Actual source code: matimpl.h
1: #pragma once
3: #include <petscmat.h>
4: #include <petscmatcoarsen.h>
5: #include <petsc/private/petscimpl.h>
7: PETSC_EXTERN PetscBool MatRegisterAllCalled;
8: PETSC_EXTERN PetscBool MatSeqAIJRegisterAllCalled;
9: PETSC_EXTERN PetscBool MatOrderingRegisterAllCalled;
10: PETSC_EXTERN PetscBool MatColoringRegisterAllCalled;
11: PETSC_EXTERN PetscBool MatPartitioningRegisterAllCalled;
12: PETSC_EXTERN PetscBool MatCoarsenRegisterAllCalled;
13: PETSC_EXTERN PetscErrorCode MatRegisterAll(void);
14: PETSC_EXTERN PetscErrorCode MatOrderingRegisterAll(void);
15: PETSC_EXTERN PetscErrorCode MatColoringRegisterAll(void);
16: PETSC_EXTERN PetscErrorCode MatPartitioningRegisterAll(void);
17: PETSC_EXTERN PetscErrorCode MatCoarsenRegisterAll(void);
18: PETSC_EXTERN PetscErrorCode MatSeqAIJRegisterAll(void);
20: /* Gets the root type of the input matrix's type (e.g., MATAIJ for MATSEQAIJ) */
21: PETSC_EXTERN PetscErrorCode MatGetRootType_Private(Mat, MatType *);
23: /* Gets the MPI type corresponding to the input matrix's type (e.g., MATMPIAIJ for MATSEQAIJ) */
24: PETSC_EXTERN PetscErrorCode MatGetMPIMatType_Private(Mat, MatType *);
26: /*
27: This file defines the parts of the matrix data structure that are
28: shared by all matrix types.
29: */
31: /*
32: If you add entries here also add them to the MATOP enum
33: in include/petscmat.h and src/mat/f90-mod/petscmat.h
34: */
35: typedef struct _MatOps *MatOps;
36: struct _MatOps {
37: /* 0*/
38: PetscErrorCode (*setvalues)(Mat, PetscInt, const PetscInt[], PetscInt, const PetscInt[], const PetscScalar[], InsertMode);
39: PetscErrorCode (*getrow)(Mat, PetscInt, PetscInt *, PetscInt *[], PetscScalar *[]);
40: PetscErrorCode (*restorerow)(Mat, PetscInt, PetscInt *, PetscInt *[], PetscScalar *[]);
41: PetscErrorCode (*mult)(Mat, Vec, Vec);
42: PetscErrorCode (*multadd)(Mat, Vec, Vec, Vec);
43: /* 5*/
44: PetscErrorCode (*multtranspose)(Mat, Vec, Vec);
45: PetscErrorCode (*multtransposeadd)(Mat, Vec, Vec, Vec);
46: PetscErrorCode (*solve)(Mat, Vec, Vec);
47: PetscErrorCode (*solveadd)(Mat, Vec, Vec, Vec);
48: PetscErrorCode (*solvetranspose)(Mat, Vec, Vec);
49: /*10*/
50: PetscErrorCode (*solvetransposeadd)(Mat, Vec, Vec, Vec);
51: PetscErrorCode (*lufactor)(Mat, IS, IS, const MatFactorInfo *);
52: PetscErrorCode (*choleskyfactor)(Mat, IS, const MatFactorInfo *);
53: PetscErrorCode (*sor)(Mat, Vec, PetscReal, MatSORType, PetscReal, PetscInt, PetscInt, Vec);
54: PetscErrorCode (*transpose)(Mat, MatReuse, Mat *);
55: /*15*/
56: PetscErrorCode (*getinfo)(Mat, MatInfoType, MatInfo *);
57: PetscErrorCode (*equal)(Mat, Mat, PetscBool *);
58: PetscErrorCode (*getdiagonal)(Mat, Vec);
59: PetscErrorCode (*diagonalscale)(Mat, Vec, Vec);
60: PetscErrorCode (*norm)(Mat, NormType, PetscReal *);
61: /*20*/
62: PetscErrorCode (*assemblybegin)(Mat, MatAssemblyType);
63: PetscErrorCode (*assemblyend)(Mat, MatAssemblyType);
64: PetscErrorCode (*setoption)(Mat, MatOption, PetscBool);
65: PetscErrorCode (*zeroentries)(Mat);
66: /*24*/
67: PetscErrorCode (*zerorows)(Mat, PetscInt, const PetscInt[], PetscScalar, Vec, Vec);
68: PetscErrorCode (*lufactorsymbolic)(Mat, Mat, IS, IS, const MatFactorInfo *);
69: PetscErrorCode (*lufactornumeric)(Mat, Mat, const MatFactorInfo *);
70: PetscErrorCode (*choleskyfactorsymbolic)(Mat, Mat, IS, const MatFactorInfo *);
71: PetscErrorCode (*choleskyfactornumeric)(Mat, Mat, const MatFactorInfo *);
72: /*29*/
73: PetscErrorCode (*setup)(Mat);
74: PetscErrorCode (*ilufactorsymbolic)(Mat, Mat, IS, IS, const MatFactorInfo *);
75: PetscErrorCode (*iccfactorsymbolic)(Mat, Mat, IS, const MatFactorInfo *);
76: PetscErrorCode (*getdiagonalblock)(Mat, Mat *);
77: PetscErrorCode (*setinf)(Mat);
78: /*34*/
79: PetscErrorCode (*duplicate)(Mat, MatDuplicateOption, Mat *);
80: PetscErrorCode (*forwardsolve)(Mat, Vec, Vec);
81: PetscErrorCode (*backwardsolve)(Mat, Vec, Vec);
82: PetscErrorCode (*ilufactor)(Mat, IS, IS, const MatFactorInfo *);
83: PetscErrorCode (*iccfactor)(Mat, IS, const MatFactorInfo *);
84: /*39*/
85: PetscErrorCode (*axpy)(Mat, PetscScalar, Mat, MatStructure);
86: PetscErrorCode (*createsubmatrices)(Mat, PetscInt, const IS[], const IS[], MatReuse, Mat *[]);
87: PetscErrorCode (*increaseoverlap)(Mat, PetscInt, IS[], PetscInt);
88: PetscErrorCode (*getvalues)(Mat, PetscInt, const PetscInt[], PetscInt, const PetscInt[], PetscScalar[]);
89: PetscErrorCode (*copy)(Mat, Mat, MatStructure);
90: /*44*/
91: PetscErrorCode (*getrowmax)(Mat, Vec, PetscInt[]);
92: PetscErrorCode (*scale)(Mat, PetscScalar);
93: PetscErrorCode (*shift)(Mat, PetscScalar);
94: PetscErrorCode (*diagonalset)(Mat, Vec, InsertMode);
95: PetscErrorCode (*zerorowscolumns)(Mat, PetscInt, const PetscInt[], PetscScalar, Vec, Vec);
96: /*49*/
97: PetscErrorCode (*setrandom)(Mat, PetscRandom);
98: PetscErrorCode (*getrowij)(Mat, PetscInt, PetscBool, PetscBool, PetscInt *, const PetscInt *[], const PetscInt *[], PetscBool *);
99: PetscErrorCode (*restorerowij)(Mat, PetscInt, PetscBool, PetscBool, PetscInt *, const PetscInt *[], const PetscInt *[], PetscBool *);
100: PetscErrorCode (*getcolumnij)(Mat, PetscInt, PetscBool, PetscBool, PetscInt *, const PetscInt *[], const PetscInt *[], PetscBool *);
101: PetscErrorCode (*restorecolumnij)(Mat, PetscInt, PetscBool, PetscBool, PetscInt *, const PetscInt *[], const PetscInt *[], PetscBool *);
102: /*54*/
103: PetscErrorCode (*fdcoloringcreate)(Mat, ISColoring, MatFDColoring);
104: PetscErrorCode (*coloringpatch)(Mat, PetscInt, PetscInt, ISColoringValue[], ISColoring *);
105: PetscErrorCode (*setunfactored)(Mat);
106: PetscErrorCode (*permute)(Mat, IS, IS, Mat *);
107: PetscErrorCode (*setvaluesblocked)(Mat, PetscInt, const PetscInt[], PetscInt, const PetscInt[], const PetscScalar[], InsertMode);
108: /*59*/
109: PetscErrorCode (*createsubmatrix)(Mat, IS, IS, MatReuse, Mat *);
110: PetscErrorCode (*destroy)(Mat);
111: PetscErrorCode (*view)(Mat, PetscViewer);
112: PetscErrorCode (*convertfrom)(Mat, MatType, MatReuse, Mat *);
113: PetscErrorCode (*placeholder_63)(void);
114: /*64*/
115: PetscErrorCode (*matmatmultsymbolic)(Mat, Mat, Mat, PetscReal, Mat);
116: PetscErrorCode (*matmatmultnumeric)(Mat, Mat, Mat, Mat);
117: PetscErrorCode (*setlocaltoglobalmapping)(Mat, ISLocalToGlobalMapping, ISLocalToGlobalMapping);
118: PetscErrorCode (*setvalueslocal)(Mat, PetscInt, const PetscInt[], PetscInt, const PetscInt[], const PetscScalar[], InsertMode);
119: PetscErrorCode (*zerorowslocal)(Mat, PetscInt, const PetscInt[], PetscScalar, Vec, Vec);
120: /*69*/
121: PetscErrorCode (*getrowmaxabs)(Mat, Vec, PetscInt[]);
122: PetscErrorCode (*getrowminabs)(Mat, Vec, PetscInt[]);
123: PetscErrorCode (*convert)(Mat, MatType, MatReuse, Mat *);
124: PetscErrorCode (*hasoperation)(Mat, MatOperation, PetscBool *);
125: PetscErrorCode (*placeholder_73)(void);
126: /*74*/
127: PetscErrorCode (*setvaluesadifor)(Mat, PetscInt, void *);
128: PetscErrorCode (*fdcoloringapply)(Mat, MatFDColoring, Vec, void *);
129: PetscErrorCode (*setfromoptions)(Mat, PetscOptionItems *);
130: PetscErrorCode (*placeholder_77)(void);
131: PetscErrorCode (*placeholder_78)(void);
132: /*79*/
133: PetscErrorCode (*findzerodiagonals)(Mat, IS *);
134: PetscErrorCode (*mults)(Mat, Vecs, Vecs);
135: PetscErrorCode (*solves)(Mat, Vecs, Vecs);
136: PetscErrorCode (*getinertia)(Mat, PetscInt *, PetscInt *, PetscInt *);
137: PetscErrorCode (*load)(Mat, PetscViewer);
138: /*84*/
139: PetscErrorCode (*issymmetric)(Mat, PetscReal, PetscBool *);
140: PetscErrorCode (*ishermitian)(Mat, PetscReal, PetscBool *);
141: PetscErrorCode (*isstructurallysymmetric)(Mat, PetscBool *);
142: PetscErrorCode (*setvaluesblockedlocal)(Mat, PetscInt, const PetscInt[], PetscInt, const PetscInt[], const PetscScalar[], InsertMode);
143: PetscErrorCode (*getvecs)(Mat, Vec *, Vec *);
144: /*89*/
145: PetscErrorCode (*placeholder_89)(void);
146: PetscErrorCode (*matmultsymbolic)(Mat, Mat, PetscReal, Mat);
147: PetscErrorCode (*matmultnumeric)(Mat, Mat, Mat);
148: PetscErrorCode (*placeholder_92)(void);
149: PetscErrorCode (*ptapsymbolic)(Mat, Mat, PetscReal, Mat); /* double dispatch wrapper routine */
150: /*94*/
151: PetscErrorCode (*ptapnumeric)(Mat, Mat, Mat); /* double dispatch wrapper routine */
152: PetscErrorCode (*placeholder_95)(void);
153: PetscErrorCode (*mattransposemultsymbolic)(Mat, Mat, PetscReal, Mat);
154: PetscErrorCode (*mattransposemultnumeric)(Mat, Mat, Mat);
155: PetscErrorCode (*bindtocpu)(Mat, PetscBool);
156: /*99*/
157: PetscErrorCode (*productsetfromoptions)(Mat);
158: PetscErrorCode (*productsymbolic)(Mat);
159: PetscErrorCode (*productnumeric)(Mat);
160: PetscErrorCode (*conjugate)(Mat); /* complex conjugate */
161: PetscErrorCode (*viewnative)(Mat, PetscViewer);
162: /*104*/
163: PetscErrorCode (*setvaluesrow)(Mat, PetscInt, const PetscScalar[]);
164: PetscErrorCode (*realpart)(Mat);
165: PetscErrorCode (*imaginarypart)(Mat);
166: PetscErrorCode (*getrowuppertriangular)(Mat);
167: PetscErrorCode (*restorerowuppertriangular)(Mat);
168: /*109*/
169: PetscErrorCode (*matsolve)(Mat, Mat, Mat);
170: PetscErrorCode (*matsolvetranspose)(Mat, Mat, Mat);
171: PetscErrorCode (*getrowmin)(Mat, Vec, PetscInt[]);
172: PetscErrorCode (*getcolumnvector)(Mat, Vec, PetscInt);
173: PetscErrorCode (*missingdiagonal)(Mat, PetscBool *, PetscInt *);
174: /*114*/
175: PetscErrorCode (*getseqnonzerostructure)(Mat, Mat *);
176: PetscErrorCode (*create)(Mat);
177: PetscErrorCode (*getghosts)(Mat, PetscInt *, const PetscInt *[]);
178: PetscErrorCode (*getlocalsubmatrix)(Mat, IS, IS, Mat *);
179: PetscErrorCode (*restorelocalsubmatrix)(Mat, IS, IS, Mat *);
180: /*119*/
181: PetscErrorCode (*multdiagonalblock)(Mat, Vec, Vec);
182: PetscErrorCode (*hermitiantranspose)(Mat, MatReuse, Mat *);
183: PetscErrorCode (*multhermitiantranspose)(Mat, Vec, Vec);
184: PetscErrorCode (*multhermitiantransposeadd)(Mat, Vec, Vec, Vec);
185: PetscErrorCode (*getmultiprocblock)(Mat, MPI_Comm, MatReuse, Mat *);
186: /*124*/
187: PetscErrorCode (*findnonzerorows)(Mat, IS *);
188: PetscErrorCode (*getcolumnreductions)(Mat, PetscInt, PetscReal *);
189: PetscErrorCode (*invertblockdiagonal)(Mat, const PetscScalar **);
190: PetscErrorCode (*invertvariableblockdiagonal)(Mat, PetscInt, const PetscInt *, PetscScalar *);
191: PetscErrorCode (*createsubmatricesmpi)(Mat, PetscInt, const IS[], const IS[], MatReuse, Mat **);
192: /*129*/
193: PetscErrorCode (*setvaluesbatch)(Mat, PetscInt, PetscInt, PetscInt *, const PetscScalar *);
