Actual source code: matimpl.h
petsc-3.13.6 2020-09-29
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 src/mat/f90-mod/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 (*placeholder_63)(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 (*placeholder_89)(Mat,void*);
142: PetscErrorCode (*matmultsymbolic)(Mat,Mat,PetscReal,Mat);
143: PetscErrorCode (*matmultnumeric)(Mat,Mat,Mat);
144: PetscErrorCode (*placeholder_92)(Mat,void*);
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 (*placeholder_95)(Mat,void*);
149: PetscErrorCode (*mattransposemultsymbolic)(Mat,Mat,PetscReal,Mat);
150: PetscErrorCode (*mattransposemultnumeric)(Mat,Mat,Mat);
151: PetscErrorCode (*bindtocpu)(Mat,PetscBool);
152: /*99*/
153: PetscErrorCode (*productsetfromoptions)(Mat);
154: PetscErrorCode (*productsymbolic)(Mat);
155: PetscErrorCode (*productnumeric)(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 (*placeholder_130)(Mat,void*);
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 (*placeholder_136)(Mat,void*);
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: PetscErrorCode (*creatempimatconcatenateseqmat)(MPI_Comm,Mat,PetscInt,MatReuse,Mat*);
207: /*145*/
208: PetscErrorCode (*destroysubmatrices)(PetscInt,Mat*[]);
209: PetscErrorCode (*mattransposesolve)(Mat,Mat,Mat);
210: PetscErrorCode (*getvalueslocal)(Mat,PetscInt,const PetscInt[],PetscInt,const PetscInt[],PetscScalar[]);
211: };
212: /*
213: If you add MatOps entries above also add them to the MATOP enum
214: in include/petscmat.h and src/mat/f90-mod/petscmat.h
215: */
217: #include <petscsys.h>
218: PETSC_EXTERN PetscErrorCode MatRegisterOp(MPI_Comm, const char[], PetscVoidFunction, const char[], PetscInt, ...);
219: PETSC_EXTERN PetscErrorCode MatQueryOp(MPI_Comm, PetscVoidFunction*, const char[], PetscInt, ...);
221: typedef struct _p_MatRootName* MatRootName;
222: struct _p_MatRootName {
223: char *rname,*sname,*mname;
224: MatRootName next;
225: };
227: PETSC_EXTERN MatRootName MatRootNameList;
229: /*
230: Utility private matrix routines
231: */
232: PETSC_INTERN PetscErrorCode MatFindNonzeroRowsOrCols_Basic(Mat,PetscBool,PetscReal,IS*);
233: PETSC_INTERN PetscErrorCode MatConvert_Basic(Mat,MatType,MatReuse,Mat*);
234: PETSC_INTERN PetscErrorCode MatConvert_Shell(Mat,MatType,MatReuse,Mat*);
235: PETSC_INTERN PetscErrorCode MatConvertFrom_Shell(Mat,MatType,MatReuse,Mat*);
236: PETSC_INTERN PetscErrorCode MatCopy_Basic(Mat,Mat,MatStructure);
237: PETSC_INTERN PetscErrorCode MatDiagonalSet_Default(Mat,Vec,InsertMode);
239: PETSC_INTERN PetscErrorCode MatProductSymbolic_Basic(Mat);
240: PETSC_EXTERN PetscErrorCode MatProductSymbolic_AB(Mat);
241: PETSC_EXTERN PetscErrorCode MatProductNumeric_AB(Mat);
242: PETSC_EXTERN PetscErrorCode MatProductSymbolic_AtB(Mat);
243: PETSC_EXTERN PetscErrorCode MatProductNumeric_AtB(Mat);
244: PETSC_EXTERN PetscErrorCode MatProductSymbolic_ABt(Mat);
245: PETSC_EXTERN PetscErrorCode MatProductNumeric_ABt(Mat);
246: PETSC_EXTERN PetscErrorCode MatProductNumeric_PtAP(Mat);
247: PETSC_EXTERN PetscErrorCode MatProductNumeric_RARt(Mat);
248: PETSC_EXTERN PetscErrorCode MatProductSymbolic_ABC(Mat);
249: PETSC_EXTERN PetscErrorCode MatProductNumeric_ABC(Mat);
250: PETSC_EXTERN PetscErrorCode MatProductCreate_Private(Mat,Mat,Mat,Mat);
252: #if defined(PETSC_USE_DEBUG)
253: # define MatCheckPreallocated(A,arg) do { \
254: 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); \
255: } while (0)
256: #else
257: # define MatCheckPreallocated(A,arg) do {} while (0)
258: #endif
260: /*
261: The stash is used to temporarily store inserted matrix values that
262: belong to another processor. During the assembly phase the stashed
263: values are moved to the correct processor and
264: */
266: typedef struct _MatStashSpace *PetscMatStashSpace;
268: struct _MatStashSpace {
269: PetscMatStashSpace next;
270: PetscScalar *space_head,*val;
271: PetscInt *idx,*idy;
272: PetscInt total_space_size;
273: PetscInt local_used;
274: PetscInt local_remaining;
275: };
277: PETSC_EXTERN PetscErrorCode PetscMatStashSpaceGet(PetscInt,PetscInt,PetscMatStashSpace *);
278: PETSC_EXTERN PetscErrorCode PetscMatStashSpaceContiguous(PetscInt,PetscMatStashSpace *,PetscScalar *,PetscInt *,PetscInt *);
279: PETSC_EXTERN PetscErrorCode PetscMatStashSpaceDestroy(PetscMatStashSpace*);
281: typedef struct {
282: PetscInt count;
283: } MatStashHeader;
285: typedef struct {
286: void *buffer; /* Of type blocktype, dynamically constructed */
287: PetscInt count;
288: char pending;
289: } MatStashFrame;
291: typedef struct _MatStash MatStash;
292: struct _MatStash {
293: PetscInt nmax; /* maximum stash size */
294: PetscInt umax; /* user specified max-size */
295: PetscInt oldnmax; /* the nmax value used previously */
296: PetscInt n; /* stash size */
297: PetscInt bs; /* block size of the stash */
298: PetscInt reallocs; /* preserve the no of mallocs invoked */
299: PetscMatStashSpace space_head,space; /* linked list to hold stashed global row/column numbers and matrix values */
301: PetscErrorCode (*ScatterBegin)(Mat,MatStash*,PetscInt*);
302: PetscErrorCode (*ScatterGetMesg)(MatStash*,PetscMPIInt*,PetscInt**,PetscInt**,PetscScalar**,PetscInt*);
303: PetscErrorCode (*ScatterEnd)(MatStash*);
304: PetscErrorCode (*ScatterDestroy)(MatStash*);
306: /* The following variables are used for communication */
307: MPI_Comm comm;
308: PetscMPIInt size,rank;
309: PetscMPIInt tag1,tag2;
310: MPI_Request *send_waits; /* array of send requests */
311: MPI_Request *recv_waits; /* array of receive requests */
312: MPI_Status *send_status; /* array of send status */
313: PetscInt nsends,nrecvs; /* numbers of sends and receives */
314: PetscScalar *svalues; /* sending data */
315: PetscInt *sindices;
316: PetscScalar **rvalues; /* receiving data (values) */
317: PetscInt **rindices; /* receiving data (indices) */
318: PetscInt nprocessed; /* number of messages already processed */
319: PetscMPIInt *flg_v; /* indicates what messages have arrived so far and from whom */
320: PetscBool reproduce;
321: PetscInt reproduce_count;
323: /* The following variables are used for BTS communication */
324: PetscBool first_assembly_done; /* Is the first time matrix assembly done? */
325: PetscBool use_status; /* Use MPI_Status to determine number of items in each message */
326: PetscMPIInt nsendranks;
327: PetscMPIInt nrecvranks;
328: PetscMPIInt *sendranks;
329: PetscMPIInt *recvranks;
330: MatStashHeader *sendhdr,*recvhdr;
331: MatStashFrame *sendframes; /* pointers to the main messages */
332: MatStashFrame *recvframes;
333: MatStashFrame *recvframe_active;
334: PetscInt recvframe_i; /* index of block within active frame */
335: PetscMPIInt recvframe_count; /* Count actually sent for current frame */
336: PetscInt recvcount; /* Number of receives processed so far */
337: PetscMPIInt *some_indices; /* From last call to MPI_Waitsome */
338: MPI_Status *some_statuses; /* Statuses from last call to MPI_Waitsome */
339: PetscMPIInt some_count; /* Number of requests completed in last call to MPI_Waitsome */
340: PetscMPIInt some_i; /* Index of request currently being processed */
341: MPI_Request *sendreqs;
342: MPI_Request *recvreqs;
343: PetscSegBuffer segsendblocks;
344: PetscSegBuffer segrecvframe;
345: PetscSegBuffer segrecvblocks;
346: MPI_Datatype blocktype;
347: size_t blocktype_size;
348: InsertMode *insertmode; /* Pointer to check mat->insertmode and set upon message arrival in case no local values have been set. */
349: };
351: #if !defined(PETSC_HAVE_MPIUNI)
352: PETSC_INTERN PetscErrorCode MatStashScatterDestroy_BTS(MatStash*);
353: #endif
354: PETSC_INTERN PetscErrorCode MatStashCreate_Private(MPI_Comm,PetscInt,MatStash*);
355: PETSC_INTERN PetscErrorCode MatStashDestroy_Private(MatStash*);
356: PETSC_INTERN PetscErrorCode MatStashScatterEnd_Private(MatStash*);
357: PETSC_INTERN PetscErrorCode MatStashSetInitialSize_Private(MatStash*,PetscInt);
