Actual source code: matimpl.h
petsc-3.11.4 2019-09-28
2: #ifndef __MATIMPL_H
5: #include <petscmat.h>
6: #include <petscmatcoarsen.h>
7: #include <petsc/private/petscimpl.h>
9: PETSC_EXTERN PetscBool MatRegisterAllCalled;
10: PETSC_EXTERN PetscBool MatSeqAIJRegisterAllCalled;
11: PETSC_EXTERN PetscBool MatOrderingRegisterAllCalled;
12: PETSC_EXTERN PetscBool MatColoringRegisterAllCalled;
13: PETSC_EXTERN PetscBool MatPartitioningRegisterAllCalled;
14: PETSC_EXTERN PetscBool MatCoarsenRegisterAllCalled;
15: PETSC_EXTERN PetscErrorCode MatRegisterAll(void);
16: PETSC_EXTERN PetscErrorCode MatOrderingRegisterAll(void);
17: PETSC_EXTERN PetscErrorCode MatColoringRegisterAll(void);
18: PETSC_EXTERN PetscErrorCode MatPartitioningRegisterAll(void);
19: PETSC_EXTERN PetscErrorCode MatCoarsenRegisterAll(void);
20: PETSC_EXTERN PetscErrorCode MatSeqAIJRegisterAll(void);
22: /*
23: This file defines the parts of the matrix data structure that are
24: shared by all matrix types.
25: */
27: /*
28: If you add entries here also add them to the MATOP enum
29: in include/petscmat.h and include/petsc/finclude/petscmat.h
30: */
31: typedef struct _MatOps *MatOps;
32: struct _MatOps {
33: /* 0*/
34: PetscErrorCode (*setvalues)(Mat,PetscInt,const PetscInt[],PetscInt,const PetscInt[],const PetscScalar[],InsertMode);
35: PetscErrorCode (*getrow)(Mat,PetscInt,PetscInt *,PetscInt*[],PetscScalar*[]);
36: PetscErrorCode (*restorerow)(Mat,PetscInt,PetscInt *,PetscInt *[],PetscScalar *[]);
37: PetscErrorCode (*mult)(Mat,Vec,Vec);
38: PetscErrorCode (*multadd)(Mat,Vec,Vec,Vec);
39: /* 5*/
40: PetscErrorCode (*multtranspose)(Mat,Vec,Vec);
41: PetscErrorCode (*multtransposeadd)(Mat,Vec,Vec,Vec);
42: PetscErrorCode (*solve)(Mat,Vec,Vec);
43: PetscErrorCode (*solveadd)(Mat,Vec,Vec,Vec);
44: PetscErrorCode (*solvetranspose)(Mat,Vec,Vec);
45: /*10*/
46: PetscErrorCode (*solvetransposeadd)(Mat,Vec,Vec,Vec);
47: PetscErrorCode (*lufactor)(Mat,IS,IS,const MatFactorInfo*);
48: PetscErrorCode (*choleskyfactor)(Mat,IS,const MatFactorInfo*);
49: PetscErrorCode (*sor)(Mat,Vec,PetscReal,MatSORType,PetscReal,PetscInt,PetscInt,Vec);
50: PetscErrorCode (*transpose)(Mat,MatReuse,Mat *);
51: /*15*/
52: PetscErrorCode (*getinfo)(Mat,MatInfoType,MatInfo*);
53: PetscErrorCode (*equal)(Mat,Mat,PetscBool *);
54: PetscErrorCode (*getdiagonal)(Mat,Vec);
55: PetscErrorCode (*diagonalscale)(Mat,Vec,Vec);
56: PetscErrorCode (*norm)(Mat,NormType,PetscReal*);
57: /*20*/
58: PetscErrorCode (*assemblybegin)(Mat,MatAssemblyType);
59: PetscErrorCode (*assemblyend)(Mat,MatAssemblyType);
60: PetscErrorCode (*setoption)(Mat,MatOption,PetscBool );
61: PetscErrorCode (*zeroentries)(Mat);
62: /*24*/
63: PetscErrorCode (*zerorows)(Mat,PetscInt,const PetscInt[],PetscScalar,Vec,Vec);
64: PetscErrorCode (*lufactorsymbolic)(Mat,Mat,IS,IS,const MatFactorInfo*);
65: PetscErrorCode (*lufactornumeric)(Mat,Mat,const MatFactorInfo*);
66: PetscErrorCode (*choleskyfactorsymbolic)(Mat,Mat,IS,const MatFactorInfo*);
67: PetscErrorCode (*choleskyfactornumeric)(Mat,Mat,const MatFactorInfo*);
68: /*29*/
69: PetscErrorCode (*setup)(Mat);
70: PetscErrorCode (*ilufactorsymbolic)(Mat,Mat,IS,IS,const MatFactorInfo*);
71: PetscErrorCode (*iccfactorsymbolic)(Mat,Mat,IS,const MatFactorInfo*);
72: PetscErrorCode (*getdiagonalblock)(Mat,Mat*);
73: PetscErrorCode (*freeintermediatedatastructures)(Mat);
74: /*34*/
75: PetscErrorCode (*duplicate)(Mat,MatDuplicateOption,Mat*);
76: PetscErrorCode (*forwardsolve)(Mat,Vec,Vec);
77: PetscErrorCode (*backwardsolve)(Mat,Vec,Vec);
78: PetscErrorCode (*ilufactor)(Mat,IS,IS,const MatFactorInfo*);
79: PetscErrorCode (*iccfactor)(Mat,IS,const MatFactorInfo*);
80: /*39*/
81: PetscErrorCode (*axpy)(Mat,PetscScalar,Mat,MatStructure);
82: PetscErrorCode (*createsubmatrices)(Mat,PetscInt,const IS[],const IS[],MatReuse,Mat *[]);
83: PetscErrorCode (*increaseoverlap)(Mat,PetscInt,IS[],PetscInt);
84: PetscErrorCode (*getvalues)(Mat,PetscInt,const PetscInt[],PetscInt,const PetscInt[],PetscScalar []);
85: PetscErrorCode (*copy)(Mat,Mat,MatStructure);
86: /*44*/
87: PetscErrorCode (*getrowmax)(Mat,Vec,PetscInt[]);
88: PetscErrorCode (*scale)(Mat,PetscScalar);
89: PetscErrorCode (*shift)(Mat,PetscScalar);
90: PetscErrorCode (*diagonalset)(Mat,Vec,InsertMode);
91: PetscErrorCode (*zerorowscolumns)(Mat,PetscInt,const PetscInt[],PetscScalar,Vec,Vec);
92: /*49*/
93: PetscErrorCode (*setrandom)(Mat,PetscRandom);
94: PetscErrorCode (*getrowij)(Mat,PetscInt,PetscBool ,PetscBool ,PetscInt*,const PetscInt *[],const PetscInt *[],PetscBool *);
95: PetscErrorCode (*restorerowij)(Mat,PetscInt,PetscBool ,PetscBool ,PetscInt *,const PetscInt *[],const PetscInt *[],PetscBool *);
96: PetscErrorCode (*getcolumnij)(Mat,PetscInt,PetscBool ,PetscBool ,PetscInt*,const PetscInt *[],const PetscInt *[],PetscBool *);
97: PetscErrorCode (*restorecolumnij)(Mat,PetscInt,PetscBool ,PetscBool ,PetscInt*,const PetscInt *[],const PetscInt *[],PetscBool *);
98: /*54*/
99: PetscErrorCode (*fdcoloringcreate)(Mat,ISColoring,MatFDColoring);
100: PetscErrorCode (*coloringpatch)(Mat,PetscInt,PetscInt,ISColoringValue[],ISColoring*);
101: PetscErrorCode (*setunfactored)(Mat);
102: PetscErrorCode (*permute)(Mat,IS,IS,Mat*);
103: PetscErrorCode (*setvaluesblocked)(Mat,PetscInt,const PetscInt[],PetscInt,const PetscInt[],const PetscScalar[],InsertMode);
104: /*59*/
105: PetscErrorCode (*createsubmatrix)(Mat,IS,IS,MatReuse,Mat*);
106: PetscErrorCode (*destroy)(Mat);
107: PetscErrorCode (*view)(Mat,PetscViewer);
108: PetscErrorCode (*convertfrom)(Mat, MatType,MatReuse,Mat*);
109: PetscErrorCode (*matmatmult)(Mat,Mat,Mat,MatReuse,PetscReal,Mat*);
110: /*64*/
111: PetscErrorCode (*matmatmultsymbolic)(Mat,Mat,Mat,PetscReal,Mat*);
112: PetscErrorCode (*matmatmultnumeric)(Mat,Mat,Mat,Mat);
113: PetscErrorCode (*setlocaltoglobalmapping)(Mat,ISLocalToGlobalMapping,ISLocalToGlobalMapping);
114: PetscErrorCode (*setvalueslocal)(Mat,PetscInt,const PetscInt[],PetscInt,const PetscInt[],const PetscScalar[],InsertMode);
115: PetscErrorCode (*zerorowslocal)(Mat,PetscInt,const PetscInt[],PetscScalar,Vec,Vec);
116: /*69*/
117: PetscErrorCode (*getrowmaxabs)(Mat,Vec,PetscInt[]);
118: PetscErrorCode (*getrowminabs)(Mat,Vec,PetscInt[]);
119: PetscErrorCode (*convert)(Mat, MatType,MatReuse,Mat*);
120: PetscErrorCode (*hasoperation)(Mat,MatOperation,PetscBool*);
121: PetscErrorCode (*placeholder_73)(Mat,void*);
122: /*74*/
123: PetscErrorCode (*setvaluesadifor)(Mat,PetscInt,void*);
124: PetscErrorCode (*fdcoloringapply)(Mat,MatFDColoring,Vec,void*);
125: PetscErrorCode (*setfromoptions)(PetscOptionItems*,Mat);
126: PetscErrorCode (*multconstrained)(Mat,Vec,Vec);
127: PetscErrorCode (*multtransposeconstrained)(Mat,Vec,Vec);
128: /*79*/
129: PetscErrorCode (*findzerodiagonals)(Mat,IS*);
130: PetscErrorCode (*mults)(Mat, Vecs, Vecs);
131: PetscErrorCode (*solves)(Mat, Vecs, Vecs);
132: PetscErrorCode (*getinertia)(Mat,PetscInt*,PetscInt*,PetscInt*);
133: PetscErrorCode (*load)(Mat, PetscViewer);
134: /*84*/
135: PetscErrorCode (*issymmetric)(Mat,PetscReal,PetscBool *);
136: PetscErrorCode (*ishermitian)(Mat,PetscReal,PetscBool *);
137: PetscErrorCode (*isstructurallysymmetric)(Mat,PetscBool *);
138: PetscErrorCode (*setvaluesblockedlocal)(Mat,PetscInt,const PetscInt[],PetscInt,const PetscInt[],const PetscScalar[],InsertMode);
139: PetscErrorCode (*getvecs)(Mat,Vec*,Vec*);
140: /*89*/
141: PetscErrorCode (*matmult)(Mat,Mat,MatReuse,PetscReal,Mat*);
142: PetscErrorCode (*matmultsymbolic)(Mat,Mat,PetscReal,Mat*);
143: PetscErrorCode (*matmultnumeric)(Mat,Mat,Mat);
144: PetscErrorCode (*ptap)(Mat,Mat,MatReuse,PetscReal,Mat*);
145: PetscErrorCode (*ptapsymbolic)(Mat,Mat,PetscReal,Mat*); /* double dispatch wrapper routine */
146: /*94*/
147: PetscErrorCode (*ptapnumeric)(Mat,Mat,Mat); /* double dispatch wrapper routine */
148: PetscErrorCode (*mattransposemult)(Mat,Mat,MatReuse,PetscReal,Mat*);
149: PetscErrorCode (*mattransposemultsymbolic)(Mat,Mat,PetscReal,Mat*);
150: PetscErrorCode (*mattransposemultnumeric)(Mat,Mat,Mat);
151: PetscErrorCode (*placeholder_98)(Mat);
152: /*99*/
153: PetscErrorCode (*placeholder_99)(Mat);
154: PetscErrorCode (*placeholder_100)(Mat);
155: PetscErrorCode (*placeholder_101)(Mat);
156: PetscErrorCode (*conjugate)(Mat); /* complex conjugate */