194: PetscErrorCode (*placeholder_130)(void);
195: PetscErrorCode (*transposematmultsymbolic)(Mat, Mat, PetscReal, Mat);
196: PetscErrorCode (*transposematmultnumeric)(Mat, Mat, Mat);
197: PetscErrorCode (*transposecoloringcreate)(Mat, ISColoring, MatTransposeColoring);
198: /*134*/
199: PetscErrorCode (*transcoloringapplysptoden)(MatTransposeColoring, Mat, Mat);
200: PetscErrorCode (*transcoloringapplydentosp)(MatTransposeColoring, Mat, Mat);
201: PetscErrorCode (*placeholder_136)(void);
202: PetscErrorCode (*rartsymbolic)(Mat, Mat, PetscReal, Mat); /* double dispatch wrapper routine */
203: PetscErrorCode (*rartnumeric)(Mat, Mat, Mat); /* double dispatch wrapper routine */
204: /*139*/
205: PetscErrorCode (*setblocksizes)(Mat, PetscInt, PetscInt);
206: PetscErrorCode (*aypx)(Mat, PetscScalar, Mat, MatStructure);
207: PetscErrorCode (*residual)(Mat, Vec, Vec, Vec);
208: PetscErrorCode (*fdcoloringsetup)(Mat, ISColoring, MatFDColoring);
209: PetscErrorCode (*findoffblockdiagonalentries)(Mat, IS *);
210: PetscErrorCode (*creatempimatconcatenateseqmat)(MPI_Comm, Mat, PetscInt, MatReuse, Mat *);
211: /*145*/
212: PetscErrorCode (*destroysubmatrices)(PetscInt, Mat *[]);
213: PetscErrorCode (*mattransposesolve)(Mat, Mat, Mat);
214: PetscErrorCode (*getvalueslocal)(Mat, PetscInt, const PetscInt[], PetscInt, const PetscInt[], PetscScalar[]);
215: PetscErrorCode (*creategraph)(Mat, PetscBool, PetscBool, PetscReal, PetscInt, PetscInt[], Mat *);
216: PetscErrorCode (*dummy)(Mat);
217: /*150*/
218: PetscErrorCode (*transposesymbolic)(Mat, Mat *);
219: PetscErrorCode (*eliminatezeros)(Mat, PetscBool);
220: PetscErrorCode (*getrowsumabs)(Mat, Vec);
221: PetscErrorCode (*getfactor)(Mat, MatSolverType, MatFactorType, Mat *);
222: PetscErrorCode (*getblockdiagonal)(Mat, Mat *); // NOTE: the caller of get{block, vblock}diagonal owns the returned matrix;
223: /*155*/
224: PetscErrorCode (*getvblockdiagonal)(Mat, Mat *); // they must destroy it after use
225: PetscErrorCode (*copyhashtoxaij)(Mat, Mat);
226: };
227: /*
228: If you add MatOps entries above also add them to the MATOP enum
229: in include/petscmat.h and src/mat/f90-mod/petscmat.h
230: */
232: #include <petscsys.h>
234: typedef struct _p_MatRootName *MatRootName;
235: struct _p_MatRootName {
236: char *rname, *sname, *mname;
237: MatRootName next;
238: };
240: PETSC_EXTERN MatRootName MatRootNameList;
242: /*
243: Utility private matrix routines used outside Mat
244: */
245: PETSC_SINGLE_LIBRARY_INTERN PetscErrorCode MatFindNonzeroRowsOrCols_Basic(Mat, PetscBool, PetscReal, IS *);
246: PETSC_EXTERN PetscErrorCode MatShellGetScalingShifts(Mat, PetscScalar *, PetscScalar *, Vec *, Vec *, Vec *, Mat *, IS *, IS *);
248: #define MAT_SHELL_NOT_ALLOWED (void *)-1
250: /*
251: Utility private matrix routines
252: */
253: PETSC_INTERN PetscErrorCode MatConvert_Basic(Mat, MatType, MatReuse, Mat *);
254: PETSC_INTERN PetscErrorCode MatConvert_Shell(Mat, MatType, MatReuse, Mat *);
255: PETSC_INTERN PetscErrorCode MatConvertFrom_Shell(Mat, MatType, MatReuse, Mat *);
256: PETSC_INTERN PetscErrorCode MatShellSetContext_Immutable(Mat, void *);
257: PETSC_INTERN PetscErrorCode MatShellSetContextDestroy_Immutable(Mat, PetscCtxDestroyFn *);
258: PETSC_INTERN PetscErrorCode MatShellSetManageScalingShifts_Immutable(Mat);
259: PETSC_INTERN PetscErrorCode MatCopy_Basic(Mat, Mat, MatStructure);
260: PETSC_INTERN PetscErrorCode MatDiagonalSet_Default(Mat, Vec, InsertMode);
261: #if defined(PETSC_HAVE_SCALAPACK)
262: PETSC_INTERN PetscErrorCode MatConvert_Dense_ScaLAPACK(Mat, MatType, MatReuse, Mat *);
263: #endif
264: PETSC_INTERN PetscErrorCode MatSetPreallocationCOO_Basic(Mat, PetscCount, PetscInt[], PetscInt[]);
265: PETSC_INTERN PetscErrorCode MatSetValuesCOO_Basic(Mat, const PetscScalar[], InsertMode);
267: /* This can be moved to the public header after implementing some missing MatProducts */
268: PETSC_INTERN PetscErrorCode MatCreateFromISLocalToGlobalMapping(ISLocalToGlobalMapping, Mat, PetscBool, PetscBool, MatType, Mat *);
270: /* these callbacks rely on the old matrix function pointers for
271: matmat operations. They are unsafe, and should be removed.
272: However, the amount of work needed to clean up all the
273: implementations is not negligible */
274: PETSC_INTERN PetscErrorCode MatProductSymbolic_AB(Mat);
275: PETSC_INTERN PetscErrorCode MatProductNumeric_AB(Mat);
276: PETSC_INTERN PetscErrorCode MatProductSymbolic_AtB(Mat);
277: PETSC_INTERN PetscErrorCode MatProductNumeric_AtB(Mat);
278: PETSC_INTERN PetscErrorCode MatProductSymbolic_ABt(Mat);
279: PETSC_INTERN PetscErrorCode MatProductNumeric_ABt(Mat);
280: PETSC_INTERN PetscErrorCode MatProductNumeric_PtAP(Mat);
281: PETSC_INTERN PetscErrorCode MatProductNumeric_RARt(Mat);
282: PETSC_INTERN PetscErrorCode MatProductSymbolic_ABC(Mat);
283: PETSC_INTERN PetscErrorCode MatProductNumeric_ABC(Mat);
285: PETSC_INTERN PetscErrorCode MatProductCreate_Private(Mat, Mat, Mat, Mat);
286: /* this callback handles all the different triple products and
287: does not rely on the function pointers; used by cuSPARSE/hipSPARSE and KOKKOS-KERNELS */
288: PETSC_INTERN PetscErrorCode MatProductSymbolic_ABC_Basic(Mat);
290: /* CreateGraph is common to AIJ seq and mpi */
291: PETSC_INTERN PetscErrorCode MatCreateGraph_Simple_AIJ(Mat, PetscBool, PetscBool, PetscReal, PetscInt, PetscInt[], Mat *);
293: #if defined(PETSC_CLANG_STATIC_ANALYZER)
294: template <typename Tm>
295: extern void MatCheckPreallocated(Tm, int);
296: template <typename Tm>
297: extern void MatCheckProduct(Tm, int);
298: #else /* PETSC_CLANG_STATIC_ANALYZER */
299: #define MatCheckPreallocated(A, arg) \
300: do { \
301: if (!(A)->preallocated) PetscCall(MatSetUp(A)); \
302: } while (0)
304: #if defined(PETSC_USE_DEBUG)
305: #define MatCheckProduct(A, arg) \
306: do { \
307: PetscCheck((A)->product, PETSC_COMM_SELF, PETSC_ERR_ARG_WRONGSTATE, "Argument %d \"%s\" is not a matrix obtained from MatProductCreate()", (arg), #A); \
308: } while (0)
309: #else
310: #define MatCheckProduct(A, arg) \
311: do { \
312: } while (0)
313: #endif
314: #endif /* PETSC_CLANG_STATIC_ANALYZER */
316: /*
317: The stash is used to temporarily store inserted matrix values that
318: belong to another processor. During the assembly phase the stashed
319: values are moved to the correct processor and
320: */
322: typedef struct _MatStashSpace *PetscMatStashSpace;
324: struct _MatStashSpace {
325: PetscMatStashSpace next;
326: PetscScalar *space_head, *val;
327: PetscInt *idx, *idy;
328: PetscInt total_space_size;
329: PetscInt local_used;
330: PetscInt local_remaining;
331: };
333: PETSC_EXTERN PetscErrorCode PetscMatStashSpaceGet(PetscInt, PetscInt, PetscMatStashSpace *);
334: PETSC_EXTERN PetscErrorCode PetscMatStashSpaceContiguous(PetscInt, PetscMatStashSpace *, PetscScalar *, PetscInt *, PetscInt *);
335: PETSC_EXTERN PetscErrorCode PetscMatStashSpaceDestroy(PetscMatStashSpace *);
337: typedef struct {
338: PetscInt count;
339: } MatStashHeader;
341: typedef struct {
342: void *buffer; /* Of type blocktype, dynamically constructed */
343: PetscInt count;
344: char pending;
345: } MatStashFrame;
347: typedef struct _MatStash MatStash;
348: struct _MatStash {
349: PetscInt nmax; /* maximum stash size */
350: PetscInt umax; /* user specified max-size */
351: PetscInt oldnmax; /* the nmax value used previously */
352: PetscInt n; /* stash size */
353: PetscInt bs; /* block size of the stash */
354: PetscInt reallocs; /* preserve the no of mallocs invoked */
355: PetscMatStashSpace space_head, space; /* linked list to hold stashed global row/column numbers and matrix values */
357: PetscErrorCode (*ScatterBegin)(Mat, MatStash *, PetscInt *);
358: PetscErrorCode (*ScatterGetMesg)(MatStash *, PetscMPIInt *, PetscInt **, PetscInt **, PetscScalar **, PetscInt *);
359: PetscErrorCode (*ScatterEnd)(MatStash *);
360: PetscErrorCode (*ScatterDestroy)(MatStash *);
362: /* The following variables are used for communication */
363: MPI_Comm comm;
364: PetscMPIInt size, rank;
365: PetscMPIInt tag1, tag2;
366: MPI_Request *send_waits; /* array of send requests */
367: MPI_Request *recv_waits; /* array of receive requests */
368: MPI_Status *send_status; /* array of send status */
369: PetscMPIInt nsends, nrecvs; /* numbers of sends and receives */
370: PetscScalar *svalues; /* sending data */
371: PetscInt *sindices;
372: PetscScalar **rvalues; /* receiving data (values) */
373: PetscInt **rindices; /* receiving data (indices) */
374: PetscMPIInt nprocessed; /* number of messages already processed */
375: PetscMPIInt *flg_v; /* indicates what messages have arrived so far and from whom */
376: PetscBool reproduce;
377: PetscMPIInt reproduce_count;
379: /* The following variables are used for BTS communication */
380: PetscBool first_assembly_done; /* Is the first time matrix assembly done? */
381: PetscBool use_status; /* Use MPI_Status to determine number of items in each message */
382: PetscMPIInt nsendranks;
383: PetscMPIInt nrecvranks;
384: PetscMPIInt *sendranks;
385: PetscMPIInt *recvranks;
386: MatStashHeader *sendhdr, *recvhdr;
387: MatStashFrame *sendframes; /* pointers to the main messages */
388: MatStashFrame *recvframes;
389: MatStashFrame *recvframe_active;
390: PetscInt recvframe_i; /* index of block within active frame */
391: PetscInt recvframe_count; /* Count actually sent for current frame */
392: PetscMPIInt recvcount; /* Number of receives processed so far */
393: PetscMPIInt *some_indices; /* From last call to MPI_Waitsome */
394: MPI_Status *some_statuses; /* Statuses from last call to MPI_Waitsome */
395: PetscMPIInt some_count; /* Number of requests completed in last call to MPI_Waitsome */
396: PetscMPIInt some_i; /* Index of request currently being processed */
397: MPI_Request *sendreqs;
398: MPI_Request *recvreqs;
399: PetscSegBuffer segsendblocks;
400: PetscSegBuffer segrecvframe;
401: PetscSegBuffer segrecvblocks;
402: MPI_Datatype blocktype;
403: size_t blocktype_size;
404: InsertMode *insertmode; /* Pointer to check mat->insertmode and set upon message arrival in case no local values have been set. */
405: };
407: #if !defined(PETSC_HAVE_MPIUNI)
408: PETSC_INTERN PetscErrorCode MatStashScatterDestroy_BTS(MatStash *);
409: #endif
410: PETSC_INTERN PetscErrorCode MatStashCreate_Private(MPI_Comm, PetscInt, MatStash *);
411: PETSC_INTERN PetscErrorCode MatStashDestroy_Private(MatStash *);
412: PETSC_INTERN PetscErrorCode MatStashScatterEnd_Private(MatStash *);