358: PETSC_INTERN PetscErrorCode MatStashGetInfo_Private(MatStash*,PetscInt*,PetscInt*);
359: PETSC_INTERN PetscErrorCode MatStashValuesRow_Private(MatStash*,PetscInt,PetscInt,const PetscInt[],const PetscScalar[],PetscBool );
360: PETSC_INTERN PetscErrorCode MatStashValuesCol_Private(MatStash*,PetscInt,PetscInt,const PetscInt[],const PetscScalar[],PetscInt,PetscBool );
361: PETSC_INTERN PetscErrorCode MatStashValuesRowBlocked_Private(MatStash*,PetscInt,PetscInt,const PetscInt[],const PetscScalar[],PetscInt,PetscInt,PetscInt);
362: PETSC_INTERN PetscErrorCode MatStashValuesColBlocked_Private(MatStash*,PetscInt,PetscInt,const PetscInt[],const PetscScalar[],PetscInt,PetscInt,PetscInt);
363: PETSC_INTERN PetscErrorCode MatStashScatterBegin_Private(Mat,MatStash*,PetscInt*);
364: PETSC_INTERN PetscErrorCode MatStashScatterGetMesg_Private(MatStash*,PetscMPIInt*,PetscInt**,PetscInt**,PetscScalar**,PetscInt*);
365: PETSC_INTERN PetscErrorCode MatGetInfo_External(Mat,MatInfoType,MatInfo*);
367: typedef struct {
368: PetscInt dim;
369: PetscInt dims[4];
370: PetscInt starts[4];
371: PetscBool noc; /* this is a single component problem, hence user will not set MatStencil.c */
372: } MatStencilInfo;
374: /* Info about using compressed row format */
375: typedef struct {
376: PetscBool use; /* indicates compressed rows have been checked and will be used */
377: PetscInt nrows; /* number of non-zero rows */
378: PetscInt *i; /* compressed row pointer */
379: PetscInt *rindex; /* compressed row index */
380: } Mat_CompressedRow;
381: PETSC_EXTERN PetscErrorCode MatCheckCompressedRow(Mat,PetscInt,Mat_CompressedRow*,PetscInt*,PetscInt,PetscReal);
383: typedef struct { /* used by MatCreateRedundantMatrix() for reusing matredundant */
384: PetscInt nzlocal,nsends,nrecvs;
385: PetscMPIInt *send_rank,*recv_rank;
386: PetscInt *sbuf_nz,*rbuf_nz,*sbuf_j,**rbuf_j;
387: PetscScalar *sbuf_a,**rbuf_a;
388: MPI_Comm subcomm; /* when user does not provide a subcomm */
389: IS isrow,iscol;
390: Mat *matseq;
391: } Mat_Redundant;
393: typedef struct { /* used by MatProduct() */
394: MatProductType type;
395: MatProductAlgorithm alg;
396: Mat A,B,C,Dwork;
397: PetscReal fill;
398: PetscBool Areplaced,Breplaced; /* if an internal implementation changes user's input A or B, these matrices cannot be called by MatProductReplaceMats(). */
399: PetscBool api_user; /* used by MatProductSetFromOptions_xxx() */
400: } Mat_Product;
402: struct _p_Mat {
403: PETSCHEADER(struct _MatOps);
404: PetscLayout rmap,cmap;
405: void *data; /* implementation-specific data */
406: MatFactorType factortype; /* MAT_FACTOR_LU, ILU, CHOLESKY or ICC */
407: PetscBool assembled; /* is the matrix assembled? */
408: PetscBool was_assembled; /* new values inserted into assembled mat */
409: PetscInt num_ass; /* number of times matrix has been assembled */
410: PetscObjectState nonzerostate; /* each time new nonzeros locations are introduced into the matrix this is updated */
411: PetscObjectState ass_nonzerostate; /* nonzero state at last assembly */
412: MatInfo info; /* matrix information */
413: InsertMode insertmode; /* have values been inserted in matrix or added? */
414: MatStash stash,bstash; /* used for assembling off-proc mat emements */
415: MatNullSpace nullsp; /* null space (operator is singular) */
416: MatNullSpace transnullsp; /* null space of transpose of operator */
417: MatNullSpace nearnullsp; /* near null space to be used by multigrid methods */
418: PetscInt congruentlayouts; /* are the rows and columns layouts congruent? */
419: PetscBool preallocated;
420: MatStencilInfo stencil; /* information for structured grid */
421: PetscBool symmetric,hermitian,structurally_symmetric,spd;
422: PetscBool symmetric_set,hermitian_set,structurally_symmetric_set,spd_set; /* if true, then corresponding flag is correct*/
423: PetscBool symmetric_eternal;
424: PetscBool nooffprocentries,nooffproczerorows;
425: PetscBool assembly_subset; /* set by MAT_SUBSET_OFF_PROC_ENTRIES */
426: PetscBool submat_singleis; /* for efficient PCSetUp_ASM() */
427: PetscBool structure_only;
428: PetscBool sortedfull; /* full, sorted rows are inserted */
429: #if defined(PETSC_HAVE_VIENNACL) || defined(PETSC_HAVE_CUDA)
430: PetscOffloadMask offloadmask; /* a mask which indicates where the valid matrix data is (GPU, CPU or both) */
431: PetscBool boundtocpu;
432: #endif
433: void *spptr; /* pointer for special library like SuperLU */
434: char *solvertype;
435: PetscBool checksymmetryonassembly,checknullspaceonassembly;
436: PetscReal checksymmetrytol;
437: Mat schur; /* Schur complement matrix */
438: MatFactorSchurStatus schur_status; /* status of the Schur complement matrix */
439: Mat_Redundant *redundant; /* used by MatCreateRedundantMatrix() */
440: PetscBool erroriffailure; /* Generate an error if detected (for example a zero pivot) instead of returning */
441: MatFactorError factorerrortype; /* type of error in factorization */
442: PetscReal factorerror_zeropivot_value; /* If numerical zero pivot was detected this is the computed value */
443: PetscInt factorerror_zeropivot_row; /* Row where zero pivot was detected */
444: PetscInt nblocks,*bsizes; /* support for MatSetVariableBlockSizes() */
445: char *defaultvectype;
446: Mat_Product *product;
447: };
449: PETSC_INTERN PetscErrorCode MatAXPY_Basic(Mat,PetscScalar,Mat,MatStructure);
450: PETSC_INTERN PetscErrorCode MatAXPY_BasicWithPreallocation(Mat,Mat,PetscScalar,Mat,MatStructure);
451: PETSC_INTERN PetscErrorCode MatAXPY_Basic_Preallocate(Mat,Mat,Mat*);
453: /*
454: Utility for MatFactor (Schur complement)
455: */
456: PETSC_INTERN PetscErrorCode MatFactorFactorizeSchurComplement_Private(Mat);
457: PETSC_INTERN PetscErrorCode MatFactorInvertSchurComplement_Private(Mat);
458: PETSC_INTERN PetscErrorCode MatFactorUpdateSchurStatus_Private(Mat);
459: PETSC_INTERN PetscErrorCode MatFactorSetUpInPlaceSchur_Private(Mat);
461: /*
462: Utility for MatZeroRows
463: */
464: PETSC_INTERN PetscErrorCode MatZeroRowsMapLocal_Private(Mat,PetscInt,const PetscInt*,PetscInt*,PetscInt**);
466: /*
467: Utility for MatView/MatLoad
468: */
469: PETSC_INTERN PetscErrorCode MatView_Binary_BlockSizes(Mat,PetscViewer);
470: PETSC_INTERN PetscErrorCode MatLoad_Binary_BlockSizes(Mat,PetscViewer);
473: /*
474: Object for partitioning graphs
475: */
477: typedef struct _MatPartitioningOps *MatPartitioningOps;
478: struct _MatPartitioningOps {
479: PetscErrorCode (*apply)(MatPartitioning,IS*);
480: PetscErrorCode (*applynd)(MatPartitioning,IS*);
481: PetscErrorCode (*setfromoptions)(PetscOptionItems*,MatPartitioning);
482: PetscErrorCode (*destroy)(MatPartitioning);
483: PetscErrorCode (*view)(MatPartitioning,PetscViewer);
484: PetscErrorCode (*improve)(MatPartitioning,IS*);
485: };
487: struct _p_MatPartitioning {
488: PETSCHEADER(struct _MatPartitioningOps);
489: Mat adj;
490: PetscInt *vertex_weights;
491: PetscReal *part_weights;
492: PetscInt n; /* number of partitions */
493: void *data;
494: PetscInt setupcalled;
495: PetscBool use_edge_weights; /* A flag indicates whether or not to use edge weights */
496: };
498: /* needed for parallel nested dissection by ParMetis and PTSCOTCH */
499: PETSC_INTERN PetscErrorCode MatPartitioningSizesToSep_Private(PetscInt,PetscInt[],PetscInt[],PetscInt[]);
501: /*
502: Object for coarsen graphs
503: */
504: typedef struct _MatCoarsenOps *MatCoarsenOps;
505: struct _MatCoarsenOps {
506: PetscErrorCode (*apply)(MatCoarsen);
507: PetscErrorCode (*setfromoptions)(PetscOptionItems*,MatCoarsen);
508: PetscErrorCode (*destroy)(MatCoarsen);
509: PetscErrorCode (*view)(MatCoarsen,PetscViewer);
510: };
512: struct _p_MatCoarsen {
513: PETSCHEADER(struct _MatCoarsenOps);
514: Mat graph;
515: PetscInt setupcalled;
516: void *subctx;
517: /* */
518: PetscBool strict_aggs;
519: IS perm;
520: PetscCoarsenData *agg_lists;
521: };
523: /*
524: MatFDColoring is used to compute Jacobian matrices efficiently
525: via coloring. The data structure is explained below in an example.
