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_INTERN 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
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 (*matmatmultsymbolic)(Mat, Mat, Mat, PetscReal, Mat);
114: /*64*/
115: PetscErrorCode (*matmatmultnumeric)(Mat, Mat, Mat, Mat);
116: PetscErrorCode (*setlocaltoglobalmapping)(Mat, ISLocalToGlobalMapping, ISLocalToGlobalMapping);
117: PetscErrorCode (*setvalueslocal)(Mat, PetscInt, const PetscInt[], PetscInt, const PetscInt[], const PetscScalar[], InsertMode);
118: PetscErrorCode (*zerorowslocal)(Mat, PetscInt, const PetscInt[], PetscScalar, Vec, Vec);
119: PetscErrorCode (*getrowmaxabs)(Mat, Vec, PetscInt[]);
120: /*69*/
121: PetscErrorCode (*getrowminabs)(Mat, Vec, PetscInt[]);
122: PetscErrorCode (*convert)(Mat, MatType, MatReuse, Mat *);
123: PetscErrorCode (*hasoperation)(Mat, MatOperation, PetscBool *);
124: PetscErrorCode (*fdcoloringapply)(Mat, MatFDColoring, Vec, void *);
125: PetscErrorCode (*setfromoptions)(Mat, PetscOptionItems);
126: /*74*/
127: PetscErrorCode (*findzerodiagonals)(Mat, IS *);
128: PetscErrorCode (*mults)(Mat, Vecs, Vecs);
129: PetscErrorCode (*solves)(Mat, Vecs, Vecs);
130: PetscErrorCode (*getinertia)(Mat, PetscInt *, PetscInt *, PetscInt *);
131: PetscErrorCode (*load)(Mat, PetscViewer);
132: /*79*/
133: PetscErrorCode (*issymmetric)(Mat, PetscReal, PetscBool *);
134: PetscErrorCode (*ishermitian)(Mat, PetscReal, PetscBool *);
135: PetscErrorCode (*isstructurallysymmetric)(Mat, PetscBool *);
136: PetscErrorCode (*setvaluesblockedlocal)(Mat, PetscInt, const PetscInt[], PetscInt, const PetscInt[], const PetscScalar[], InsertMode);
137: PetscErrorCode (*getvecs)(Mat, Vec *, Vec *);
138: /*84*/
139: PetscErrorCode (*matmultsymbolic)(Mat, Mat, PetscReal, Mat);
140: PetscErrorCode (*matmultnumeric)(Mat, Mat, Mat);
141: PetscErrorCode (*ptapnumeric)(Mat, Mat, Mat); /* double dispatch wrapper routine */
142: PetscErrorCode (*mattransposemultsymbolic)(Mat, Mat, PetscReal, Mat);
143: PetscErrorCode (*mattransposemultnumeric)(Mat, Mat, Mat);
144: /*89*/
145: PetscErrorCode (*bindtocpu)(Mat, PetscBool);
146: PetscErrorCode (*productsetfromoptions)(Mat);
147: PetscErrorCode (*productsymbolic)(Mat);
148: PetscErrorCode (*productnumeric)(Mat);
149: PetscErrorCode (*conjugate)(Mat); /* complex conjugate */
150: /*94*/
151: PetscErrorCode (*viewnative)(Mat, PetscViewer);
152: PetscErrorCode (*setvaluesrow)(Mat, PetscInt, const PetscScalar[]);
153: PetscErrorCode (*realpart)(Mat);
154: PetscErrorCode (*imaginarypart)(Mat);
155: PetscErrorCode (*getrowuppertriangular)(Mat);
156: /*99*/
157: PetscErrorCode (*restorerowuppertriangular)(Mat);
158: PetscErrorCode (*matsolve)(Mat, Mat, Mat);
159: PetscErrorCode (*matsolvetranspose)(Mat, Mat, Mat);
160: PetscErrorCode (*getrowmin)(Mat, Vec, PetscInt[]);
161: PetscErrorCode (*getcolumnvector)(Mat, Vec, PetscInt);
162: /*104*/
163: PetscErrorCode (*missingdiagonal)(Mat, PetscBool *, PetscInt *);
164: PetscErrorCode (*getseqnonzerostructure)(Mat, Mat *);
165: PetscErrorCode (*create)(Mat);
166: PetscErrorCode (*getghosts)(Mat, PetscInt *, const PetscInt *[]);
167: PetscErrorCode (*getlocalsubmatrix)(Mat, IS, IS, Mat *);
168: /*109*/
169: PetscErrorCode (*restorelocalsubmatrix)(Mat, IS, IS, Mat *);
170: PetscErrorCode (*multdiagonalblock)(Mat, Vec, Vec);
171: PetscErrorCode (*hermitiantranspose)(Mat, MatReuse, Mat *);
172: PetscErrorCode (*multhermitiantranspose)(Mat, Vec, Vec);
173: PetscErrorCode (*multhermitiantransposeadd)(Mat, Vec, Vec, Vec);
174: /*114*/
175: PetscErrorCode (*getmultiprocblock)(Mat, MPI_Comm, MatReuse, Mat *);
176: PetscErrorCode (*findnonzerorows)(Mat, IS *);
177: PetscErrorCode (*getcolumnreductions)(Mat, PetscInt, PetscReal *);
178: PetscErrorCode (*invertblockdiagonal)(Mat, const PetscScalar **);
179: PetscErrorCode (*invertvariableblockdiagonal)(Mat, PetscInt, const PetscInt *, PetscScalar *);
180: /*119*/
181: PetscErrorCode (*createsubmatricesmpi)(Mat, PetscInt, const IS[], const IS[], MatReuse, Mat **);
182: PetscErrorCode (*setvaluesbatch)(Mat, PetscInt, PetscInt, PetscInt *, const PetscScalar *);
183: PetscErrorCode (*transposematmultsymbolic)(Mat, Mat, PetscReal, Mat);
184: PetscErrorCode (*transposematmultnumeric)(Mat, Mat, Mat);
185: PetscErrorCode (*transposecoloringcreate)(Mat, ISColoring, MatTransposeColoring);
186: /*124*/
187: PetscErrorCode (*transcoloringapplysptoden)(MatTransposeColoring, Mat, Mat);
188: PetscErrorCode (*transcoloringapplydentosp)(MatTransposeColoring, Mat, Mat);
189: PetscErrorCode (*rartnumeric)(Mat, Mat, Mat); /* double dispatch wrapper routine */
190: PetscErrorCode (*setblocksizes)(Mat, PetscInt, PetscInt);
191: PetscErrorCode (*residual)(Mat, Vec, Vec, Vec);
192: /*129*/
193: PetscErrorCode (*fdcoloringsetup)(Mat, ISColoring, MatFDColoring);
194: PetscErrorCode (*findoffblockdiagonalentries)(Mat, IS *);
195: PetscErrorCode (*creatempimatconcatenateseqmat)(MPI_Comm, Mat, PetscInt, MatReuse, Mat *);
196: PetscErrorCode (*destroysubmatrices)(PetscInt, Mat *[]);
197: PetscErrorCode (*mattransposesolve)(Mat, Mat, Mat);
198: /*134*/
199: PetscErrorCode (*getvalueslocal)(Mat, PetscInt, const PetscInt[], PetscInt, const PetscInt[], PetscScalar[]);
200: PetscErrorCode (*creategraph)(Mat, PetscBool, PetscBool, PetscReal, PetscInt, PetscInt[], Mat *);
201: PetscErrorCode (*transposesymbolic)(Mat, Mat *);
202: PetscErrorCode (*eliminatezeros)(Mat, PetscBool);
203: PetscErrorCode (*getrowsumabs)(Mat, Vec);
204: /*139*/
205: PetscErrorCode (*getfactor)(Mat, MatSolverType, MatFactorType, Mat *);
206: PetscErrorCode (*getblockdiagonal)(Mat, Mat *); // NOTE: the caller of get{block, vblock}diagonal owns the returned matrix;
207: PetscErrorCode (*getvblockdiagonal)(Mat, Mat *); // they must destroy it after use
208: PetscErrorCode (*copyhashtoxaij)(Mat, Mat);
209: PetscErrorCode (*getcurrentmemtype)(Mat, PetscMemType *);
210: /*144*/
211: PetscErrorCode (*zerorowscolumnslocal)(Mat, PetscInt, const PetscInt[], PetscScalar, Vec, Vec);
212: };
213: /*
214: If you add MatOps entries above also add them to the MATOP enum
215: in include/petscmat.h
216: */
218: #include <petscsys.h>
220: typedef struct _p_MatRootName *MatRootName;
221: struct _p_MatRootName {
222: char *rname, *sname, *mname;
223: MatRootName next;
224: };
226: PETSC_EXTERN MatRootName MatRootNameList;
228: /*
229: Utility private matrix routines used outside Mat
230: */
231: PETSC_SINGLE_LIBRARY_INTERN PetscErrorCode MatFindNonzeroRowsOrCols_Basic(Mat, PetscBool, PetscReal, IS *);
232: PETSC_EXTERN PetscErrorCode MatShellGetScalingShifts(Mat, PetscScalar *, PetscScalar *, Vec *, Vec *, Vec *, Mat *, IS *, IS *);
234: #define MAT_SHELL_NOT_ALLOWED (void *)-1
236: /*
237: Utility private matrix routines
238: */
239: PETSC_INTERN PetscErrorCode MatConvert_Basic(Mat, MatType, MatReuse, Mat *);
240: PETSC_INTERN PetscErrorCode MatConvert_Shell(Mat, MatType, MatReuse, Mat *);
241: PETSC_INTERN PetscErrorCode MatConvertFrom_Shell(Mat, MatType, MatReuse, Mat *);
242: PETSC_INTERN PetscErrorCode MatShellSetContext_Immutable(Mat, void *);
243: PETSC_INTERN PetscErrorCode MatShellSetContextDestroy_Immutable(Mat, PetscCtxDestroyFn *);
244: PETSC_INTERN PetscErrorCode MatShellSetManageScalingShifts_Immutable(Mat);
245: PETSC_INTERN PetscErrorCode MatCopy_Basic(Mat, Mat, MatStructure);
246: PETSC_INTERN PetscErrorCode MatDiagonalSet_Default(Mat, Vec, InsertMode);
247: #if defined(PETSC_HAVE_SCALAPACK)
248: PETSC_INTERN PetscErrorCode MatConvert_Dense_ScaLAPACK(Mat, MatType, MatReuse, Mat *);
249: #endif
250: PETSC_INTERN PetscErrorCode MatSetPreallocationCOO_Basic(Mat, PetscCount, PetscInt[], PetscInt[]);
251: PETSC_INTERN PetscErrorCode MatSetValuesCOO_Basic(Mat, const PetscScalar[], InsertMode);
253: /* This can be moved to the public header after implementing some missing MatProducts */
254: PETSC_INTERN PetscErrorCode MatCreateFromISLocalToGlobalMapping(ISLocalToGlobalMapping, Mat, PetscBool, PetscBool, MatType, Mat *);
256: /* these callbacks rely on the old matrix function pointers for
257: matmat operations. They are unsafe, and should be removed.