157: PetscErrorCode (*viewnative)(Mat,PetscViewer);
158: /*104*/
159: PetscErrorCode (*setvaluesrow)(Mat,PetscInt,const PetscScalar[]);
160: PetscErrorCode (*realpart)(Mat);
161: PetscErrorCode (*imaginarypart)(Mat);
162: PetscErrorCode (*getrowuppertriangular)(Mat);
163: PetscErrorCode (*restorerowuppertriangular)(Mat);
164: /*109*/
165: PetscErrorCode (*matsolve)(Mat,Mat,Mat);
166: PetscErrorCode (*matsolvetranspose)(Mat,Mat,Mat);
167: PetscErrorCode (*getrowmin)(Mat,Vec,PetscInt[]);
168: PetscErrorCode (*getcolumnvector)(Mat,Vec,PetscInt);
169: PetscErrorCode (*missingdiagonal)(Mat,PetscBool *,PetscInt*);
170: /*114*/
171: PetscErrorCode (*getseqnonzerostructure)(Mat,Mat *);
172: PetscErrorCode (*create)(Mat);
173: PetscErrorCode (*getghosts)(Mat,PetscInt*,const PetscInt *[]);
174: PetscErrorCode (*getlocalsubmatrix)(Mat,IS,IS,Mat*);
175: PetscErrorCode (*restorelocalsubmatrix)(Mat,IS,IS,Mat*);
176: /*119*/
177: PetscErrorCode (*multdiagonalblock)(Mat,Vec,Vec);
178: PetscErrorCode (*hermitiantranspose)(Mat,MatReuse,Mat*);
179: PetscErrorCode (*multhermitiantranspose)(Mat,Vec,Vec);
180: PetscErrorCode (*multhermitiantransposeadd)(Mat,Vec,Vec,Vec);
181: PetscErrorCode (*getmultiprocblock)(Mat,MPI_Comm,MatReuse,Mat*);
182: /*124*/
183: PetscErrorCode (*findnonzerorows)(Mat,IS*);
184: PetscErrorCode (*getcolumnnorms)(Mat,NormType,PetscReal*);
185: PetscErrorCode (*invertblockdiagonal)(Mat,const PetscScalar**);
186: PetscErrorCode (*invertvariableblockdiagonal)(Mat,PetscInt,const PetscInt*,PetscScalar*);
187: PetscErrorCode (*createsubmatricesmpi)(Mat,PetscInt,const IS[], const IS[], MatReuse, Mat**);
188: /*129*/
189: PetscErrorCode (*setvaluesbatch)(Mat,PetscInt,PetscInt,PetscInt*,const PetscScalar*);
190: PetscErrorCode (*transposematmult)(Mat,Mat,MatReuse,PetscReal,Mat*);
191: PetscErrorCode (*transposematmultsymbolic)(Mat,Mat,PetscReal,Mat*);
192: PetscErrorCode (*transposematmultnumeric)(Mat,Mat,Mat);
193: PetscErrorCode (*transposecoloringcreate)(Mat,ISColoring,MatTransposeColoring);
194: /*134*/
195: PetscErrorCode (*transcoloringapplysptoden)(MatTransposeColoring,Mat,Mat);
196: PetscErrorCode (*transcoloringapplydentosp)(MatTransposeColoring,Mat,Mat);
197: PetscErrorCode (*rart)(Mat,Mat,MatReuse,PetscReal,Mat*);
198: PetscErrorCode (*rartsymbolic)(Mat,Mat,PetscReal,Mat*); /* double dispatch wrapper routine */
199: PetscErrorCode (*rartnumeric)(Mat,Mat,Mat); /* double dispatch wrapper routine */
200: /*139*/
201: PetscErrorCode (*setblocksizes)(Mat,PetscInt,PetscInt);
202: PetscErrorCode (*aypx)(Mat,PetscScalar,Mat,MatStructure);
203: PetscErrorCode (*residual)(Mat,Vec,Vec,Vec);
204: PetscErrorCode (*fdcoloringsetup)(Mat,ISColoring,MatFDColoring);
205: PetscErrorCode (*findoffblockdiagonalentries)(Mat,IS*);
206: /*144*/
207: PetscErrorCode (*creatempimatconcatenateseqmat)(MPI_Comm,Mat,PetscInt,MatReuse,Mat*);
208: PetscErrorCode (*destroysubmatrices)(PetscInt,Mat*[]);
209: PetscErrorCode (*mattransposesolve)(Mat,Mat,Mat);
210: };
211: /*
212: If you add MatOps entries above also add them to the MATOP enum
213: in include/petscmat.h and include/petsc/finclude/petscmat.h
214: */
216: #include <petscsys.h>
217: PETSC_EXTERN PetscErrorCode MatRegisterOp(MPI_Comm, const char[], PetscVoidFunction, const char[], PetscInt, ...);
218: PETSC_EXTERN PetscErrorCode MatQueryOp(MPI_Comm, PetscVoidFunction*, const char[], PetscInt, ...);
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
230: */
231: PETSC_INTERN PetscErrorCode MatFindNonzeroRowsOrCols_Basic(Mat,PetscBool,PetscReal,IS*);
232: PETSC_INTERN PetscErrorCode MatConvert_Basic(Mat,MatType,MatReuse,Mat*);
233: PETSC_INTERN PetscErrorCode MatConvert_Shell(Mat, MatType,MatReuse,Mat*);
234: PETSC_INTERN PetscErrorCode MatCopy_Basic(Mat,Mat,MatStructure);
235: PETSC_INTERN PetscErrorCode MatDiagonalSet_Default(Mat,Vec,InsertMode);
237: #if defined(PETSC_USE_DEBUG)
238: # define MatCheckPreallocated(A,arg) do { \
239: if (PetscUnlikely(!(A)->preallocated)) SETERRQ3(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Must call MatXXXSetPreallocation() or MatSetUp() on argument %D \"%s\" before %s()",(arg),#A,PETSC_FUNCTION_NAME); \
240: } while (0)
241: #else
242: # define MatCheckPreallocated(A,arg) do {} while (0)
243: #endif
245: /*
246: The stash is used to temporarily store inserted matrix values that
247: belong to another processor. During the assembly phase the stashed
248: values are moved to the correct processor and
249: */
251: typedef struct _MatStashSpace *PetscMatStashSpace;
253: struct _MatStashSpace {
254: PetscMatStashSpace next;
255: PetscScalar *space_head,*val;
256: PetscInt *idx,*idy;
257: PetscInt total_space_size;
258: PetscInt local_used;
259: PetscInt local_remaining;
260: };
262: PETSC_EXTERN PetscErrorCode PetscMatStashSpaceGet(PetscInt,PetscInt,PetscMatStashSpace *);
263: PETSC_EXTERN PetscErrorCode PetscMatStashSpaceContiguous(PetscInt,PetscMatStashSpace *,PetscScalar *,PetscInt *,PetscInt *);
264: PETSC_EXTERN PetscErrorCode PetscMatStashSpaceDestroy(PetscMatStashSpace*);
266: typedef struct {
267: PetscInt count;
268: } MatStashHeader;
270: typedef struct {
271: void *buffer; /* Of type blocktype, dynamically constructed */
272: PetscInt count;
273: char pending;
274: } MatStashFrame;
276: typedef struct _MatStash MatStash;
277: struct _MatStash {
278: PetscInt nmax; /* maximum stash size */
279: PetscInt umax; /* user specified max-size */
280: PetscInt oldnmax; /* the nmax value used previously */
281: PetscInt n; /* stash size */
282: PetscInt bs; /* block size of the stash */
283: PetscInt reallocs; /* preserve the no of mallocs invoked */
284: PetscMatStashSpace space_head,space; /* linked list to hold stashed global row/column numbers and matrix values */
286: PetscErrorCode (*ScatterBegin)(Mat,MatStash*,PetscInt*);
287: PetscErrorCode (*ScatterGetMesg)(MatStash*,PetscMPIInt*,PetscInt**,PetscInt**,PetscScalar**,PetscInt*);
288: PetscErrorCode (*ScatterEnd)(MatStash*);
289: PetscErrorCode (*ScatterDestroy)(MatStash*);
291: /* The following variables are used for communication */
292: MPI_Comm comm;
293: PetscMPIInt size,rank;
294: PetscMPIInt tag1,tag2;
295: MPI_Request *send_waits; /* array of send requests */
296: MPI_Request *recv_waits; /* array of receive requests */
297: MPI_Status *send_status; /* array of send status */
298: PetscInt nsends,nrecvs; /* numbers of sends and receives */
299: PetscScalar *svalues; /* sending data */
300: PetscInt *sindices;
301: PetscScalar **rvalues; /* receiving data (values) */
302: PetscInt **rindices; /* receiving data (indices) */
303: PetscInt nprocessed; /* number of messages already processed */
304: PetscMPIInt *flg_v; /* indicates what messages have arrived so far and from whom */
305: PetscBool reproduce;
306: PetscInt reproduce_count;
308: /* The following variables are used for BTS communication */
309: PetscBool first_assembly_done; /* Is the first time matrix assembly done? */
310: PetscBool use_status; /* Use MPI_Status to determine number of items in each message */
311: PetscMPIInt nsendranks;
312: PetscMPIInt nrecvranks;
313: PetscMPIInt *sendranks;
314: PetscMPIInt *recvranks;
315: MatStashHeader *sendhdr,*recvhdr;
316: MatStashFrame *sendframes; /* pointers to the main messages */
317: MatStashFrame *recvframes;
318: MatStashFrame *recvframe_active;
319: PetscInt recvframe_i; /* index of block within active frame */
320: PetscMPIInt recvframe_count; /* Count actually sent for current frame */
321: PetscInt recvcount; /* Number of receives processed so far */
322: PetscMPIInt *some_indices; /* From last call to MPI_Waitsome */
323: MPI_Status *some_statuses; /* Statuses from last call to MPI_Waitsome */
324: PetscMPIInt some_count; /* Number of requests completed in last call to MPI_Waitsome */
325: PetscMPIInt some_i; /* Index of request currently being processed */
326: MPI_Request *sendreqs;
327: MPI_Request *recvreqs;
328: PetscSegBuffer segsendblocks;
329: PetscSegBuffer segrecvframe;
330: PetscSegBuffer segrecvblocks;
331: MPI_Datatype blocktype;
332: size_t blocktype_size;
333: InsertMode *insertmode; /* Pointer to check mat->insertmode and set upon message arrival in case no local values have been set. */
334: };
336: #if !defined(PETSC_HAVE_MPIUNI)
337: PETSC_INTERN PetscErrorCode MatStashScatterDestroy_BTS(MatStash*);