413: PETSC_INTERN PetscErrorCode MatStashSetInitialSize_Private(MatStash *, PetscInt);
414: PETSC_INTERN PetscErrorCode MatStashGetInfo_Private(MatStash *, PetscInt *, PetscInt *);
415: PETSC_INTERN PetscErrorCode MatStashValuesRow_Private(MatStash *, PetscInt, PetscInt, const PetscInt[], const PetscScalar[], PetscBool);
416: PETSC_INTERN PetscErrorCode MatStashValuesCol_Private(MatStash *, PetscInt, PetscInt, const PetscInt[], const PetscScalar[], PetscInt, PetscBool);
417: PETSC_INTERN PetscErrorCode MatStashValuesRowBlocked_Private(MatStash *, PetscInt, PetscInt, const PetscInt[], const PetscScalar[], PetscInt, PetscInt, PetscInt);
418: PETSC_INTERN PetscErrorCode MatStashValuesColBlocked_Private(MatStash *, PetscInt, PetscInt, const PetscInt[], const PetscScalar[], PetscInt, PetscInt, PetscInt);
419: PETSC_INTERN PetscErrorCode MatStashScatterBegin_Private(Mat, MatStash *, PetscInt *);
420: PETSC_INTERN PetscErrorCode MatStashScatterGetMesg_Private(MatStash *, PetscMPIInt *, PetscInt **, PetscInt **, PetscScalar **, PetscInt *);
421: PETSC_INTERN PetscErrorCode MatGetInfo_External(Mat, MatInfoType, MatInfo *);
423: typedef struct {
424: PetscInt dim;
425: PetscInt dims[4];
426: PetscInt starts[4];
427: PetscBool noc; /* this is a single component problem, hence user will not set MatStencil.c */
428: } MatStencilInfo;
430: /* Info about using compressed row format */
431: typedef struct {
432: PetscBool use; /* indicates compressed rows have been checked and will be used */
433: PetscInt nrows; /* number of non-zero rows */
434: PetscInt *i; /* compressed row pointer */
435: PetscInt *rindex; /* compressed row index */
436: } Mat_CompressedRow;
437: PETSC_EXTERN PetscErrorCode MatCheckCompressedRow(Mat, PetscInt, Mat_CompressedRow *, PetscInt *, PetscInt, PetscReal);
439: typedef struct { /* used by MatCreateRedundantMatrix() for reusing matredundant */
440: PetscInt nzlocal, nsends, nrecvs;
441: PetscMPIInt *send_rank, *recv_rank;
442: PetscInt *sbuf_nz, *rbuf_nz, *sbuf_j, **rbuf_j;
443: PetscScalar *sbuf_a, **rbuf_a;
444: MPI_Comm subcomm; /* when user does not provide a subcomm */
445: IS isrow, iscol;
446: Mat *matseq;
447: } Mat_Redundant;
449: typedef struct { /* used by MatProduct() */
450: MatProductType type;
451: char *alg;
452: Mat A, B, C, Dwork;
453: PetscBool symbolic_used_the_fact_A_is_symmetric; /* Symbolic phase took advantage of the fact that A is symmetric, and optimized e.g. AtB as AB. Then, .. */
454: PetscBool symbolic_used_the_fact_B_is_symmetric; /* .. in the numeric phase, if a new A is not symmetric (but has the same sparsity as the old A therefore .. */
455: PetscBool symbolic_used_the_fact_C_is_symmetric; /* MatMatMult(A,B,MAT_REUSE_MATRIX,..&C) is still legitimate), we need to redo symbolic! */
456: PetscObjectParameterDeclare(PetscReal, fill);
457: PetscBool api_user; /* used to distinguish command line options and to indicate the matrix values are ready to be consumed at symbolic phase if needed */
458: PetscBool setfromoptionscalled;
460: /* Some products may display the information on the algorithm used */
461: PetscErrorCode (*view)(Mat, PetscViewer);
463: /* many products have intermediate data structures, each specific to Mat types and product type */
464: PetscBool clear; /* whether or not to clear the data structures after MatProductNumeric has been called */
465: void *data; /* where to stash those structures */
466: PetscErrorCode (*destroy)(void *); /* destroy routine */
467: } Mat_Product;
469: struct _p_Mat {
470: PETSCHEADER(struct _MatOps);
471: PetscLayout rmap, cmap;
472: void *data; /* implementation-specific data */
473: MatFactorType factortype; /* MAT_FACTOR_LU, ILU, CHOLESKY or ICC */
474: PetscBool trivialsymbolic; /* indicates the symbolic factorization doesn't actually do a symbolic factorization, it is delayed to the numeric factorization */
475: PetscBool canuseordering; /* factorization can use ordering provide to routine (most PETSc implementations) */
476: MatOrderingType preferredordering[MAT_FACTOR_NUM_TYPES]; /* what is the preferred (or default) ordering for the matrix solver type */
477: PetscBool assembled; /* is the matrix assembled? */
478: PetscBool was_assembled; /* new values inserted into assembled mat */
479: PetscInt num_ass; /* number of times matrix has been assembled */
480: PetscObjectState nonzerostate; /* each time new nonzeros locations are introduced into the matrix this is updated */
481: PetscObjectState ass_nonzerostate; /* nonzero state at last assembly */
482: MatInfo info; /* matrix information */
483: InsertMode insertmode; /* have values been inserted in matrix or added? */
484: MatStash stash, bstash; /* used for assembling off-proc mat emements */
485: MatNullSpace nullsp; /* null space (operator is singular) */
486: MatNullSpace transnullsp; /* null space of transpose of operator */
487: MatNullSpace nearnullsp; /* near null space to be used by multigrid methods */
488: PetscInt congruentlayouts; /* are the rows and columns layouts congruent? */
489: PetscBool preallocated;
490: MatStencilInfo stencil; /* information for structured grid */
491: PetscBool3 symmetric, hermitian, structurally_symmetric, spd;
492: PetscBool symmetry_eternal, structural_symmetry_eternal, spd_eternal;
493: PetscBool nooffprocentries, nooffproczerorows;
494: PetscBool assembly_subset; /* set by MAT_SUBSET_OFF_PROC_ENTRIES */
495: PetscBool submat_singleis; /* for efficient PCSetUp_ASM() */
496: PetscBool structure_only;
497: PetscBool sortedfull; /* full, sorted rows are inserted */
498: PetscBool force_diagonals; /* set by MAT_FORCE_DIAGONAL_ENTRIES */
499: #if defined(PETSC_HAVE_DEVICE)
500: PetscOffloadMask offloadmask; /* a mask which indicates where the valid matrix data is (GPU, CPU or both) */
501: PetscBool boundtocpu;
502: PetscBool bindingpropagates;
503: #endif
504: char *defaultrandtype;
505: void *spptr; /* pointer for special library like SuperLU */
506: char *solvertype;
507: PetscBool checksymmetryonassembly, checknullspaceonassembly;
508: PetscReal checksymmetrytol;
509: Mat schur; /* Schur complement matrix */
510: MatFactorSchurStatus schur_status; /* status of the Schur complement matrix */
511: Mat_Redundant *redundant; /* used by MatCreateRedundantMatrix() */
512: PetscBool erroriffailure; /* Generate an error if detected (for example a zero pivot) instead of returning */
513: MatFactorError factorerrortype; /* type of error in factorization */
514: PetscReal factorerror_zeropivot_value; /* If numerical zero pivot was detected this is the computed value */
515: PetscInt factorerror_zeropivot_row; /* Row where zero pivot was detected */
516: PetscInt nblocks, *bsizes; /* support for MatSetVariableBlockSizes() */
517: PetscInt p_cstart, p_rank, p_cend, n_rank; /* Information from parallel MatComputeVariableBlockEnvelope() */
518: PetscBool p_parallel;
519: char *defaultvectype;
520: Mat_Product *product;
521: PetscBool form_explicit_transpose; /* hint to generate an explicit mat tranpsose for operations like MatMultTranspose() */
522: PetscBool transupdated; /* whether or not the explicitly generated transpose is up-to-date */
523: char *factorprefix; /* the prefix to use with factored matrix that is created */
524: PetscBool hash_active; /* indicates MatSetValues() is being handled by hashing */
525: };
527: PETSC_INTERN PetscErrorCode MatAXPY_Basic(Mat, PetscScalar, Mat, MatStructure);
528: PETSC_INTERN PetscErrorCode MatAXPY_BasicWithPreallocation(Mat, Mat, PetscScalar, Mat, MatStructure);
529: PETSC_INTERN PetscErrorCode MatAXPY_Basic_Preallocate(Mat, Mat, Mat *);
530: PETSC_INTERN PetscErrorCode MatAXPY_Dense_Nest(Mat, PetscScalar, Mat);
532: PETSC_INTERN PetscErrorCode MatSetUp_Default(Mat);
534: /*
535: Utility for MatZeroRows
536: */
537: PETSC_INTERN PetscErrorCode MatZeroRowsMapLocal_Private(Mat, PetscInt, const PetscInt *, PetscInt *, PetscInt **);
539: /*
540: Utility for MatView/MatLoad
541: */
542: PETSC_INTERN PetscErrorCode MatView_Binary_BlockSizes(Mat, PetscViewer);
543: PETSC_INTERN PetscErrorCode MatLoad_Binary_BlockSizes(Mat, PetscViewer);
545: /*
546: Object for partitioning graphs
547: */
549: typedef struct _MatPartitioningOps *MatPartitioningOps;
550: struct _MatPartitioningOps {
551: PetscErrorCode (*apply)(MatPartitioning, IS *);
552: PetscErrorCode (*applynd)(MatPartitioning, IS *);
553: PetscErrorCode (*setfromoptions)(MatPartitioning, PetscOptionItems *);
554: PetscErrorCode (*destroy)(MatPartitioning);
555: PetscErrorCode (*view)(MatPartitioning, PetscViewer);
556: PetscErrorCode (*improve)(MatPartitioning, IS *);
557: };
559: struct _p_MatPartitioning {
560: PETSCHEADER(struct _MatPartitioningOps);
561: Mat adj;
562: PetscInt *vertex_weights;
563: PetscReal *part_weights;
564: PetscInt n; /* number of partitions */
565: PetscInt ncon; /* number of vertex weights per vertex */
566: void *data;
567: PetscInt setupcalled;
568: PetscBool use_edge_weights; /* A flag indicates whether or not to use edge weights */
569: };
571: /* needed for parallel nested dissection by ParMetis and PTSCOTCH */
572: PETSC_INTERN PetscErrorCode MatPartitioningSizesToSep_Private(PetscInt, PetscInt[], PetscInt[], PetscInt[]);
574: /*
575: Object for coarsen graphs
576: */
577: typedef struct _MatCoarsenOps *MatCoarsenOps;
578: struct _MatCoarsenOps {
579: PetscErrorCode (*apply)(MatCoarsen);
580: PetscErrorCode (*setfromoptions)(MatCoarsen, PetscOptionItems *);
581: PetscErrorCode (*destroy)(MatCoarsen);
582: PetscErrorCode (*view)(MatCoarsen, PetscViewer);
583: };
585: #define MAT_COARSEN_STRENGTH_INDEX_SIZE 3
586: struct _p_MatCoarsen {
587: PETSCHEADER(struct _MatCoarsenOps);
588: Mat graph;
589: void *subctx;
590: /* */
591: PetscBool strict_aggs;
592: IS perm;
593: PetscCoarsenData *agg_lists;
594: PetscInt max_it; /* number of iterations in HEM */
595: PetscReal threshold; /* HEM can filter interim graphs */
596: PetscInt strength_index_size;
597: PetscInt strength_index[MAT_COARSEN_STRENGTH_INDEX_SIZE];
598: };
600: PETSC_EXTERN PetscErrorCode MatCoarsenMISKSetDistance(MatCoarsen, PetscInt);
601: PETSC_EXTERN PetscErrorCode MatCoarsenMISKGetDistance(MatCoarsen, PetscInt *);
603: /*
604: Used in aijdevice.h
605: */
606: typedef struct {
607: PetscInt *i;
608: PetscInt *j;
609: PetscScalar *a;
610: PetscInt n;
611: PetscInt ignorezeroentries;
612: } PetscCSRDataStructure;
614: /*
615: MatFDColoring is used to compute Jacobian matrices efficiently
616: via coloring. The data structure is explained below in an example.