258: However, the amount of work needed to clean up all the
259: implementations is not negligible */
260: PETSC_INTERN PetscErrorCode MatProductSymbolic_AB(Mat);
261: PETSC_INTERN PetscErrorCode MatProductNumeric_AB(Mat);
262: PETSC_INTERN PetscErrorCode MatProductSymbolic_AtB(Mat);
263: PETSC_INTERN PetscErrorCode MatProductNumeric_AtB(Mat);
264: PETSC_INTERN PetscErrorCode MatProductSymbolic_ABt(Mat);
265: PETSC_INTERN PetscErrorCode MatProductNumeric_ABt(Mat);
266: PETSC_INTERN PetscErrorCode MatProductNumeric_PtAP(Mat);
267: PETSC_INTERN PetscErrorCode MatProductNumeric_RARt(Mat);
268: PETSC_INTERN PetscErrorCode MatProductSymbolic_ABC(Mat);
269: PETSC_INTERN PetscErrorCode MatProductNumeric_ABC(Mat);
271: PETSC_INTERN PetscErrorCode MatProductCreate_Private(Mat, Mat, Mat, Mat);
272: /* this callback handles all the different triple products and
273: does not rely on the function pointers; used by cuSPARSE/hipSPARSE and KOKKOS-KERNELS */
274: PETSC_INTERN PetscErrorCode MatProductSymbolic_ABC_Basic(Mat);
276: /* CreateGraph is common to AIJ seq and mpi */
277: PETSC_INTERN PetscErrorCode MatCreateGraph_Simple_AIJ(Mat, PetscBool, PetscBool, PetscReal, PetscInt, PetscInt[], Mat *);
279: #if defined(PETSC_CLANG_STATIC_ANALYZER)
280: template <typename Tm>
281: extern void MatCheckPreallocated(Tm, int);
282: template <typename Tm>
283: extern void MatCheckProduct(Tm, int);
284: #else /* PETSC_CLANG_STATIC_ANALYZER */
285: #define MatCheckPreallocated(A, arg) \
286: do { \
287: if (!(A)->preallocated) PetscCall(MatSetUp(A)); \
288: } while (0)
290: #if defined(PETSC_USE_DEBUG)
291: #define MatCheckProduct(A, arg) \
292: do { \
293: PetscCheck((A)->product, PETSC_COMM_SELF, PETSC_ERR_ARG_WRONGSTATE, "Argument %d \"%s\" is not a matrix obtained from MatProductCreate()", (arg), #A); \
294: } while (0)
295: #else
296: #define MatCheckProduct(A, arg) \
297: do { \
298: } while (0)
299: #endif
300: #endif /* PETSC_CLANG_STATIC_ANALYZER */
302: /*
303: The stash is used to temporarily store inserted matrix values that
304: belong to another processor. During the assembly phase the stashed
305: values are moved to the correct processor and
306: */
308: typedef struct _MatStashSpace *PetscMatStashSpace;
310: struct _MatStashSpace {
311: PetscMatStashSpace next;
312: PetscScalar *space_head, *val;
313: PetscInt *idx, *idy;
314: PetscInt total_space_size;
315: PetscInt local_used;
316: PetscInt local_remaining;
317: };
319: PETSC_EXTERN PetscErrorCode PetscMatStashSpaceGet(PetscInt, PetscInt, PetscMatStashSpace *);
320: PETSC_EXTERN PetscErrorCode PetscMatStashSpaceContiguous(PetscInt, PetscMatStashSpace *, PetscScalar *, PetscInt *, PetscInt *);
321: PETSC_EXTERN PetscErrorCode PetscMatStashSpaceDestroy(PetscMatStashSpace *);
323: typedef struct {
324: PetscInt count;
325: } MatStashHeader;
327: typedef struct {
328: void *buffer; /* Of type blocktype, dynamically constructed */
329: PetscInt count;
330: char pending;
331: } MatStashFrame;
333: typedef struct _MatStash MatStash;
334: struct _MatStash {
335: PetscInt nmax; /* maximum stash size */
336: PetscInt umax; /* user specified max-size */
337: PetscInt oldnmax; /* the nmax value used previously */
338: PetscInt n; /* stash size */
339: PetscInt bs; /* block size of the stash */
340: PetscInt reallocs; /* preserve the no of mallocs invoked */
341: PetscMatStashSpace space_head, space; /* linked list to hold stashed global row/column numbers and matrix values */
343: PetscErrorCode (*ScatterBegin)(Mat, MatStash *, PetscInt *);
344: PetscErrorCode (*ScatterGetMesg)(MatStash *, PetscMPIInt *, PetscInt **, PetscInt **, PetscScalar **, PetscInt *);
345: PetscErrorCode (*ScatterEnd)(MatStash *);
346: PetscErrorCode (*ScatterDestroy)(MatStash *);
348: /* The following variables are used for communication */
349: MPI_Comm comm;
350: PetscMPIInt size, rank;
351: PetscMPIInt tag1, tag2;
352: MPI_Request *send_waits; /* array of send requests */
353: MPI_Request *recv_waits; /* array of receive requests */
354: MPI_Status *send_status; /* array of send status */
355: PetscMPIInt nsends, nrecvs; /* numbers of sends and receives */
356: PetscScalar *svalues; /* sending data */
357: PetscInt *sindices;
358: PetscScalar **rvalues; /* receiving data (values) */
359: PetscInt **rindices; /* receiving data (indices) */
360: PetscMPIInt nprocessed; /* number of messages already processed */
361: PetscMPIInt *flg_v; /* indicates what messages have arrived so far and from whom */
362: PetscBool reproduce;
363: PetscMPIInt reproduce_count;
365: /* The following variables are used for BTS communication */
366: PetscBool first_assembly_done; /* Is the first time matrix assembly done? */
367: PetscBool use_status; /* Use MPI_Status to determine number of items in each message */
368: PetscMPIInt nsendranks;
369: PetscMPIInt nrecvranks;
370: PetscMPIInt *sendranks;
371: PetscMPIInt *recvranks;
372: MatStashHeader *sendhdr, *recvhdr;
373: MatStashFrame *sendframes; /* pointers to the main messages */
374: MatStashFrame *recvframes;
375: MatStashFrame *recvframe_active;
376: PetscInt recvframe_i; /* index of block within active frame */
377: PetscInt recvframe_count; /* Count actually sent for current frame */
378: PetscMPIInt recvcount; /* Number of receives processed so far */
379: PetscMPIInt *some_indices; /* From last call to MPI_Waitsome */
380: MPI_Status *some_statuses; /* Statuses from last call to MPI_Waitsome */
381: PetscMPIInt some_count; /* Number of requests completed in last call to MPI_Waitsome */
382: PetscMPIInt some_i; /* Index of request currently being processed */
383: MPI_Request *sendreqs;
384: MPI_Request *recvreqs;
385: PetscSegBuffer segsendblocks;
386: PetscSegBuffer segrecvframe;
387: PetscSegBuffer segrecvblocks;
388: MPI_Datatype blocktype;
389: size_t blocktype_size;
390: InsertMode *insertmode; /* Pointer to check mat->insertmode and set upon message arrival in case no local values have been set. */
391: };
393: #if !defined(PETSC_HAVE_MPIUNI)
394: PETSC_INTERN PetscErrorCode MatStashScatterDestroy_BTS(MatStash *);
395: #endif
396: PETSC_INTERN PetscErrorCode MatStashCreate_Private(MPI_Comm, PetscInt, MatStash *);
397: PETSC_INTERN PetscErrorCode MatStashDestroy_Private(MatStash *);
398: PETSC_INTERN PetscErrorCode MatStashScatterEnd_Private(MatStash *);
399: PETSC_INTERN PetscErrorCode MatStashSetInitialSize_Private(MatStash *, PetscInt);
400: PETSC_INTERN PetscErrorCode MatStashGetInfo_Private(MatStash *, PetscInt *, PetscInt *);
401: PETSC_INTERN PetscErrorCode MatStashValuesRow_Private(MatStash *, PetscInt, PetscInt, const PetscInt[], const PetscScalar[], PetscBool);
402: PETSC_INTERN PetscErrorCode MatStashValuesCol_Private(MatStash *, PetscInt, PetscInt, const PetscInt[], const PetscScalar[], PetscInt, PetscBool);
403: PETSC_INTERN PetscErrorCode MatStashValuesRowBlocked_Private(MatStash *, PetscInt, PetscInt, const PetscInt[], const PetscScalar[], PetscInt, PetscInt, PetscInt);
404: PETSC_INTERN PetscErrorCode MatStashValuesColBlocked_Private(MatStash *, PetscInt, PetscInt, const PetscInt[], const PetscScalar[], PetscInt, PetscInt, PetscInt);
405: PETSC_INTERN PetscErrorCode MatStashScatterBegin_Private(Mat, MatStash *, PetscInt *);
406: PETSC_INTERN PetscErrorCode MatStashScatterGetMesg_Private(MatStash *, PetscMPIInt *, PetscInt **, PetscInt **, PetscScalar **, PetscInt *);
407: PETSC_INTERN PetscErrorCode MatGetInfo_External(Mat, MatInfoType, MatInfo *);
409: typedef struct {
410: PetscInt dim;
411: PetscInt dims[4];
412: PetscInt starts[4];
413: PetscBool noc; /* this is a single component problem, hence user will not set MatStencil.c */
414: } MatStencilInfo;
416: /* Info about using compressed row format */
417: typedef struct {
418: PetscBool use; /* indicates compressed rows have been checked and will be used */
419: PetscInt nrows; /* number of non-zero rows */
420: PetscInt *i; /* compressed row pointer */
421: PetscInt *rindex; /* compressed row index */
422: } Mat_CompressedRow;
423: PETSC_EXTERN PetscErrorCode MatCheckCompressedRow(Mat, PetscInt, Mat_CompressedRow *, PetscInt *, PetscInt, PetscReal);
425: typedef struct { /* used by MatCreateRedundantMatrix() for reusing matredundant */
426: PetscInt nzlocal, nsends, nrecvs;
427: PetscMPIInt *send_rank, *recv_rank;
428: PetscInt *sbuf_nz, *rbuf_nz, *sbuf_j, **rbuf_j;
429: PetscScalar *sbuf_a, **rbuf_a;
430: MPI_Comm subcomm; /* when user does not provide a subcomm */
431: IS isrow, iscol;
432: Mat *matseq;
433: } Mat_Redundant;
435: typedef struct { /* used by MatProduct() */
436: MatProductType type;
437: char *alg;
438: Mat A, B, C, Dwork;
439: 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, .. */
