WRONGSTATE, "Not for unassembled matrix");
1028: PetscCheck(!D->factortype, PetscObjectComm((PetscObject)D), PETSC_ERR_ARG_WRONGSTATE, "Not for factored matrix");
1030: /* Create a supporting struct and attach it to D */
1031: PetscCall(MatProductClear(D));
1032: PetscCall(MatProductCreate_Private(A, B, C, D));
1033: PetscFunctionReturn(PETSC_SUCCESS);
1034: }
1036: /*@
1037: MatProductCreate - create a matrix to hold the result of a matrix-matrix product operation
1039: Collective
1041: Input Parameters:
1042: + A - the first matrix
1043: . B - the second matrix
1044: - C - the third matrix (or `NULL`)
1046: Output Parameter:
1047: . D - the matrix whose values are to be computed via a matrix-matrix product operation
1049: Level: intermediate
1051: Example:
1052: .vb
1053: MatProductCreate(A,B,C,&D); or MatProductCreateWithMat(A,B,C,D)
1054: MatProductSetType(D, MATPRODUCT_AB or MATPRODUCT_AtB or MATPRODUCT_ABt or MATPRODUCT_PtAP or MATPRODUCT_RARt or MATPRODUCT_ABC)
1055: MatProductSetAlgorithm(D, alg)
1056: MatProductSetFill(D,fill)
1057: MatProductSetFromOptions(D)
1058: MatProductSymbolic(D)
1059: MatProductNumeric(D)
1060: Change numerical values in some of the matrices
1061: MatProductNumeric(D)
1062: .ve
1064: Notes:
1065: Use `MatProductCreateWithMat()` if the matrix you wish computed, the `D` matrix, already exists.
1067: The information computed during the symbolic stage can be reused for new numerical computations with the same non-zero structure
1069: Developer Notes:
1070: It is undocumented what happens if the nonzero structure of the input matrices changes. Is the symbolic stage automatically redone? Does it crash?
1071: Is there error checking for it?
1073: .seealso: [](ch_matrices), `MatProduct`, `Mat`, `MatProductCreateWithMat()`, `MatProductSetType()`, `MatProductSetAlgorithm()`, `MatProductClear()`
1074: @*/
1075: PetscErrorCode MatProductCreate(Mat A, Mat B, Mat C, Mat *D)
1076: {
1077: PetscFunctionBegin;
1082: PetscCheck(!A->factortype, PetscObjectComm((PetscObject)A), PETSC_ERR_ARG_WRONGSTATE, "Not for factored matrix A");
1083: PetscCheck(!B->factortype, PetscObjectComm((PetscObject)B), PETSC_ERR_ARG_WRONGSTATE, "Not for factored matrix B");
1085: if (C) {
1088: PetscCheck(!C->factortype, PetscObjectComm((PetscObject)C), PETSC_ERR_ARG_WRONGSTATE, "Not for factored matrix C");
1089: }
1091: PetscAssertPointer(D, 4);
1092: PetscCall(MatCreate(PetscObjectComm((PetscObject)A), D));
1093: /* Delay setting type of D to the MatProduct symbolic phase, as we allow sparse A and dense B */
1094: PetscCall(MatProductCreate_Private(A, B, C, *D));
1095: PetscFunctionReturn(PETSC_SUCCESS);
1096: }
1098: /*
1099: These are safe basic implementations of ABC, RARt and PtAP
1100: that do not rely on mat->ops->matmatop function pointers.