618: Color = 0 1 0 2 | 2 3 0
619: ---------------------------------------------------
620: 00 01 | 05
621: 10 11 | 14 15 Processor 0
622: 22 23 | 25
623: 32 33 |
624: ===================================================
625: | 44 45 46
626: 50 | 55 Processor 1
627: | 64 66
628: ---------------------------------------------------
630: ncolors = 4;
632: ncolumns = {2,1,1,0}
633: columns = {{0,2},{1},{3},{}}
634: nrows = {4,2,3,3}
635: rows = {{0,1,2,3},{0,1},{1,2,3},{0,1,2}}
636: vwscale = {dx(0),dx(1),dx(2),dx(3)} MPI Vec
637: vscale = {dx(0),dx(1),dx(2),dx(3),dx(4),dx(5)} Seq Vec
639: ncolumns = {1,0,1,1}
640: columns = {{6},{},{4},{5}}
641: nrows = {3,0,2,2}
642: rows = {{0,1,2},{},{1,2},{1,2}}
643: vwscale = {dx(4),dx(5),dx(6)} MPI Vec
644: vscale = {dx(0),dx(4),dx(5),dx(6)} Seq Vec
646: See the routine MatFDColoringApply() for how this data is used
647: to compute the Jacobian.
649: */
650: typedef struct {
651: PetscInt row;
652: PetscInt col;
653: PetscScalar *valaddr; /* address of value */
654: } MatEntry;
656: typedef struct {
657: PetscInt row;
658: PetscScalar *valaddr; /* address of value */
659: } MatEntry2;
661: struct _p_MatFDColoring {
662: PETSCHEADER(int);
663: PetscInt M, N, m; /* total rows, columns; local rows */
664: PetscInt rstart; /* first row owned by local processor */
665: PetscInt ncolors; /* number of colors */
666: PetscInt *ncolumns; /* number of local columns for a color */
667: PetscInt **columns; /* lists the local columns of each color (using global column numbering) */
668: IS *isa; /* these are the IS that contain the column values given in columns */
669: PetscInt *nrows; /* number of local rows for each color */
670: MatEntry *matentry; /* holds (row, column, address of value) for Jacobian matrix entry */
671: MatEntry2 *matentry2; /* holds (row, address of value) for Jacobian matrix entry */
672: PetscScalar *dy; /* store a block of F(x+dx)-F(x) when J is in BAIJ format */
673: PetscReal error_rel; /* square root of relative error in computing function */
674: PetscReal umin; /* minimum allowable u'dx value */
675: Vec w1, w2, w3; /* work vectors used in computing Jacobian */
676: PetscBool fset; /* indicates that the initial function value F(X) is set */
677: PetscErrorCode (*f)(void); /* function that defines Jacobian */
678: void *fctx; /* optional user-defined context for use by the function f */
679: Vec vscale; /* holds FD scaling, i.e. 1/dx for each perturbed column */
680: PetscInt currentcolor; /* color for which function evaluation is being done now */
681: const char *htype; /* "wp" or "ds" */
682: ISColoringType ctype; /* IS_COLORING_GLOBAL or IS_COLORING_LOCAL */
683: PetscInt brows, bcols; /* number of block rows or columns for speedup inserting the dense matrix into sparse Jacobian */
684: PetscBool setupcalled; /* true if setup has been called */
685: PetscBool viewed; /* true if the -mat_fd_coloring_view has been triggered already */
686: void (*ftn_func_pointer)(void), *ftn_func_cntx; /* serve the same purpose as *fortran_func_pointers in PETSc objects */
687: PetscObjectId matid; /* matrix this object was created with, must always be the same */
688: };
690: typedef struct _MatColoringOps *MatColoringOps;
691: struct _MatColoringOps {
692: PetscErrorCode (*destroy)(MatColoring);
693: PetscErrorCode (*setfromoptions)(MatColoring, PetscOptionItems *);
694: PetscErrorCode (*view)(MatColoring, PetscViewer);
695: PetscErrorCode (*apply)(MatColoring, ISColoring *);
696: PetscErrorCode (*weights)(MatColoring, PetscReal **, PetscInt **);
697: };
699: struct _p_MatColoring {
700: PETSCHEADER(struct _MatColoringOps);
701: Mat mat;
702: PetscInt dist; /* distance of the coloring */
703: PetscInt maxcolors; /* the maximum number of colors returned, maxcolors=1 for MIS */
704: void *data; /* inner context */
705: PetscBool valid; /* check to see if what is produced is a valid coloring */
706: MatColoringWeightType weight_type; /* type of weight computation to be performed */
707: PetscReal *user_weights; /* custom weights and permutation */
708: PetscInt *user_lperm;
709: PetscBool valid_iscoloring; /* check to see if matcoloring is produced a valid iscoloring */
710: };
712: struct _p_MatTransposeColoring {
713: PETSCHEADER(int);
714: PetscInt M, N, m; /* total rows, columns; local rows */
715: PetscInt rstart; /* first row owned by local processor */
716: PetscInt ncolors; /* number of colors */
717: PetscInt *ncolumns; /* number of local columns for a color */
718: PetscInt *nrows; /* number of local rows for each color */
719: PetscInt currentcolor; /* color for which function evaluation is being done now */
720: ISColoringType ctype; /* IS_COLORING_GLOBAL or IS_COLORING_LOCAL */
722: PetscInt *colorforrow, *colorforcol; /* pointer to rows and columns */
723: PetscInt *rows; /* lists the local rows for each color (using the local row numbering) */
724: PetscInt *den2sp; /* maps (row,color) in the dense matrix to index of sparse matrix array a->a */
725: PetscInt *columns; /* lists the local columns of each color (using global column numbering) */
726: PetscInt brows; /* number of rows for efficient implementation of MatTransColoringApplyDenToSp() */
727: PetscInt *lstart; /* array used for loop over row blocks of Csparse */
728: };
730: /*
731: Null space context for preconditioner/operators
732: */
733: struct _p_MatNullSpace {
734: PETSCHEADER(int);
735: PetscBool has_cnst;
736: PetscInt n;
737: Vec *vecs;
738: PetscScalar *alpha; /* for projections */
739: PetscErrorCode (*remove)(MatNullSpace, Vec, void *); /* for user provided removal function */
740: void *rmctx; /* context for remove() function */
741: };
743: /*
744: Checking zero pivot for LU, ILU preconditioners.
745: */
746: typedef struct {
747: PetscInt nshift, nshift_max;
748: PetscReal shift_amount, shift_lo, shift_hi, shift_top, shift_fraction;
749: PetscBool newshift;
750: PetscReal rs; /* active row sum of abs(off-diagonals) */
751: PetscScalar pv; /* pivot of the active row */
752: } FactorShiftCtx;
754: PETSC_EXTERN PetscErrorCode MatTransposeCheckNonzeroState_Private(Mat, Mat);
756: /*
757: Used by MatTranspose() and potentially other functions to track the matrix used in the generation of another matrix
758: */
759: typedef struct {
760: PetscObjectId id;
761: PetscObjectState state;
762: PetscObjectState nonzerostate;
763: } MatParentState;
765: PETSC_EXTERN PetscErrorCode MatFactorDumpMatrix(Mat);
766: PETSC_INTERN PetscErrorCode MatSetBlockSizes_Default(Mat, PetscInt, PetscInt);
768: PETSC_SINGLE_LIBRARY_INTERN PetscErrorCode MatShift_Basic(Mat, PetscScalar);
770: static inline PetscErrorCode MatPivotCheck_nz(PETSC_UNUSED Mat mat, const MatFactorInfo *info, FactorShiftCtx *sctx, PETSC_UNUSED PetscInt row)
771: {
772: PetscReal _rs = sctx->rs;
773: PetscReal _zero = info->zeropivot * _rs;
775: PetscFunctionBegin;
776: if (PetscAbsScalar(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)) {
777: /* force |diag| > zeropivot*rs */
778: if (!sctx->nshift) sctx->shift_amount = info->shiftamount;
779: else sctx->shift_amount *= 2.0;
780: sctx->newshift = PETSC_TRUE;
781: (sctx->nshift)++;
782: } else {
783: sctx->newshift = PETSC_FALSE;
784: }
785: PetscFunctionReturn(PETSC_SUCCESS);
786: }
788: static inline PetscErrorCode MatPivotCheck_pd(PETSC_UNUSED Mat mat, const MatFactorInfo *info, FactorShiftCtx *sctx, PETSC_UNUSED PetscInt row)
789: {
790: PetscReal _rs = sctx->rs;
791: PetscReal _zero = info->zeropivot * _rs;
793: PetscFunctionBegin;
794: if (PetscRealPart(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)) {
795: /* force matfactor to be diagonally dominant */
796: if (sctx->nshift == sctx->nshift_max) {
797: sctx->shift_fraction = sctx->shift_hi;
798: } else {
799: sctx->shift_lo = sctx->shift_fraction;
800: sctx->shift_fraction = (sctx->shift_hi + sctx->shift_lo) / (PetscReal)2.;
801: }
802: sctx->shift_amount = sctx->shift_fraction * sctx->shift_top;
803: sctx->nshift++;
804: sctx->newshift = PETSC_TRUE;
805: } else {
806: sctx->newshift = PETSC_FALSE;
807: }
808: PetscFunctionReturn(PETSC_SUCCESS);
809: }
811: static inline PetscErrorCode MatPivotCheck_inblocks(PETSC_UNUSED Mat mat, const MatFactorInfo *info, FactorShiftCtx *sctx, PETSC_UNUSED PetscInt row)
812: {
813: PetscReal _zero = info->zeropivot;
815: PetscFunctionBegin;
816: if (PetscAbsScalar(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)) {
817: sctx->pv += info->shiftamount;
818: sctx->shift_amount = 0.0;
819: sctx->nshift++;
820: }
821: sctx->newshift = PETSC_FALSE;
822: PetscFunctionReturn(PETSC_SUCCESS);
823: }
825: static inline PetscErrorCode MatPivotCheck_none(Mat fact, Mat mat, const MatFactorInfo *info, FactorShiftCtx *sctx, PetscInt row)
826: {
827: PetscReal _zero = info->zeropivot;
829: PetscFunctionBegin;
830: sctx->newshift = PETSC_FALSE;
831: if (PetscAbsScalar(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)) {
832: PetscCheck(!mat->erroriffailure, PETSC_COMM_SELF, PETSC_ERR_MAT_LU_ZRPVT, "Zero pivot row %" PetscInt_FMT " value %g tolerance %g", row, (double)PetscAbsScalar(sctx->pv), (double)_zero);
833: PetscCall(PetscInfo(mat, "Detected zero pivot in factorization in row %" PetscInt_FMT " value %g tolerance %g\n", row, (double)PetscAbsScalar(sctx->pv), (double)_zero));
834: fact->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT;
835: fact->factorerror_zeropivot_value = PetscAbsScalar(sctx->pv);
836: fact->factorerror_zeropivot_row = row;
837: }
838: PetscFunctionReturn(PETSC_SUCCESS);
839: }
841: static inline PetscErrorCode MatPivotCheck(Mat fact, Mat mat, const MatFactorInfo *info, FactorShiftCtx *sctx, PetscInt row)
842: {
843: PetscFunctionBegin;
844: if (info->shifttype == (PetscReal)MAT_SHIFT_NONZERO) PetscCall(MatPivotCheck_nz(mat, info, sctx, row));
845: else if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) PetscCall(MatPivotCheck_pd(mat, info, sctx, row));
846: else if (info->shifttype == (PetscReal)MAT_SHIFT_INBLOCKS) PetscCall(MatPivotCheck_inblocks(mat, info, sctx, row));
847: else PetscCall(MatPivotCheck_none(fact, mat, info, sctx, row));
848: PetscFunctionReturn(PETSC_SUCCESS);
849: }
851: #include <petscbt.h>
852: /*
853: Create and initialize a linked list
854: Input Parameters:
855: idx_start - starting index of the list
856: lnk_max - max value of lnk indicating the end of the list
857: nlnk - max length of the list
858: Output Parameters:
859: lnk - list initialized
860: bt - PetscBT (bitarray) with all bits set to false
861: lnk_empty - flg indicating the list is empty
862: */
863: #define PetscLLCreate(idx_start, lnk_max, nlnk, lnk, bt) ((PetscErrorCode)(PetscMalloc1(nlnk, &lnk) || PetscBTCreate(nlnk, &(bt)) || (lnk[idx_start] = lnk_max, PETSC_SUCCESS)))
865: #define PetscLLCreate_new(idx_start, lnk_max, nlnk, lnk, bt, lnk_empty) ((PetscErrorCode)(PetscMalloc1(nlnk, &lnk) || PetscBTCreate(nlnk, &(bt)) || (lnk_empty = PETSC_TRUE, 0) || (lnk[idx_start] = lnk_max, PETSC_SUCCESS)))