440: 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 .. */
441: PetscBool symbolic_used_the_fact_C_is_symmetric; /* MatMatMult(A,B,MAT_REUSE_MATRIX,..&C) is still legitimate), we need to redo symbolic! */
442: PetscObjectParameterDeclare(PetscReal, fill);
443: 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 */
444: PetscBool setfromoptionscalled;
446: /* Some products may display the information on the algorithm used */
447: PetscErrorCode (*view)(Mat, PetscViewer);
449: /* many products have intermediate data structures, each specific to Mat types and product type */
450: PetscBool clear; /* whether or not to clear the data structures after MatProductNumeric has been called */
451: void *data; /* where to stash those structures */
452: PetscErrorCode (*destroy)(void *); /* destroy routine */
453: } Mat_Product;
455: struct _p_Mat {
456: PETSCHEADER(struct _MatOps);
457: PetscLayout rmap, cmap;
458: void *data; /* implementation-specific data */
459: MatFactorType factortype; /* MAT_FACTOR_LU, ILU, CHOLESKY or ICC */
460: PetscBool trivialsymbolic; /* indicates the symbolic factorization doesn't actually do a symbolic factorization, it is delayed to the numeric factorization */
461: PetscBool canuseordering; /* factorization can use ordering provide to routine (most PETSc implementations) */
462: MatOrderingType preferredordering[MAT_FACTOR_NUM_TYPES]; /* what is the preferred (or default) ordering for the matrix solver type */
463: PetscBool assembled; /* is the matrix assembled? */
464: PetscBool was_assembled; /* new values inserted into assembled mat */
465: PetscInt num_ass; /* number of times matrix has been assembled */
466: PetscObjectState nonzerostate; /* each time new nonzeros locations are introduced into the matrix this is updated */
467: PetscObjectState ass_nonzerostate; /* nonzero state at last assembly */
468: MatInfo info; /* matrix information */
469: InsertMode insertmode; /* have values been inserted in matrix or added? */
470: MatStash stash, bstash; /* used for assembling off-proc mat emements */
471: MatNullSpace nullsp; /* null space (operator is singular) */
472: MatNullSpace transnullsp; /* null space of transpose of operator */
473: MatNullSpace nearnullsp; /* near null space to be used by multigrid methods */
474: PetscInt congruentlayouts; /* are the rows and columns layouts congruent? */
475: PetscBool preallocated;
476: MatStencilInfo stencil; /* information for structured grid */
477: PetscBool3 symmetric, hermitian, structurally_symmetric, spd;
478: PetscBool symmetry_eternal, structural_symmetry_eternal, spd_eternal;
479: PetscBool nooffprocentries, nooffproczerorows;
480: PetscBool assembly_subset; /* set by MAT_SUBSET_OFF_PROC_ENTRIES */
481: PetscBool submat_singleis; /* for efficient PCSetUp_ASM() */
482: PetscBool structure_only;
483: PetscBool sortedfull; /* full, sorted rows are inserted */
484: PetscBool force_diagonals; /* set by MAT_FORCE_DIAGONAL_ENTRIES */
485: #if defined(PETSC_HAVE_DEVICE)
486: PetscOffloadMask offloadmask; /* a mask which indicates where the valid matrix data is (GPU, CPU or both) */
487: PetscBool boundtocpu;
488: PetscBool bindingpropagates;
489: #endif
490: char *defaultrandtype;
491: void *spptr; /* pointer for special library like SuperLU */
492: char *solvertype;
493: PetscBool checksymmetryonassembly, checknullspaceonassembly;
494: PetscReal checksymmetrytol;
495: Mat schur; /* Schur complement matrix */
496: MatFactorSchurStatus schur_status; /* status of the Schur complement matrix */
497: Mat_Redundant *redundant; /* used by MatCreateRedundantMatrix() */
498: PetscBool erroriffailure; /* Generate an error if detected (for example a zero pivot) instead of returning */
499: MatFactorError factorerrortype; /* type of error in factorization */
500: PetscReal factorerror_zeropivot_value; /* If numerical zero pivot was detected this is the computed value */
501: PetscInt factorerror_zeropivot_row; /* Row where zero pivot was detected */
502: PetscInt nblocks, *bsizes; /* support for MatSetVariableBlockSizes() */
503: PetscInt p_cstart, p_rank, p_cend, n_rank; /* Information from parallel MatComputeVariableBlockEnvelope() */
504: PetscBool p_parallel;
505: char *defaultvectype;
506: Mat_Product *product;
507: PetscBool form_explicit_transpose; /* hint to generate an explicit mat tranpsose for operations like MatMultTranspose() */
508: PetscBool transupdated; /* whether or not the explicitly generated transpose is up-to-date */
509: char *factorprefix; /* the prefix to use with factored matrix that is created */
510: PetscBool hash_active; /* indicates MatSetValues() is being handled by hashing */
511: };
513: PETSC_INTERN PetscErrorCode MatAXPY_Basic(Mat, PetscScalar, Mat, MatStructure);
514: PETSC_INTERN PetscErrorCode MatAXPY_BasicWithPreallocation(Mat, Mat, PetscScalar, Mat, MatStructure);
515: PETSC_INTERN PetscErrorCode MatAXPY_Basic_Preallocate(Mat, Mat, Mat *);
516: PETSC_INTERN PetscErrorCode MatAXPY_Dense_Nest(Mat, PetscScalar, Mat);
518: PETSC_INTERN PetscErrorCode MatSetUp_Default(Mat);
520: /*
521: Utility for MatZeroRows
522: */
523: PETSC_INTERN PetscErrorCode MatZeroRowsMapLocal_Private(Mat, PetscInt, const PetscInt *, PetscInt *, PetscInt **);
525: /*
526: Utility for MatView/MatLoad
527: */
528: PETSC_INTERN PetscErrorCode MatView_Binary_BlockSizes(Mat, PetscViewer);
529: PETSC_INTERN PetscErrorCode MatLoad_Binary_BlockSizes(Mat, PetscViewer);
531: /*
532: Object for partitioning graphs
533: */
535: typedef struct _MatPartitioningOps *MatPartitioningOps;
536: struct _MatPartitioningOps {
537: PetscErrorCode (*apply)(MatPartitioning, IS *);
538: PetscErrorCode (*applynd)(MatPartitioning, IS *);
539: PetscErrorCode (*setfromoptions)(MatPartitioning, PetscOptionItems);
540: PetscErrorCode (*destroy)(MatPartitioning);
541: PetscErrorCode (*view)(MatPartitioning, PetscViewer);
542: PetscErrorCode (*improve)(MatPartitioning, IS *);
543: };
545: struct _p_MatPartitioning {
546: PETSCHEADER(struct _MatPartitioningOps);
547: Mat adj;
548: PetscInt *vertex_weights;
549: PetscReal *part_weights;
550: PetscInt n; /* number of partitions */
551: PetscInt ncon; /* number of vertex weights per vertex */
552: void *data;
553: PetscBool use_edge_weights; /* A flag indicates whether or not to use edge weights */
554: };
556: /* needed for parallel nested dissection by ParMetis and PTSCOTCH */
557: PETSC_INTERN PetscErrorCode MatPartitioningSizesToSep_Private(PetscInt, PetscInt[], PetscInt[], PetscInt[]);
559: /*
560: Object for coarsen graphs
561: */
562: typedef struct _MatCoarsenOps *MatCoarsenOps;
563: struct _MatCoarsenOps {
564: PetscErrorCode (*apply)(MatCoarsen);
565: PetscErrorCode (*setfromoptions)(MatCoarsen, PetscOptionItems);
566: PetscErrorCode (*destroy)(MatCoarsen);
567: PetscErrorCode (*view)(MatCoarsen, PetscViewer);
568: };
570: #define MAT_COARSEN_STRENGTH_INDEX_SIZE 3
571: struct _p_MatCoarsen {
572: PETSCHEADER(struct _MatCoarsenOps);
573: Mat graph;
574: void *subctx;
575: /* */
576: PetscBool strict_aggs;
577: IS perm;
578: PetscCoarsenData *agg_lists;
579: PetscInt max_it; /* number of iterations in HEM */
580: PetscReal threshold; /* HEM can filter interim graphs */
581: PetscInt strength_index_size;
582: PetscInt strength_index[MAT_COARSEN_STRENGTH_INDEX_SIZE];
583: };
585: PETSC_EXTERN PetscErrorCode MatCoarsenMISKSetDistance(MatCoarsen, PetscInt);
586: PETSC_EXTERN PetscErrorCode MatCoarsenMISKGetDistance(MatCoarsen, PetscInt *);
588: /*
589: Used in aijdevice.h
590: */
591: typedef struct {
592: PetscInt *i;
593: PetscInt *j;
594: PetscScalar *a;
595: PetscInt n;
596: PetscInt ignorezeroentries;
597: } PetscCSRDataStructure;
599: /*
600: MatFDColoring is used to compute Jacobian matrices efficiently
601: via coloring. The data structure is explained below in an example.
603: Color = 0 1 0 2 | 2 3 0
604: ---------------------------------------------------
605: 00 01 | 05
606: 10 11 | 14 15 Processor 0
607: 22 23 | 25
608: 32 33 |
609: ===================================================
610: | 44 45 46
611: 50 | 55 Processor 1
612: | 64 66
613: ---------------------------------------------------
615: ncolors = 4;
617: ncolumns = {2,1,1,0}
618: columns = {{0,2},{1},{3},{}}
619: nrows = {4,2,3,3}
620: rows = {{0,1,2,3},{0,1},{1,2,3},{0,1,2}}
621: vwscale = {dx(0),dx(1),dx(2),dx(3)} MPI Vec
622: vscale = {dx(0),dx(1),dx(2),dx(3),dx(4),dx(5)} Seq Vec
624: ncolumns = {1,0,1,1}
625: columns = {{6},{},{4},{5}}
626: nrows = {3,0,2,2}
627: rows = {{0,1,2},{},{1,2},{1,2}}
628: vwscale = {dx(4),dx(5),dx(6)} MPI Vec
629: vscale = {dx(0),dx(4),dx(5),dx(6)} Seq Vec
631: See the routine MatFDColoringApply() for how this data is used
632: to compute the Jacobian.