338: #endif
339: PETSC_INTERN PetscErrorCode MatStashCreate_Private(MPI_Comm,PetscInt,MatStash*);
340: PETSC_INTERN PetscErrorCode MatStashDestroy_Private(MatStash*);
341: PETSC_INTERN PetscErrorCode MatStashScatterEnd_Private(MatStash*);
342: PETSC_INTERN PetscErrorCode MatStashSetInitialSize_Private(MatStash*,PetscInt);
343: PETSC_INTERN PetscErrorCode MatStashGetInfo_Private(MatStash*,PetscInt*,PetscInt*);
344: PETSC_INTERN PetscErrorCode MatStashValuesRow_Private(MatStash*,PetscInt,PetscInt,const PetscInt[],const PetscScalar[],PetscBool );
345: PETSC_INTERN PetscErrorCode MatStashValuesCol_Private(MatStash*,PetscInt,PetscInt,const PetscInt[],const PetscScalar[],PetscInt,PetscBool );
346: PETSC_INTERN PetscErrorCode MatStashValuesRowBlocked_Private(MatStash*,PetscInt,PetscInt,const PetscInt[],const PetscScalar[],PetscInt,PetscInt,PetscInt);
347: PETSC_INTERN PetscErrorCode MatStashValuesColBlocked_Private(MatStash*,PetscInt,PetscInt,const PetscInt[],const PetscScalar[],PetscInt,PetscInt,PetscInt);
348: PETSC_INTERN PetscErrorCode MatStashScatterBegin_Private(Mat,MatStash*,PetscInt*);
349: PETSC_INTERN PetscErrorCode MatStashScatterGetMesg_Private(MatStash*,PetscMPIInt*,PetscInt**,PetscInt**,PetscScalar**,PetscInt*);
350: PETSC_INTERN PetscErrorCode MatGetInfo_External(Mat,MatInfoType,MatInfo*);
352: typedef struct {
353: PetscInt dim;
354: PetscInt dims[4];
355: PetscInt starts[4];
356: PetscBool noc; /* this is a single component problem, hence user will not set MatStencil.c */
357: } MatStencilInfo;
359: /* Info about using compressed row format */
360: typedef struct {
361: PetscBool use; /* indicates compressed rows have been checked and will be used */
362: PetscInt nrows; /* number of non-zero rows */
363: PetscInt *i; /* compressed row pointer */
364: PetscInt *rindex; /* compressed row index */
365: } Mat_CompressedRow;
366: PETSC_EXTERN PetscErrorCode MatCheckCompressedRow(Mat,PetscInt,Mat_CompressedRow*,PetscInt*,PetscInt,PetscReal);
368: typedef struct { /* used by MatCreateRedundantMatrix() for reusing matredundant */
369: PetscInt nzlocal,nsends,nrecvs;
370: PetscMPIInt *send_rank,*recv_rank;
371: PetscInt *sbuf_nz,*rbuf_nz,*sbuf_j,**rbuf_j;
372: PetscScalar *sbuf_a,**rbuf_a;
373: MPI_Comm subcomm; /* when user does not provide a subcomm */
374: IS isrow,iscol;
375: Mat *matseq;
376: } Mat_Redundant;
378: struct _p_Mat {
379: PETSCHEADER(struct _MatOps);
380: PetscLayout rmap,cmap;
381: void *data; /* implementation-specific data */
382: MatFactorType factortype; /* MAT_FACTOR_LU, ILU, CHOLESKY or ICC */
383: PetscBool assembled; /* is the matrix assembled? */
384: PetscBool was_assembled; /* new values inserted into assembled mat */
385: PetscInt num_ass; /* number of times matrix has been assembled */
386: PetscObjectState nonzerostate; /* each time new nonzeros locations are introduced into the matrix this is updated */
387: MatInfo info; /* matrix information */
388: InsertMode insertmode; /* have values been inserted in matrix or added? */
389: MatStash stash,bstash; /* used for assembling off-proc mat emements */
390: MatNullSpace nullsp; /* null space (operator is singular) */
391: MatNullSpace transnullsp; /* null space of transpose of operator */
392: MatNullSpace nearnullsp; /* near null space to be used by multigrid methods */
393: PetscInt congruentlayouts; /* are the rows and columns layouts congruent? */
394: PetscBool preallocated;
395: MatStencilInfo stencil; /* information for structured grid */
396: PetscBool symmetric,hermitian,structurally_symmetric,spd;
397: PetscBool symmetric_set,hermitian_set,structurally_symmetric_set,spd_set; /* if true, then corresponding flag is correct*/
398: PetscBool symmetric_eternal;
399: PetscBool nooffprocentries,nooffproczerorows;
400: PetscBool assembly_subset; /* set by MAT_SUBSET_OFF_PROC_ENTRIES */
401: PetscBool submat_singleis; /* for efficient PCSetUP_ASM() */
402: PetscBool structure_only;
403: #if defined(PETSC_HAVE_VIENNACL) || defined(PETSC_HAVE_CUDA)
404: PetscOffloadFlag valid_GPU_matrix; /* flag pointing to the matrix on the gpu*/
405: #endif
406: void *spptr; /* pointer for special library like SuperLU */
407: char *solvertype;
408: PetscBool checksymmetryonassembly,checknullspaceonassembly;
409: PetscReal checksymmetrytol;
410: Mat schur; /* Schur complement matrix */
411: MatFactorSchurStatus schur_status; /* status of the Schur complement matrix */
412: Mat_Redundant *redundant; /* used by MatCreateRedundantMatrix() */
413: PetscBool erroriffailure; /* Generate an error if detected (for example a zero pivot) instead of returning */
414: MatFactorError factorerrortype; /* type of error in factorization */
415: PetscReal factorerror_zeropivot_value; /* If numerical zero pivot was detected this is the computed value */
416: PetscInt factorerror_zeropivot_row; /* Row where zero pivot was detected */
417: PetscInt nblocks,*bsizes; /* support for MatSetVariableBlockSizes() */
418: char *defaultvectype;
419: };
421: PETSC_INTERN PetscErrorCode MatAXPY_Basic(Mat,PetscScalar,Mat,MatStructure);
422: PETSC_INTERN PetscErrorCode MatAXPY_BasicWithPreallocation(Mat,Mat,PetscScalar,Mat,MatStructure);
423: PETSC_INTERN PetscErrorCode MatAXPY_Basic_Preallocate(Mat,Mat,Mat*);
425: /*
426: Utility for MatFactor (Schur complement)
427: */
428: PETSC_INTERN PetscErrorCode MatFactorFactorizeSchurComplement_Private(Mat);
429: PETSC_INTERN PetscErrorCode MatFactorInvertSchurComplement_Private(Mat);
430: PETSC_INTERN PetscErrorCode MatFactorUpdateSchurStatus_Private(Mat);
431: PETSC_INTERN PetscErrorCode MatFactorSetUpInPlaceSchur_Private(Mat);
433: /*
434: Utility for MatZeroRows
435: */
436: PETSC_INTERN PetscErrorCode MatZeroRowsMapLocal_Private(Mat,PetscInt,const PetscInt*,PetscInt*,PetscInt**);
438: /*
439: Object for partitioning graphs
440: */
442: typedef struct _MatPartitioningOps *MatPartitioningOps;
443: struct _MatPartitioningOps {
444: PetscErrorCode (*apply)(MatPartitioning,IS*);
445: PetscErrorCode (*applynd)(MatPartitioning,IS*);
446: PetscErrorCode (*setfromoptions)(PetscOptionItems*,MatPartitioning);
447: PetscErrorCode (*destroy)(MatPartitioning);
448: PetscErrorCode (*view)(MatPartitioning,PetscViewer);
449: PetscErrorCode (*improve)(MatPartitioning,IS*);
450: };
452: struct _p_MatPartitioning {
453: PETSCHEADER(struct _MatPartitioningOps);
454: Mat adj;
455: PetscInt *vertex_weights;
456: PetscReal *part_weights;
457: PetscInt n; /* number of partitions */
458: void *data;
459: PetscInt setupcalled;
460: };
462: /* needed for parallel nested dissection by ParMetis and PTSCOTCH */
463: PETSC_INTERN PetscErrorCode MatPartitioningSizesToSep_Private(PetscInt,PetscInt[],PetscInt[],PetscInt[]);
465: /*
466: Object for coarsen graphs
467: */
468: typedef struct _MatCoarsenOps *MatCoarsenOps;
469: struct _MatCoarsenOps {
470: PetscErrorCode (*apply)(MatCoarsen);
471: PetscErrorCode (*setfromoptions)(PetscOptionItems*,MatCoarsen);
472: PetscErrorCode (*destroy)(MatCoarsen);
473: PetscErrorCode (*view)(MatCoarsen,PetscViewer);
474: };
476: struct _p_MatCoarsen {
477: PETSCHEADER(struct _MatCoarsenOps);
478: Mat graph;
479: PetscInt setupcalled;
480: void *subctx;
481: /* */
482: PetscBool strict_aggs;
483: IS perm;
484: PetscCoarsenData *agg_lists;
485: };
487: /*
488: MatFDColoring is used to compute Jacobian matrices efficiently
489: via coloring. The data structure is explained below in an example.
491: Color = 0 1 0 2 | 2 3 0
492: ---------------------------------------------------
493: 00 01 | 05
494: 10 11 | 14 15 Processor 0
495: 22 23 | 25
496: 32 33 |
497: ===================================================
498: | 44 45 46
499: 50 | 55 Processor 1
500: | 64 66
501: ---------------------------------------------------
503: ncolors = 4;
505: ncolumns = {2,1,1,0}
506: columns = {{0,2},{1},{3},{}}
507: nrows = {4,2,3,3}
508: rows = {{0,1,2,3},{0,1},{1,2,3},{0,1,2}}
509: vwscale = {dx(0),dx(1),dx(2),dx(3)} MPI Vec
510: vscale = {dx(0),dx(1),dx(2),dx(3),dx(4),dx(5)} Seq Vec
512: ncolumns = {1,0,1,1}
513: columns = {{6},{},{4},{5}}
514: nrows = {3,0,2,2}
515: rows = {{0,1,2},{},{1,2},{1,2}}
516: vwscale = {dx(4),dx(5),dx(6)} MPI Vec
517: vscale = {dx(0),dx(4),dx(5),dx(6)} Seq Vec
519: See the routine MatFDColoringApply() for how this data is used
520: to compute the Jacobian.