867: static inline PetscErrorCode PetscLLInsertLocation_Private(PetscBool assume_sorted, PetscInt k, PetscInt idx_start, PetscInt entry, PetscInt *PETSC_RESTRICT nlnk, PetscInt *PETSC_RESTRICT lnkdata, PetscInt *PETSC_RESTRICT lnk)
868: {
869: PetscInt location;
871: PetscFunctionBegin;
872: /* start from the beginning if entry < previous entry */
873: if (!assume_sorted && k && entry < *lnkdata) *lnkdata = idx_start;
874: /* search for insertion location */
875: do {
876: location = *lnkdata;
877: *lnkdata = lnk[location];
878: } while (entry > *lnkdata);
879: /* insertion location is found, add entry into lnk */
880: lnk[location] = entry;
881: lnk[entry] = *lnkdata;
882: ++(*nlnk);
883: *lnkdata = entry; /* next search starts from here if next_entry > entry */
884: PetscFunctionReturn(PETSC_SUCCESS);
885: }
887: static inline PetscErrorCode PetscLLAdd_Private(PetscInt nidx, const PetscInt *PETSC_RESTRICT indices, PetscInt idx_start, PetscInt *PETSC_RESTRICT nlnk, PetscInt *PETSC_RESTRICT lnk, PetscBT bt, PetscBool assume_sorted)
888: {
889: PetscFunctionBegin;
890: *nlnk = 0;
891: for (PetscInt k = 0, lnkdata = idx_start; k < nidx; ++k) {
892: const PetscInt entry = indices[k];
894: if (!PetscBTLookupSet(bt, entry)) PetscCall(PetscLLInsertLocation_Private(assume_sorted, k, idx_start, entry, nlnk, &lnkdata, lnk));
895: }
896: PetscFunctionReturn(PETSC_SUCCESS);
897: }
899: /*
900: Add an index set into a sorted linked list
901: Input Parameters:
902: nidx - number of input indices
903: indices - integer array
904: idx_start - starting index of the list
905: lnk - linked list(an integer array) that is created
906: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
907: output Parameters:
908: nlnk - number of newly added indices
909: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
910: bt - updated PetscBT (bitarray)
911: */
912: static inline PetscErrorCode PetscLLAdd(PetscInt nidx, const PetscInt *PETSC_RESTRICT indices, PetscInt idx_start, PetscInt *PETSC_RESTRICT nlnk, PetscInt *PETSC_RESTRICT lnk, PetscBT bt)
913: {
914: PetscFunctionBegin;
915: PetscCall(PetscLLAdd_Private(nidx, indices, idx_start, nlnk, lnk, bt, PETSC_FALSE));
916: PetscFunctionReturn(PETSC_SUCCESS);
917: }
919: /*
920: Add a SORTED ascending index set into a sorted linked list - same as PetscLLAdd() bus skip 'if (_k && _entry < _lnkdata) _lnkdata = idx_start;'
921: Input Parameters:
922: nidx - number of input indices
923: indices - sorted integer array
924: idx_start - starting index of the list
925: lnk - linked list(an integer array) that is created
926: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
927: output Parameters:
928: nlnk - number of newly added indices
929: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
930: bt - updated PetscBT (bitarray)
931: */
932: static inline PetscErrorCode PetscLLAddSorted(PetscInt nidx, const PetscInt *PETSC_RESTRICT indices, PetscInt idx_start, PetscInt *PETSC_RESTRICT nlnk, PetscInt *PETSC_RESTRICT lnk, PetscBT bt)
933: {
934: PetscFunctionBegin;
935: PetscCall(PetscLLAdd_Private(nidx, indices, idx_start, nlnk, lnk, bt, PETSC_TRUE));
936: PetscFunctionReturn(PETSC_SUCCESS);
937: }
939: /*
940: Add a permuted index set into a sorted linked list
941: Input Parameters:
942: nidx - number of input indices
943: indices - integer array
944: perm - permutation of indices
945: idx_start - starting index of the list
946: lnk - linked list(an integer array) that is created
947: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
948: output Parameters:
949: nlnk - number of newly added indices
950: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
951: bt - updated PetscBT (bitarray)
952: */
953: static inline PetscErrorCode PetscLLAddPerm(PetscInt nidx, const PetscInt *PETSC_RESTRICT indices, const PetscInt *PETSC_RESTRICT perm, PetscInt idx_start, PetscInt *PETSC_RESTRICT nlnk, PetscInt *PETSC_RESTRICT lnk, PetscBT bt)
954: {
955: PetscFunctionBegin;
956: *nlnk = 0;
957: for (PetscInt k = 0, lnkdata = idx_start; k < nidx; ++k) {
958: const PetscInt entry = perm[indices[k]];
960: if (!PetscBTLookupSet(bt, entry)) PetscCall(PetscLLInsertLocation_Private(PETSC_FALSE, k, idx_start, entry, nlnk, &lnkdata, lnk));
961: }
962: PetscFunctionReturn(PETSC_SUCCESS);
963: }
965: #if 0
966: /* this appears to be unused? */
967: static inline PetscErrorCode PetscLLAddSorted_new(PetscInt nidx, PetscInt *indices, PetscInt idx_start, PetscBool *lnk_empty, PetscInt *nlnk, PetscInt *lnk, PetscBT bt)
968: {
969: PetscInt lnkdata = idx_start;
971: PetscFunctionBegin;
972: if (*lnk_empty) {
973: for (PetscInt k = 0; k < nidx; ++k) {
974: const PetscInt entry = indices[k], location = lnkdata;
976: PetscCall(PetscBTSet(bt,entry)); /* mark the new entry */
977: lnkdata = lnk[location];
978: /* insertion location is found, add entry into lnk */
979: lnk[location] = entry;
980: lnk[entry] = lnkdata;
981: lnkdata = entry; /* next search starts from here */
982: }
983: /* lnk[indices[nidx-1]] = lnk[idx_start];
984: lnk[idx_start] = indices[0];
985: PetscCall(PetscBTSet(bt,indices[0]));
986: for (_k=1; _k<nidx; _k++) {
987: PetscCall(PetscBTSet(bt,indices[_k]));
988: lnk[indices[_k-1]] = indices[_k];
989: }
990: */
991: *nlnk = nidx;
992: *lnk_empty = PETSC_FALSE;
993: } else {
994: *nlnk = 0;
995: for (PetscInt k = 0; k < nidx; ++k) {
996: const PetscInt entry = indices[k];
998: if (!PetscBTLookupSet(bt,entry)) PetscCall(PetscLLInsertLocation_Private(PETSC_TRUE,k,idx_start,entry,nlnk,&lnkdata,lnk));
999: }
1000: }
1001: PetscFunctionReturn(PETSC_SUCCESS);
1002: }
1003: #endif
1005: /*
1006: Add a SORTED index set into a sorted linked list used for LUFactorSymbolic()
1007: Same as PetscLLAddSorted() with an additional operation:
1008: count the number of input indices that are no larger than 'diag'
1009: Input Parameters:
1010: indices - sorted integer array
1011: idx_start - starting index of the list, index of pivot row
1012: lnk - linked list(an integer array) that is created
1013: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1014: diag - index of the active row in LUFactorSymbolic
1015: nzbd - number of input indices with indices <= idx_start
1016: im - im[idx_start] is initialized as num of nonzero entries in row=idx_start
1017: output Parameters:
1018: nlnk - number of newly added indices
1019: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
1020: bt - updated PetscBT (bitarray)
1021: im - im[idx_start]: unchanged if diag is not an entry
1022: : num of entries with indices <= diag if diag is an entry
1023: */
1024: static inline PetscErrorCode PetscLLAddSortedLU(const PetscInt *PETSC_RESTRICT indices, PetscInt idx_start, PetscInt *PETSC_RESTRICT nlnk, PetscInt *PETSC_RESTRICT lnk, PetscBT bt, PetscInt diag, PetscInt nzbd, PetscInt *PETSC_RESTRICT im)
1025: {
1026: const PetscInt nidx = im[idx_start] - nzbd; /* num of entries with idx_start < index <= diag */
1028: PetscFunctionBegin;
1029: *nlnk = 0;
1030: for (PetscInt k = 0, lnkdata = idx_start; k < nidx; ++k) {
1031: const PetscInt entry = indices[k];
1033: ++nzbd;
1034: if (entry == diag) im[idx_start] = nzbd;
1035: if (!PetscBTLookupSet(bt, entry)) PetscCall(PetscLLInsertLocation_Private(PETSC_TRUE, k, idx_start, entry, nlnk, &lnkdata, lnk));
1036: }
1037: PetscFunctionReturn(PETSC_SUCCESS);
1038: }
1040: /*
1041: Copy data on the list into an array, then initialize the list
1042: Input Parameters:
1043: idx_start - starting index of the list
1044: lnk_max - max value of lnk indicating the end of the list
1045: nlnk - number of data on the list to be copied
1046: lnk - linked list
1047: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1048: output Parameters:
1049: indices - array that contains the copied data
1050: lnk - linked list that is cleaned and initialize
1051: bt - PetscBT (bitarray) with all bits set to false
1052: */
1053: static inline PetscErrorCode PetscLLClean(PetscInt idx_start, PetscInt lnk_max, PetscInt nlnk, PetscInt *PETSC_RESTRICT lnk, PetscInt *PETSC_RESTRICT indices, PetscBT bt)
1054: {
1055: PetscFunctionBegin;
1056: for (PetscInt j = 0, idx = idx_start; j < nlnk; ++j) {
1057: idx = lnk[idx];
1058: indices[j] = idx;
1059: PetscCall(PetscBTClear(bt, idx));
1060: }
1061: lnk[idx_start] = lnk_max;
1062: PetscFunctionReturn(PETSC_SUCCESS);
1063: }
1065: /*
1066: Free memories used by the list
1067: */
1068: #define PetscLLDestroy(lnk, bt) ((PetscErrorCode)(PetscFree(lnk) || PetscBTDestroy(&(bt))))
1070: /* Routines below are used for incomplete matrix factorization */
1071: /*
1072: Create and initialize a linked list and its levels
1073: Input Parameters:
1074: idx_start - starting index of the list
1075: lnk_max - max value of lnk indicating the end of the list
1076: nlnk - max length of the list
1077: Output Parameters:
1078: lnk - list initialized
1079: lnk_lvl - array of size nlnk for storing levels of lnk
1080: bt - PetscBT (bitarray) with all bits set to false
1081: */
1082: #define PetscIncompleteLLCreate(idx_start, lnk_max, nlnk, lnk, lnk_lvl, bt) \
1083: ((PetscErrorCode)(PetscIntMultError(2, nlnk, NULL) || PetscMalloc1(2 * nlnk, &lnk) || PetscBTCreate(nlnk, &(bt)) || (lnk[idx_start] = lnk_max, lnk_lvl = lnk + nlnk, PETSC_SUCCESS)))
1085: static inline PetscErrorCode PetscIncompleteLLInsertLocation_Private(PetscBool assume_sorted, PetscInt k, PetscInt idx_start, PetscInt entry, PetscInt *PETSC_RESTRICT nlnk, PetscInt *PETSC_RESTRICT lnkdata, PetscInt *PETSC_RESTRICT lnk, PetscInt *PETSC_RESTRICT lnklvl, PetscInt newval)
1086: {
1087: PetscFunctionBegin;
1088: PetscCall(PetscLLInsertLocation_Private(assume_sorted, k, idx_start, entry, nlnk, lnkdata, lnk));
1089: lnklvl[entry] = newval;
1090: PetscFunctionReturn(PETSC_SUCCESS);
1091: }
1093: /*
1094: Initialize a sorted linked list used for ILU and ICC
1095: Input Parameters:
1096: nidx - number of input idx
1097: idx - integer array used for storing column indices
1098: idx_start - starting index of the list
1099: perm - indices of an IS
1100: lnk - linked list(an integer array) that is created
1101: lnklvl - levels of lnk
1102: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1103: output Parameters:
1104: nlnk - number of newly added idx
1105: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1106: lnklvl - levels of lnk
1107: bt - updated PetscBT (bitarray)
1108: */
1109: static inline PetscErrorCode PetscIncompleteLLInit(PetscInt nidx, const PetscInt *PETSC_RESTRICT idx, PetscInt idx_start, const PetscInt *PETSC_RESTRICT perm, PetscInt *PETSC_RESTRICT nlnk, PetscInt *PETSC_RESTRICT lnk, PetscInt *PETSC_RESTRICT lnklvl, PetscBT bt)
1110: {
1111: PetscFunctionBegin;
1112: *nlnk = 0;
1113: for (PetscInt k = 0, lnkdata = idx_start; k < nidx; ++k) {
1114: const PetscInt entry = perm[idx[k]];
1116: if (!PetscBTLookupSet(bt, entry)) PetscCall(PetscIncompleteLLInsertLocation_Private(PETSC_FALSE, k, idx_start, entry, nlnk, &lnkdata, lnk, lnklvl, 0));
1117: }
1118: PetscFunctionReturn(PETSC_SUCCESS);
1119: }