634: */
635: typedef struct {
636: PetscInt row;
637: PetscInt col;
638: PetscScalar *valaddr; /* address of value */
639: } MatEntry;
641: typedef struct {
642: PetscInt row;
643: PetscScalar *valaddr; /* address of value */
644: } MatEntry2;
646: struct _p_MatFDColoring {
647: PETSCHEADER(int);
648: PetscInt M, N, m; /* total rows, columns; local rows */
649: PetscInt rstart; /* first row owned by local processor */
650: PetscInt ncolors; /* number of colors */
651: PetscInt *ncolumns; /* number of local columns for a color */
652: PetscInt **columns; /* lists the local columns of each color (using global column numbering) */
653: IS *isa; /* these are the IS that contain the column values given in columns */
654: PetscInt *nrows; /* number of local rows for each color */
655: MatEntry *matentry; /* holds (row, column, address of value) for Jacobian matrix entry */
656: MatEntry2 *matentry2; /* holds (row, address of value) for Jacobian matrix entry */
657: PetscScalar *dy; /* store a block of F(x+dx)-F(x) when J is in BAIJ format */
658: PetscReal error_rel; /* square root of relative error in computing function */
659: PetscReal umin; /* minimum allowable u'dx value */
660: Vec w1, w2, w3; /* work vectors used in computing Jacobian */
661: PetscBool fset; /* indicates that the initial function value F(X) is set */
662: MatFDColoringFn *f; /* function that defines Jacobian */
663: void *fctx; /* optional user-defined context for use by the function f */
664: Vec vscale; /* holds FD scaling, i.e. 1/dx for each perturbed column */
665: PetscInt currentcolor; /* color for which function evaluation is being done now */
666: const char *htype; /* "wp" or "ds" */
667: ISColoringType ctype; /* IS_COLORING_GLOBAL or IS_COLORING_LOCAL */
668: PetscInt brows, bcols; /* number of block rows or columns for speedup inserting the dense matrix into sparse Jacobian */
669: PetscBool setupcalled; /* true if setup has been called */
670: PetscBool viewed; /* true if the -mat_fd_coloring_view has been triggered already */
671: void (*ftn_func_pointer)(void), *ftn_func_cntx; /* serve the same purpose as *fortran_func_pointers in PETSc objects */
672: PetscObjectId matid; /* matrix this object was created with, must always be the same */
673: };
675: typedef struct _MatColoringOps *MatColoringOps;
676: struct _MatColoringOps {
677: PetscErrorCode (*destroy)(MatColoring);
678: PetscErrorCode (*setfromoptions)(MatColoring, PetscOptionItems);
679: PetscErrorCode (*view)(MatColoring, PetscViewer);
680: PetscErrorCode (*apply)(MatColoring, ISColoring *);
681: PetscErrorCode (*weights)(MatColoring, PetscReal **, PetscInt **);
682: };
684: struct _p_MatColoring {
685: PETSCHEADER(struct _MatColoringOps);
686: Mat mat;
687: PetscInt dist; /* distance of the coloring */
688: PetscInt maxcolors; /* the maximum number of colors returned, maxcolors=1 for MIS */
689: void *data; /* inner context */
690: PetscBool valid; /* check to see if what is produced is a valid coloring */
691: MatColoringWeightType weight_type; /* type of weight computation to be performed */
692: PetscReal *user_weights; /* custom weights and permutation */
693: PetscInt *user_lperm;
694: PetscBool valid_iscoloring; /* check to see if matcoloring is produced a valid iscoloring */
695: };
697: struct _p_MatTransposeColoring {
698: PETSCHEADER(int);
699: PetscInt M, N, m; /* total rows, columns; local rows */
700: PetscInt rstart; /* first row owned by local processor */
701: PetscInt ncolors; /* number of colors */
702: PetscInt *ncolumns; /* number of local columns for a color */
703: PetscInt *nrows; /* number of local rows for each color */
704: PetscInt currentcolor; /* color for which function evaluation is being done now */
705: ISColoringType ctype; /* IS_COLORING_GLOBAL or IS_COLORING_LOCAL */
707: PetscInt *colorforrow, *colorforcol; /* pointer to rows and columns */
708: PetscInt *rows; /* lists the local rows for each color (using the local row numbering) */
709: PetscInt *den2sp; /* maps (row,color) in the dense matrix to index of sparse matrix array a->a */
710: PetscInt *columns; /* lists the local columns of each color (using global column numbering) */
711: PetscInt brows; /* number of rows for efficient implementation of MatTransColoringApplyDenToSp() */
712: PetscInt *lstart; /* array used for loop over row blocks of Csparse */
713: };
715: /*
716: Null space context for preconditioner/operators
717: */
718: struct _p_MatNullSpace {
719: PETSCHEADER(int);
720: PetscBool has_cnst;
721: PetscInt n;
722: Vec *vecs;
723: PetscScalar *alpha; /* for projections */
724: MatNullSpaceRemoveFn *remove; /* for user provided removal function */
725: void *rmctx; /* context for remove() function */
726: };
728: /*
729: Internal data structure for MATMPIDENSE
730: */
731: typedef struct {
732: Mat A; /* local submatrix */
734: /* The following variables are used for matrix assembly */
735: PetscBool donotstash; /* Flag indicating if values should be stashed */
736: MPI_Request *send_waits; /* array of send requests */
737: MPI_Request *recv_waits; /* array of receive requests */
738: PetscInt nsends, nrecvs; /* numbers of sends and receives */
739: PetscScalar *svalues, *rvalues; /* sending and receiving data */
740: PetscInt rmax; /* maximum message length */
742: /* The following variables are used for matrix-vector products */
743: Vec lvec; /* local vector */
744: PetscSF Mvctx; /* for mat-mult communications */
745: PetscBool roworiented; /* if true, row-oriented input (default) */
747: /* Support for MatDenseGetColumnVec and MatDenseGetSubMatrix */
748: Mat cmat; /* matrix representation of a given subset of columns */
749: Vec cvec; /* vector representation of a given column */
750: const PetscScalar *ptrinuse; /* holds array to be restored (just a placeholder) */
751: PetscInt vecinuse; /* if cvec is in use (col = vecinuse-1) */
752: PetscInt matinuse; /* if cmat is in use (cbegin = matinuse-1) */
753: /* if this is from MatDenseGetSubMatrix, which columns and rows does it correspond to? */
754: PetscInt sub_rbegin;
755: PetscInt sub_rend;
756: PetscInt sub_cbegin;
757: PetscInt sub_cend;
758: } Mat_MPIDense;
760: /*
761: Checking zero pivot for LU, ILU preconditioners.
762: */
763: typedef struct {
764: PetscInt nshift, nshift_max;
765: PetscReal shift_amount, shift_lo, shift_hi, shift_top, shift_fraction;
766: PetscBool newshift;
767: PetscReal rs; /* active row sum of abs(off-diagonals) */
768: PetscScalar pv; /* pivot of the active row */
769: } FactorShiftCtx;
771: PETSC_SINGLE_LIBRARY_INTERN PetscErrorCode MatTransposeCheckNonzeroState_Private(Mat, Mat);
773: /*
774: Used by MatTranspose() and potentially other functions to track the matrix used in the generation of another matrix
775: */
776: typedef struct {
777: PetscObjectId id;
778: PetscObjectState state;
779: PetscObjectState nonzerostate;
780: } MatParentState;
782: PETSC_EXTERN PetscErrorCode MatFactorDumpMatrix(Mat);
783: PETSC_INTERN PetscErrorCode MatSetBlockSizes_Default(Mat, PetscInt, PetscInt);
785: PETSC_SINGLE_LIBRARY_INTERN PetscErrorCode MatShift_Basic(Mat, PetscScalar);
787: static inline PetscErrorCode MatPivotCheck_nz(PETSC_UNUSED Mat mat, const MatFactorInfo *info, FactorShiftCtx *sctx, PETSC_UNUSED PetscInt row)
788: {
789: PetscReal _rs = sctx->rs;
790: PetscReal _zero = info->zeropivot * _rs;
792: PetscFunctionBegin;
793: if (PetscAbsScalar(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)) {
794: /* force |diag| > zeropivot*rs */
795: if (!sctx->nshift) sctx->shift_amount = info->shiftamount;
796: else sctx->shift_amount *= 2.0;
797: sctx->newshift = PETSC_TRUE;
798: (sctx->nshift)++;
799: } else {
800: sctx->newshift = PETSC_FALSE;
801: }
802: PetscFunctionReturn(PETSC_SUCCESS);
803: }
805: static inline PetscErrorCode MatPivotCheck_pd(PETSC_UNUSED Mat mat, const MatFactorInfo *info, FactorShiftCtx *sctx, PETSC_UNUSED PetscInt row)
806: {
807: PetscReal _rs = sctx->rs;
808: PetscReal _zero = info->zeropivot * _rs;
810: PetscFunctionBegin;
811: if (PetscRealPart(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)) {
812: /* force matfactor to be diagonally dominant */
813: if (sctx->nshift == sctx->nshift_max) {
814: sctx->shift_fraction = sctx->shift_hi;
815: } else {
816: sctx->shift_lo = sctx->shift_fraction;
817: sctx->shift_fraction = (sctx->shift_hi + sctx->shift_lo) / (PetscReal)2.;
818: }
819: sctx->shift_amount = sctx->shift_fraction * sctx->shift_top;
820: sctx->nshift++;
821: sctx->newshift = PETSC_TRUE;
822: } else {
823: sctx->newshift = PETSC_FALSE;
824: }
825: PetscFunctionReturn(PETSC_SUCCESS);
826: }
828: static inline PetscErrorCode MatPivotCheck_inblocks(PETSC_UNUSED Mat mat, const MatFactorInfo *info, FactorShiftCtx *sctx, PETSC_UNUSED PetscInt row)
829: {
830: PetscReal _zero = info->zeropivot;
832: PetscFunctionBegin;
833: if (PetscAbsScalar(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)) {
834: sctx->pv += info->shiftamount;
835: sctx->shift_amount = 0.0;
836: sctx->nshift++;
837: }
838: sctx->newshift = PETSC_FALSE;
839: PetscFunctionReturn(PETSC_SUCCESS);
840: }
842: static inline PetscErrorCode MatPivotCheck_none(Mat fact, Mat mat, const MatFactorInfo *info, FactorShiftCtx *sctx, PetscInt row)
843: {
844: PetscReal _zero = info->zeropivot;
846: PetscFunctionBegin;
847: sctx->newshift = PETSC_FALSE;
848: if (PetscAbsScalar(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)) {
849: 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);
850: PetscCall(PetscInfo(mat, "Detected zero pivot in factorization in row %" PetscInt_FMT " value %g tolerance %g\n", row, (double)PetscAbsScalar(sctx->pv), (double)_zero));
851: fact->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT;
852: fact->factorerror_zeropivot_value = PetscAbsScalar(sctx->pv);
853: fact->factorerror_zeropivot_row = row;
854: }
855: PetscFunctionReturn(PETSC_SUCCESS);
856: }
858: static inline PetscErrorCode MatPivotCheck(Mat fact, Mat mat, const MatFactorInfo *info, FactorShiftCtx *sctx, PetscInt row)
859: {
860: PetscFunctionBegin;
861: if (info->shifttype == (PetscReal)MAT_SHIFT_NONZERO) PetscCall(MatPivotCheck_nz(mat, info, sctx, row));