527: Color = 0 1 0 2 | 2 3 0
528: ---------------------------------------------------
529: 00 01 | 05
530: 10 11 | 14 15 Processor 0
531: 22 23 | 25
532: 32 33 |
533: ===================================================
534: | 44 45 46
535: 50 | 55 Processor 1
536: | 64 66
537: ---------------------------------------------------
539: ncolors = 4;
541: ncolumns = {2,1,1,0}
542: columns = {{0,2},{1},{3},{}}
543: nrows = {4,2,3,3}
544: rows = {{0,1,2,3},{0,1},{1,2,3},{0,1,2}}
545: vwscale = {dx(0),dx(1),dx(2),dx(3)} MPI Vec
546: vscale = {dx(0),dx(1),dx(2),dx(3),dx(4),dx(5)} Seq Vec
548: ncolumns = {1,0,1,1}
549: columns = {{6},{},{4},{5}}
550: nrows = {3,0,2,2}
551: rows = {{0,1,2},{},{1,2},{1,2}}
552: vwscale = {dx(4),dx(5),dx(6)} MPI Vec
553: vscale = {dx(0),dx(4),dx(5),dx(6)} Seq Vec
555: See the routine MatFDColoringApply() for how this data is used
556: to compute the Jacobian.
558: */
559: typedef struct {
560: PetscInt row;
561: PetscInt col;
562: PetscScalar *valaddr; /* address of value */
563: } MatEntry;
565: typedef struct {
566: PetscInt row;
567: PetscScalar *valaddr; /* address of value */
568: } MatEntry2;
570: struct _p_MatFDColoring{
571: PETSCHEADER(int);
572: PetscInt M,N,m; /* total rows, columns; local rows */
573: PetscInt rstart; /* first row owned by local processor */
574: PetscInt ncolors; /* number of colors */
575: PetscInt *ncolumns; /* number of local columns for a color */
576: PetscInt **columns; /* lists the local columns of each color (using global column numbering) */
577: IS *isa; /* these are the IS that contain the column values given in columns */
578: PetscInt *nrows; /* number of local rows for each color */
579: MatEntry *matentry; /* holds (row, column, address of value) for Jacobian matrix entry */
580: MatEntry2 *matentry2; /* holds (row, address of value) for Jacobian matrix entry */
581: PetscScalar *dy; /* store a block of F(x+dx)-F(x) when J is in BAIJ format */
582: PetscReal error_rel; /* square root of relative error in computing function */
583: PetscReal umin; /* minimum allowable u'dx value */
584: Vec w1,w2,w3; /* work vectors used in computing Jacobian */
585: PetscBool fset; /* indicates that the initial function value F(X) is set */
586: PetscErrorCode (*f)(void); /* function that defines Jacobian */
587: void *fctx; /* optional user-defined context for use by the function f */
588: Vec vscale; /* holds FD scaling, i.e. 1/dx for each perturbed column */
589: PetscInt currentcolor; /* color for which function evaluation is being done now */
590: const char *htype; /* "wp" or "ds" */
591: ISColoringType ctype; /* IS_COLORING_GLOBAL or IS_COLORING_LOCAL */
592: PetscInt brows,bcols; /* number of block rows or columns for speedup inserting the dense matrix into sparse Jacobian */
593: PetscBool setupcalled; /* true if setup has been called */
594: PetscBool viewed; /* true if the -mat_fd_coloring_view has been triggered already */
595: void (*ftn_func_pointer)(void),*ftn_func_cntx; /* serve the same purpose as *fortran_func_pointers in PETSc objects */
596: PetscObjectId matid; /* matrix this object was created with, must always be the same */
597: };
599: typedef struct _MatColoringOps *MatColoringOps;
600: struct _MatColoringOps {
601: PetscErrorCode (*destroy)(MatColoring);
602: PetscErrorCode (*setfromoptions)(PetscOptionItems*,MatColoring);
603: PetscErrorCode (*view)(MatColoring,PetscViewer);
604: PetscErrorCode (*apply)(MatColoring,ISColoring*);
605: PetscErrorCode (*weights)(MatColoring,PetscReal**,PetscInt**);
606: };
608: struct _p_MatColoring {
609: PETSCHEADER(struct _MatColoringOps);
610: Mat mat;
611: PetscInt dist; /* distance of the coloring */
612: PetscInt maxcolors; /* the maximum number of colors returned, maxcolors=1 for MIS */
613: void *data; /* inner context */
614: PetscBool valid; /* check to see if what is produced is a valid coloring */
615: MatColoringWeightType weight_type; /* type of weight computation to be performed */
616: PetscReal *user_weights; /* custom weights and permutation */
617: PetscInt *user_lperm;
618: PetscBool valid_iscoloring; /* check to see if matcoloring is produced a valid iscoloring */
619: };
621: struct _p_MatTransposeColoring{
622: PETSCHEADER(int);
623: PetscInt M,N,m; /* total rows, columns; local rows */
624: PetscInt rstart; /* first row owned by local processor */
625: PetscInt ncolors; /* number of colors */
626: PetscInt *ncolumns; /* number of local columns for a color */
627: PetscInt *nrows; /* number of local rows for each color */
628: PetscInt currentcolor; /* color for which function evaluation is being done now */
629: ISColoringType ctype; /* IS_COLORING_GLOBAL or IS_COLORING_LOCAL */
631: PetscInt *colorforrow,*colorforcol; /* pointer to rows and columns */
632: PetscInt *rows; /* lists the local rows for each color (using the local row numbering) */
633: PetscInt *den2sp; /* maps (row,color) in the dense matrix to index of sparse matrix array a->a */
634: PetscInt *columns; /* lists the local columns of each color (using global column numbering) */
635: PetscInt brows; /* number of rows for efficient implementation of MatTransColoringApplyDenToSp() */
636: PetscInt *lstart; /* array used for loop over row blocks of Csparse */
637: };
639: /*
640: Null space context for preconditioner/operators
641: */
642: struct _p_MatNullSpace {
643: PETSCHEADER(int);
644: PetscBool has_cnst;
645: PetscInt n;
646: Vec* vecs;
647: PetscScalar* alpha; /* for projections */
648: PetscErrorCode (*remove)(MatNullSpace,Vec,void*); /* for user provided removal function */
649: void* rmctx; /* context for remove() function */
650: };
652: /*
653: Checking zero pivot for LU, ILU preconditioners.