1121: static inline PetscErrorCode PetscIncompleteLLAdd_Private(PetscInt nidx, const PetscInt *PETSC_RESTRICT idx, PetscReal level, const PetscInt *PETSC_RESTRICT idxlvl, PetscInt idx_start, PetscInt *PETSC_RESTRICT nlnk, PetscInt *PETSC_RESTRICT lnk, PetscInt *PETSC_RESTRICT lnklvl, PetscBT bt, PetscInt prow_offset, PetscBool assume_sorted)
1122: {
1123: PetscFunctionBegin;
1124: *nlnk = 0;
1125: for (PetscInt k = 0, lnkdata = idx_start; k < nidx; ++k) {
1126: const PetscInt incrlev = idxlvl[k] + prow_offset + 1;
1128: if (incrlev <= level) {
1129: const PetscInt entry = idx[k];
1131: if (!PetscBTLookupSet(bt, entry)) PetscCall(PetscIncompleteLLInsertLocation_Private(assume_sorted, k, idx_start, entry, nlnk, &lnkdata, lnk, lnklvl, incrlev));
1132: else if (lnklvl[entry] > incrlev) lnklvl[entry] = incrlev; /* existing entry */
1133: }
1134: }
1135: PetscFunctionReturn(PETSC_SUCCESS);
1136: }
1138: /*
1139: Add a SORTED index set into a sorted linked list for ICC
1140: Input Parameters:
1141: nidx - number of input indices
1142: idx - sorted integer array used for storing column indices
1143: level - level of fill, e.g., ICC(level)
1144: idxlvl - level of idx
1145: idx_start - starting index of the list
1146: lnk - linked list(an integer array) that is created
1147: lnklvl - levels of lnk
1148: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1149: idxlvl_prow - idxlvl[prow], where prow is the row number of the idx
1150: output Parameters:
1151: nlnk - number of newly added indices
1152: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1153: lnklvl - levels of lnk
1154: bt - updated PetscBT (bitarray)
1155: Note: the level of U(i,j) is set as lvl(i,j) = min{ lvl(i,j), lvl(prow,i)+lvl(prow,j)+1)
1156: where idx = non-zero columns of U(prow,prow+1:n-1), prow<i
1157: */
1158: static inline PetscErrorCode PetscICCLLAddSorted(PetscInt nidx, const PetscInt *PETSC_RESTRICT idx, PetscReal level, const PetscInt *PETSC_RESTRICT idxlvl, PetscInt idx_start, PetscInt *PETSC_RESTRICT nlnk, PetscInt *PETSC_RESTRICT lnk, PetscInt *PETSC_RESTRICT lnklvl, PetscBT bt, PetscInt idxlvl_prow)
1159: {
1160: PetscFunctionBegin;
1161: PetscCall(PetscIncompleteLLAdd_Private(nidx, idx, level, idxlvl, idx_start, nlnk, lnk, lnklvl, bt, idxlvl_prow, PETSC_TRUE));
1162: PetscFunctionReturn(PETSC_SUCCESS);
1163: }
1165: /*
1166: Add a SORTED index set into a sorted linked list for ILU
1167: Input Parameters:
1168: nidx - number of input indices
1169: idx - sorted integer array used for storing column indices
1170: level - level of fill, e.g., ICC(level)
1171: idxlvl - level of idx
1172: idx_start - starting index of the list
1173: lnk - linked list(an integer array) that is created
1174: lnklvl - levels of lnk
1175: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1176: prow - the row number of idx
1177: output Parameters:
1178: nlnk - number of newly added idx
1179: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1180: lnklvl - levels of lnk
1181: bt - updated PetscBT (bitarray)
1183: Note: the level of factor(i,j) is set as lvl(i,j) = min{ lvl(i,j), lvl(i,prow)+lvl(prow,j)+1)
1184: where idx = non-zero columns of U(prow,prow+1:n-1), prow<i
1185: */
1186: static inline PetscErrorCode PetscILULLAddSorted(PetscInt nidx, const PetscInt *PETSC_RESTRICT idx, PetscInt level, const PetscInt *PETSC_RESTRICT idxlvl, PetscInt idx_start, PetscInt *PETSC_RESTRICT nlnk, PetscInt *PETSC_RESTRICT lnk, PetscInt *PETSC_RESTRICT lnklvl, PetscBT bt, PetscInt prow)
1187: {
1188: PetscFunctionBegin;
1189: PetscCall(PetscIncompleteLLAdd_Private(nidx, idx, level, idxlvl, idx_start, nlnk, lnk, lnklvl, bt, lnklvl[prow], PETSC_TRUE));
1190: PetscFunctionReturn(PETSC_SUCCESS);
1191: }
1193: /*
1194: Add a index set into a sorted linked list
1195: Input Parameters:
1196: nidx - number of input idx
1197: idx - integer array used for storing column indices
1198: level - level of fill, e.g., ICC(level)
1199: idxlvl - level of idx
1200: idx_start - starting index of the list
1201: lnk - linked list(an integer array) that is created
1202: lnklvl - levels of lnk
1203: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1204: output Parameters:
1205: nlnk - number of newly added idx
1206: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1207: lnklvl - levels of lnk
1208: bt - updated PetscBT (bitarray)
1209: */
1210: static inline PetscErrorCode PetscIncompleteLLAdd(PetscInt nidx, const PetscInt *PETSC_RESTRICT idx, PetscReal level, const PetscInt *PETSC_RESTRICT idxlvl, PetscInt idx_start, PetscInt *PETSC_RESTRICT nlnk, PetscInt *PETSC_RESTRICT lnk, PetscInt *PETSC_RESTRICT lnklvl, PetscBT bt)
1211: {
1212: PetscFunctionBegin;
1213: PetscCall(PetscIncompleteLLAdd_Private(nidx, idx, level, idxlvl, idx_start, nlnk, lnk, lnklvl, bt, 0, PETSC_FALSE));
1214: PetscFunctionReturn(PETSC_SUCCESS);
1215: }
1217: /*
1218: Add a SORTED index set into a sorted linked list
1219: Input Parameters:
1220: nidx - number of input indices
1221: idx - sorted integer array used for storing column indices
1222: level - level of fill, e.g., ICC(level)
1223: idxlvl - level of idx
1224: idx_start - starting index of the list
1225: lnk - linked list(an integer array) that is created
1226: lnklvl - levels of lnk
1227: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1228: output Parameters:
1229: nlnk - number of newly added idx
1230: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1231: lnklvl - levels of lnk
1232: bt - updated PetscBT (bitarray)
1233: */
1234: static inline PetscErrorCode PetscIncompleteLLAddSorted(PetscInt nidx, const PetscInt *PETSC_RESTRICT idx, PetscReal level, const PetscInt *PETSC_RESTRICT idxlvl, PetscInt idx_start, PetscInt *PETSC_RESTRICT nlnk, PetscInt *PETSC_RESTRICT lnk, PetscInt *PETSC_RESTRICT lnklvl, PetscBT bt)
1235: {
1236: PetscFunctionBegin;
1237: PetscCall(PetscIncompleteLLAdd_Private(nidx, idx, level, idxlvl, idx_start, nlnk, lnk, lnklvl, bt, 0, PETSC_TRUE));
1238: PetscFunctionReturn(PETSC_SUCCESS);
1239: }
1241: /*
1242: Copy data on the list into an array, then initialize the list
1243: Input Parameters:
1244: idx_start - starting index of the list
1245: lnk_max - max value of lnk indicating the end of the list
1246: nlnk - number of data on the list to be copied
1247: lnk - linked list
1248: lnklvl - level of lnk
1249: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1250: output Parameters:
1251: indices - array that contains the copied data
1252: lnk - linked list that is cleaned and initialize
1253: lnklvl - level of lnk that is reinitialized
1254: bt - PetscBT (bitarray) with all bits set to false
1255: */
1256: static inline PetscErrorCode PetscIncompleteLLClean(PetscInt idx_start, PetscInt lnk_max, PetscInt nlnk, PetscInt *PETSC_RESTRICT lnk, PetscInt *PETSC_RESTRICT lnklvl, PetscInt *PETSC_RESTRICT indices, PetscInt *PETSC_RESTRICT indiceslvl, PetscBT bt)
1257: {
1258: PetscFunctionBegin;
1259: for (PetscInt j = 0, idx = idx_start; j < nlnk; ++j) {
1260: idx = lnk[idx];
1261: indices[j] = idx;
1262: indiceslvl[j] = lnklvl[idx];
1263: lnklvl[idx] = -1;
1264: PetscCall(PetscBTClear(bt, idx));
1265: }
1266: lnk[idx_start] = lnk_max;
1267: PetscFunctionReturn(PETSC_SUCCESS);
1268: }
1270: /*
1271: Free memories used by the list
1272: */
1273: #define PetscIncompleteLLDestroy(lnk, bt) ((PetscErrorCode)(PetscFree(lnk) || PetscBTDestroy(&(bt))))
1275: #if !defined(PETSC_CLANG_STATIC_ANALYZER)
1276: #define MatCheckSameLocalSize(A, ar1, B, ar2) \
1277: do { \
1278: PetscCheckSameComm(A, ar1, B, ar2); \
1279: PetscCheck(((A)->rmap->n == (B)->rmap->n) && ((A)->cmap->n == (B)->cmap->n), PETSC_COMM_SELF, PETSC_ERR_ARG_INCOMP, "Incompatible matrix local sizes: parameter # %d (%" PetscInt_FMT " x %" PetscInt_FMT ") != parameter # %d (%" PetscInt_FMT " x %" PetscInt_FMT ")", ar1, \
1280: (A)->rmap->n, (A)->cmap->n, ar2, (B)->rmap->n, (B)->cmap->n); \
1281: } while (0)
1282: #define MatCheckSameSize(A, ar1, B, ar2) \
1283: do { \
1284: PetscCheck(((A)->rmap->N == (B)->rmap->N) && ((A)->cmap->N == (B)->cmap->N), PetscObjectComm((PetscObject)(A)), PETSC_ERR_ARG_INCOMP, "Incompatible matrix global sizes: parameter # %d (%" PetscInt_FMT " x %" PetscInt_FMT ") != parameter # %d (%" PetscInt_FMT " x %" PetscInt_FMT ")", ar1, \
1285: (A)->rmap->N, (A)->cmap->N, ar2, (B)->rmap->N, (B)->cmap->N); \
1286: MatCheckSameLocalSize(A, ar1, B, ar2); \
1287: } while (0)
1288: #else
1289: template <typename Tm>
1290: extern void MatCheckSameLocalSize(Tm, int, Tm, int);
1291: template <typename Tm>
1292: extern void MatCheckSameSize(Tm, int, Tm, int);
1293: #endif
1295: #define VecCheckMatCompatible(M, x, ar1, b, ar2) \
1296: do { \
1297: PetscCheck((M)->cmap->N == (x)->map->N, PetscObjectComm((PetscObject)(M)), PETSC_ERR_ARG_SIZ, "Vector global length incompatible with matrix: parameter # %d global size %" PetscInt_FMT " != matrix column global size %" PetscInt_FMT, ar1, (x)->map->N, \
1298: (M)->cmap->N); \
1299: PetscCheck((M)->rmap->N == (b)->map->N, PetscObjectComm((PetscObject)(M)), PETSC_ERR_ARG_SIZ, "Vector global length incompatible with matrix: parameter # %d global size %" PetscInt_FMT " != matrix row global size %" PetscInt_FMT, ar2, (b)->map->N, \
1300: (M)->rmap->N); \
1301: } while (0)
1303: /* -------------------------------------------------------------------------------------------------------*/
1304: /*
1305: Create and initialize a condensed linked list -
1306: same as PetscLLCreate(), but uses a scalable array 'lnk' with size of max number of entries, not O(N).
1307: Barry suggested this approach (Dec. 6, 2011):
1308: I've thought of an alternative way of representing a linked list that is efficient but doesn't have the O(N) scaling issue
1309: (it may be faster than the O(N) even sequentially due to less crazy memory access).
1311: Instead of having some like a 2 -> 4 -> 11 -> 22 list that uses slot 2 4 11 and 22 in a big array use a small array with two slots
1312: for each entry for example [ 2 1 | 4 3 | 22 -1 | 11 2] so the first number (of the pair) is the value while the second tells you where
1313: in the list the next entry is. Inserting a new link means just append another pair at the end. For example say we want to insert 13 into the
1314: list it would then become [2 1 | 4 3 | 22 -1 | 11 4 | 13 2 ] you just add a pair at the end and fix the point for the one that points to it.
1315: That is 11 use to point to the 2 slot, after the change 11 points to the 4th slot which has the value 13. Note that values are always next
1316: to each other so memory access is much better than using the big array.
1318: Example:
1319: nlnk_max=5, lnk_max=36:
1320: Initial list: [0, 0 | 36, 2 | 0, 0 | 0, 0 | 0, 0 | 0, 0 | 0, 0]
1321: here, head_node has index 2 with value lnk[2]=lnk_max=36,
1322: 0-th entry is used to store the number of entries in the list,
1323: The initial lnk represents head -> tail(marked by 36) with number of entries = lnk[0]=0.
1325: Now adding a sorted set {2,4}, the list becomes
1326: [2, 0 | 36, 4 |2, 6 | 4, 2 | 0, 0 | 0, 0 | 0, 0 ]
1327: represents head -> 2 -> 4 -> tail with number of entries = lnk[0]=2.