862: else if (info->shifttype == (PetscReal)MAT_SHIFT_POSITIVE_DEFINITE) PetscCall(MatPivotCheck_pd(mat, info, sctx, row));
863: else if (info->shifttype == (PetscReal)MAT_SHIFT_INBLOCKS) PetscCall(MatPivotCheck_inblocks(mat, info, sctx, row));
864: else PetscCall(MatPivotCheck_none(fact, mat, info, sctx, row));
865: PetscFunctionReturn(PETSC_SUCCESS);
866: }
868: #include <petscbt.h>
869: /*
870: Create and initialize a linked list
871: Input Parameters:
872: idx_start - starting index of the list
873: lnk_max - max value of lnk indicating the end of the list
874: nlnk - max length of the list
875: Output Parameters:
876: lnk - list initialized
877: bt - PetscBT (bitarray) with all bits set to false
878: lnk_empty - flg indicating the list is empty
879: */
880: #define PetscLLCreate(idx_start, lnk_max, nlnk, lnk, bt) ((PetscErrorCode)(PetscMalloc1(nlnk, &lnk) || PetscBTCreate(nlnk, &(bt)) || (lnk[idx_start] = lnk_max, PETSC_SUCCESS)))
882: #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)))
884: 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)
885: {
886: PetscInt location;
888: PetscFunctionBegin;
889: /* start from the beginning if entry < previous entry */
890: if (!assume_sorted && k && entry < *lnkdata) *lnkdata = idx_start;
891: /* search for insertion location */
892: do {
893: location = *lnkdata;
894: *lnkdata = lnk[location];
895: } while (entry > *lnkdata);
896: /* insertion location is found, add entry into lnk */
897: lnk[location] = entry;
898: lnk[entry] = *lnkdata;
899: ++(*nlnk);
900: *lnkdata = entry; /* next search starts from here if next_entry > entry */
901: PetscFunctionReturn(PETSC_SUCCESS);
902: }
904: 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)
905: {
906: PetscFunctionBegin;
907: *nlnk = 0;
908: for (PetscInt k = 0, lnkdata = idx_start; k < nidx; ++k) {
909: const PetscInt entry = indices[k];
911: if (!PetscBTLookupSet(bt, entry)) PetscCall(PetscLLInsertLocation_Private(assume_sorted, k, idx_start, entry, nlnk, &lnkdata, lnk));
912: }
913: PetscFunctionReturn(PETSC_SUCCESS);
914: }
916: /*
917: Add an index set into a sorted linked list
918: Input Parameters:
919: nidx - number of input indices
920: indices - integer array
921: idx_start - starting index of the list
922: lnk - linked list(an integer array) that is created
923: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
924: output Parameters:
925: nlnk - number of newly added indices
926: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
927: bt - updated PetscBT (bitarray)
928: */
929: static inline PetscErrorCode PetscLLAdd(PetscInt nidx, const PetscInt *PETSC_RESTRICT indices, PetscInt idx_start, PetscInt *PETSC_RESTRICT nlnk, PetscInt *PETSC_RESTRICT lnk, PetscBT bt)
930: {
931: PetscFunctionBegin;
932: PetscCall(PetscLLAdd_Private(nidx, indices, idx_start, nlnk, lnk, bt, PETSC_FALSE));
933: PetscFunctionReturn(PETSC_SUCCESS);
934: }
936: /*
937: Add a SORTED ascending index set into a sorted linked list - same as PetscLLAdd() bus skip 'if (_k && _entry < _lnkdata) _lnkdata = idx_start;'
938: Input Parameters:
939: nidx - number of input indices
940: indices - sorted integer array
941: idx_start - starting index of the list
942: lnk - linked list(an integer array) that is created
943: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
944: output Parameters:
945: nlnk - number of newly added indices
946: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
947: bt - updated PetscBT (bitarray)
948: */
949: static inline PetscErrorCode PetscLLAddSorted(PetscInt nidx, const PetscInt *PETSC_RESTRICT indices, PetscInt idx_start, PetscInt *PETSC_RESTRICT nlnk, PetscInt *PETSC_RESTRICT lnk, PetscBT bt)
950: {
951: PetscFunctionBegin;
952: PetscCall(PetscLLAdd_Private(nidx, indices, idx_start, nlnk, lnk, bt, PETSC_TRUE));
953: PetscFunctionReturn(PETSC_SUCCESS);
954: }
956: /*
957: Add a permuted index set into a sorted linked list
958: Input Parameters:
959: nidx - number of input indices
960: indices - integer array
961: perm - permutation of indices
962: idx_start - starting index of the list
963: lnk - linked list(an integer array) that is created
964: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
965: output Parameters:
966: nlnk - number of newly added indices
967: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
968: bt - updated PetscBT (bitarray)
969: */
970: 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)
971: {
972: PetscFunctionBegin;
973: *nlnk = 0;
974: for (PetscInt k = 0, lnkdata = idx_start; k < nidx; ++k) {
975: const PetscInt entry = perm[indices[k]];
977: if (!PetscBTLookupSet(bt, entry)) PetscCall(PetscLLInsertLocation_Private(PETSC_FALSE, k, idx_start, entry, nlnk, &lnkdata, lnk));
978: }
979: PetscFunctionReturn(PETSC_SUCCESS);
980: }
982: #if 0
983: /* this appears to be unused? */
984: static inline PetscErrorCode PetscLLAddSorted_new(PetscInt nidx, PetscInt *indices, PetscInt idx_start, PetscBool *lnk_empty, PetscInt *nlnk, PetscInt *lnk, PetscBT bt)
985: {
986: PetscInt lnkdata = idx_start;
988: PetscFunctionBegin;
989: if (*lnk_empty) {
990: for (PetscInt k = 0; k < nidx; ++k) {
991: const PetscInt entry = indices[k], location = lnkdata;
993: PetscCall(PetscBTSet(bt,entry)); /* mark the new entry */
994: lnkdata = lnk[location];
995: /* insertion location is found, add entry into lnk */
996: lnk[location] = entry;
997: lnk[entry] = lnkdata;
998: lnkdata = entry; /* next search starts from here */
999: }
1000: /* lnk[indices[nidx-1]] = lnk[idx_start];
1001: lnk[idx_start] = indices[0];
1002: PetscCall(PetscBTSet(bt,indices[0]));
1003: for (_k=1; _k<nidx; _k++) {
1004: PetscCall(PetscBTSet(bt,indices[_k]));
1005: lnk[indices[_k-1]] = indices[_k];
1006: }
1007: */
1008: *nlnk = nidx;
1009: *lnk_empty = PETSC_FALSE;
1010: } else {
1011: *nlnk = 0;
1012: for (PetscInt k = 0; k < nidx; ++k) {
1013: const PetscInt entry = indices[k];
1015: if (!PetscBTLookupSet(bt,entry)) PetscCall(PetscLLInsertLocation_Private(PETSC_TRUE,k,idx_start,entry,nlnk,&lnkdata,lnk));
1016: }
1017: }
1018: PetscFunctionReturn(PETSC_SUCCESS);
1019: }
1020: #endif
1022: /*
1023: Add a SORTED index set into a sorted linked list used for LUFactorSymbolic()
1024: Same as PetscLLAddSorted() with an additional operation:
1025: count the number of input indices that are no larger than 'diag'
1026: Input Parameters:
1027: indices - sorted integer array
1028: idx_start - starting index of the list, index of pivot row
1029: lnk - linked list(an integer array) that is created
1030: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1031: diag - index of the active row in LUFactorSymbolic
1032: nzbd - number of input indices with indices <= idx_start
1033: im - im[idx_start] is initialized as num of nonzero entries in row=idx_start
1034: output Parameters:
1035: nlnk - number of newly added indices
1036: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
1037: bt - updated PetscBT (bitarray)
1038: im - im[idx_start]: unchanged if diag is not an entry
1039: : num of entries with indices <= diag if diag is an entry
1040: */
1041: 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)
1042: {
1043: const PetscInt nidx = im[idx_start] - nzbd; /* num of entries with idx_start < index <= diag */
1045: PetscFunctionBegin;
1046: *nlnk = 0;
1047: for (PetscInt k = 0, lnkdata = idx_start; k < nidx; ++k) {
1048: const PetscInt entry = indices[k];
1050: ++nzbd;
1051: if (entry == diag) im[idx_start] = nzbd;
1052: if (!PetscBTLookupSet(bt, entry)) PetscCall(PetscLLInsertLocation_Private(PETSC_TRUE, k, idx_start, entry, nlnk, &lnkdata, lnk));
1053: }
1054: PetscFunctionReturn(PETSC_SUCCESS);
1055: }
1057: /*
1058: Copy data on the list into an array, then initialize the list
1059: Input Parameters:
1060: idx_start - starting index of the list
1061: lnk_max - max value of lnk indicating the end of the list
1062: nlnk - number of data on the list to be copied
1063: lnk - linked list
1064: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1065: output Parameters:
1066: indices - array that contains the copied data
1067: lnk - linked list that is cleaned and initialize
1068: bt - PetscBT (bitarray) with all bits set to false
1069: */
1070: static inline PetscErrorCode PetscLLClean(PetscInt idx_start, PetscInt lnk_max, PetscInt nlnk, PetscInt *PETSC_RESTRICT lnk, PetscInt *PETSC_RESTRICT indices, PetscBT bt)
1071: {
1072: PetscFunctionBegin;
1073: for (PetscInt j = 0, idx = idx_start; j < nlnk; ++j) {
1074: idx = lnk[idx];
1075: indices[j] = idx;
1076: PetscCall(PetscBTClear(bt, idx));
1077: }
1078: lnk[idx_start] = lnk_max;
1079: PetscFunctionReturn(PETSC_SUCCESS);
1080: }
1082: /*
1083: Free memories used by the list
1084: */
1085: #define PetscLLDestroy(lnk, bt) ((PetscErrorCode)(PetscFree(lnk) || PetscBTDestroy(&(bt))))
1087: /* Routines below are used for incomplete matrix factorization */
1088: /*
1089: Create and initialize a linked list and its levels
1090: Input Parameters:
1091: idx_start - starting index of the list
1092: lnk_max - max value of lnk indicating the end of the list
1093: nlnk - max length of the list
1094: Output Parameters:
1095: lnk - list initialized
1096: lnk_lvl - array of size nlnk for storing levels of lnk
1097: bt - PetscBT (bitarray) with all bits set to false
1098: */
1099: #define PetscIncompleteLLCreate(idx_start, lnk_max, nlnk, lnk, lnk_lvl, bt) \
1100: ((PetscErrorCode)(PetscIntMultError(2, nlnk, NULL) || PetscMalloc1(2 * nlnk, &lnk) || PetscBTCreate(nlnk, &(bt)) || (lnk[idx_start] = lnk_max, lnk_lvl = lnk + nlnk, PETSC_SUCCESS)))
1102: 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)
1103: {
1104: PetscFunctionBegin;
1105: PetscCall(PetscLLInsertLocation_Private(assume_sorted, k, idx_start, entry, nlnk, lnkdata, lnk));
1106: lnklvl[entry] = newval;
1107: PetscFunctionReturn(PETSC_SUCCESS);
1108: }
1110: /*
1111: Initialize a sorted linked list used for ILU and ICC
1112: Input Parameters:
1113: nidx - number of input idx
1114: idx - integer array used for storing column indices
1115: idx_start - starting index of the list
1116: perm - indices of an IS
1117: lnk - linked list(an integer array) that is created
1118: lnklvl - levels of lnk