522: */
523: typedef struct {
524: PetscInt row;
525: PetscInt col;
526: PetscScalar *valaddr; /* address of value */
527: } MatEntry;
529: typedef struct {
530: PetscInt row;
531: PetscScalar *valaddr; /* address of value */
532: } MatEntry2;
534: struct _p_MatFDColoring{
535: PETSCHEADER(int);
536: PetscInt M,N,m; /* total rows, columns; local rows */
537: PetscInt rstart; /* first row owned by local processor */
538: PetscInt ncolors; /* number of colors */
539: PetscInt *ncolumns; /* number of local columns for a color */
540: PetscInt **columns; /* lists the local columns of each color (using global column numbering) */
541: PetscInt *nrows; /* number of local rows for each color */
542: MatEntry *matentry; /* holds (row, column, address of value) for Jacobian matrix entry */
543: MatEntry2 *matentry2; /* holds (row, address of value) for Jacobian matrix entry */
544: PetscScalar *dy; /* store a block of F(x+dx)-F(x) when J is in BAIJ format */
545: PetscReal error_rel; /* square root of relative error in computing function */
546: PetscReal umin; /* minimum allowable u'dx value */
547: Vec w1,w2,w3; /* work vectors used in computing Jacobian */
548: PetscBool fset; /* indicates that the initial function value F(X) is set */
549: PetscErrorCode (*f)(void); /* function that defines Jacobian */
550: void *fctx; /* optional user-defined context for use by the function f */
551: Vec vscale; /* holds FD scaling, i.e. 1/dx for each perturbed column */
552: PetscInt currentcolor; /* color for which function evaluation is being done now */
553: const char *htype; /* "wp" or "ds" */
554: ISColoringType ctype; /* IS_COLORING_GLOBAL or IS_COLORING_LOCAL */
555: PetscInt brows,bcols; /* number of block rows or columns for speedup inserting the dense matrix into sparse Jacobian */
556: PetscBool setupcalled; /* true if setup has been called */
557: PetscBool viewed; /* true if the -mat_fd_coloring_view has been triggered already */
558: void (*ftn_func_pointer)(void),*ftn_func_cntx; /* serve the same purpose as *fortran_func_pointers in PETSc objects */
559: };
561: typedef struct _MatColoringOps *MatColoringOps;
562: struct _MatColoringOps {
563: PetscErrorCode (*destroy)(MatColoring);
564: PetscErrorCode (*setfromoptions)(PetscOptionItems*,MatColoring);
565: PetscErrorCode (*view)(MatColoring,PetscViewer);
566: PetscErrorCode (*apply)(MatColoring,ISColoring*);
567: PetscErrorCode (*weights)(MatColoring,PetscReal**,PetscInt**);
568: };
570: struct _p_MatColoring {
571: PETSCHEADER(struct _MatColoringOps);
572: Mat mat;
573: PetscInt dist; /* distance of the coloring */
574: PetscInt maxcolors; /* the maximum number of colors returned, maxcolors=1 for MIS */
575: void *data; /* inner context */
576: PetscBool valid; /* check to see if what is produced is a valid coloring */
577: MatColoringWeightType weight_type; /* type of weight computation to be performed */
578: PetscReal *user_weights; /* custom weights and permutation */
579: PetscInt *user_lperm;
580: PetscBool valid_iscoloring; /* check to see if matcoloring is produced a valid iscoloring */
581: };
583: struct _p_MatTransposeColoring{
584: PETSCHEADER(int);
585: PetscInt M,N,m; /* total rows, columns; local rows */
586: PetscInt rstart; /* first row owned by local processor */
587: PetscInt ncolors; /* number of colors */
588: PetscInt *ncolumns; /* number of local columns for a color */
589: PetscInt *nrows; /* number of local rows for each color */
590: PetscInt currentcolor; /* color for which function evaluation is being done now */
591: ISColoringType ctype; /* IS_COLORING_GLOBAL or IS_COLORING_LOCAL */
593: PetscInt *colorforrow,*colorforcol; /* pointer to rows and columns */
594: PetscInt *rows; /* lists the local rows for each color (using the local row numbering) */
595: PetscInt *den2sp; /* maps (row,color) in the dense matrix to index of sparse matrix array a->a */
596: PetscInt *columns; /* lists the local columns of each color (using global column numbering) */
597: PetscInt brows; /* number of rows for efficient implementation of MatTransColoringApplyDenToSp() */
598: PetscInt *lstart; /* array used for loop over row blocks of Csparse */
599: };
601: /*
602: Null space context for preconditioner/operators
603: */
604: struct _p_MatNullSpace {
605: PETSCHEADER(int);
606: PetscBool has_cnst;
607: PetscInt n;
608: Vec* vecs;
609: PetscScalar* alpha; /* for projections */
610: PetscErrorCode (*remove)(MatNullSpace,Vec,void*); /* for user provided removal function */
611: void* rmctx; /* context for remove() function */
612: };
614: /*
615: Checking zero pivot for LU, ILU preconditioners.
616: */
617: typedef struct {
618: PetscInt nshift,nshift_max;
619: PetscReal shift_amount,shift_lo,shift_hi,shift_top,shift_fraction;
620: PetscBool newshift;
621: PetscReal rs; /* active row sum of abs(offdiagonals) */
622: PetscScalar pv; /* pivot of the active row */
623: } FactorShiftCtx;
625: /*
626: Used by MatCreateSubMatrices_MPIXAIJ_Local()
627: */
628: #include <petscctable.h>
629: typedef struct { /* used by MatCreateSubMatrices_MPIAIJ_SingleIS_Local() and MatCreateSubMatrices_MPIAIJ_Local */
630: PetscInt id; /* index of submats, only submats[0] is responsible for deleting some arrays below */
631: PetscInt nrqs,nrqr;
632: PetscInt **rbuf1,**rbuf2,**rbuf3,**sbuf1,**sbuf2;
633: PetscInt **ptr;
634: PetscInt *tmp;
635: PetscInt *ctr;
636: PetscInt *pa; /* proc array */
637: PetscInt *req_size,*req_source1,*req_source2;
638: PetscBool allcolumns,allrows;
639: PetscBool singleis;
640: PetscInt *row2proc; /* row to proc map */
641: PetscInt nstages;
642: #if defined(PETSC_USE_CTABLE)
643: PetscTable cmap,rmap;
644: PetscInt *cmap_loc,*rmap_loc;
645: #else
646: PetscInt *cmap,*rmap;
647: #endif
649: PetscErrorCode (*destroy)(Mat);
650: } Mat_SubSppt;
652: PETSC_EXTERN PetscErrorCode MatFactorDumpMatrix(Mat);
653: PETSC_INTERN PetscErrorCode MatShift_Basic(Mat,PetscScalar);
654: PETSC_INTERN PetscErrorCode MatSetBlockSizes_Default(Mat,PetscInt,PetscInt);
656: PETSC_STATIC_INLINE PetscErrorCode MatPivotCheck_nz(Mat mat,const MatFactorInfo *info,FactorShiftCtx *sctx,PetscInt row)
657: {
658: PetscReal _rs = sctx->rs;
659: PetscReal _zero = info->zeropivot*_rs;
662: if (PetscAbsScalar(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)){
663: /* force |diag| > zeropivot*rs */
664: if (!sctx->nshift) sctx->shift_amount = info->shiftamount;
665: else sctx->shift_amount *= 2.0;
666: sctx->newshift = PETSC_TRUE;
667: (sctx->nshift)++;
668: } else {
669: sctx->newshift = PETSC_FALSE;
670: }
671: return(0);
672: }
674: PETSC_STATIC_INLINE PetscErrorCode MatPivotCheck_pd(Mat mat,const MatFactorInfo *info,FactorShiftCtx *sctx,PetscInt row)
675: {
676: PetscReal _rs = sctx->rs;
677: PetscReal _zero = info->zeropivot*_rs;
680: if (PetscRealPart(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)){
681: /* force matfactor to be diagonally dominant */
682: if (sctx->nshift == sctx->nshift_max) {
683: sctx->shift_fraction = sctx->shift_hi;
684: } else {
685: sctx->shift_lo = sctx->shift_fraction;
686: sctx->shift_fraction = (sctx->shift_hi+sctx->shift_lo)/2.;
687: }
688: sctx->shift_amount = sctx->shift_fraction * sctx->shift_top;
689: sctx->nshift++;
690: sctx->newshift = PETSC_TRUE;
691: } else {
692: sctx->newshift = PETSC_FALSE;
693: }
694: return(0);
695: }
697: PETSC_STATIC_INLINE PetscErrorCode MatPivotCheck_inblocks(Mat mat,const MatFactorInfo *info,FactorShiftCtx *sctx,PetscInt row)
698: {
699: PetscReal _zero = info->zeropivot;
702: if (PetscAbsScalar(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)){
703: sctx->pv += info->shiftamount;
704: sctx->shift_amount = 0.0;
705: sctx->nshift++;
706: }
707: sctx->newshift = PETSC_FALSE;
708: return(0);
709: }
711: PETSC_STATIC_INLINE PetscErrorCode MatPivotCheck_none(Mat fact,Mat mat,const MatFactorInfo *info,FactorShiftCtx *sctx,PetscInt row)
712: {
713: PetscReal _zero = info->zeropivot;
717: sctx->newshift = PETSC_FALSE;
718: if (PetscAbsScalar(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)) {
719: if (!mat->erroriffailure) {
720: PetscInfo3(mat,"Detected zero pivot in factorization in row %D value %g tolerance %g\n",row,(double)PetscAbsScalar(sctx->pv),(double)_zero);
721: fact->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT;
722: fact->factorerror_zeropivot_value = PetscAbsScalar(sctx->pv);
723: fact->factorerror_zeropivot_row = row;
724: } else SETERRQ3(PETSC_COMM_SELF,PETSC_ERR_MAT_LU_ZRPVT,"Zero pivot row %D value %g tolerance %g\n",row,(double)PetscAbsScalar(sctx->pv),(double)_zero);
725: }
726: return(0);
727: }
729: PETSC_STATIC_INLINE PetscErrorCode MatPivotCheck(Mat fact,Mat mat,const MatFactorInfo *info,FactorShiftCtx *sctx,PetscInt row)
730: {
734: if (info->shifttype == (PetscReal) MAT_SHIFT_NONZERO){
735: MatPivotCheck_nz(mat,info,sctx,row);
736: } else if (info->shifttype == (PetscReal) MAT_SHIFT_POSITIVE_DEFINITE){
737: MatPivotCheck_pd(mat,info,sctx,row);
738: } else if (info->shifttype == (PetscReal) MAT_SHIFT_INBLOCKS){
739: MatPivotCheck_inblocks(mat,info,sctx,row);
740: } else {
741: MatPivotCheck_none(fact,mat,info,sctx,row);
742: }
743: return(0);
744: }
746: /*
747: Create and initialize a linked list
748: Input Parameters:
749: idx_start - starting index of the list
750: lnk_max - max value of lnk indicating the end of the list
751: nlnk - max length of the list
752: Output Parameters:
753: lnk - list initialized
754: bt - PetscBT (bitarray) with all bits set to false
755: lnk_empty - flg indicating the list is empty
756: */
757: #define PetscLLCreate(idx_start,lnk_max,nlnk,lnk,bt) \
758: (PetscMalloc1(nlnk,&lnk) || PetscBTCreate(nlnk,&(bt)) || (lnk[idx_start] = lnk_max,0))
760: #define PetscLLCreate_new(idx_start,lnk_max,nlnk,lnk,bt,lnk_empty)\
761: (PetscMalloc1(nlnk,&lnk) || PetscBTCreate(nlnk,&(bt)) || (lnk_empty = PETSC_TRUE,0) ||(lnk[idx_start] = lnk_max,0))