654: */
655: typedef struct {
656: PetscInt nshift,nshift_max;
657: PetscReal shift_amount,shift_lo,shift_hi,shift_top,shift_fraction;
658: PetscBool newshift;
659: PetscReal rs; /* active row sum of abs(offdiagonals) */
660: PetscScalar pv; /* pivot of the active row */
661: } FactorShiftCtx;
663: /*
664: Used by MatCreateSubMatrices_MPIXAIJ_Local()
665: */
666: #include <petscctable.h>
667: typedef struct { /* used by MatCreateSubMatrices_MPIAIJ_SingleIS_Local() and MatCreateSubMatrices_MPIAIJ_Local */
668: PetscInt id; /* index of submats, only submats[0] is responsible for deleting some arrays below */
669: PetscInt nrqs,nrqr;
670: PetscInt **rbuf1,**rbuf2,**rbuf3,**sbuf1,**sbuf2;
671: PetscInt **ptr;
672: PetscInt *tmp;
673: PetscInt *ctr;
674: PetscInt *pa; /* proc array */
675: PetscInt *req_size,*req_source1,*req_source2;
676: PetscBool allcolumns,allrows;
677: PetscBool singleis;
678: PetscInt *row2proc; /* row to proc map */
679: PetscInt nstages;
680: #if defined(PETSC_USE_CTABLE)
681: PetscTable cmap,rmap;
682: PetscInt *cmap_loc,*rmap_loc;
683: #else
684: PetscInt *cmap,*rmap;
685: #endif
687: PetscErrorCode (*destroy)(Mat);
688: } Mat_SubSppt;
690: PETSC_EXTERN PetscErrorCode MatFactorDumpMatrix(Mat);
691: PETSC_INTERN PetscErrorCode MatShift_Basic(Mat,PetscScalar);
692: PETSC_INTERN PetscErrorCode MatSetBlockSizes_Default(Mat,PetscInt,PetscInt);
694: PETSC_STATIC_INLINE PetscErrorCode MatPivotCheck_nz(Mat mat,const MatFactorInfo *info,FactorShiftCtx *sctx,PetscInt row)
695: {
696: PetscReal _rs = sctx->rs;
697: PetscReal _zero = info->zeropivot*_rs;
700: if (PetscAbsScalar(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)){
701: /* force |diag| > zeropivot*rs */
702: if (!sctx->nshift) sctx->shift_amount = info->shiftamount;
703: else sctx->shift_amount *= 2.0;
704: sctx->newshift = PETSC_TRUE;
705: (sctx->nshift)++;
706: } else {
707: sctx->newshift = PETSC_FALSE;
708: }
709: return(0);
710: }
712: PETSC_STATIC_INLINE PetscErrorCode MatPivotCheck_pd(Mat mat,const MatFactorInfo *info,FactorShiftCtx *sctx,PetscInt row)
713: {
714: PetscReal _rs = sctx->rs;
715: PetscReal _zero = info->zeropivot*_rs;
718: if (PetscRealPart(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)){
719: /* force matfactor to be diagonally dominant */
720: if (sctx->nshift == sctx->nshift_max) {
721: sctx->shift_fraction = sctx->shift_hi;
722: } else {
723: sctx->shift_lo = sctx->shift_fraction;
724: sctx->shift_fraction = (sctx->shift_hi+sctx->shift_lo)/2.;
725: }
726: sctx->shift_amount = sctx->shift_fraction * sctx->shift_top;
727: sctx->nshift++;
728: sctx->newshift = PETSC_TRUE;
729: } else {
730: sctx->newshift = PETSC_FALSE;
731: }
732: return(0);
733: }
735: PETSC_STATIC_INLINE PetscErrorCode MatPivotCheck_inblocks(Mat mat,const MatFactorInfo *info,FactorShiftCtx *sctx,PetscInt row)
736: {
737: PetscReal _zero = info->zeropivot;
740: if (PetscAbsScalar(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)){
741: sctx->pv += info->shiftamount;
742: sctx->shift_amount = 0.0;
743: sctx->nshift++;
744: }
745: sctx->newshift = PETSC_FALSE;
746: return(0);
747: }
749: PETSC_STATIC_INLINE PetscErrorCode MatPivotCheck_none(Mat fact,Mat mat,const MatFactorInfo *info,FactorShiftCtx *sctx,PetscInt row)
750: {
751: PetscReal _zero = info->zeropivot;
755: sctx->newshift = PETSC_FALSE;
756: if (PetscAbsScalar(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)) {
757: if (!mat->erroriffailure) {
758: PetscInfo3(mat,"Detected zero pivot in factorization in row %D value %g tolerance %g\n",row,(double)PetscAbsScalar(sctx->pv),(double)_zero);
759: fact->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT;
760: fact->factorerror_zeropivot_value = PetscAbsScalar(sctx->pv);
761: fact->factorerror_zeropivot_row = row;
762: } 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);
763: }
764: return(0);
765: }
767: PETSC_STATIC_INLINE PetscErrorCode MatPivotCheck(Mat fact,Mat mat,const MatFactorInfo *info,FactorShiftCtx *sctx,PetscInt row)
768: {
772: if (info->shifttype == (PetscReal) MAT_SHIFT_NONZERO){
773: MatPivotCheck_nz(mat,info,sctx,row);
774: } else if (info->shifttype == (PetscReal) MAT_SHIFT_POSITIVE_DEFINITE){
775: MatPivotCheck_pd(mat,info,sctx,row);
776: } else if (info->shifttype == (PetscReal) MAT_SHIFT_INBLOCKS){
777: MatPivotCheck_inblocks(mat,info,sctx,row);
778: } else {
779: MatPivotCheck_none(fact,mat,info,sctx,row);
780: }
781: return(0);
782: }
784: /*
785: Create and initialize a linked list
786: Input Parameters:
787: idx_start - starting index of the list
788: lnk_max - max value of lnk indicating the end of the list
789: nlnk - max length of the list
790: Output Parameters:
791: lnk - list initialized
792: bt - PetscBT (bitarray) with all bits set to false
793: lnk_empty - flg indicating the list is empty
794: */
795: #define PetscLLCreate(idx_start,lnk_max,nlnk,lnk,bt) \
796: (PetscMalloc1(nlnk,&lnk) || PetscBTCreate(nlnk,&(bt)) || (lnk[idx_start] = lnk_max,0))
798: #define PetscLLCreate_new(idx_start,lnk_max,nlnk,lnk,bt,lnk_empty)\
799: (PetscMalloc1(nlnk,&lnk) || PetscBTCreate(nlnk,&(bt)) || (lnk_empty = PETSC_TRUE,0) ||(lnk[idx_start] = lnk_max,0))
801: /*
802: Add an index set into a sorted linked list
803: Input Parameters:
804: nidx - number of input indices
805: indices - integer array
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 PetscLLAdd(nidx,indices,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 = 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 permuted index set into a sorted linked list
840: Input Parameters:
841: nidx - number of input indices
842: indices - integer array
843: perm - permutation of indices
844: idx_start - starting index of the list
845: lnk - linked list(an integer array) that is created
846: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
847: output Parameters:
848: nlnk - number of newly added indices
849: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
850: bt - updated PetscBT (bitarray)
851: */
852: #define PetscLLAddPerm(nidx,indices,perm,idx_start,nlnk,lnk,bt) 0;\
853: {\
854: PetscInt _k,_entry,_location,_lnkdata;\
855: nlnk = 0;\
856: _lnkdata = idx_start;\
857: for (_k=0; _k<nidx; _k++){\
858: _entry = perm[indices[_k]];\
859: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
860: /* search for insertion location */\
861: /* start from the beginning if _entry < previous _entry */\
862: if (_k && _entry < _lnkdata) _lnkdata = idx_start;\
863: do {\
864: _location = _lnkdata;\
865: _lnkdata = lnk[_location];\
866: } while (_entry > _lnkdata);\
867: /* insertion location is found, add entry into lnk */\
868: lnk[_location] = _entry;\
869: lnk[_entry] = _lnkdata;\
870: nlnk++;\
871: _lnkdata = _entry; /* next search starts from here if next_entry > _entry */\
872: }\
873: }\
874: }
876: /*
877: Add a SORTED ascending index set into a sorted linked list - same as PetscLLAdd() bus skip 'if (_k && _entry < _lnkdata) _lnkdata = idx_start;'
878: Input Parameters:
879: nidx - number of input indices
880: indices - sorted integer array
881: idx_start - starting index of the list
882: lnk - linked list(an integer array) that is created
883: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
884: output Parameters:
885: nlnk - number of newly added indices
886: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
887: bt - updated PetscBT (bitarray)
888: */
889: #define PetscLLAddSorted(nidx,indices,idx_start,nlnk,lnk,bt) 0;\
890: {\
891: PetscInt _k,_entry,_location,_lnkdata;\
892: nlnk = 0;\
893: _lnkdata = idx_start;\
894: for (_k=0; _k<nidx; _k++){\
895: _entry = indices[_k];\
896: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
897: /* search for insertion location */\
898: do {\
899: _location = _lnkdata;\
900: _lnkdata = lnk[_location];\
901: } while (_entry > _lnkdata);\
902: /* insertion location is found, add entry into lnk */\
903: lnk[_location] = _entry;\
904: lnk[_entry] = _lnkdata;\
905: nlnk++;\
906: _lnkdata = _entry; /* next search starts from here */\
907: }\
908: }\
909: }
911: #define PetscLLAddSorted_new(nidx,indices,idx_start,lnk_empty,nlnk,lnk,bt) 0; \
912: {\