1119: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1120: output Parameters:
1121: nlnk - number of newly added idx
1122: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1123: lnklvl - levels of lnk
1124: bt - updated PetscBT (bitarray)
1125: */
1126: 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)
1127: {
1128: PetscFunctionBegin;
1129: *nlnk = 0;
1130: for (PetscInt k = 0, lnkdata = idx_start; k < nidx; ++k) {
1131: const PetscInt entry = perm[idx[k]];
1133: if (!PetscBTLookupSet(bt, entry)) PetscCall(PetscIncompleteLLInsertLocation_Private(PETSC_FALSE, k, idx_start, entry, nlnk, &lnkdata, lnk, lnklvl, 0));
1134: }
1135: PetscFunctionReturn(PETSC_SUCCESS);
1136: }
1138: 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)
1139: {
1140: PetscFunctionBegin;
1141: *nlnk = 0;
1142: for (PetscInt k = 0, lnkdata = idx_start; k < nidx; ++k) {
1143: const PetscInt incrlev = idxlvl[k] + prow_offset + 1;
1145: if (incrlev <= level) {
1146: const PetscInt entry = idx[k];
1148: if (!PetscBTLookupSet(bt, entry)) PetscCall(PetscIncompleteLLInsertLocation_Private(assume_sorted, k, idx_start, entry, nlnk, &lnkdata, lnk, lnklvl, incrlev));
1149: else if (lnklvl[entry] > incrlev) lnklvl[entry] = incrlev; /* existing entry */
1150: }
1151: }
1152: PetscFunctionReturn(PETSC_SUCCESS);
1153: }
1155: /*
1156: Add a SORTED index set into a sorted linked list for ICC
1157: Input Parameters:
1158: nidx - number of input indices
1159: idx - sorted integer array used for storing column indices
1160: level - level of fill, e.g., ICC(level)
1161: idxlvl - level of idx
1162: idx_start - starting index of the list
1163: lnk - linked list(an integer array) that is created
1164: lnklvl - levels of lnk
1165: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1166: idxlvl_prow - idxlvl[prow], where prow is the row number of the idx
1167: output Parameters:
1168: nlnk - number of newly added indices
1169: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1170: lnklvl - levels of lnk
1171: bt - updated PetscBT (bitarray)
1172: Note: the level of U(i,j) is set as lvl(i,j) = min{ lvl(i,j), lvl(prow,i)+lvl(prow,j)+1)
1173: where idx = non-zero columns of U(prow,prow+1:n-1), prow<i
1174: */
1175: 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)
1176: {
1177: PetscFunctionBegin;
1178: PetscCall(PetscIncompleteLLAdd_Private(nidx, idx, level, idxlvl, idx_start, nlnk, lnk, lnklvl, bt, idxlvl_prow, PETSC_TRUE));
1179: PetscFunctionReturn(PETSC_SUCCESS);
1180: }
1182: /*
1183: Add a SORTED index set into a sorted linked list for ILU
1184: Input Parameters:
1185: nidx - number of input indices
1186: idx - sorted integer array used for storing column indices
1187: level - level of fill, e.g., ICC(level)
1188: idxlvl - level of idx
1189: idx_start - starting index of the list
1190: lnk - linked list(an integer array) that is created
1191: lnklvl - levels of lnk
1192: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1193: prow - the row number of idx
1194: output Parameters:
1195: nlnk - number of newly added idx
1196: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1197: lnklvl - levels of lnk
1198: bt - updated PetscBT (bitarray)
1200: Note: the level of factor(i,j) is set as lvl(i,j) = min{ lvl(i,j), lvl(i,prow)+lvl(prow,j)+1)
1201: where idx = non-zero columns of U(prow,prow+1:n-1), prow<i
1202: */
1203: 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)
1204: {
1205: PetscFunctionBegin;
1206: PetscCall(PetscIncompleteLLAdd_Private(nidx, idx, level, idxlvl, idx_start, nlnk, lnk, lnklvl, bt, lnklvl[prow], PETSC_TRUE));
1207: PetscFunctionReturn(PETSC_SUCCESS);
1208: }
1210: /*
1211: Add a index set into a sorted linked list
1212: Input Parameters:
1213: nidx - number of input idx
1214: idx - integer array used for storing column indices
1215: level - level of fill, e.g., ICC(level)
1216: idxlvl - level of idx
1217: idx_start - starting index of the list
1218: lnk - linked list(an integer array) that is created
1219: lnklvl - levels of lnk
1220: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1221: output Parameters:
1222: nlnk - number of newly added idx
1223: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1224: lnklvl - levels of lnk
1225: bt - updated PetscBT (bitarray)
1226: */
1227: 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)
1228: {
1229: PetscFunctionBegin;
1230: PetscCall(PetscIncompleteLLAdd_Private(nidx, idx, level, idxlvl, idx_start, nlnk, lnk, lnklvl, bt, 0, PETSC_FALSE));
1231: PetscFunctionReturn(PETSC_SUCCESS);
1232: }
1234: /*
1235: Add a SORTED index set into a sorted linked list
1236: Input Parameters:
1237: nidx - number of input indices
1238: idx - sorted integer array used for storing column indices
1239: level - level of fill, e.g., ICC(level)
1240: idxlvl - level of idx
1241: idx_start - starting index of the list
1242: lnk - linked list(an integer array) that is created
1243: lnklvl - levels of lnk
1244: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1245: output Parameters:
1246: nlnk - number of newly added idx
1247: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1248: lnklvl - levels of lnk
1249: bt - updated PetscBT (bitarray)
1250: */
1251: 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)
1252: {
1253: PetscFunctionBegin;
1254: PetscCall(PetscIncompleteLLAdd_Private(nidx, idx, level, idxlvl, idx_start, nlnk, lnk, lnklvl, bt, 0, PETSC_TRUE));
1255: PetscFunctionReturn(PETSC_SUCCESS);
1256: }
1258: /*
1259: Copy data on the list into an array, then initialize the list
1260: Input Parameters:
1261: idx_start - starting index of the list
1262: lnk_max - max value of lnk indicating the end of the list
1263: nlnk - number of data on the list to be copied
1264: lnk - linked list
1265: lnklvl - level of lnk
1266: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1267: output Parameters:
1268: indices - array that contains the copied data
1269: lnk - linked list that is cleaned and initialize
1270: lnklvl - level of lnk that is reinitialized
1271: bt - PetscBT (bitarray) with all bits set to false
1272: */
1273: 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)
1274: {
1275: PetscFunctionBegin;
1276: for (PetscInt j = 0, idx = idx_start; j < nlnk; ++j) {
1277: idx = lnk[idx];
1278: indices[j] = idx;
1279: indiceslvl[j] = lnklvl[idx];
1280: lnklvl[idx] = -1;
1281: PetscCall(PetscBTClear(bt, idx));
1282: }
1283: lnk[idx_start] = lnk_max;
1284: PetscFunctionReturn(PETSC_SUCCESS);
1285: }
1287: /*
1288: Free memories used by the list
1289: */
1290: #define PetscIncompleteLLDestroy(lnk, bt) ((PetscErrorCode)(PetscFree(lnk) || PetscBTDestroy(&(bt))))
1292: #if !defined(PETSC_CLANG_STATIC_ANALYZER)
1293: #define MatCheckSameLocalSize(A, ar1, B, ar2) \
1294: do { \
1295: PetscCheckSameComm(A, ar1, B, ar2); \
1296: 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, \
1297: (A)->rmap->n, (A)->cmap->n, ar2, (B)->rmap->n, (B)->cmap->n); \
1298: } while (0)
1299: #define MatCheckSameSize(A, ar1, B, ar2) \
1300: do { \
1301: 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, \
1302: (A)->rmap->N, (A)->cmap->N, ar2, (B)->rmap->N, (B)->cmap->N); \
1303: MatCheckSameLocalSize(A, ar1, B, ar2); \
1304: } while (0)
1305: #else
1306: template <typename Tm>
1307: extern void MatCheckSameLocalSize(Tm, int, Tm, int);
1308: template <typename Tm>
1309: extern void MatCheckSameSize(Tm, int, Tm, int);
1310: #endif
1312: #define VecCheckMatCompatible(M, x, ar1, b, ar2) \
1313: do { \
1314: 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, \
1315: (M)->cmap->N); \
1316: 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, \
1317: (M)->rmap->N); \
1318: } while (0)
1320: /*
1321: Create and initialize a condensed linked list -
1322: same as PetscLLCreate(), but uses a scalable array 'lnk' with size of max number of entries, not O(N).
1323: Barry suggested this approach (Dec. 6, 2011):
1324: I've thought of an alternative way of representing a linked list that is efficient but doesn't have the O(N) scaling issue
1325: (it may be faster than the O(N) even sequentially due to less crazy memory access).
1327: 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
1328: 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
1329: 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
1330: 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.
1331: 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
1332: to each other so memory access is much better than using the big array.
1334: Example:
1335: nlnk_max=5, lnk_max=36:
1336: Initial list: [0, 0 | 36, 2 | 0, 0 | 0, 0 | 0, 0 | 0, 0 | 0, 0]
1337: here, head_node has index 2 with value lnk[2]=lnk_max=36,
1338: 0-th entry is used to store the number of entries in the list,
1339: The initial lnk represents head -> tail(marked by 36) with number of entries = lnk[0]=0.
1341: Now adding a sorted set {2,4}, the list becomes
1342: [2, 0 | 36, 4 |2, 6 | 4, 2 | 0, 0 | 0, 0 | 0, 0 ]
1343: represents head -> 2 -> 4 -> tail with number of entries = lnk[0]=2.
1345: Then adding a sorted set {0,3,35}, the list
1346: [5, 0 | 36, 8 | 2, 10 | 4, 12 | 0, 4 | 3, 6 | 35, 2 ]
1347: represents head -> 0 -> 2 -> 3 -> 4 -> 35 -> tail with number of entries = lnk[0]=5.
1349: Input Parameters:
1350: nlnk_max - max length of the list
1351: lnk_max - max value of the entries
1352: Output Parameters:
1353: lnk - list created and initialized
1354: bt - PetscBT (bitarray) with all bits set to false. Note: bt has size lnk_max, not nln_max!
1355: */
1356: static inline PetscErrorCode PetscLLCondensedCreate(PetscInt nlnk_max, PetscInt lnk_max, PetscInt **lnk, PetscBT *bt)
1357: {
1358: PetscInt *llnk, lsize = 0;
1360: PetscFunctionBegin;
1361: PetscCall(PetscIntMultError(2, nlnk_max + 2, &lsize));
1362: PetscCall(PetscMalloc1(lsize, lnk));
1363: PetscCall(PetscBTCreate(lnk_max, bt));
1364: llnk = *lnk;
1365: llnk[0] = 0; /* number of entries on the list */
1366: llnk[2] = lnk_max; /* value in the head node */
1367: llnk[3] = 2; /* next for the head node */
1368: PetscFunctionReturn(PETSC_SUCCESS);
1369: }
1371: /*
1372: Add a SORTED ascending index set into a sorted linked list. See PetscLLCondensedCreate() for detailed description.