763: /*
764: Add an index set into a sorted linked list
765: Input Parameters:
766: nidx - number of input indices
767: indices - integer array
768: idx_start - starting index of the list
769: lnk - linked list(an integer array) that is created
770: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
771: output Parameters:
772: nlnk - number of newly added indices
773: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
774: bt - updated PetscBT (bitarray)
775: */
776: #define PetscLLAdd(nidx,indices,idx_start,nlnk,lnk,bt) 0;\
777: {\
778: PetscInt _k,_entry,_location,_lnkdata;\
779: nlnk = 0;\
780: _lnkdata = idx_start;\
781: for (_k=0; _k<nidx; _k++){\
782: _entry = indices[_k];\
783: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
784: /* search for insertion location */\
785: /* start from the beginning if _entry < previous _entry */\
786: if (_k && _entry < _lnkdata) _lnkdata = idx_start;\
787: do {\
788: _location = _lnkdata;\
789: _lnkdata = lnk[_location];\
790: } while (_entry > _lnkdata);\
791: /* insertion location is found, add entry into lnk */\
792: lnk[_location] = _entry;\
793: lnk[_entry] = _lnkdata;\
794: nlnk++;\
795: _lnkdata = _entry; /* next search starts from here if next_entry > _entry */\
796: }\
797: }\
798: }
800: /*
801: Add a permuted index set into a sorted linked list
802: Input Parameters:
803: nidx - number of input indices
804: indices - integer array
805: perm - permutation of indices
806: idx_start - starting index of the list
807: lnk - linked list(an integer array) that is created
808: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
809: output Parameters:
810: nlnk - number of newly added indices
811: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
812: bt - updated PetscBT (bitarray)
813: */
814: #define PetscLLAddPerm(nidx,indices,perm,idx_start,nlnk,lnk,bt) 0;\
815: {\
816: PetscInt _k,_entry,_location,_lnkdata;\
817: nlnk = 0;\
818: _lnkdata = idx_start;\
819: for (_k=0; _k<nidx; _k++){\
820: _entry = perm[indices[_k]];\
821: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
822: /* search for insertion location */\
823: /* start from the beginning if _entry < previous _entry */\
824: if (_k && _entry < _lnkdata) _lnkdata = idx_start;\
825: do {\
826: _location = _lnkdata;\
827: _lnkdata = lnk[_location];\
828: } while (_entry > _lnkdata);\
829: /* insertion location is found, add entry into lnk */\
830: lnk[_location] = _entry;\
831: lnk[_entry] = _lnkdata;\
832: nlnk++;\
833: _lnkdata = _entry; /* next search starts from here if next_entry > _entry */\
834: }\
835: }\
836: }
838: /*
839: Add a SORTED ascending index set into a sorted linked list - same as PetscLLAdd() bus skip 'if (_k && _entry < _lnkdata) _lnkdata = idx_start;'
840: Input Parameters:
841: nidx - number of input indices
842: indices - sorted integer array
843: idx_start - starting index of the list
844: lnk - linked list(an integer array) that is created
845: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
846: output Parameters:
847: nlnk - number of newly added indices
848: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
849: bt - updated PetscBT (bitarray)
850: */
851: #define PetscLLAddSorted(nidx,indices,idx_start,nlnk,lnk,bt) 0;\
852: {\
853: PetscInt _k,_entry,_location,_lnkdata;\
854: nlnk = 0;\
855: _lnkdata = idx_start;\
856: for (_k=0; _k<nidx; _k++){\
857: _entry = indices[_k];\
858: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
859: /* search for insertion location */\
860: do {\
861: _location = _lnkdata;\
862: _lnkdata = lnk[_location];\
863: } while (_entry > _lnkdata);\
864: /* insertion location is found, add entry into lnk */\
865: lnk[_location] = _entry;\
866: lnk[_entry] = _lnkdata;\
867: nlnk++;\
868: _lnkdata = _entry; /* next search starts from here */\
869: }\
870: }\
871: }
873: #define PetscLLAddSorted_new(nidx,indices,idx_start,lnk_empty,nlnk,lnk,bt) 0; \
874: {\
875: PetscInt _k,_entry,_location,_lnkdata;\
876: if (lnk_empty){\
877: _lnkdata = idx_start; \
878: for (_k=0; _k<nidx; _k++){ \
879: _entry = indices[_k]; \
880: PetscBTSet(bt,_entry); /* mark the new entry */ \
881: _location = _lnkdata; \
882: _lnkdata = lnk[_location]; \
883: /* insertion location is found, add entry into lnk */ \
884: lnk[_location] = _entry; \
885: lnk[_entry] = _lnkdata; \
886: _lnkdata = _entry; /* next search starts from here */ \
887: } \
888: /*\
889: lnk[indices[nidx-1]] = lnk[idx_start];\
890: lnk[idx_start] = indices[0];\
891: PetscBTSet(bt,indices[0]); \
892: for (_k=1; _k<nidx; _k++){ \
893: PetscBTSet(bt,indices[_k]); \
894: lnk[indices[_k-1]] = indices[_k]; \
895: } \
896: */\
897: nlnk = nidx;\
898: lnk_empty = PETSC_FALSE;\
899: } else {\
900: nlnk = 0; \
901: _lnkdata = idx_start; \
902: for (_k=0; _k<nidx; _k++){ \
903: _entry = indices[_k]; \
904: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */ \
905: /* search for insertion location */ \
906: do { \
907: _location = _lnkdata; \
908: _lnkdata = lnk[_location]; \
909: } while (_entry > _lnkdata); \
910: /* insertion location is found, add entry into lnk */ \
911: lnk[_location] = _entry; \
912: lnk[_entry] = _lnkdata; \
913: nlnk++; \
914: _lnkdata = _entry; /* next search starts from here */ \
915: } \
916: } \
917: } \
918: }
920: /*
921: Add a SORTED index set into a sorted linked list used for LUFactorSymbolic()
922: Same as PetscLLAddSorted() with an additional operation:
923: count the number of input indices that are no larger than 'diag'
924: Input Parameters:
925: indices - sorted integer array
926: idx_start - starting index of the list, index of pivot row
927: lnk - linked list(an integer array) that is created
928: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
929: diag - index of the active row in LUFactorSymbolic
930: nzbd - number of input indices with indices <= idx_start
931: im - im[idx_start] is initialized as num of nonzero entries in row=idx_start
932: output Parameters:
933: nlnk - number of newly added indices
934: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
935: bt - updated PetscBT (bitarray)
936: im - im[idx_start]: unchanged if diag is not an entry
937: : num of entries with indices <= diag if diag is an entry
938: */
939: #define PetscLLAddSortedLU(indices,idx_start,nlnk,lnk,bt,diag,nzbd,im) 0;\
940: {\
941: PetscInt _k,_entry,_location,_lnkdata,_nidx;\
942: nlnk = 0;\
943: _lnkdata = idx_start;\
944: _nidx = im[idx_start] - nzbd; /* num of entries with idx_start < index <= diag */\
945: for (_k=0; _k<_nidx; _k++){\
946: _entry = indices[_k];\
947: nzbd++;\
948: if ( _entry== diag) im[idx_start] = nzbd;\
949: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
950: /* search for insertion location */\
951: do {\
952: _location = _lnkdata;\
953: _lnkdata = lnk[_location];\
954: } while (_entry > _lnkdata);\
955: /* insertion location is found, add entry into lnk */\
956: lnk[_location] = _entry;\
957: lnk[_entry] = _lnkdata;\
958: nlnk++;\
959: _lnkdata = _entry; /* next search starts from here */\
960: }\
961: }\
962: }
964: /*
965: Copy data on the list into an array, then initialize the list
966: Input Parameters:
967: idx_start - starting index of the list
968: lnk_max - max value of lnk indicating the end of the list
969: nlnk - number of data on the list to be copied
970: lnk - linked list
971: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
972: output Parameters:
973: indices - array that contains the copied data
974: lnk - linked list that is cleaned and initialize
975: bt - PetscBT (bitarray) with all bits set to false
976: */
977: #define PetscLLClean(idx_start,lnk_max,nlnk,lnk,indices,bt) 0;\
978: {\
979: PetscInt _j,_idx=idx_start;\
980: for (_j=0; _j<nlnk; _j++){\
981: _idx = lnk[_idx];\
982: indices[_j] = _idx;\
983: PetscBTClear(bt,_idx);\
984: }\
985: lnk[idx_start] = lnk_max;\
986: }
987: /*
988: Free memories used by the list
989: */
990: #define PetscLLDestroy(lnk,bt) (PetscFree(lnk) || PetscBTDestroy(&(bt)))
992: /* Routines below are used for incomplete matrix factorization */
993: /*
994: Create and initialize a linked list and its levels
995: Input Parameters:
996: idx_start - starting index of the list
997: lnk_max - max value of lnk indicating the end of the list
998: nlnk - max length of the list
999: Output Parameters:
1000: lnk - list initialized
1001: lnk_lvl - array of size nlnk for storing levels of lnk
1002: bt - PetscBT (bitarray) with all bits set to false
1003: */
1004: #define PetscIncompleteLLCreate(idx_start,lnk_max,nlnk,lnk,lnk_lvl,bt)\
1005: (PetscIntMultError(2,nlnk,NULL) || PetscMalloc1(2*nlnk,&lnk) || PetscBTCreate(nlnk,&(bt)) || (lnk[idx_start] = lnk_max,lnk_lvl = lnk + nlnk,0))
1007: /*
1008: Initialize a sorted linked list used for ILU and ICC
1009: Input Parameters:
1010: nidx - number of input idx
1011: idx - integer array used for storing column indices
1012: idx_start - starting index of the list
1013: perm - indices of an IS
1014: lnk - linked list(an integer array) that is created
1015: lnklvl - levels of lnk
1016: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1017: output Parameters:
1018: nlnk - number of newly added idx
1019: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1020: lnklvl - levels of lnk
1021: bt - updated PetscBT (bitarray)
1022: */
1023: #define PetscIncompleteLLInit(nidx,idx,idx_start,perm,nlnk,lnk,lnklvl,bt) 0;\
1024: {\
1025: PetscInt _k,_entry,_location,_lnkdata;\
1026: nlnk = 0;\
1027: _lnkdata = idx_start;\
1028: for (_k=0; _k<nidx; _k++){\
1029: _entry = perm[idx[_k]];\
1030: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
1031: /* search for insertion location */\
1032: if (_k && _entry < _lnkdata) _lnkdata = idx_start;\
1033: do {\
1034: _location = _lnkdata;\
1035: _lnkdata = lnk[_location];\
1036: } while (_entry > _lnkdata);\