913: PetscInt _k,_entry,_location,_lnkdata;\
914: if (lnk_empty){\
915: _lnkdata = idx_start; \
916: for (_k=0; _k<nidx; _k++){ \
917: _entry = indices[_k]; \
918: PetscBTSet(bt,_entry); /* mark the new entry */ \
919: _location = _lnkdata; \
920: _lnkdata = lnk[_location]; \
921: /* insertion location is found, add entry into lnk */ \
922: lnk[_location] = _entry; \
923: lnk[_entry] = _lnkdata; \
924: _lnkdata = _entry; /* next search starts from here */ \
925: } \
926: /*\
927: lnk[indices[nidx-1]] = lnk[idx_start];\
928: lnk[idx_start] = indices[0];\
929: PetscBTSet(bt,indices[0]); \
930: for (_k=1; _k<nidx; _k++){ \
931: PetscBTSet(bt,indices[_k]); \
932: lnk[indices[_k-1]] = indices[_k]; \
933: } \
934: */\
935: nlnk = nidx;\
936: lnk_empty = PETSC_FALSE;\
937: } else {\
938: nlnk = 0; \
939: _lnkdata = idx_start; \
940: for (_k=0; _k<nidx; _k++){ \
941: _entry = indices[_k]; \
942: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */ \
943: /* search for insertion location */ \
944: do { \
945: _location = _lnkdata; \
946: _lnkdata = lnk[_location]; \
947: } while (_entry > _lnkdata); \
948: /* insertion location is found, add entry into lnk */ \
949: lnk[_location] = _entry; \
950: lnk[_entry] = _lnkdata; \
951: nlnk++; \
952: _lnkdata = _entry; /* next search starts from here */ \
953: } \
954: } \
955: } \
956: }
958: /*
959: Add a SORTED index set into a sorted linked list used for LUFactorSymbolic()
960: Same as PetscLLAddSorted() with an additional operation:
961: count the number of input indices that are no larger than 'diag'
962: Input Parameters:
963: indices - sorted integer array
964: idx_start - starting index of the list, index of pivot row
965: lnk - linked list(an integer array) that is created
966: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
967: diag - index of the active row in LUFactorSymbolic
968: nzbd - number of input indices with indices <= idx_start
969: im - im[idx_start] is initialized as num of nonzero entries in row=idx_start
970: output Parameters:
971: nlnk - number of newly added indices
972: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
973: bt - updated PetscBT (bitarray)
974: im - im[idx_start]: unchanged if diag is not an entry
975: : num of entries with indices <= diag if diag is an entry
976: */
977: #define PetscLLAddSortedLU(indices,idx_start,nlnk,lnk,bt,diag,nzbd,im) 0;\
978: {\
979: PetscInt _k,_entry,_location,_lnkdata,_nidx;\
980: nlnk = 0;\
981: _lnkdata = idx_start;\
982: _nidx = im[idx_start] - nzbd; /* num of entries with idx_start < index <= diag */\
983: for (_k=0; _k<_nidx; _k++){\
984: _entry = indices[_k];\
985: nzbd++;\
986: if ( _entry== diag) im[idx_start] = nzbd;\
987: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
988: /* search for insertion location */\
989: do {\
990: _location = _lnkdata;\
991: _lnkdata = lnk[_location];\
992: } while (_entry > _lnkdata);\
993: /* insertion location is found, add entry into lnk */\
994: lnk[_location] = _entry;\
995: lnk[_entry] = _lnkdata;\
996: nlnk++;\
997: _lnkdata = _entry; /* next search starts from here */\
998: }\
999: }\
1000: }
1002: /*
1003: Copy data on the list into an array, then initialize the list
1004: Input Parameters:
1005: idx_start - starting index of the list
1006: lnk_max - max value of lnk indicating the end of the list
1007: nlnk - number of data on the list to be copied
1008: lnk - linked list
1009: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1010: output Parameters:
1011: indices - array that contains the copied data
1012: lnk - linked list that is cleaned and initialize
1013: bt - PetscBT (bitarray) with all bits set to false
1014: */
1015: #define PetscLLClean(idx_start,lnk_max,nlnk,lnk,indices,bt) 0;\
1016: {\
1017: PetscInt _j,_idx=idx_start;\
1018: for (_j=0; _j<nlnk; _j++){\
1019: _idx = lnk[_idx];\
1020: indices[_j] = _idx;\
1021: PetscBTClear(bt,_idx);\
1022: }\
1023: lnk[idx_start] = lnk_max;\
1024: }
1025: /*
1026: Free memories used by the list
1027: */
1028: #define PetscLLDestroy(lnk,bt) (PetscFree(lnk) || PetscBTDestroy(&(bt)))
1030: /* Routines below are used for incomplete matrix factorization */
1031: /*
1032: Create and initialize a linked list and its levels
1033: Input Parameters:
1034: idx_start - starting index of the list
1035: lnk_max - max value of lnk indicating the end of the list
1036: nlnk - max length of the list
1037: Output Parameters:
1038: lnk - list initialized
1039: lnk_lvl - array of size nlnk for storing levels of lnk
1040: bt - PetscBT (bitarray) with all bits set to false
1041: */
1042: #define PetscIncompleteLLCreate(idx_start,lnk_max,nlnk,lnk,lnk_lvl,bt)\
1043: (PetscIntMultError(2,nlnk,NULL) || PetscMalloc1(2*nlnk,&lnk) || PetscBTCreate(nlnk,&(bt)) || (lnk[idx_start] = lnk_max,lnk_lvl = lnk + nlnk,0))
1045: /*
1046: Initialize a sorted linked list used for ILU and ICC
1047: Input Parameters:
1048: nidx - number of input idx
1049: idx - integer array used for storing column indices
1050: idx_start - starting index of the list
1051: perm - indices of an IS
1052: lnk - linked list(an integer array) that is created
1053: lnklvl - levels of lnk
1054: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1055: output Parameters:
1056: nlnk - number of newly added idx
1057: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1058: lnklvl - levels of lnk
1059: bt - updated PetscBT (bitarray)
1060: */
1061: #define PetscIncompleteLLInit(nidx,idx,idx_start,perm,nlnk,lnk,lnklvl,bt) 0;\
1062: {\
1063: PetscInt _k,_entry,_location,_lnkdata;\
1064: nlnk = 0;\
1065: _lnkdata = idx_start;\
1066: for (_k=0; _k<nidx; _k++){\
1067: _entry = perm[idx[_k]];\
1068: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
1069: /* search for insertion location */\
1070: if (_k && _entry < _lnkdata) _lnkdata = idx_start;\
1071: do {\
1072: _location = _lnkdata;\
1073: _lnkdata = lnk[_location];\
1074: } while (_entry > _lnkdata);\
1075: /* insertion location is found, add entry into lnk */\
1076: lnk[_location] = _entry;\
1077: lnk[_entry] = _lnkdata;\
1078: lnklvl[_entry] = 0;\
1079: nlnk++;\
1080: _lnkdata = _entry; /* next search starts from here if next_entry > _entry */\
1081: }\
1082: }\
1083: }
1085: /*
1086: Add a SORTED index set into a sorted linked list for ILU
1087: Input Parameters:
1088: nidx - number of input indices
1089: idx - sorted integer array used for storing column indices
1090: level - level of fill, e.g., ICC(level)
1091: idxlvl - level of idx
1092: idx_start - starting index of the list
1093: lnk - linked list(an integer array) that is created
1094: lnklvl - levels of lnk
1095: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1096: prow - the row number of idx
1097: output Parameters:
1098: nlnk - number of newly added idx
1099: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1100: lnklvl - levels of lnk
1101: bt - updated PetscBT (bitarray)
1103: Note: the level of factor(i,j) is set as lvl(i,j) = min{ lvl(i,j), lvl(i,prow)+lvl(prow,j)+1)
1104: where idx = non-zero columns of U(prow,prow+1:n-1), prow<i
1105: */
1106: #define PetscILULLAddSorted(nidx,idx,level,idxlvl,idx_start,nlnk,lnk,lnklvl,bt,lnklvl_prow) 0;\
1107: {\
1108: PetscInt _k,_entry,_location,_lnkdata,_incrlev,_lnklvl_prow=lnklvl[prow];\
1109: nlnk = 0;\
1110: _lnkdata = idx_start;\
1111: for (_k=0; _k<nidx; _k++){\
1112: _incrlev = idxlvl[_k] + _lnklvl_prow + 1;\
1113: if (_incrlev > level) continue;\
1114: _entry = idx[_k];\
1115: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
1116: /* search for insertion location */\
1117: do {\
1118: _location = _lnkdata;\
1119: _lnkdata = lnk[_location];\
1120: } while (_entry > _lnkdata);\
1121: /* insertion location is found, add entry into lnk */\
1122: lnk[_location] = _entry;\
1123: lnk[_entry] = _lnkdata;\
1124: lnklvl[_entry] = _incrlev;\
1125: nlnk++;\
1126: _lnkdata = _entry; /* next search starts from here if next_entry > _entry */\
1127: } else { /* existing entry: update lnklvl */\
1128: if (lnklvl[_entry] > _incrlev) lnklvl[_entry] = _incrlev;\
1129: }\
1130: }\
1131: }
1133: /*
1134: Add a index set into a sorted linked list
1135: Input Parameters:
1136: nidx - number of input idx
1137: idx - integer array used for storing column indices
1138: level - level of fill, e.g., ICC(level)
1139: idxlvl - level of idx
1140: idx_start - starting index of the list
1141: lnk - linked list(an integer array) that is created
1142: lnklvl - levels of lnk
1143: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1144: output Parameters:
1145: nlnk - number of newly added idx
1146: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1147: lnklvl - levels of lnk
1148: bt - updated PetscBT (bitarray)
1149: */
1150: #define PetscIncompleteLLAdd(nidx,idx,level,idxlvl,idx_start,nlnk,lnk,lnklvl,bt) 0;\
1151: {\
1152: PetscInt _k,_entry,_location,_lnkdata,_incrlev;\
1153: nlnk = 0;\
1154: _lnkdata = idx_start;\
1155: for (_k=0; _k<nidx; _k++){\
1156: _incrlev = idxlvl[_k] + 1;\
1157: if (_incrlev > level) continue;\
1158: _entry = idx[_k];\
1159: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
1160: /* search for insertion location */\
1161: if (_k && _entry < _lnkdata) _lnkdata = idx_start;\
1162: do {\
1163: _location = _lnkdata;\
1164: _lnkdata = lnk[_location];\
1165: } while (_entry > _lnkdata);\
1166: /* insertion location is found, add entry into lnk */\
1167: lnk[_location] = _entry;\
1168: lnk[_entry] = _lnkdata;\
1169: lnklvl[_entry] = _incrlev;\
1170: nlnk++;\
1171: _lnkdata = _entry; /* next search starts from here if next_entry > _entry */\
1172: } else { /* existing entry: update lnklvl */\
1173: if (lnklvl[_entry] > _incrlev) lnklvl[_entry] = _incrlev;\
1174: }\
1175: }\
1176: }
1178: /*
1179: Add a SORTED index set into a sorted linked list
1180: Input Parameters:
1181: nidx - number of input indices
1182: idx - sorted integer array used for storing column indices
1183: level - level of fill, e.g., ICC(level)
1184: idxlvl - level of idx
1185: idx_start - starting index of the list
1186: lnk - linked list(an integer array) that is created
1187: lnklvl - levels of lnk
1188: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1189: output Parameters:
1190: nlnk - number of newly added idx
1191: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1192: lnklvl - levels of lnk
1193: bt - updated PetscBT (bitarray)
1194: */
1195: #define PetscIncompleteLLAddSorted(nidx,idx,level,idxlvl,idx_start,nlnk,lnk,lnklvl,bt) 0;\
1196: {\
1197: PetscInt _k,_entry,_location,_lnkdata,_incrlev;\
1198: nlnk = 0;\
1199: _lnkdata = idx_start;\
1200: for (_k=0; _k<nidx; _k++){\
1201: _incrlev = idxlvl[_k] + 1;\
1202: if (_incrlev > level) continue;\
1203: _entry = idx[_k];\
1204: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
1205: /* search for insertion location */\
1206: do {\
1207: _location = _lnkdata;\
1208: _lnkdata = lnk[_location];\
1209: } while (_entry > _lnkdata);\
1210: /* insertion location is found, add entry into lnk */\
1211: lnk[_location] = _entry;\
1212: lnk[_entry] = _lnkdata;\
1213: lnklvl[_entry] = _incrlev;\
1214: nlnk++;\
1215: _lnkdata = _entry; /* next search starts from here */\
1216: } else { /* existing entry: update lnklvl */\
1217: if (lnklvl[_entry] > _incrlev) lnklvl[_entry] = _incrlev;\
1218: }\
1219: }\
1220: }
1222: /*
1223: Add a SORTED index set into a sorted linked list for ICC
1224: Input Parameters:
1225: nidx - number of input indices
1226: idx - sorted integer array used for storing column indices
1227: level - level of fill, e.g., ICC(level)
1228: idxlvl - level of idx
1229: idx_start - starting index of the list
1230: lnk - linked list(an integer array) that is created
1231: lnklvl - levels of lnk
1232: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1233: idxlvl_prow - idxlvl[prow], where prow is the row number of the idx
1234: output Parameters:
1235: nlnk - number of newly added indices
1236: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1237: lnklvl - levels of lnk
1238: bt - updated PetscBT (bitarray)
1239: Note: the level of U(i,j) is set as lvl(i,j) = min{ lvl(i,j), lvl(prow,i)+lvl(prow,j)+1)
1240: where idx = non-zero columns of U(prow,prow+1:n-1), prow<i
1241: */
1242: #define PetscICCLLAddSorted(nidx,idx,level,idxlvl,idx_start,nlnk,lnk,lnklvl,bt,idxlvl_prow) 0;\
1243: {\
1244: PetscInt _k,_entry,_location,_lnkdata,_incrlev;\
1245: nlnk = 0;\
1246: _lnkdata = idx_start;\
1247: for (_k=0; _k<nidx; _k++){\
1248: _incrlev = idxlvl[_k] + idxlvl_prow + 1;\
1249: if (_incrlev > level) continue;\
1250: _entry = idx[_k];\
1251: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
1252: /* search for insertion location */\
1253: do {\
1254: _location = _lnkdata;\
1255: _lnkdata = lnk[_location];\
1256: } while (_entry > _lnkdata);\
1257: /* insertion location is found, add entry into lnk */\
1258: lnk[_location] = _entry;\
1259: lnk[_entry] = _lnkdata;\
1260: lnklvl[_entry] = _incrlev;\
1261: nlnk++;\
1262: _lnkdata = _entry; /* next search starts from here */\
1263: } else { /* existing entry: update lnklvl */\
1264: if (lnklvl[_entry] > _incrlev) lnklvl[_entry] = _incrlev;\
1265: }\
1266: }\
1267: }
1269: /*
1270: Copy data on the list into an array, then initialize the list
1271: Input Parameters:
1272: idx_start - starting index of the list
1273: lnk_max - max value of lnk indicating the end of the list
1274: nlnk - number of data on the list to be copied
1275: lnk - linked list
1276: lnklvl - level of lnk
1277: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1278: output Parameters:
1279: indices - array that contains the copied data
1280: lnk - linked list that is cleaned and initialize
1281: lnklvl - level of lnk that is reinitialized
1282: bt - PetscBT (bitarray) with all bits set to false
1283: */
1284: #define PetscIncompleteLLClean(idx_start,lnk_max,nlnk,lnk,lnklvl,indices,indiceslvl,bt) 0;\
1285: do {\
1286: PetscInt _j,_idx=idx_start;\
1287: for (_j=0; _j<nlnk; _j++){\
1288: _idx = lnk[_idx];\
1289: *(indices+_j) = _idx;\
1290: *(indiceslvl+_j) = lnklvl[_idx];\
1291: lnklvl[_idx] = -1;\
1292: PetscBTClear(bt,_idx);\
1293: }\
1294: lnk[idx_start] = lnk_max;\
1295: } while (0)
1296: /*
1297: Free memories used by the list
1298: */
1299: #define PetscIncompleteLLDestroy(lnk,bt) (PetscFree(lnk) || PetscBTDestroy(&(bt)))
1301: #define MatCheckSameLocalSize(A,ar1,B,ar2) do { \
1303: 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);} while (0)
1305: #define MatCheckSameSize(A,ar1,B,ar2) do { \
1306: 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);\
1307: MatCheckSameLocalSize(A,ar1,B,ar2);} while (0)
1309: #define VecCheckMatCompatible(M,x,ar1,b,ar2) do { \
1310: 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); \
1311: 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);} while (0)
1313: /* -------------------------------------------------------------------------------------------------------*/
1314: #include <petscbt.h>
1315: /*
1316: Create and initialize a condensed linked list -
1317: same as PetscLLCreate(), but uses a scalable array 'lnk' with size of max number of entries, not O(N).
1318: Barry suggested this approach (Dec. 6, 2011):
1319: I've thought of an alternative way of representing a linked list that is efficient but doesn't have the O(N) scaling issue
1320: (it may be faster than the O(N) even sequentially due to less crazy memory access).
1322: 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
1323: 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
1324: 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
1325: 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.
1326: 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
1327: to each other so memory access is much better than using the big array.
1329: Example:
1330: nlnk_max=5, lnk_max=36:
1331: Initial list: [0, 0 | 36, 2 | 0, 0 | 0, 0 | 0, 0 | 0, 0 | 0, 0]
1332: here, head_node has index 2 with value lnk[2]=lnk_max=36,
1333: 0-th entry is used to store the number of entries in the list,
1334: The initial lnk represents head -> tail(marked by 36) with number of entries = lnk[0]=0.
1336: Now adding a sorted set {2,4}, the list becomes
1337: [2, 0 | 36, 4 |2, 6 | 4, 2 | 0, 0 | 0, 0 | 0, 0 ]
1338: represents head -> 2 -> 4 -> tail with number of entries = lnk[0]=2.