1373: Input Parameters:
1374: nidx - number of input indices
1375: indices - sorted integer array
1376: lnk - condensed linked list(an integer array) that is created
1377: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1378: output Parameters:
1379: lnk - the sorted(increasing order) linked list containing previous and newly added non-redundate indices
1380: bt - updated PetscBT (bitarray)
1381: */
1382: static inline PetscErrorCode PetscLLCondensedAddSorted(PetscInt nidx, const PetscInt indices[], PetscInt lnk[], PetscBT bt)
1383: {
1384: PetscInt location = 2; /* head */
1385: PetscInt nlnk = lnk[0]; /* num of entries on the input lnk */
1387: PetscFunctionBegin;
1388: for (PetscInt k = 0; k < nidx; k++) {
1389: const PetscInt entry = indices[k];
1390: if (!PetscBTLookupSet(bt, entry)) { /* new entry */
1391: PetscInt next, lnkdata;
1393: /* search for insertion location */
1394: do {
1395: next = location + 1; /* link from previous node to next node */
1396: location = lnk[next]; /* idx of next node */
1397: lnkdata = lnk[location]; /* value of next node */
1398: } while (entry > lnkdata);
1399: /* insertion location is found, add entry into lnk */
1400: const PetscInt newnode = 2 * (nlnk + 2); /* index for this new node */
1401: lnk[next] = newnode; /* connect previous node to the new node */
1402: lnk[newnode] = entry; /* set value of the new node */
1403: lnk[newnode + 1] = location; /* connect new node to next node */
1404: location = newnode; /* next search starts from the new node */
1405: nlnk++;
1406: }
1407: }
1408: lnk[0] = nlnk; /* number of entries in the list */
1409: PetscFunctionReturn(PETSC_SUCCESS);
1410: }
1412: static inline PetscErrorCode PetscLLCondensedClean(PetscInt lnk_max, PETSC_UNUSED PetscInt nidx, PetscInt *indices, PetscInt lnk[], PetscBT bt)
1413: {
1414: const PetscInt nlnk = lnk[0]; /* num of entries on the list */
1415: PetscInt next = lnk[3]; /* head node */
1417: PetscFunctionBegin;
1418: for (PetscInt k = 0; k < nlnk; k++) {
1419: indices[k] = lnk[next];
1420: next = lnk[next + 1];
1421: PetscCall(PetscBTClear(bt, indices[k]));
1422: }
1423: lnk[0] = 0; /* num of entries on the list */
1424: lnk[2] = lnk_max; /* initialize head node */
1425: lnk[3] = 2; /* head node */
1426: PetscFunctionReturn(PETSC_SUCCESS);
1427: }
1429: static inline PetscErrorCode PetscLLCondensedView(PetscInt *lnk)
1430: {
1431: PetscFunctionBegin;
1432: PetscCall(PetscPrintf(PETSC_COMM_SELF, "LLCondensed of size %" PetscInt_FMT ", (val, next)\n", lnk[0]));
1433: 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]));
1434: PetscFunctionReturn(PETSC_SUCCESS);
1435: }
1437: /*
1438: Free memories used by the list
1439: */
1440: static inline PetscErrorCode PetscLLCondensedDestroy(PetscInt *lnk, PetscBT bt)
1441: {
1442: PetscFunctionBegin;
1443: PetscCall(PetscFree(lnk));
1444: PetscCall(PetscBTDestroy(&bt));
1445: PetscFunctionReturn(PETSC_SUCCESS);
1446: }
1448: /*
1449: Same as PetscLLCondensedCreate(), but does not use non-scalable O(lnk_max) bitarray
1450: Input Parameters:
1451: nlnk_max - max length of the list
1452: Output Parameters:
1453: lnk - list created and initialized
1454: */
1455: static inline PetscErrorCode PetscLLCondensedCreate_Scalable(PetscInt nlnk_max, PetscInt **lnk)
1456: {
1457: PetscInt *llnk, lsize = 0;
1459: PetscFunctionBegin;
1460: PetscCall(PetscIntMultError(2, nlnk_max + 2, &lsize));
1461: PetscCall(PetscMalloc1(lsize, lnk));
1462: llnk = *lnk;
1463: llnk[0] = 0; /* number of entries on the list */
1464: llnk[2] = PETSC_INT_MAX; /* value in the head node */
1465: llnk[3] = 2; /* next for the head node */
1466: PetscFunctionReturn(PETSC_SUCCESS);
1467: }
1469: static inline PetscErrorCode PetscLLCondensedExpand_Scalable(PetscInt nlnk_max, PetscInt **lnk)
1470: {
1471: PetscInt lsize = 0;
1473: PetscFunctionBegin;
1474: PetscCall(PetscIntMultError(2, nlnk_max + 2, &lsize));
1475: PetscCall(PetscRealloc(lsize * sizeof(PetscInt), lnk));
1476: PetscFunctionReturn(PETSC_SUCCESS);
1477: }
1479: static inline PetscErrorCode PetscLLCondensedAddSorted_Scalable(PetscInt nidx, const PetscInt indices[], PetscInt lnk[])
1480: {
1481: PetscInt location = 2; /* head */
1482: PetscInt nlnk = lnk[0]; /* num of entries on the input lnk */
1484: for (PetscInt k = 0; k < nidx; k++) {
1485: const PetscInt entry = indices[k];
1486: PetscInt next, lnkdata;
1488: /* search for insertion location */
1489: do {
1490: next = location + 1; /* link from previous node to next node */
1491: location = lnk[next]; /* idx of next node */
1492: lnkdata = lnk[location]; /* value of next node */
1493: } while (entry > lnkdata);
1494: if (entry < lnkdata) {
1495: /* insertion location is found, add entry into lnk */
1496: const PetscInt newnode = 2 * (nlnk + 2); /* index for this new node */
1497: lnk[next] = newnode; /* connect previous node to the new node */
1498: lnk[newnode] = entry; /* set value of the new node */
1499: lnk[newnode + 1] = location; /* connect new node to next node */
1500: location = newnode; /* next search starts from the new node */
1501: nlnk++;
1502: }
1503: }
1504: lnk[0] = nlnk; /* number of entries in the list */
1505: return PETSC_SUCCESS;
1506: }
1508: static inline PetscErrorCode PetscLLCondensedClean_Scalable(PETSC_UNUSED PetscInt nidx, PetscInt *indices, PetscInt *lnk)
1509: {
1510: const PetscInt nlnk = lnk[0];
1511: PetscInt next = lnk[3]; /* head node */
1513: for (PetscInt k = 0; k < nlnk; k++) {
1514: indices[k] = lnk[next];
1515: next = lnk[next + 1];
1516: }
1517: lnk[0] = 0; /* num of entries on the list */
1518: lnk[3] = 2; /* head node */
1519: return PETSC_SUCCESS;
1520: }
1522: static inline PetscErrorCode PetscLLCondensedDestroy_Scalable(PetscInt *lnk)
1523: {
1524: return PetscFree(lnk);
1525: }
1527: /*
1528: lnk[0] number of links
1529: lnk[1] number of entries
1530: lnk[3n] value
1531: lnk[3n+1] len
1532: lnk[3n+2] link to next value
1534: The next three are always the first link
1536: lnk[3] PETSC_INT_MIN+1
1537: lnk[4] 1
1538: lnk[5] link to first real entry
1540: The next three are always the last link
1542: lnk[6] PETSC_INT_MAX - 1
1543: lnk[7] 1
1544: lnk[8] next valid link (this is the same as lnk[0] but without the decreases)
1545: */
1547: static inline PetscErrorCode PetscLLCondensedCreate_fast(PetscInt nlnk_max, PetscInt **lnk)
1548: {
1549: PetscInt *llnk;
1550: PetscInt lsize = 0;
1552: PetscFunctionBegin;
1553: PetscCall(PetscIntMultError(3, nlnk_max + 3, &lsize));
1554: PetscCall(PetscMalloc1(lsize, lnk));
1555: llnk = *lnk;
1556: llnk[0] = 0; /* nlnk: number of entries on the list */
1557: llnk[1] = 0; /* number of integer entries represented in list */
1558: llnk[3] = PETSC_INT_MIN + 1; /* value in the first node */
1559: llnk[4] = 1; /* count for the first node */
1560: llnk[5] = 6; /* next for the first node */
1561: llnk[6] = PETSC_INT_MAX - 1; /* value in the last node */
1562: llnk[7] = 1; /* count for the last node */
1563: llnk[8] = 0; /* next valid node to be used */
1564: PetscFunctionReturn(PETSC_SUCCESS);
1565: }
1567: static inline PetscErrorCode PetscLLCondensedAddSorted_fast(PetscInt nidx, const PetscInt indices[], PetscInt lnk[])
1568: {
1569: for (PetscInt k = 0, prev = 3 /* first value */; k < nidx; k++) {
1570: const PetscInt entry = indices[k];
1571: PetscInt next = lnk[prev + 2];
1573: /* search for insertion location */
1574: while (entry >= lnk[next]) {
1575: prev = next;
1576: next = lnk[next + 2];
1577: }
1578: /* entry is in range of previous list */
1579: if (entry < lnk[prev] + lnk[prev + 1]) continue;
1580: lnk[1]++;
1581: /* entry is right after previous list */
1582: if (entry == lnk[prev] + lnk[prev + 1]) {
1583: lnk[prev + 1]++;
1584: if (lnk[next] == entry + 1) { /* combine two contiguous strings */
1585: lnk[prev + 1] += lnk[next + 1];
1586: lnk[prev + 2] = lnk[next + 2];
1587: next = lnk[next + 2];
1588: lnk[0]--;
1589: }
1590: continue;
1591: }
1592: /* entry is right before next list */
1593: if (entry == lnk[next] - 1) {
1594: lnk[next]--;
1595: lnk[next + 1]++;
1596: prev = next;
1597: next = lnk[prev + 2];
1598: continue;
1599: }
1600: /* add entry into lnk */
1601: lnk[prev + 2] = 3 * ((lnk[8]++) + 3); /* connect previous node to the new node */
1602: prev = lnk[prev + 2];
1603: lnk[prev] = entry; /* set value of the new node */
1604: lnk[prev + 1] = 1; /* number of values in contiguous string is one to start */
1605: lnk[prev + 2] = next; /* connect new node to next node */
1606: lnk[0]++;
1607: }
1608: return PETSC_SUCCESS;
1609: }
1611: static inline PetscErrorCode PetscLLCondensedClean_fast(PETSC_UNUSED PetscInt nidx, PetscInt *indices, PetscInt *lnk)
1612: {
1613: const PetscInt nlnk = lnk[0];
1614: PetscInt next = lnk[5]; /* first node */
1616: for (PetscInt k = 0, cnt = 0; k < nlnk; k++) {
1617: for (PetscInt j = 0; j < lnk[next + 1]; j++) indices[cnt++] = lnk[next] + j;