1037: /* insertion location is found, add entry into lnk */\
1038: lnk[_location] = _entry;\
1039: lnk[_entry] = _lnkdata;\
1040: lnklvl[_entry] = 0;\
1041: nlnk++;\
1042: _lnkdata = _entry; /* next search starts from here if next_entry > _entry */\
1043: }\
1044: }\
1045: }
1047: /*
1048: Add a SORTED index set into a sorted linked list for ILU
1049: Input Parameters:
1050: nidx - number of input indices
1051: idx - sorted integer array used for storing column indices
1052: level - level of fill, e.g., ICC(level)
1053: idxlvl - level of idx
1054: idx_start - starting index of the list
1055: lnk - linked list(an integer array) that is created
1056: lnklvl - levels of lnk
1057: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1058: prow - the row number of idx
1059: output Parameters:
1060: nlnk - number of newly added idx
1061: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1062: lnklvl - levels of lnk
1063: bt - updated PetscBT (bitarray)
1065: Note: the level of factor(i,j) is set as lvl(i,j) = min{ lvl(i,j), lvl(i,prow)+lvl(prow,j)+1)
1066: where idx = non-zero columns of U(prow,prow+1:n-1), prow<i
1067: */
1068: #define PetscILULLAddSorted(nidx,idx,level,idxlvl,idx_start,nlnk,lnk,lnklvl,bt,lnklvl_prow) 0;\
1069: {\
1070: PetscInt _k,_entry,_location,_lnkdata,_incrlev,_lnklvl_prow=lnklvl[prow];\
1071: nlnk = 0;\
1072: _lnkdata = idx_start;\
1073: for (_k=0; _k<nidx; _k++){\
1074: _incrlev = idxlvl[_k] + _lnklvl_prow + 1;\
1075: if (_incrlev > level) continue;\
1076: _entry = idx[_k];\
1077: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
1078: /* search for insertion location */\
1079: do {\
1080: _location = _lnkdata;\
1081: _lnkdata = lnk[_location];\
1082: } while (_entry > _lnkdata);\
1083: /* insertion location is found, add entry into lnk */\
1084: lnk[_location] = _entry;\
1085: lnk[_entry] = _lnkdata;\
1086: lnklvl[_entry] = _incrlev;\
1087: nlnk++;\
1088: _lnkdata = _entry; /* next search starts from here if next_entry > _entry */\
1089: } else { /* existing entry: update lnklvl */\
1090: if (lnklvl[_entry] > _incrlev) lnklvl[_entry] = _incrlev;\
1091: }\
1092: }\
1093: }
1095: /*
1096: Add a index set into a sorted linked list
1097: Input Parameters:
1098: nidx - number of input idx
1099: idx - integer array used for storing column indices
1100: level - level of fill, e.g., ICC(level)
1101: idxlvl - level of idx
1102: idx_start - starting index of the list
1103: lnk - linked list(an integer array) that is created
1104: lnklvl - levels of lnk
1105: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1106: output Parameters:
1107: nlnk - number of newly added idx
1108: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1109: lnklvl - levels of lnk
1110: bt - updated PetscBT (bitarray)
1111: */
1112: #define PetscIncompleteLLAdd(nidx,idx,level,idxlvl,idx_start,nlnk,lnk,lnklvl,bt) 0;\
1113: {\
1114: PetscInt _k,_entry,_location,_lnkdata,_incrlev;\
1115: nlnk = 0;\
1116: _lnkdata = idx_start;\
1117: for (_k=0; _k<nidx; _k++){\
1118: _incrlev = idxlvl[_k] + 1;\
1119: if (_incrlev > level) continue;\
1120: _entry = idx[_k];\
1121: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
1122: /* search for insertion location */\
1123: if (_k && _entry < _lnkdata) _lnkdata = idx_start;\
1124: do {\
1125: _location = _lnkdata;\
1126: _lnkdata = lnk[_location];\
1127: } while (_entry > _lnkdata);\
1128: /* insertion location is found, add entry into lnk */\
1129: lnk[_location] = _entry;\
1130: lnk[_entry] = _lnkdata;\
1131: lnklvl[_entry] = _incrlev;\
1132: nlnk++;\
1133: _lnkdata = _entry; /* next search starts from here if next_entry > _entry */\
1134: } else { /* existing entry: update lnklvl */\
1135: if (lnklvl[_entry] > _incrlev) lnklvl[_entry] = _incrlev;\
1136: }\
1137: }\
1138: }
1140: /*
1141: Add a SORTED index set into a sorted linked list
1142: Input Parameters:
1143: nidx - number of input indices
1144: idx - sorted integer array used for storing column indices
1145: level - level of fill, e.g., ICC(level)
1146: idxlvl - level of idx
1147: idx_start - starting index of the list
1148: lnk - linked list(an integer array) that is created
1149: lnklvl - levels of lnk
1150: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1151: output Parameters:
1152: nlnk - number of newly added idx
1153: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1154: lnklvl - levels of lnk
1155: bt - updated PetscBT (bitarray)
1156: */
1157: #define PetscIncompleteLLAddSorted(nidx,idx,level,idxlvl,idx_start,nlnk,lnk,lnklvl,bt) 0;\
1158: {\
1159: PetscInt _k,_entry,_location,_lnkdata,_incrlev;\
1160: nlnk = 0;\
1161: _lnkdata = idx_start;\
1162: for (_k=0; _k<nidx; _k++){\
1163: _incrlev = idxlvl[_k] + 1;\
1164: if (_incrlev > level) continue;\
1165: _entry = idx[_k];\
1166: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
1167: /* search for insertion location */\
1168: do {\
1169: _location = _lnkdata;\
1170: _lnkdata = lnk[_location];\
1171: } while (_entry > _lnkdata);\
1172: /* insertion location is found, add entry into lnk */\
1173: lnk[_location] = _entry;\
1174: lnk[_entry] = _lnkdata;\
1175: lnklvl[_entry] = _incrlev;\
1176: nlnk++;\
1177: _lnkdata = _entry; /* next search starts from here */\
1178: } else { /* existing entry: update lnklvl */\
1179: if (lnklvl[_entry] > _incrlev) lnklvl[_entry] = _incrlev;\
1180: }\
1181: }\
1182: }
1184: /*
1185: Add a SORTED index set into a sorted linked list for ICC
1186: Input Parameters:
1187: nidx - number of input indices
1188: idx - sorted integer array used for storing column indices
1189: level - level of fill, e.g., ICC(level)
1190: idxlvl - level of idx
1191: idx_start - starting index of the list
1192: lnk - linked list(an integer array) that is created
1193: lnklvl - levels of lnk
1194: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1195: idxlvl_prow - idxlvl[prow], where prow is the row number of the idx
1196: output Parameters:
1197: nlnk - number of newly added indices
1198: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1199: lnklvl - levels of lnk
1200: bt - updated PetscBT (bitarray)
1201: Note: the level of U(i,j) is set as lvl(i,j) = min{ lvl(i,j), lvl(prow,i)+lvl(prow,j)+1)
1202: where idx = non-zero columns of U(prow,prow+1:n-1), prow<i
1203: */
1204: #define PetscICCLLAddSorted(nidx,idx,level,idxlvl,idx_start,nlnk,lnk,lnklvl,bt,idxlvl_prow) 0;\
1205: {\
1206: PetscInt _k,_entry,_location,_lnkdata,_incrlev;\
1207: nlnk = 0;\
1208: _lnkdata = idx_start;\
1209: for (_k=0; _k<nidx; _k++){\
1210: _incrlev = idxlvl[_k] + idxlvl_prow + 1;\
1211: if (_incrlev > level) continue;\
1212: _entry = idx[_k];\
1213: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
1214: /* search for insertion location */\
1215: do {\
1216: _location = _lnkdata;\
1217: _lnkdata = lnk[_location];\
1218: } while (_entry > _lnkdata);\
1219: /* insertion location is found, add entry into lnk */\
1220: lnk[_location] = _entry;\
1221: lnk[_entry] = _lnkdata;\
1222: lnklvl[_entry] = _incrlev;\
1223: nlnk++;\
1224: _lnkdata = _entry; /* next search starts from here */\
1225: } else { /* existing entry: update lnklvl */\
1226: if (lnklvl[_entry] > _incrlev) lnklvl[_entry] = _incrlev;\
1227: }\
1228: }\
1229: }
1231: /*
1232: Copy data on the list into an array, then initialize the list
1233: Input Parameters:
1234: idx_start - starting index of the list
1235: lnk_max - max value of lnk indicating the end of the list
1236: nlnk - number of data on the list to be copied
1237: lnk - linked list
1238: lnklvl - level of lnk
1239: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1240: output Parameters:
1241: indices - array that contains the copied data
1242: lnk - linked list that is cleaned and initialize
1243: lnklvl - level of lnk that is reinitialized
1244: bt - PetscBT (bitarray) with all bits set to false
1245: */
1246: #define PetscIncompleteLLClean(idx_start,lnk_max,nlnk,lnk,lnklvl,indices,indiceslvl,bt) 0;\
1247: {\
1248: PetscInt _j,_idx=idx_start;\
1249: for (_j=0; _j<nlnk; _j++){\
1250: _idx = lnk[_idx];\
1251: *(indices+_j) = _idx;\
1252: *(indiceslvl+_j) = lnklvl[_idx];\
1253: lnklvl[_idx] = -1;\
1254: PetscBTClear(bt,_idx);\
1255: }\
1256: lnk[idx_start] = lnk_max;\
1257: }
1258: /*
1259: Free memories used by the list
1260: */
1261: #define PetscIncompleteLLDestroy(lnk,bt) (PetscFree(lnk) || PetscBTDestroy(&(bt)))
1263: #define MatCheckSameLocalSize(A,ar1,B,ar2) \
1265: if ((A->rmap->n != B->rmap->n) || (A->cmap->n != B->cmap->n)) SETERRQ6(PETSC_COMM_SELF,PETSC_ERR_ARG_INCOMP,"Incompatible matrix local sizes: parameter # %d (%D x %D) != parameter # %d (%D x %D)",ar1,A->rmap->n,A->cmap->n,ar2,B->rmap->n,B->cmap->n);
1266:
1267: #define MatCheckSameSize(A,ar1,B,ar2) \
1268: if ((A->rmap->N != B->rmap->N) || (A->cmap->N != B->cmap->N)) SETERRQ6(PetscObjectComm((PetscObject)A),PETSC_ERR_ARG_INCOMP,"Incompatible matrix global sizes: parameter # %d (%D x %D) != parameter # %d (%D x %D)",ar1,A->rmap->N,A->cmap->N,ar2,B->rmap->N,B->cmap->N);\
1269: MatCheckSameLocalSize(A,ar1,B,ar2);
1270:
1271: #define VecCheckMatCompatible(M,x,ar1,b,ar2) \
1272: if (M->cmap->N != x->map->N) SETERRQ3(PetscObjectComm((PetscObject)M),PETSC_ERR_ARG_SIZ,"Vector global length incompatible with matrix: parameter # %d global size %D != matrix column global size %D",ar1,x->map->N,M->cmap->N);\
1273: if (M->rmap->N != b->map->N) SETERRQ3(PetscObjectComm((PetscObject)M),PETSC_ERR_ARG_SIZ,"Vector global length incompatible with matrix: parameter # %d global size %D != matrix row global size %D",ar2,b->map->N,M->rmap->N);
1275: /* -------------------------------------------------------------------------------------------------------*/
1276: #include <petscbt.h>
1277: /*
1278: Create and initialize a condensed linked list -
1279: same as PetscLLCreate(), but uses a scalable array 'lnk' with size of max number of entries, not O(N).