1329: Then adding a sorted set {0,3,35}, the list
1330: [5, 0 | 36, 8 | 2, 10 | 4, 12 | 0, 4 | 3, 6 | 35, 2 ]
1331: represents head -> 0 -> 2 -> 3 -> 4 -> 35 -> tail with number of entries = lnk[0]=5.
1333: Input Parameters:
1334: nlnk_max - max length of the list
1335: lnk_max - max value of the entries
1336: Output Parameters:
1337: lnk - list created and initialized
1338: bt - PetscBT (bitarray) with all bits set to false. Note: bt has size lnk_max, not nln_max!
1339: */
1340: static inline PetscErrorCode PetscLLCondensedCreate(PetscInt nlnk_max, PetscInt lnk_max, PetscInt **lnk, PetscBT *bt)
1341: {
1342: PetscInt *llnk, lsize = 0;
1344: PetscFunctionBegin;
1345: PetscCall(PetscIntMultError(2, nlnk_max + 2, &lsize));
1346: PetscCall(PetscMalloc1(lsize, lnk));
1347: PetscCall(PetscBTCreate(lnk_max, bt));
1348: llnk = *lnk;
1349: llnk[0] = 0; /* number of entries on the list */
1350: llnk[2] = lnk_max; /* value in the head node */
1351: llnk[3] = 2; /* next for the head node */
1352: PetscFunctionReturn(PETSC_SUCCESS);
1353: }
1355: /*
1356: Add a SORTED ascending index set into a sorted linked list. See PetscLLCondensedCreate() for detailed description.
1357: Input Parameters:
1358: nidx - number of input indices
1359: indices - sorted integer array
1360: lnk - condensed linked list(an integer array) that is created
1361: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1362: output Parameters:
1363: lnk - the sorted(increasing order) linked list containing previous and newly added non-redundate indices
1364: bt - updated PetscBT (bitarray)
1365: */
1366: static inline PetscErrorCode PetscLLCondensedAddSorted(PetscInt nidx, const PetscInt indices[], PetscInt lnk[], PetscBT bt)
1367: {
1368: PetscInt location = 2; /* head */
1369: PetscInt nlnk = lnk[0]; /* num of entries on the input lnk */
1371: PetscFunctionBegin;
1372: for (PetscInt k = 0; k < nidx; k++) {
1373: const PetscInt entry = indices[k];
1374: if (!PetscBTLookupSet(bt, entry)) { /* new entry */
1375: PetscInt next, lnkdata;
1377: /* search for insertion location */
1378: do {
1379: next = location + 1; /* link from previous node to next node */
1380: location = lnk[next]; /* idx of next node */
1381: lnkdata = lnk[location]; /* value of next node */
1382: } while (entry > lnkdata);
1383: /* insertion location is found, add entry into lnk */
1384: const PetscInt newnode = 2 * (nlnk + 2); /* index for this new node */
1385: lnk[next] = newnode; /* connect previous node to the new node */
1386: lnk[newnode] = entry; /* set value of the new node */
1387: lnk[newnode + 1] = location; /* connect new node to next node */
1388: location = newnode; /* next search starts from the new node */
1389: nlnk++;
1390: }
1391: }
1392: lnk[0] = nlnk; /* number of entries in the list */
1393: PetscFunctionReturn(PETSC_SUCCESS);
1394: }
1396: static inline PetscErrorCode PetscLLCondensedClean(PetscInt lnk_max, PETSC_UNUSED PetscInt nidx, PetscInt *indices, PetscInt lnk[], PetscBT bt)
1397: {
1398: const PetscInt nlnk = lnk[0]; /* num of entries on the list */
1399: PetscInt next = lnk[3]; /* head node */
1401: PetscFunctionBegin;
1402: for (PetscInt k = 0; k < nlnk; k++) {
1403: indices[k] = lnk[next];
1404: next = lnk[next + 1];
1405: PetscCall(PetscBTClear(bt, indices[k]));
1406: }
1407: lnk[0] = 0; /* num of entries on the list */
1408: lnk[2] = lnk_max; /* initialize head node */
1409: lnk[3] = 2; /* head node */
1410: PetscFunctionReturn(PETSC_SUCCESS);
1411: }
1413: static inline PetscErrorCode PetscLLCondensedView(PetscInt *lnk)
1414: {
1415: PetscFunctionBegin;
1416: PetscCall(PetscPrintf(PETSC_COMM_SELF, "LLCondensed of size %" PetscInt_FMT ", (val, next)\n", lnk[0]));
1417: for (PetscInt k = 2; k < lnk[0] + 2; ++k) PetscCall(PetscPrintf(PETSC_COMM_SELF, " %" PetscInt_FMT ": (%" PetscInt_FMT ", %" PetscInt_FMT ")\n", 2 * k, lnk[2 * k], lnk[2 * k + 1]));
1418: PetscFunctionReturn(PETSC_SUCCESS);
1419: }
1421: /*
1422: Free memories used by the list
1423: */
1424: static inline PetscErrorCode PetscLLCondensedDestroy(PetscInt *lnk, PetscBT bt)
1425: {
1426: PetscFunctionBegin;
1427: PetscCall(PetscFree(lnk));
1428: PetscCall(PetscBTDestroy(&bt));
1429: PetscFunctionReturn(PETSC_SUCCESS);
1430: }
1432: /* -------------------------------------------------------------------------------------------------------*/
1433: /*
1434: Same as PetscLLCondensedCreate(), but does not use non-scalable O(lnk_max) bitarray
1435: Input Parameters:
1436: nlnk_max - max length of the list
1437: Output Parameters:
1438: lnk - list created and initialized
1439: */
1440: static inline PetscErrorCode PetscLLCondensedCreate_Scalable(PetscInt nlnk_max, PetscInt **lnk)
1441: {
1442: PetscInt *llnk, lsize = 0;
1444: PetscFunctionBegin;
1445: PetscCall(PetscIntMultError(2, nlnk_max + 2, &lsize));
1446: PetscCall(PetscMalloc1(lsize, lnk));
1447: llnk = *lnk;
1448: llnk[0] = 0; /* number of entries on the list */
1449: llnk[2] = PETSC_INT_MAX; /* value in the head node */
1450: llnk[3] = 2; /* next for the head node */
1451: PetscFunctionReturn(PETSC_SUCCESS);
1452: }
1454: static inline PetscErrorCode PetscLLCondensedExpand_Scalable(PetscInt nlnk_max, PetscInt **lnk)
1455: {
1456: PetscInt lsize = 0;
1458: PetscFunctionBegin;
1459: PetscCall(PetscIntMultError(2, nlnk_max + 2, &lsize));
1460: PetscCall(PetscRealloc(lsize * sizeof(PetscInt), lnk));
1461: PetscFunctionReturn(PETSC_SUCCESS);
1462: }
1464: static inline PetscErrorCode PetscLLCondensedAddSorted_Scalable(PetscInt nidx, const PetscInt indices[], PetscInt lnk[])
1465: {
1466: PetscInt location = 2; /* head */
1467: PetscInt nlnk = lnk[0]; /* num of entries on the input lnk */
1469: for (PetscInt k = 0; k < nidx; k++) {
1470: const PetscInt entry = indices[k];
1471: PetscInt next, lnkdata;
1473: /* search for insertion location */
1474: do {
1475: next = location + 1; /* link from previous node to next node */
1476: location = lnk[next]; /* idx of next node */
1477: lnkdata = lnk[location]; /* value of next node */
1478: } while (entry > lnkdata);
1479: if (entry < lnkdata) {
1480: /* insertion location is found, add entry into lnk */
1481: const PetscInt newnode = 2 * (nlnk + 2); /* index for this new node */
1482: lnk[next] = newnode; /* connect previous node to the new node */
1483: lnk[newnode] = entry; /* set value of the new node */
1484: lnk[newnode + 1] = location; /* connect new node to next node */
1485: location = newnode; /* next search starts from the new node */
1486: nlnk++;
1487: }
1488: }
1489: lnk[0] = nlnk; /* number of entries in the list */
1490: return PETSC_SUCCESS;
1491: }
1493: static inline PetscErrorCode PetscLLCondensedClean_Scalable(PETSC_UNUSED PetscInt nidx, PetscInt *indices, PetscInt *lnk)
1494: {
1495: const PetscInt nlnk = lnk[0];
1496: PetscInt next = lnk[3]; /* head node */
1498: for (PetscInt k = 0; k < nlnk; k++) {
1499: indices[k] = lnk[next];
1500: next = lnk[next + 1];
1501: }
1502: lnk[0] = 0; /* num of entries on the list */
1503: lnk[3] = 2; /* head node */
1504: return PETSC_SUCCESS;
1505: }
1507: static inline PetscErrorCode PetscLLCondensedDestroy_Scalable(PetscInt *lnk)
1508: {
1509: return PetscFree(lnk);
1510: }
1512: /* -------------------------------------------------------------------------------------------------------*/
1513: /*
1514: lnk[0] number of links
1515: lnk[1] number of entries
1516: lnk[3n] value
1517: lnk[3n+1] len
1518: lnk[3n+2] link to next value
1520: The next three are always the first link
1522: lnk[3] PETSC_INT_MIN+1
1523: lnk[4] 1
1524: lnk[5] link to first real entry
1526: The next three are always the last link
1528: lnk[6] PETSC_INT_MAX - 1
1529: lnk[7] 1
1530: lnk[8] next valid link (this is the same as lnk[0] but without the decreases)
1531: */
1533: static inline PetscErrorCode PetscLLCondensedCreate_fast(PetscInt nlnk_max, PetscInt **lnk)
1534: {
1535: PetscInt *llnk;
1536: PetscInt lsize = 0;
1538: PetscFunctionBegin;
1539: PetscCall(PetscIntMultError(3, nlnk_max + 3, &lsize));
1540: PetscCall(PetscMalloc1(lsize, lnk));
1541: llnk = *lnk;
1542: llnk[0] = 0; /* nlnk: number of entries on the list */
1543: llnk[1] = 0; /* number of integer entries represented in list */
1544: llnk[3] = PETSC_INT_MIN + 1; /* value in the first node */
1545: llnk[4] = 1; /* count for the first node */
1546: llnk[5] = 6; /* next for the first node */
1547: llnk[6] = PETSC_INT_MAX - 1; /* value in the last node */
1548: llnk[7] = 1; /* count for the last node */
1549: llnk[8] = 0; /* next valid node to be used */
1550: PetscFunctionReturn(PETSC_SUCCESS);
1551: }
1553: static inline PetscErrorCode PetscLLCondensedAddSorted_fast(PetscInt nidx, const PetscInt indices[], PetscInt lnk[])
1554: {
1555: for (PetscInt k = 0, prev = 3 /* first value */; k < nidx; k++) {
1556: const PetscInt entry = indices[k];
1557: PetscInt next = lnk[prev + 2];
1559: /* search for insertion location */
1560: while (entry >= lnk[next]) {
1561: prev = next;
1562: next = lnk[next + 2];
1563: }
1564: /* entry is in range of previous list */
1565: if (entry < lnk[prev] + lnk[prev + 1]) continue;
1566: lnk[1]++;
1567: /* entry is right after previous list */
1568: if (entry == lnk[prev] + lnk[prev + 1]) {
1569: lnk[prev + 1]++;
1570: if (lnk[next] == entry + 1) { /* combine two contiguous strings */
1571: lnk[prev + 1] += lnk[next + 1];
1572: lnk[prev + 2] = lnk[next + 2];
1573: next = lnk[next + 2];
1574: lnk[0]--;
1575: }
1576: continue;
1577: }
1578: /* entry is right before next list */
1579: if (entry == lnk[next] - 1) {
1580: lnk[next]--;
1581: lnk[next + 1]++;
1582: prev = next;
1583: next = lnk[prev + 2];
1584: continue;
1585: }