1618: next = lnk[next + 2];
1619: }
1620: lnk[0] = 0; /* nlnk: number of links */
1621: lnk[1] = 0; /* number of integer entries represented in list */
1622: lnk[3] = PETSC_INT_MIN + 1; /* value in the first node */
1623: lnk[4] = 1; /* count for the first node */
1624: lnk[5] = 6; /* next for the first node */
1625: lnk[6] = PETSC_INT_MAX - 1; /* value in the last node */
1626: lnk[7] = 1; /* count for the last node */
1627: lnk[8] = 0; /* next valid location to make link */
1628: return PETSC_SUCCESS;
1629: }
1631: static inline PetscErrorCode PetscLLCondensedView_fast(const PetscInt *lnk)
1632: {
1633: const PetscInt nlnk = lnk[0];
1634: PetscInt next = lnk[5]; /* first node */
1636: for (PetscInt k = 0; k < nlnk; k++) {
1637: #if 0 /* Debugging code */
1638: printf("%d value %d len %d next %d\n", next, lnk[next], lnk[next + 1], lnk[next + 2]);
1639: #endif
1640: next = lnk[next + 2];
1641: }
1642: return PETSC_SUCCESS;
1643: }
1645: static inline PetscErrorCode PetscLLCondensedDestroy_fast(PetscInt *lnk)
1646: {
1647: return PetscFree(lnk);
1648: }
1650: PETSC_EXTERN PetscErrorCode PetscCDCreate(PetscInt, PetscCoarsenData **);
1651: PETSC_EXTERN PetscErrorCode PetscCDDestroy(PetscCoarsenData *);
1652: PETSC_EXTERN PetscErrorCode PetscCDIntNdSetID(PetscCDIntNd *, PetscInt);
1653: PETSC_EXTERN PetscErrorCode PetscCDIntNdGetID(const PetscCDIntNd *, PetscInt *);
1654: PETSC_EXTERN PetscErrorCode PetscCDAppendID(PetscCoarsenData *, PetscInt, PetscInt);
1655: PETSC_EXTERN PetscErrorCode PetscCDMoveAppend(PetscCoarsenData *, PetscInt, PetscInt);
1656: PETSC_EXTERN PetscErrorCode PetscCDAppendNode(PetscCoarsenData *, PetscInt, PetscCDIntNd *);
1657: PETSC_EXTERN PetscErrorCode PetscCDRemoveNextNode(PetscCoarsenData *, PetscInt, PetscCDIntNd *);
1658: PETSC_EXTERN PetscErrorCode PetscCDCountAt(const PetscCoarsenData *, PetscInt, PetscInt *);
1659: PETSC_EXTERN PetscErrorCode PetscCDIsEmptyAt(const PetscCoarsenData *, PetscInt, PetscBool *);
1660: PETSC_EXTERN PetscErrorCode PetscCDSetChunkSize(PetscCoarsenData *, PetscInt);
1661: PETSC_EXTERN PetscErrorCode PetscCDPrint(const PetscCoarsenData *, PetscInt, MPI_Comm);
1662: PETSC_EXTERN PetscErrorCode PetscCDGetNonemptyIS(PetscCoarsenData *, IS *);
1663: PETSC_EXTERN PetscErrorCode PetscCDGetMat(PetscCoarsenData *, Mat *);
1664: PETSC_EXTERN PetscErrorCode PetscCDSetMat(PetscCoarsenData *, Mat);
1665: PETSC_EXTERN PetscErrorCode PetscCDClearMat(PetscCoarsenData *);
1666: PETSC_EXTERN PetscErrorCode PetscCDRemoveAllAt(PetscCoarsenData *, PetscInt);
1667: PETSC_EXTERN PetscErrorCode PetscCDCount(const PetscCoarsenData *, PetscInt *_sz);
1669: PETSC_EXTERN PetscErrorCode PetscCDGetHeadPos(const PetscCoarsenData *, PetscInt, PetscCDIntNd **);
1670: PETSC_EXTERN PetscErrorCode PetscCDGetNextPos(const PetscCoarsenData *, PetscInt, PetscCDIntNd **);
1671: PETSC_EXTERN PetscErrorCode PetscCDGetASMBlocks(const PetscCoarsenData *, const PetscInt, PetscInt *, IS **);
1673: PETSC_SINGLE_LIBRARY_VISIBILITY_INTERNAL PetscErrorCode MatFDColoringApply_AIJ(Mat, MatFDColoring, Vec, void *);
1675: PETSC_EXTERN PetscLogEvent MAT_Mult;
1676: PETSC_EXTERN PetscLogEvent MAT_MultAdd;
1677: PETSC_EXTERN PetscLogEvent MAT_MultTranspose;
1678: PETSC_EXTERN PetscLogEvent MAT_MultHermitianTranspose;
1679: PETSC_EXTERN PetscLogEvent MAT_MultTransposeAdd;
1680: PETSC_EXTERN PetscLogEvent MAT_MultHermitianTransposeAdd;
1681: PETSC_EXTERN PetscLogEvent MAT_Solve;
1682: PETSC_EXTERN PetscLogEvent MAT_Solves;
1683: PETSC_EXTERN PetscLogEvent MAT_SolveAdd;
1684: PETSC_EXTERN PetscLogEvent MAT_SolveTranspose;
1685: PETSC_EXTERN PetscLogEvent MAT_SolveTransposeAdd;
1686: PETSC_EXTERN PetscLogEvent MAT_SOR;
1687: PETSC_EXTERN PetscLogEvent MAT_ForwardSolve;
1688: PETSC_EXTERN PetscLogEvent MAT_BackwardSolve;
1689: PETSC_EXTERN PetscLogEvent MAT_LUFactor;
1690: PETSC_EXTERN PetscLogEvent MAT_LUFactorSymbolic;
1691: PETSC_EXTERN PetscLogEvent MAT_LUFactorNumeric;
1692: PETSC_EXTERN PetscLogEvent MAT_QRFactor;
1693: PETSC_EXTERN PetscLogEvent MAT_QRFactorSymbolic;
1694: PETSC_EXTERN PetscLogEvent MAT_QRFactorNumeric;
1695: PETSC_EXTERN PetscLogEvent MAT_CholeskyFactor;
1696: PETSC_EXTERN PetscLogEvent MAT_CholeskyFactorSymbolic;
1697: PETSC_EXTERN PetscLogEvent MAT_CholeskyFactorNumeric;
1698: PETSC_EXTERN PetscLogEvent MAT_ILUFactor;
1699: PETSC_EXTERN PetscLogEvent MAT_ILUFactorSymbolic;
1700: PETSC_EXTERN PetscLogEvent MAT_ICCFactorSymbolic;
1701: PETSC_EXTERN PetscLogEvent MAT_Copy;
1702: PETSC_EXTERN PetscLogEvent MAT_Convert;
1703: PETSC_EXTERN PetscLogEvent MAT_Scale;
1704: PETSC_EXTERN PetscLogEvent MAT_AssemblyBegin;
1705: PETSC_EXTERN PetscLogEvent MAT_AssemblyEnd;
1706: PETSC_EXTERN PetscLogEvent MAT_SetValues;
1707: PETSC_EXTERN PetscLogEvent MAT_GetValues;
1708: PETSC_EXTERN PetscLogEvent MAT_GetRow;
1709: PETSC_EXTERN PetscLogEvent MAT_GetRowIJ;
1710: PETSC_EXTERN PetscLogEvent MAT_CreateSubMats;
1711: PETSC_EXTERN PetscLogEvent MAT_GetOrdering;
1712: PETSC_EXTERN PetscLogEvent MAT_RedundantMat;
1713: PETSC_EXTERN PetscLogEvent MAT_IncreaseOverlap;
1714: PETSC_EXTERN PetscLogEvent MAT_Partitioning;
1715: PETSC_EXTERN PetscLogEvent MAT_PartitioningND;
1716: PETSC_EXTERN PetscLogEvent MAT_Coarsen;
1717: PETSC_EXTERN PetscLogEvent MAT_ZeroEntries;
1718: PETSC_EXTERN PetscLogEvent MAT_Load;
1719: PETSC_EXTERN PetscLogEvent MAT_View;
1720: PETSC_EXTERN PetscLogEvent MAT_AXPY;
1721: PETSC_EXTERN PetscLogEvent MAT_FDColoringCreate;
1722: PETSC_EXTERN PetscLogEvent MAT_TransposeColoringCreate;
1723: PETSC_EXTERN PetscLogEvent MAT_FDColoringSetUp;
1724: PETSC_EXTERN PetscLogEvent MAT_FDColoringApply;
1725: PETSC_EXTERN PetscLogEvent MAT_Transpose;
1726: PETSC_EXTERN PetscLogEvent MAT_FDColoringFunction;
1727: PETSC_EXTERN PetscLogEvent MAT_CreateSubMat;
1728: PETSC_EXTERN PetscLogEvent MAT_MatSolve;
1729: PETSC_EXTERN PetscLogEvent MAT_MatTrSolve;
1730: PETSC_EXTERN PetscLogEvent MAT_MatMultSymbolic;
1731: PETSC_EXTERN PetscLogEvent MAT_MatMultNumeric;
1732: PETSC_EXTERN PetscLogEvent MAT_Getlocalmatcondensed;
1733: PETSC_EXTERN PetscLogEvent MAT_GetBrowsOfAcols;
1734: PETSC_EXTERN PetscLogEvent MAT_GetBrowsOfAocols;
1735: PETSC_EXTERN PetscLogEvent MAT_PtAPSymbolic;
1736: PETSC_EXTERN PetscLogEvent MAT_PtAPNumeric;
1737: PETSC_EXTERN PetscLogEvent MAT_Seqstompinum;
1738: PETSC_EXTERN PetscLogEvent MAT_Seqstompisym;
1739: PETSC_EXTERN PetscLogEvent MAT_Seqstompi;
1740: PETSC_EXTERN PetscLogEvent MAT_Getlocalmat;
1741: PETSC_EXTERN PetscLogEvent MAT_RARtSymbolic;
1742: PETSC_EXTERN PetscLogEvent MAT_RARtNumeric;
1743: PETSC_EXTERN PetscLogEvent MAT_MatTransposeMultSymbolic;
1744: PETSC_EXTERN PetscLogEvent MAT_MatTransposeMultNumeric;
1745: PETSC_EXTERN PetscLogEvent MAT_TransposeMatMultSymbolic;
1746: PETSC_EXTERN PetscLogEvent MAT_TransposeMatMultNumeric;
1747: PETSC_EXTERN PetscLogEvent MAT_MatMatMultSymbolic;
1748: PETSC_EXTERN PetscLogEvent MAT_MatMatMultNumeric;
1749: PETSC_EXTERN PetscLogEvent MAT_Getsymtransreduced;
1750: PETSC_EXTERN PetscLogEvent MAT_GetSeqNonzeroStructure;
1751: PETSC_EXTERN PetscLogEvent MATMFFD_Mult;
1752: PETSC_EXTERN PetscLogEvent MAT_GetMultiProcBlock;
1753: PETSC_EXTERN PetscLogEvent MAT_CUSPARSECopyToGPU;
1754: PETSC_EXTERN PetscLogEvent MAT_CUSPARSECopyFromGPU;
1755: PETSC_EXTERN PetscLogEvent MAT_CUSPARSEGenerateTranspose;
1756: PETSC_EXTERN PetscLogEvent MAT_CUSPARSESolveAnalysis;
1757: PETSC_EXTERN PetscLogEvent MAT_HIPSPARSECopyToGPU;
1758: PETSC_EXTERN PetscLogEvent MAT_HIPSPARSECopyFromGPU;
1759: PETSC_EXTERN PetscLogEvent MAT_HIPSPARSEGenerateTranspose;
1760: PETSC_EXTERN PetscLogEvent MAT_HIPSPARSESolveAnalysis;
1761: PETSC_EXTERN PetscLogEvent MAT_SetValuesBatch;
1762: PETSC_EXTERN PetscLogEvent MAT_CreateGraph;
1763: PETSC_EXTERN PetscLogEvent MAT_ViennaCLCopyToGPU;
1764: PETSC_EXTERN PetscLogEvent MAT_DenseCopyToGPU;
1765: PETSC_EXTERN PetscLogEvent MAT_DenseCopyFromGPU;
1766: PETSC_EXTERN PetscLogEvent MAT_Merge;
1767: PETSC_EXTERN PetscLogEvent MAT_Residual;
1768: PETSC_EXTERN PetscLogEvent MAT_SetRandom;
1769: PETSC_EXTERN PetscLogEvent MAT_FactorFactS;
1770: PETSC_EXTERN PetscLogEvent MAT_FactorInvS;
1771: PETSC_EXTERN PetscLogEvent MAT_PreallCOO;
1772: PETSC_EXTERN PetscLogEvent MAT_SetVCOO;
1773: PETSC_EXTERN PetscLogEvent MATCOLORING_Apply;
1774: PETSC_EXTERN PetscLogEvent MATCOLORING_Comm;
1775: PETSC_EXTERN PetscLogEvent MATCOLORING_Local;
1776: PETSC_EXTERN PetscLogEvent MATCOLORING_ISCreate;
1777: PETSC_EXTERN PetscLogEvent MATCOLORING_SetUp;
1778: PETSC_EXTERN PetscLogEvent MATCOLORING_Weights;
1779: PETSC_EXTERN PetscLogEvent MAT_H2Opus_Build;
1780: PETSC_EXTERN PetscLogEvent MAT_H2Opus_Compress;
1781: PETSC_EXTERN PetscLogEvent MAT_H2Opus_Orthog;
1782: PETSC_EXTERN PetscLogEvent MAT_H2Opus_LR;
1783: PETSC_EXTERN PetscLogEvent MAT_CUDACopyToGPU;
1784: PETSC_EXTERN PetscLogEvent MAT_HIPCopyToGPU;