1280: Barry suggested this approach (Dec. 6, 2011):
1281: I've thought of an alternative way of representing a linked list that is efficient but doesn't have the O(N) scaling issue
1282: (it may be faster than the O(N) even sequentially due to less crazy memory access).
1284: 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
1285: 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
1286: 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
1287: 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.
1288: 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
1289: to each other so memory access is much better than using the big array.
1291: Example:
1292: nlnk_max=5, lnk_max=36:
1293: Initial list: [0, 0 | 36, 2 | 0, 0 | 0, 0 | 0, 0 | 0, 0 | 0, 0]
1294: here, head_node has index 2 with value lnk[2]=lnk_max=36,
1295: 0-th entry is used to store the number of entries in the list,
1296: The initial lnk represents head -> tail(marked by 36) with number of entries = lnk[0]=0.
1298: Now adding a sorted set {2,4}, the list becomes
1299: [2, 0 | 36, 4 |2, 6 | 4, 2 | 0, 0 | 0, 0 | 0, 0 ]
1300: represents head -> 2 -> 4 -> tail with number of entries = lnk[0]=2.
1302: Then adding a sorted set {0,3,35}, the list
1303: [5, 0 | 36, 8 | 2, 10 | 4, 12 | 0, 4 | 3, 6 | 35, 2 ]
1304: represents head -> 0 -> 2 -> 3 -> 4 -> 35 -> tail with number of entries = lnk[0]=5.
1306: Input Parameters:
1307: nlnk_max - max length of the list
1308: lnk_max - max value of the entries
1309: Output Parameters:
1310: lnk - list created and initialized
1311: bt - PetscBT (bitarray) with all bits set to false. Note: bt has size lnk_max, not nln_max!
1312: */
1313: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedCreate(PetscInt nlnk_max,PetscInt lnk_max,PetscInt **lnk,PetscBT *bt)
1314: {
1316: PetscInt *llnk,lsize = 0;
1319: PetscIntMultError(2,nlnk_max+2,&lsize);
1320: PetscMalloc1(lsize,lnk);
1321: PetscBTCreate(lnk_max,bt);
1322: llnk = *lnk;
1323: llnk[0] = 0; /* number of entries on the list */
1324: llnk[2] = lnk_max; /* value in the head node */
1325: llnk[3] = 2; /* next for the head node */
1326: return(0);
1327: }
1329: /*
1330: Add a SORTED ascending index set into a sorted linked list. See PetscLLCondensedCreate() for detailed description.
1331: Input Parameters:
1332: nidx - number of input indices
1333: indices - sorted integer array
1334: lnk - condensed linked list(an integer array) that is created
1335: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1336: output Parameters:
1337: lnk - the sorted(increasing order) linked list containing previous and newly added non-redundate indices
1338: bt - updated PetscBT (bitarray)
1339: */
1340: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedAddSorted(PetscInt nidx,const PetscInt indices[],PetscInt lnk[],PetscBT bt)
1341: {
1342: PetscInt _k,_entry,_location,_next,_lnkdata,_nlnk,_newnode;
1345: _nlnk = lnk[0]; /* num of entries on the input lnk */
1346: _location = 2; /* head */
1347: for (_k=0; _k<nidx; _k++){
1348: _entry = indices[_k];
1349: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */
1350: /* search for insertion location */
1351: do {
1352: _next = _location + 1; /* link from previous node to next node */
1353: _location = lnk[_next]; /* idx of next node */
1354: _lnkdata = lnk[_location];/* value of next node */
1355: } while (_entry > _lnkdata);
1356: /* insertion location is found, add entry into lnk */
1357: _newnode = 2*(_nlnk+2); /* index for this new node */
1358: lnk[_next] = _newnode; /* connect previous node to the new node */
1359: lnk[_newnode] = _entry; /* set value of the new node */
1360: lnk[_newnode+1] = _location; /* connect new node to next node */
1361: _location = _newnode; /* next search starts from the new node */
1362: _nlnk++;
1363: } \
1364: }\
1365: lnk[0] = _nlnk; /* number of entries in the list */
1366: return(0);
1367: }
1369: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedClean(PetscInt lnk_max,PetscInt nidx,PetscInt *indices,PetscInt lnk[],PetscBT bt)
1370: {
1372: PetscInt _k,_next,_nlnk;
1375: _next = lnk[3]; /* head node */
1376: _nlnk = lnk[0]; /* num of entries on the list */
1377: for (_k=0; _k<_nlnk; _k++){
1378: indices[_k] = lnk[_next];
1379: _next = lnk[_next + 1];
1380: PetscBTClear(bt,indices[_k]);
1381: }
1382: lnk[0] = 0; /* num of entries on the list */
1383: lnk[2] = lnk_max; /* initialize head node */
1384: lnk[3] = 2; /* head node */
1385: return(0);
1386: }
1388: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedView(PetscInt *lnk)
1389: {
1391: PetscInt k;
1394: PetscPrintf(PETSC_COMM_SELF,"LLCondensed of size %D, (val, next)\n",lnk[0]);
1395: for (k=2; k< lnk[0]+2; k++){
1396: PetscPrintf(PETSC_COMM_SELF," %D: (%D, %D)\n",2*k,lnk[2*k],lnk[2*k+1]);
1397: }
1398: return(0);
1399: }
1401: /*
1402: Free memories used by the list
1403: */
1404: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedDestroy(PetscInt *lnk,PetscBT bt)
1405: {
1409: PetscFree(lnk);
1410: PetscBTDestroy(&bt);
1411: return(0);
1412: }
1414: /* -------------------------------------------------------------------------------------------------------*/
1415: /*
1416: Same as PetscLLCondensedCreate(), but does not use non-scalable O(lnk_max) bitarray
1417: Input Parameters:
1418: nlnk_max - max length of the list
1419: Output Parameters:
1420: lnk - list created and initialized
1421: */
1422: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedCreate_Scalable(PetscInt nlnk_max,PetscInt **lnk)
1423: {
1425: PetscInt *llnk,lsize = 0;
1428: PetscIntMultError(2,nlnk_max+2,&lsize);
1429: PetscMalloc1(lsize,lnk);
1430: llnk = *lnk;
1431: llnk[0] = 0; /* number of entries on the list */
1432: llnk[2] = PETSC_MAX_INT; /* value in the head node */
1433: llnk[3] = 2; /* next for the head node */
1434: return(0);
1435: }
1437: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedExpand_Scalable(PetscInt nlnk_max,PetscInt **lnk)
1438: {
1440: PetscInt lsize = 0;
1443: PetscIntMultError(2,nlnk_max+2,&lsize);
1444: PetscRealloc(lsize*sizeof(PetscInt),lnk);
1445: return(0);
1446: }
1448: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedAddSorted_Scalable(PetscInt nidx,const PetscInt indices[],PetscInt lnk[])
1449: {
1450: PetscInt _k,_entry,_location,_next,_lnkdata,_nlnk,_newnode;
1451: _nlnk = lnk[0]; /* num of entries on the input lnk */
1452: _location = 2; /* head */ \
1453: for (_k=0; _k<nidx; _k++){
1454: _entry = indices[_k];
1455: /* search for insertion location */
1456: do {
1457: _next = _location + 1; /* link from previous node to next node */
1458: _location = lnk[_next]; /* idx of next node */
1459: _lnkdata = lnk[_location];/* value of next node */
1460: } while (_entry > _lnkdata);
1461: if (_entry < _lnkdata) {
1462: /* insertion location is found, add entry into lnk */
1463: _newnode = 2*(_nlnk+2); /* index for this new node */
1464: lnk[_next] = _newnode; /* connect previous node to the new node */
1465: lnk[_newnode] = _entry; /* set value of the new node */
1466: lnk[_newnode+1] = _location; /* connect new node to next node */
1467: _location = _newnode; /* next search starts from the new node */
1468: _nlnk++;
1469: }
1470: }
1471: lnk[0] = _nlnk; /* number of entries in the list */
1472: return 0;
1473: }
1475: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedClean_Scalable(PetscInt nidx,PetscInt *indices,PetscInt *lnk)
1476: {
1477: PetscInt _k,_next,_nlnk;
1478: _next = lnk[3]; /* head node */
1479: _nlnk = lnk[0];
1480: for (_k=0; _k<_nlnk; _k++){
1481: indices[_k] = lnk[_next];
1482: _next = lnk[_next + 1];
1483: }
1484: lnk[0] = 0; /* num of entries on the list */
1485: lnk[3] = 2; /* head node */
1486: return 0;
1487: }
1489: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedDestroy_Scalable(PetscInt *lnk)
1490: {
1491: return PetscFree(lnk);
1492: }
1494: /* -------------------------------------------------------------------------------------------------------*/
1495: /*
1496: lnk[0] number of links
1497: lnk[1] number of entries
1498: lnk[3n] value
1499: lnk[3n+1] len
1500: lnk[3n+2] link to next value
1502: The next three are always the first link
1504: lnk[3] PETSC_MIN_INT+1
1505: lnk[4] 1
1506: lnk[5] link to first real entry
1508: The next three are always the last link
1510: lnk[6] PETSC_MAX_INT - 1
1511: lnk[7] 1
1512: lnk[8] next valid link (this is the same as lnk[0] but without the decreases)
1513: */
1515: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedCreate_fast(PetscInt nlnk_max,PetscInt **lnk)
1516: {
1518: PetscInt *llnk,lsize = 0;
1521: PetscIntMultError(3,nlnk_max+3,&lsize);
1522: PetscMalloc1(lsize,lnk);
1523: llnk = *lnk;
1524: llnk[0] = 0; /* nlnk: number of entries on the list */