1340: Then adding a sorted set {0,3,35}, the list
1341: [5, 0 | 36, 8 | 2, 10 | 4, 12 | 0, 4 | 3, 6 | 35, 2 ]
1342: represents head -> 0 -> 2 -> 3 -> 4 -> 35 -> tail with number of entries = lnk[0]=5.
1344: Input Parameters:
1345: nlnk_max - max length of the list
1346: lnk_max - max value of the entries
1347: Output Parameters:
1348: lnk - list created and initialized
1349: bt - PetscBT (bitarray) with all bits set to false. Note: bt has size lnk_max, not nln_max!
1350: */
1351: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedCreate(PetscInt nlnk_max,PetscInt lnk_max,PetscInt **lnk,PetscBT *bt)
1352: {
1354: PetscInt *llnk,lsize = 0;
1357: PetscIntMultError(2,nlnk_max+2,&lsize);
1358: PetscMalloc1(lsize,lnk);
1359: PetscBTCreate(lnk_max,bt);
1360: llnk = *lnk;
1361: llnk[0] = 0; /* number of entries on the list */
1362: llnk[2] = lnk_max; /* value in the head node */
1363: llnk[3] = 2; /* next for the head node */
1364: return(0);
1365: }
1367: /*
1368: Add a SORTED ascending index set into a sorted linked list. See PetscLLCondensedCreate() for detailed description.
1369: Input Parameters:
1370: nidx - number of input indices
1371: indices - sorted integer array
1372: lnk - condensed linked list(an integer array) that is created
1373: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1374: output Parameters:
1375: lnk - the sorted(increasing order) linked list containing previous and newly added non-redundate indices
1376: bt - updated PetscBT (bitarray)
1377: */
1378: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedAddSorted(PetscInt nidx,const PetscInt indices[],PetscInt lnk[],PetscBT bt)
1379: {
1380: PetscInt _k,_entry,_location,_next,_lnkdata,_nlnk,_newnode;
1383: _nlnk = lnk[0]; /* num of entries on the input lnk */
1384: _location = 2; /* head */
1385: for (_k=0; _k<nidx; _k++){
1386: _entry = indices[_k];
1387: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */
1388: /* search for insertion location */
1389: do {
1390: _next = _location + 1; /* link from previous node to next node */
1391: _location = lnk[_next]; /* idx of next node */
1392: _lnkdata = lnk[_location];/* value of next node */
1393: } while (_entry > _lnkdata);
1394: /* insertion location is found, add entry into lnk */
1395: _newnode = 2*(_nlnk+2); /* index for this new node */
1396: lnk[_next] = _newnode; /* connect previous node to the new node */
1397: lnk[_newnode] = _entry; /* set value of the new node */
1398: lnk[_newnode+1] = _location; /* connect new node to next node */
1399: _location = _newnode; /* next search starts from the new node */
1400: _nlnk++;
1401: } \
1402: }\
1403: lnk[0] = _nlnk; /* number of entries in the list */
1404: return(0);
1405: }
1407: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedClean(PetscInt lnk_max,PetscInt nidx,PetscInt *indices,PetscInt lnk[],PetscBT bt)
1408: {
1410: PetscInt _k,_next,_nlnk;
1413: _next = lnk[3]; /* head node */
1414: _nlnk = lnk[0]; /* num of entries on the list */
1415: for (_k=0; _k<_nlnk; _k++){
1416: indices[_k] = lnk[_next];
1417: _next = lnk[_next + 1];
1418: PetscBTClear(bt,indices[_k]);
1419: }
1420: lnk[0] = 0; /* num of entries on the list */
1421: lnk[2] = lnk_max; /* initialize head node */
1422: lnk[3] = 2; /* head node */
1423: return(0);
1424: }
1426: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedView(PetscInt *lnk)
1427: {
1429: PetscInt k;
1432: PetscPrintf(PETSC_COMM_SELF,"LLCondensed of size %D, (val, next)\n",lnk[0]);
1433: for (k=2; k< lnk[0]+2; k++){
1434: PetscPrintf(PETSC_COMM_SELF," %D: (%D, %D)\n",2*k,lnk[2*k],lnk[2*k+1]);
1435: }
1436: return(0);
1437: }
1439: /*
1440: Free memories used by the list
1441: */
1442: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedDestroy(PetscInt *lnk,PetscBT bt)
1443: {
1447: PetscFree(lnk);
1448: PetscBTDestroy(&bt);
1449: return(0);
1450: }
1452: /* -------------------------------------------------------------------------------------------------------*/
1453: /*
1454: Same as PetscLLCondensedCreate(), but does not use non-scalable O(lnk_max) bitarray
1455: Input Parameters:
1456: nlnk_max - max length of the list
1457: Output Parameters:
1458: lnk - list created and initialized
1459: */
1460: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedCreate_Scalable(PetscInt nlnk_max,PetscInt **lnk)
1461: {
1463: PetscInt *llnk,lsize = 0;
1466: PetscIntMultError(2,nlnk_max+2,&lsize);
1467: PetscMalloc1(lsize,lnk);
1468: llnk = *lnk;
1469: llnk[0] = 0; /* number of entries on the list */
1470: llnk[2] = PETSC_MAX_INT; /* value in the head node */
1471: llnk[3] = 2; /* next for the head node */
1472: return(0);
1473: }
1475: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedExpand_Scalable(PetscInt nlnk_max,PetscInt **lnk)
1476: {
1478: PetscInt lsize = 0;
1481: PetscIntMultError(2,nlnk_max+2,&lsize);
1482: PetscRealloc(lsize*sizeof(PetscInt),lnk);
1483: return(0);
1484: }
1486: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedAddSorted_Scalable(PetscInt nidx,const PetscInt indices[],PetscInt lnk[])
1487: {
1488: PetscInt _k,_entry,_location,_next,_lnkdata,_nlnk,_newnode;
1489: _nlnk = lnk[0]; /* num of entries on the input lnk */
1490: _location = 2; /* head */ \
1491: for (_k=0; _k<nidx; _k++){
1492: _entry = indices[_k];
1493: /* search for insertion location */
1494: do {
1495: _next = _location + 1; /* link from previous node to next node */
1496: _location = lnk[_next]; /* idx of next node */
1497: _lnkdata = lnk[_location];/* value of next node */
1498: } while (_entry > _lnkdata);
1499: if (_entry < _lnkdata) {
1500: /* insertion location is found, add entry into lnk */
1501: _newnode = 2*(_nlnk+2); /* index for this new node */
1502: lnk[_next] = _newnode; /* connect previous node to the new node */
1503: lnk[_newnode] = _entry; /* set value of the new node */
1504: lnk[_newnode+1] = _location; /* connect new node to next node */
1505: _location = _newnode; /* next search starts from the new node */
1506: _nlnk++;
1507: }
1508: }
1509: lnk[0] = _nlnk; /* number of entries in the list */
1510: return 0;
1511: }
1513: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedClean_Scalable(PetscInt nidx,PetscInt *indices,PetscInt *lnk)
1514: {
1515: PetscInt _k,_next,_nlnk;
1516: _next = lnk[3]; /* head node */
1517: _nlnk = lnk[0];
1518: for (_k=0; _k<_nlnk; _k++){
1519: indices[_k] = lnk[_next];
1520: _next = lnk[_next + 1];
1521: }
1522: lnk[0] = 0; /* num of entries on the list */
1523: lnk[3] = 2; /* head node */
1524: return 0;
1525: }
1527: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedDestroy_Scalable(PetscInt *lnk)
1528: {
1529: return PetscFree(lnk);
1530: }
1532: /* -------------------------------------------------------------------------------------------------------*/
1533: /*
1534: lnk[0] number of links
1535: lnk[1] number of entries
1536: lnk[3n] value
1537: lnk[3n+1] len
1538: lnk[3n+2] link to next value
1540: The next three are always the first link
1542: lnk[3] PETSC_MIN_INT+1
1543: lnk[4] 1
1544: lnk[5] link to first real entry