1586: /* add entry into lnk */
1587: lnk[prev + 2] = 3 * ((lnk[8]++) + 3); /* connect previous node to the new node */
1588: prev = lnk[prev + 2];
1589: lnk[prev] = entry; /* set value of the new node */
1590: lnk[prev + 1] = 1; /* number of values in contiguous string is one to start */
1591: lnk[prev + 2] = next; /* connect new node to next node */
1592: lnk[0]++;
1593: }
1594: return PETSC_SUCCESS;
1595: }
1597: static inline PetscErrorCode PetscLLCondensedClean_fast(PETSC_UNUSED PetscInt nidx, PetscInt *indices, PetscInt *lnk)
1598: {
1599: const PetscInt nlnk = lnk[0];
1600: PetscInt next = lnk[5]; /* first node */
1602: for (PetscInt k = 0, cnt = 0; k < nlnk; k++) {
1603: for (PetscInt j = 0; j < lnk[next + 1]; j++) indices[cnt++] = lnk[next] + j;
1604: next = lnk[next + 2];
1605: }
1606: lnk[0] = 0; /* nlnk: number of links */
1607: lnk[1] = 0; /* number of integer entries represented in list */
1608: lnk[3] = PETSC_INT_MIN + 1; /* value in the first node */
1609: lnk[4] = 1; /* count for the first node */
1610: lnk[5] = 6; /* next for the first node */
1611: lnk[6] = PETSC_INT_MAX - 1; /* value in the last node */
1612: lnk[7] = 1; /* count for the last node */
1613: lnk[8] = 0; /* next valid location to make link */
1614: return PETSC_SUCCESS;
1615: }
1617: static inline PetscErrorCode PetscLLCondensedView_fast(const PetscInt *lnk)
1618: {
1619: const PetscInt nlnk = lnk[0];
1620: PetscInt next = lnk[5]; /* first node */
1622: for (PetscInt k = 0; k < nlnk; k++) {
1623: #if 0 /* Debugging code */
1624: printf("%d value %d len %d next %d\n", next, lnk[next], lnk[next + 1], lnk[next + 2]);
1625: #endif
1626: next = lnk[next + 2];
1627: }
1628: return PETSC_SUCCESS;
1629: }
1631: static inline PetscErrorCode PetscLLCondensedDestroy_fast(PetscInt *lnk)
1632: {
1633: return PetscFree(lnk);
1634: }
1636: PETSC_EXTERN PetscErrorCode PetscCDCreate(PetscInt, PetscCoarsenData **);
1637: PETSC_EXTERN PetscErrorCode PetscCDDestroy(PetscCoarsenData *);
1638: PETSC_EXTERN PetscErrorCode PetscCDIntNdSetID(PetscCDIntNd *, PetscInt);
1639: PETSC_EXTERN PetscErrorCode PetscCDIntNdGetID(const PetscCDIntNd *, PetscInt *);
1640: PETSC_EXTERN PetscErrorCode PetscCDAppendID(PetscCoarsenData *, PetscInt, PetscInt);
1641: PETSC_EXTERN PetscErrorCode PetscCDMoveAppend(PetscCoarsenData *, PetscInt, PetscInt);
1642: PETSC_EXTERN PetscErrorCode PetscCDAppendNode(PetscCoarsenData *, PetscInt, PetscCDIntNd *);
1643: PETSC_EXTERN PetscErrorCode PetscCDRemoveNextNode(PetscCoarsenData *, PetscInt, PetscCDIntNd *);
1644: PETSC_EXTERN PetscErrorCode PetscCDCountAt(const PetscCoarsenData *, PetscInt, PetscInt *);
1645: PETSC_EXTERN PetscErrorCode PetscCDIsEmptyAt(const PetscCoarsenData *, PetscInt, PetscBool *);
1646: PETSC_EXTERN PetscErrorCode PetscCDSetChunkSize(PetscCoarsenData *, PetscInt);
1647: PETSC_EXTERN PetscErrorCode PetscCDPrint(const PetscCoarsenData *, PetscInt, MPI_Comm);
1648: PETSC_EXTERN PetscErrorCode PetscCDGetNonemptyIS(PetscCoarsenData *, IS *);
1649: PETSC_EXTERN PetscErrorCode PetscCDGetMat(PetscCoarsenData *, Mat *);
1650: PETSC_EXTERN PetscErrorCode PetscCDSetMat(PetscCoarsenData *, Mat);
1651: PETSC_EXTERN PetscErrorCode PetscCDClearMat(PetscCoarsenData *);
1652: PETSC_EXTERN PetscErrorCode PetscCDRemoveAllAt(PetscCoarsenData *, PetscInt);
1653: PETSC_EXTERN PetscErrorCode PetscCDCount(const PetscCoarsenData *, PetscInt *_sz);
1655: PETSC_EXTERN PetscErrorCode PetscCDGetHeadPos(const PetscCoarsenData *, PetscInt, PetscCDIntNd **);
1656: PETSC_EXTERN PetscErrorCode PetscCDGetNextPos(const PetscCoarsenData *, PetscInt, PetscCDIntNd **);
1657: PETSC_EXTERN PetscErrorCode PetscCDGetASMBlocks(const PetscCoarsenData *, const PetscInt, PetscInt *, IS **);
1659: /* this is extern because it is used in MatFDColoringUseDM() which is in the DM library */
1660: PETSC_EXTERN PetscErrorCode MatFDColoringApply_AIJ(Mat, MatFDColoring, Vec, void *);
1662: PETSC_EXTERN PetscLogEvent MAT_Mult;
1663: PETSC_EXTERN PetscLogEvent MAT_MultAdd;
1664: PETSC_EXTERN PetscLogEvent MAT_MultTranspose;
1665: PETSC_EXTERN PetscLogEvent MAT_MultHermitianTranspose;
1666: PETSC_EXTERN PetscLogEvent MAT_MultTransposeAdd;
1667: PETSC_EXTERN PetscLogEvent MAT_MultHermitianTransposeAdd;
1668: PETSC_EXTERN PetscLogEvent MAT_Solve;
1669: PETSC_EXTERN PetscLogEvent MAT_Solves;
1670: PETSC_EXTERN PetscLogEvent MAT_SolveAdd;
1671: PETSC_EXTERN PetscLogEvent MAT_SolveTranspose;
1672: PETSC_EXTERN PetscLogEvent MAT_SolveTransposeAdd;
1673: PETSC_EXTERN PetscLogEvent MAT_SOR;
1674: PETSC_EXTERN PetscLogEvent MAT_ForwardSolve;
1675: PETSC_EXTERN PetscLogEvent MAT_BackwardSolve;
1676: PETSC_EXTERN PetscLogEvent MAT_LUFactor;
1677: PETSC_EXTERN PetscLogEvent MAT_LUFactorSymbolic;
1678: PETSC_EXTERN PetscLogEvent MAT_LUFactorNumeric;
1679: PETSC_EXTERN PetscLogEvent MAT_QRFactor;
1680: PETSC_EXTERN PetscLogEvent MAT_QRFactorSymbolic;
1681: PETSC_EXTERN PetscLogEvent MAT_QRFactorNumeric;
1682: PETSC_EXTERN PetscLogEvent MAT_CholeskyFactor;
1683: PETSC_EXTERN PetscLogEvent MAT_CholeskyFactorSymbolic;
1684: PETSC_EXTERN PetscLogEvent MAT_CholeskyFactorNumeric;
1685: PETSC_EXTERN PetscLogEvent MAT_ILUFactor;
1686: PETSC_EXTERN PetscLogEvent MAT_ILUFactorSymbolic;
1687: PETSC_EXTERN PetscLogEvent MAT_ICCFactorSymbolic;
1688: PETSC_EXTERN PetscLogEvent MAT_Copy;
1689: PETSC_EXTERN PetscLogEvent MAT_Convert;
1690: PETSC_EXTERN PetscLogEvent MAT_Scale;
1691: PETSC_EXTERN PetscLogEvent MAT_AssemblyBegin;
1692: PETSC_EXTERN PetscLogEvent MAT_AssemblyEnd;
1693: PETSC_EXTERN PetscLogEvent MAT_SetValues;
1694: PETSC_EXTERN PetscLogEvent MAT_GetValues;
1695: PETSC_EXTERN PetscLogEvent MAT_GetRow;
1696: PETSC_EXTERN PetscLogEvent MAT_GetRowIJ;
1697: PETSC_EXTERN PetscLogEvent MAT_CreateSubMats;
1698: PETSC_EXTERN PetscLogEvent MAT_GetOrdering;
1699: PETSC_EXTERN PetscLogEvent MAT_RedundantMat;
1700: PETSC_EXTERN PetscLogEvent MAT_IncreaseOverlap;
1701: PETSC_EXTERN PetscLogEvent MAT_Partitioning;
1702: PETSC_EXTERN PetscLogEvent MAT_PartitioningND;
1703: PETSC_EXTERN PetscLogEvent MAT_Coarsen;
1704: PETSC_EXTERN PetscLogEvent MAT_ZeroEntries;
1705: PETSC_EXTERN PetscLogEvent MAT_Load;
1706: PETSC_EXTERN PetscLogEvent MAT_View;
1707: PETSC_EXTERN PetscLogEvent MAT_AXPY;
1708: PETSC_EXTERN PetscLogEvent MAT_FDColoringCreate;
1709: PETSC_EXTERN PetscLogEvent MAT_TransposeColoringCreate;
1710: PETSC_EXTERN PetscLogEvent MAT_FDColoringSetUp;
1711: PETSC_EXTERN PetscLogEvent MAT_FDColoringApply;
1712: PETSC_EXTERN PetscLogEvent MAT_Transpose;
1713: PETSC_EXTERN PetscLogEvent MAT_FDColoringFunction;
1714: PETSC_EXTERN PetscLogEvent MAT_CreateSubMat;
1715: PETSC_EXTERN PetscLogEvent MAT_MatSolve;
1716: PETSC_EXTERN PetscLogEvent MAT_MatTrSolve;
1717: PETSC_EXTERN PetscLogEvent MAT_MatMultSymbolic;
1718: PETSC_EXTERN PetscLogEvent MAT_MatMultNumeric;
1719: PETSC_EXTERN PetscLogEvent MAT_Getlocalmatcondensed;
1720: PETSC_EXTERN PetscLogEvent MAT_GetBrowsOfAcols;
1721: PETSC_EXTERN PetscLogEvent MAT_GetBrowsOfAocols;
1722: PETSC_EXTERN PetscLogEvent MAT_PtAPSymbolic;
1723: PETSC_EXTERN PetscLogEvent MAT_PtAPNumeric;
1724: PETSC_EXTERN PetscLogEvent MAT_Seqstompinum;
1725: PETSC_EXTERN PetscLogEvent MAT_Seqstompisym;
1726: PETSC_EXTERN PetscLogEvent MAT_Seqstompi;
1727: PETSC_EXTERN PetscLogEvent MAT_Getlocalmat;
1728: PETSC_EXTERN PetscLogEvent MAT_RARtSymbolic;
1729: PETSC_EXTERN PetscLogEvent MAT_RARtNumeric;
1730: PETSC_EXTERN PetscLogEvent MAT_MatTransposeMultSymbolic;
1731: PETSC_EXTERN PetscLogEvent MAT_MatTransposeMultNumeric;
1732: PETSC_EXTERN PetscLogEvent MAT_TransposeMatMultSymbolic;
1733: PETSC_EXTERN PetscLogEvent MAT_TransposeMatMultNumeric;
1734: PETSC_EXTERN PetscLogEvent MAT_MatMatMultSymbolic;
1735: PETSC_EXTERN PetscLogEvent MAT_MatMatMultNumeric;
1736: PETSC_EXTERN PetscLogEvent MAT_Getsymtransreduced;
1737: PETSC_EXTERN PetscLogEvent MAT_GetSeqNonzeroStructure;
1738: PETSC_EXTERN PetscLogEvent MATMFFD_Mult;
1739: PETSC_EXTERN PetscLogEvent MAT_GetMultiProcBlock;
1740: PETSC_EXTERN PetscLogEvent MAT_CUSPARSECopyToGPU;
1741: PETSC_EXTERN PetscLogEvent MAT_CUSPARSECopyFromGPU;
1742: PETSC_EXTERN PetscLogEvent MAT_CUSPARSEGenerateTranspose;
1743: PETSC_EXTERN PetscLogEvent MAT_CUSPARSESolveAnalysis;
1744: PETSC_EXTERN PetscLogEvent MAT_HIPSPARSECopyToGPU;
1745: PETSC_EXTERN PetscLogEvent MAT_HIPSPARSECopyFromGPU;
1746: PETSC_EXTERN PetscLogEvent MAT_HIPSPARSEGenerateTranspose;
1747: PETSC_EXTERN PetscLogEvent MAT_HIPSPARSESolveAnalysis;
1748: PETSC_EXTERN PetscLogEvent MAT_SetValuesBatch;
1749: PETSC_EXTERN PetscLogEvent MAT_CreateGraph;
1750: PETSC_EXTERN PetscLogEvent MAT_ViennaCLCopyToGPU;
1751: PETSC_EXTERN PetscLogEvent MAT_DenseCopyToGPU;
1752: PETSC_EXTERN PetscLogEvent MAT_DenseCopyFromGPU;
1753: PETSC_EXTERN PetscLogEvent MAT_Merge;
1754: PETSC_EXTERN PetscLogEvent MAT_Residual;
1755: PETSC_EXTERN PetscLogEvent MAT_SetRandom;
1756: PETSC_EXTERN PetscLogEvent MAT_FactorFactS;
1757: PETSC_EXTERN PetscLogEvent MAT_FactorInvS;
1758: PETSC_EXTERN PetscLogEvent MAT_PreallCOO;
1759: PETSC_EXTERN PetscLogEvent MAT_SetVCOO;
1760: PETSC_EXTERN PetscLogEvent MATCOLORING_Apply;
1761: PETSC_EXTERN PetscLogEvent MATCOLORING_Comm;
1762: PETSC_EXTERN PetscLogEvent MATCOLORING_Local;
1763: PETSC_EXTERN PetscLogEvent MATCOLORING_ISCreate;
1764: PETSC_EXTERN PetscLogEvent MATCOLORING_SetUp;
1765: PETSC_EXTERN PetscLogEvent MATCOLORING_Weights;
1766: PETSC_EXTERN PetscLogEvent MAT_H2Opus_Build;
1767: PETSC_EXTERN PetscLogEvent MAT_H2Opus_Compress;
1768: PETSC_EXTERN PetscLogEvent MAT_H2Opus_Orthog;
1769: PETSC_EXTERN PetscLogEvent MAT_H2Opus_LR;
1770: PETSC_EXTERN PetscLogEvent MAT_CUDACopyToGPU;
1771: PETSC_EXTERN PetscLogEvent MAT_HIPCopyToGPU;