1525: llnk[1] = 0; /* number of integer entries represented in list */
1526: llnk[3] = PETSC_MIN_INT+1; /* value in the first node */
1527: llnk[4] = 1; /* count for the first node */
1528: llnk[5] = 6; /* next for the first node */
1529: llnk[6] = PETSC_MAX_INT-1; /* value in the last node */
1530: llnk[7] = 1; /* count for the last node */
1531: llnk[8] = 0; /* next valid node to be used */
1532: return(0);
1533: }
1535: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedAddSorted_fast(PetscInt nidx,const PetscInt indices[],PetscInt lnk[])
1536: {
1537: PetscInt k,entry,prev,next;
1538: prev = 3; /* first value */
1539: next = lnk[prev+2];
1540: for (k=0; k<nidx; k++){
1541: entry = indices[k];
1542: /* search for insertion location */
1543: while (entry >= lnk[next]) {
1544: prev = next;
1545: next = lnk[next+2];
1546: }
1547: /* entry is in range of previous list */
1548: if (entry < lnk[prev]+lnk[prev+1]) continue;
1549: lnk[1]++;
1550: /* entry is right after previous list */
1551: if (entry == lnk[prev]+lnk[prev+1]) {
1552: lnk[prev+1]++;
1553: if (lnk[next] == entry+1) { /* combine two contiguous strings */
1554: lnk[prev+1] += lnk[next+1];
1555: lnk[prev+2] = lnk[next+2];
1556: next = lnk[next+2];
1557: lnk[0]--;
1558: }
1559: continue;
1560: }
1561: /* entry is right before next list */
1562: if (entry == lnk[next]-1) {
1563: lnk[next]--;
1564: lnk[next+1]++;
1565: prev = next;
1566: next = lnk[prev+2];
1567: continue;
1568: }
1569: /* add entry into lnk */
1570: lnk[prev+2] = 3*((lnk[8]++)+3); /* connect previous node to the new node */
1571: prev = lnk[prev+2];
1572: lnk[prev] = entry; /* set value of the new node */
1573: lnk[prev+1] = 1; /* number of values in contiguous string is one to start */
1574: lnk[prev+2] = next; /* connect new node to next node */
1575: lnk[0]++;
1576: }
1577: return 0;
1578: }
1580: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedClean_fast(PetscInt nidx,PetscInt *indices,PetscInt *lnk)
1581: {
1582: PetscInt _k,_next,_nlnk,cnt,j;
1583: _next = lnk[5]; /* first node */
1584: _nlnk = lnk[0];
1585: cnt = 0;
1586: for (_k=0; _k<_nlnk; _k++){
1587: for (j=0; j<lnk[_next+1]; j++) {
1588: indices[cnt++] = lnk[_next] + j;
1589: }
1590: _next = lnk[_next + 2];
1591: }
1592: lnk[0] = 0; /* nlnk: number of links */
1593: lnk[1] = 0; /* number of integer entries represented in list */
1594: lnk[3] = PETSC_MIN_INT+1; /* value in the first node */
1595: lnk[4] = 1; /* count for the first node */
1596: lnk[5] = 6; /* next for the first node */
1597: lnk[6] = PETSC_MAX_INT-1; /* value in the last node */
1598: lnk[7] = 1; /* count for the last node */
1599: lnk[8] = 0; /* next valid location to make link */
1600: return 0;
1601: }
1603: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedView_fast(PetscInt *lnk)
1604: {
1605: PetscInt k,next,nlnk;
1606: next = lnk[5]; /* first node */
1607: nlnk = lnk[0];
1608: for (k=0; k<nlnk; k++){
1609: #if 0 /* Debugging code */
1610: printf("%d value %d len %d next %d\n",next,lnk[next],lnk[next+1],lnk[next+2]);
1611: #endif
1612: next = lnk[next + 2];
1613: }
1614: return 0;
1615: }
1617: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedDestroy_fast(PetscInt *lnk)
1618: {
1619: return PetscFree(lnk);
1620: }
1622: /* this is extern because it is used in MatFDColoringUseDM() which is in the DM library */
1623: PETSC_EXTERN PetscErrorCode MatFDColoringApply_AIJ(Mat,MatFDColoring,Vec,void*);
1625: PETSC_EXTERN PetscLogEvent MAT_Mult;
1626: PETSC_EXTERN PetscLogEvent MAT_MultMatrixFree;
1627: PETSC_EXTERN PetscLogEvent MAT_Mults;
1628: PETSC_EXTERN PetscLogEvent MAT_MultConstrained;
1629: PETSC_EXTERN PetscLogEvent MAT_MultAdd;
1630: PETSC_EXTERN PetscLogEvent MAT_MultTranspose;
1631: PETSC_EXTERN PetscLogEvent MAT_MultTransposeConstrained;
1632: PETSC_EXTERN PetscLogEvent MAT_MultTransposeAdd;
1633: PETSC_EXTERN PetscLogEvent MAT_Solve;
1634: PETSC_EXTERN PetscLogEvent MAT_Solves;
1635: PETSC_EXTERN PetscLogEvent MAT_SolveAdd;
1636: PETSC_EXTERN PetscLogEvent MAT_SolveTranspose;
1637: PETSC_EXTERN PetscLogEvent MAT_SolveTransposeAdd;
1638: PETSC_EXTERN PetscLogEvent MAT_SOR;
1639: PETSC_EXTERN PetscLogEvent MAT_ForwardSolve;
1640: PETSC_EXTERN PetscLogEvent MAT_BackwardSolve;
1641: PETSC_EXTERN PetscLogEvent MAT_LUFactor;
1642: PETSC_EXTERN PetscLogEvent MAT_LUFactorSymbolic;
1643: PETSC_EXTERN PetscLogEvent MAT_LUFactorNumeric;
1644: PETSC_EXTERN PetscLogEvent MAT_CholeskyFactor;
1645: PETSC_EXTERN PetscLogEvent MAT_CholeskyFactorSymbolic;
1646: PETSC_EXTERN PetscLogEvent MAT_CholeskyFactorNumeric;
1647: PETSC_EXTERN PetscLogEvent MAT_ILUFactor;
1648: PETSC_EXTERN PetscLogEvent MAT_ILUFactorSymbolic;
1649: PETSC_EXTERN PetscLogEvent MAT_ICCFactorSymbolic;
1650: PETSC_EXTERN PetscLogEvent MAT_Copy;
1651: PETSC_EXTERN PetscLogEvent MAT_Convert;
1652: PETSC_EXTERN PetscLogEvent MAT_Scale;
1653: PETSC_EXTERN PetscLogEvent MAT_AssemblyBegin;
1654: PETSC_EXTERN PetscLogEvent MAT_AssemblyEnd;
1655: PETSC_EXTERN PetscLogEvent MAT_SetValues;
1656: PETSC_EXTERN PetscLogEvent MAT_GetValues;
1657: PETSC_EXTERN PetscLogEvent MAT_GetRow;
1658: PETSC_EXTERN PetscLogEvent MAT_GetRowIJ;
1659: PETSC_EXTERN PetscLogEvent MAT_CreateSubMats;
1660: PETSC_EXTERN PetscLogEvent MAT_GetColoring;
1661: PETSC_EXTERN PetscLogEvent MAT_GetOrdering;
1662: PETSC_EXTERN PetscLogEvent MAT_RedundantMat;
1663: PETSC_EXTERN PetscLogEvent MAT_IncreaseOverlap;
1664: PETSC_EXTERN PetscLogEvent MAT_Partitioning;
1665: PETSC_EXTERN PetscLogEvent MAT_PartitioningND;
1666: PETSC_EXTERN PetscLogEvent MAT_Coarsen;
1667: PETSC_EXTERN PetscLogEvent MAT_ZeroEntries;
1668: PETSC_EXTERN PetscLogEvent MAT_Load;
1669: PETSC_EXTERN PetscLogEvent MAT_View;
1670: PETSC_EXTERN PetscLogEvent MAT_AXPY;
1671: PETSC_EXTERN PetscLogEvent MAT_FDColoringCreate;
1672: PETSC_EXTERN PetscLogEvent MAT_TransposeColoringCreate;
1673: PETSC_EXTERN PetscLogEvent MAT_FDColoringSetUp;
1674: PETSC_EXTERN PetscLogEvent MAT_FDColoringApply;
1675: PETSC_EXTERN PetscLogEvent MAT_Transpose;
1676: PETSC_EXTERN PetscLogEvent MAT_FDColoringFunction;
1677: PETSC_EXTERN PetscLogEvent MAT_CreateSubMat;
1678: PETSC_EXTERN PetscLogEvent MAT_MatMult;
1679: PETSC_EXTERN PetscLogEvent MAT_MatSolve;
1680: PETSC_EXTERN PetscLogEvent MAT_MatTrSolve;
1681: PETSC_EXTERN PetscLogEvent MAT_MatMultSymbolic;
1682: PETSC_EXTERN PetscLogEvent MAT_MatMultNumeric;
1683: PETSC_EXTERN PetscLogEvent MAT_Getlocalmatcondensed;
1684: PETSC_EXTERN PetscLogEvent MAT_GetBrowsOfAcols;
1685: PETSC_EXTERN PetscLogEvent MAT_GetBrowsOfAocols;
1686: PETSC_EXTERN PetscLogEvent MAT_PtAP;
1687: PETSC_EXTERN PetscLogEvent MAT_PtAPSymbolic;
1688: PETSC_EXTERN PetscLogEvent MAT_PtAPNumeric;
1689: PETSC_EXTERN PetscLogEvent MAT_Seqstompinum;
1690: PETSC_EXTERN PetscLogEvent MAT_Seqstompisym;
1691: PETSC_EXTERN PetscLogEvent MAT_Seqstompi;
1692: PETSC_EXTERN PetscLogEvent MAT_Getlocalmat;
1693: PETSC_EXTERN PetscLogEvent MAT_RARt;
1694: PETSC_EXTERN PetscLogEvent MAT_RARtSymbolic;
1695: PETSC_EXTERN PetscLogEvent MAT_RARtNumeric;
1696: PETSC_EXTERN PetscLogEvent MAT_MatTransposeMult;
1697: PETSC_EXTERN PetscLogEvent MAT_MatTransposeMultSymbolic;
1698: PETSC_EXTERN PetscLogEvent MAT_MatTransposeMultNumeric;
1699: PETSC_EXTERN PetscLogEvent MAT_TransposeMatMult;
1700: PETSC_EXTERN PetscLogEvent MAT_TransposeMatMultSymbolic;
1701: PETSC_EXTERN PetscLogEvent MAT_TransposeMatMultNumeric;
1702: PETSC_EXTERN PetscLogEvent MAT_MatMatMult;
1703: PETSC_EXTERN PetscLogEvent MAT_MatMatMultSymbolic;
1704: PETSC_EXTERN PetscLogEvent MAT_MatMatMultNumeric;
1705: PETSC_EXTERN PetscLogEvent MAT_Applypapt;
1706: PETSC_EXTERN PetscLogEvent MAT_Applypapt_symbolic;
1707: PETSC_EXTERN PetscLogEvent MAT_Applypapt_numeric;
1708: PETSC_EXTERN PetscLogEvent MAT_Getsymtranspose;
1709: PETSC_EXTERN PetscLogEvent MAT_Transpose_SeqAIJ;
1710: PETSC_EXTERN PetscLogEvent MAT_Getsymtransreduced;
1711: PETSC_EXTERN PetscLogEvent MAT_GetSequentialNonzeroStructure;
1712: PETSC_EXTERN PetscLogEvent MATMFFD_Mult;
1713: PETSC_EXTERN PetscLogEvent MAT_GetMultiProcBlock;
1714: PETSC_EXTERN PetscLogEvent MAT_CUSPARSECopyToGPU;
1715: PETSC_EXTERN PetscLogEvent MAT_SetValuesBatch;
1716: PETSC_EXTERN PetscLogEvent MAT_ViennaCLCopyToGPU;
1717: PETSC_EXTERN PetscLogEvent MAT_Merge;
1718: PETSC_EXTERN PetscLogEvent MAT_Residual;
1719: PETSC_EXTERN PetscLogEvent MAT_SetRandom;
1720: PETSC_EXTERN PetscLogEvent MATCOLORING_Apply;
1721: PETSC_EXTERN PetscLogEvent MATCOLORING_Comm;
1722: PETSC_EXTERN PetscLogEvent MATCOLORING_Local;
1723: PETSC_EXTERN PetscLogEvent MATCOLORING_ISCreate;
1724: PETSC_EXTERN PetscLogEvent MATCOLORING_SetUp;
1725: PETSC_EXTERN PetscLogEvent MATCOLORING_Weights;
1727: #endif