1101: They only use the MatProduct API and are currently used by
1102: cuSPARSE and KOKKOS-KERNELS backends
1103: */
1104: typedef struct {
1105: Mat BC;
1106: Mat ABC;
1107: } MatMatMatPrivate;
1109: static PetscErrorCode MatDestroy_MatMatMatPrivate(void *data)
1110: {
1111: MatMatMatPrivate *mmdata = (MatMatMatPrivate *)data;
1113: PetscFunctionBegin;
1114: PetscCall(MatDestroy(&mmdata->BC));
1115: PetscCall(MatDestroy(&mmdata->ABC));
1116: PetscCall(PetscFree(data));
1117: PetscFunctionReturn(PETSC_SUCCESS);
1118: }
1120: static PetscErrorCode MatProductNumeric_ABC_Basic(Mat mat)
1121: {
1122: Mat_Product *product = mat->product;
1123: MatMatMatPrivate *mmabc;
1125: PetscFunctionBegin;
1126: MatCheckProduct(mat, 1);
1127: PetscCheck(mat->product->data, PetscObjectComm((PetscObject)mat), PETSC_ERR_PLIB, "Product data empty");
1128: mmabc = (MatMatMatPrivate *)mat->product->data;
1129: PetscCheck(mmabc->BC->ops->productnumeric, PetscObjectComm((PetscObject)mat), PETSC_ERR_PLIB, "Missing numeric stage");
1130: /* use function pointer directly to prevent logging */
1131: PetscCall((*mmabc->BC->ops->productnumeric)(mmabc->BC));
1132: /* swap ABC product stuff with that of ABC for the numeric phase on mat */
1133: mat->product = mmabc->ABC->product;
1134: mat->ops->productnumeric = mmabc->ABC->ops->productnumeric;
1135: /* use function pointer directly to prevent logging */
1136: PetscUseTypeMethod(mat, productnumeric);
1137: mat->ops->productnumeric = MatProductNumeric_ABC_Basic;
1138: mat->product = product;
1139: PetscFunctionReturn(PETSC_SUCCESS);
1140: }
1142: PetscErrorCode MatProductSymbolic_ABC_Basic(Mat mat)
1143: {
1144: Mat_Product *product = mat->product;
1145: Mat A, B, C;
1146: MatProductType p1, p2;
1147: MatMatMatPrivate *mmabc;
1148: const char *prefix;
1150: PetscFunctionBegin;
1151: MatCheckProduct(mat, 1);
1152: PetscCheck(!mat->product->data, PetscObjectComm((PetscObject)mat), PETSC_ERR_PLIB, "Product data not empty");
1153: PetscCall(MatGetOptionsPrefix(mat, &prefix));
1154: PetscCall(PetscNew(&mmabc));
1155: product->data = mmabc;
1156: product->destroy = MatDestroy_MatMatMatPrivate;
1157: switch (product->type) {
1158: case MATPRODUCT_PtAP:
1159: p1 = MATPRODUCT_AB;
1160: p2 = MATPRODUCT_AtB;
1161: A = product->B;
1162: B = product->A;
1163: C = product->B;
1164: break;
1165: case MATPRODUCT_RARt:
1166: p1 = MATPRODUCT_ABt;
1167: p2 = MATPRODUCT_AB;
1168: A = product->B;
1169: B = product->A;
1170: C = product->B;
1171: break;
1172: case MATPRODUCT_ABC:
1173: p1 = MATPRODUCT_AB;
1174: p2 = MATPRODUCT_AB;
1175: A = product->A;
1176: B = product->B;
1177: C = product->C;
1178: break;
1179: default:
1180: SETERRQ(PetscObjectComm((PetscObject)mat), PETSC_ERR_PLIB, "Not for ProductType %s", MatProductTypes[product->type]);
1181: }
1182: PetscCall(MatProductCreate(B, C, NULL, &mmabc->BC));
1183: PetscCall(MatSetOptionsPrefix(mmabc->BC, prefix));
1184: PetscCall(MatAppendOptionsPrefix(mmabc->BC, "P1_"));
1185: PetscCall(MatProductSetType(mmabc->BC, p1));
1186: PetscCall(MatProductSetAlgorithm(mmabc->BC, MATPRODUCTALGORITHMDEFAULT));
1187: PetscCall(MatProductSetFill(mmabc->BC, product->fill));
1188: mmabc->BC->product->api_user = product->api_user;
1189: PetscCall(MatProductSetFromOptions(mmabc->BC));
1190: PetscCheck(mmabc->BC->ops->productsymbolic, PetscObjectComm((PetscObject)mat), PETSC_ERR_SUP, "Symbolic ProductType %s not supported with %s and %s", MatProductTypes[p1], ((PetscObject)B)->type_name, ((PetscObject)C)->type_name);
1191: /* use function pointer directly to prevent logging */
1192: PetscCall((*mmabc->BC->ops->productsymbolic)(mmabc->BC));
1194: PetscCall(MatProductCreate(A, mmabc->BC, NULL, &mmabc->ABC));
1195: PetscCall(MatSetOptionsPrefix(mmabc->ABC, prefix));
1196: PetscCall(MatAppendOptionsPrefix(mmabc->ABC, "P2_"));
1197: PetscCall(MatProductSetType(mmabc->ABC, p2));
1198: PetscCall(MatProductSetAlgorithm(mmabc->ABC, MATPRODUCTALGORITHMDEFAULT));
1199: PetscCall(MatProductSetFill(mmabc->ABC, product->fill));
1200: mmabc->ABC->product->api_user = product->api_user;
1201: PetscCall(MatProductSetFromOptions(mmabc->ABC));
1202: PetscCheck(mmabc->ABC->ops->productsymbolic, PetscObjectComm((PetscObject)mat), PETSC_ERR_SUP, "Symbolic ProductType %s not supported with %s and %s", MatProductTypes[p2], ((PetscObject)A)->type_name, ((PetscObject)mmabc->BC)->type_name);
1203: /* swap ABC product stuff with that of ABC for the symbolic phase on mat */
1204: mat->product = mmabc->ABC->product;
1205: mat->ops->productsymbolic = mmabc->ABC->ops->productsymbolic;
1206: /* use function pointer directly to prevent logging */
1207: PetscUseTypeMethod(mat, productsymbolic);
1208: mmabc->ABC->ops->productnumeric = mat->ops->productnumeric;
1209: mat->ops->productsymbolic = MatProductSymbolic_ABC_Basic;
1210: mat->ops->productnumeric = MatProductNumeric_ABC_Basic;
1211: mat->product = product;
1212: PetscFunctionReturn(PETSC_SUCCESS);
1213: }
1215: /*@
1216: MatProductGetType - Returns the type of matrix-matrix product associated with computing values for the given matrix
1218: Not Collective
1220: Input Parameter:
1221: . mat - the matrix whose values are to be computed via a matrix-matrix product operation
1223: Output Parameter:
1224: . mtype - the `MatProductType`
1226: Level: intermediate
1228: .seealso: [](ch_matrices), `MatProduct`, `Mat`, `MatProductCreateWithMat()`, `MatProductSetType()`, `MatProductCreate()`, `MatProductType`, `MatProductAlgorithm`
1229: @*/
1230: PetscErrorCode MatProductGetType(Mat mat, MatProductType *mtype)
1231: {
1232: PetscFunctionBegin;
1234: PetscAssertPointer(mtype, 2);
1235: *mtype = MATPRODUCT_UNSPECIFIED;
1236: if (mat->product) *mtype = mat->product->type;
1237: PetscFunctionReturn(PETSC_SUCCESS);
1238: }
1240: /*@
1241: MatProductGetMats - Returns the matrices associated with the matrix-matrix product associated with computing values for the given matrix
1243: Not Collective
1245: Input Parameter:
1246: . mat - the matrix whose values are to be computed via a matrix-matrix product operation
1248: Output Parameters:
1249: + A - the first matrix
1250: . B - the second matrix
1251: - C - the third matrix (may be `NULL` for some `MatProductType`)
1253: Level: intermediate
1255: .seealso: [](ch_matrices), `MatProduct`, `Mat`, `MatProductCreateWithMat()`, `MatProductSetType()`, `MatProductSetAlgorithm()`, `MatProductCreate()`
1256: @*/
1257: PetscErrorCode MatProductGetMats(Mat mat, Mat *A, Mat *B, Mat *C)
1258: {
1259: PetscFunctionBegin;
1261: if (A) *A = mat->product ? mat->product->A : NULL;
1262: if (B) *B = mat->product ? mat->product->B : NULL;
1263: if (C) *C = mat->product ? mat->product->C : NULL;
1264: PetscFunctionReturn(PETSC_SUCCESS);
1265: }