1546: The next three are always the last link
1548: lnk[6] PETSC_MAX_INT - 1
1549: lnk[7] 1
1550: lnk[8] next valid link (this is the same as lnk[0] but without the decreases)
1551: */
1553: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedCreate_fast(PetscInt nlnk_max,PetscInt **lnk)
1554: {
1556: PetscInt *llnk,lsize = 0;
1559: PetscIntMultError(3,nlnk_max+3,&lsize);
1560: PetscMalloc1(lsize,lnk);
1561: llnk = *lnk;
1562: llnk[0] = 0; /* nlnk: number of entries on the list */
1563: llnk[1] = 0; /* number of integer entries represented in list */
1564: llnk[3] = PETSC_MIN_INT+1; /* value in the first node */
1565: llnk[4] = 1; /* count for the first node */
1566: llnk[5] = 6; /* next for the first node */
1567: llnk[6] = PETSC_MAX_INT-1; /* value in the last node */
1568: llnk[7] = 1; /* count for the last node */
1569: llnk[8] = 0; /* next valid node to be used */
1570: return(0);
1571: }
1573: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedAddSorted_fast(PetscInt nidx,const PetscInt indices[],PetscInt lnk[])
1574: {
1575: PetscInt k,entry,prev,next;
1576: prev = 3; /* first value */
1577: next = lnk[prev+2];
1578: for (k=0; k<nidx; k++){
1579: entry = indices[k];
1580: /* search for insertion location */
1581: while (entry >= lnk[next]) {
1582: prev = next;
1583: next = lnk[next+2];
1584: }
1585: /* entry is in range of previous list */
1586: if (entry < lnk[prev]+lnk[prev+1]) continue;
1587: lnk[1]++;
1588: /* entry is right after previous list */
1589: if (entry == lnk[prev]+lnk[prev+1]) {
1590: lnk[prev+1]++;
1591: if (lnk[next] == entry+1) { /* combine two contiguous strings */
1592: lnk[prev+1] += lnk[next+1];
1593: lnk[prev+2] = lnk[next+2];
1594: next = lnk[next+2];
1595: lnk[0]--;
1596: }
1597: continue;
1598: }
1599: /* entry is right before next list */
1600: if (entry == lnk[next]-1) {
1601: lnk[next]--;
1602: lnk[next+1]++;
1603: prev = next;
1604: next = lnk[prev+2];
1605: continue;
1606: }
1607: /* add entry into lnk */
1608: lnk[prev+2] = 3*((lnk[8]++)+3); /* connect previous node to the new node */
1609: prev = lnk[prev+2];
1610: lnk[prev] = entry; /* set value of the new node */
1611: lnk[prev+1] = 1; /* number of values in contiguous string is one to start */
1612: lnk[prev+2] = next; /* connect new node to next node */
1613: lnk[0]++;
1614: }
1615: return 0;
1616: }
1618: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedClean_fast(PetscInt nidx,PetscInt *indices,PetscInt *lnk)
1619: {
1620: PetscInt _k,_next,_nlnk,cnt,j;
1621: _next = lnk[5]; /* first node */
1622: _nlnk = lnk[0];
1623: cnt = 0;
1624: for (_k=0; _k<_nlnk; _k++){
1625: for (j=0; j<lnk[_next+1]; j++) {
1626: indices[cnt++] = lnk[_next] + j;
1627: }
1628: _next = lnk[_next + 2];
1629: }
1630: lnk[0] = 0; /* nlnk: number of links */
1631: lnk[1] = 0; /* number of integer entries represented in list */
1632: lnk[3] = PETSC_MIN_INT+1; /* value in the first node */
1633: lnk[4] = 1; /* count for the first node */
1634: lnk[5] = 6; /* next for the first node */
1635: lnk[6] = PETSC_MAX_INT-1; /* value in the last node */
1636: lnk[7] = 1; /* count for the last node */
1637: lnk[8] = 0; /* next valid location to make link */
1638: return 0;
1639: }
1641: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedView_fast(PetscInt *lnk)
1642: {
1643: PetscInt k,next,nlnk;
1644: next = lnk[5]; /* first node */
1645: nlnk = lnk[0];
1646: for (k=0; k<nlnk; k++){
1647: #if 0 /* Debugging code */
1648: printf("%d value %d len %d next %d\n",next,lnk[next],lnk[next+1],lnk[next+2]);
1649: #endif
1650: next = lnk[next + 2];
1651: }
1652: return 0;
1653: }
1655: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedDestroy_fast(PetscInt *lnk)
1656: {
1657: return PetscFree(lnk);
1658: }
1660: /* this is extern because it is used in MatFDColoringUseDM() which is in the DM library */
1661: PETSC_EXTERN PetscErrorCode MatFDColoringApply_AIJ(Mat,MatFDColoring,Vec,void*);
1663: PETSC_EXTERN PetscLogEvent MAT_Mult;
1664: PETSC_EXTERN PetscLogEvent MAT_MultMatrixFree;
1665: PETSC_EXTERN PetscLogEvent MAT_Mults;
1666: PETSC_EXTERN PetscLogEvent MAT_MultConstrained;
1667: PETSC_EXTERN PetscLogEvent MAT_MultAdd;
1668: PETSC_EXTERN PetscLogEvent MAT_MultTranspose;
1669: PETSC_EXTERN PetscLogEvent MAT_MultTransposeConstrained;
1670: PETSC_EXTERN PetscLogEvent MAT_MultTransposeAdd;
1671: PETSC_EXTERN PetscLogEvent MAT_Solve;
1672: PETSC_EXTERN PetscLogEvent MAT_Solves;
1673: PETSC_EXTERN PetscLogEvent MAT_SolveAdd;
1674: PETSC_EXTERN PetscLogEvent MAT_SolveTranspose;
1675: PETSC_EXTERN PetscLogEvent MAT_SolveTransposeAdd;
1676: PETSC_EXTERN PetscLogEvent MAT_SOR;
1677: PETSC_EXTERN PetscLogEvent MAT_ForwardSolve;
1678: PETSC_EXTERN PetscLogEvent MAT_BackwardSolve;
1679: PETSC_EXTERN PetscLogEvent MAT_LUFactor;
1680: PETSC_EXTERN PetscLogEvent MAT_LUFactorSymbolic;
1681: PETSC_EXTERN PetscLogEvent MAT_LUFactorNumeric;
1682: PETSC_EXTERN PetscLogEvent MAT_CholeskyFactor;
1683: PETSC_EXTERN PetscLogEvent MAT_CholeskyFactorSymbolic;
1684: PETSC_EXTERN PetscLogEvent MAT_CholeskyFactorNumeric;
1685: PETSC_EXTERN PetscLogEvent MAT_ILUFactor;
1686: PETSC_EXTERN PetscLogEvent MAT_ILUFactorSymbolic;
1687: PETSC_EXTERN PetscLogEvent MAT_ICCFactorSymbolic;
1688: PETSC_EXTERN PetscLogEvent MAT_Copy;
1689: PETSC_EXTERN PetscLogEvent MAT_Convert;
1690: PETSC_EXTERN PetscLogEvent MAT_Scale;
1691: PETSC_EXTERN PetscLogEvent MAT_AssemblyBegin;
1692: PETSC_EXTERN PetscLogEvent MAT_AssemblyEnd;
1693: PETSC_EXTERN PetscLogEvent MAT_SetValues;
1694: PETSC_EXTERN PetscLogEvent MAT_GetValues;
1695: PETSC_EXTERN PetscLogEvent MAT_GetRow;
1696: PETSC_EXTERN PetscLogEvent MAT_GetRowIJ;
1697: PETSC_EXTERN PetscLogEvent MAT_CreateSubMats;
1698: PETSC_EXTERN PetscLogEvent MAT_GetColoring;
1699: PETSC_EXTERN PetscLogEvent MAT_GetOrdering;
1700: PETSC_EXTERN PetscLogEvent MAT_RedundantMat;
1701: PETSC_EXTERN PetscLogEvent MAT_IncreaseOverlap;
1702: PETSC_EXTERN PetscLogEvent MAT_Partitioning;
1703: PETSC_EXTERN PetscLogEvent MAT_PartitioningND;
1704: PETSC_EXTERN PetscLogEvent MAT_Coarsen;
1705: PETSC_EXTERN PetscLogEvent MAT_ZeroEntries;
1706: PETSC_EXTERN PetscLogEvent MAT_Load;
1707: PETSC_EXTERN PetscLogEvent MAT_View;
1708: PETSC_EXTERN PetscLogEvent MAT_AXPY;
1709: PETSC_EXTERN PetscLogEvent MAT_FDColoringCreate;
1710: PETSC_EXTERN PetscLogEvent MAT_TransposeColoringCreate;
1711: PETSC_EXTERN PetscLogEvent MAT_FDColoringSetUp;
1712: PETSC_EXTERN PetscLogEvent MAT_FDColoringApply;
1713: PETSC_EXTERN PetscLogEvent MAT_Transpose;
1714: PETSC_EXTERN PetscLogEvent MAT_FDColoringFunction;
1715: PETSC_EXTERN PetscLogEvent MAT_CreateSubMat;
1716: PETSC_EXTERN PetscLogEvent MAT_MatSolve;
1717: PETSC_EXTERN PetscLogEvent MAT_MatTrSolve;
1718: PETSC_EXTERN PetscLogEvent MAT_MatMultSymbolic;
1719: PETSC_EXTERN PetscLogEvent MAT_MatMultNumeric;
1720: PETSC_EXTERN PetscLogEvent MAT_Getlocalmatcondensed;
1721: PETSC_EXTERN PetscLogEvent MAT_GetBrowsOfAcols;
1722: PETSC_EXTERN PetscLogEvent MAT_GetBrowsOfAocols;
1723: PETSC_EXTERN PetscLogEvent MAT_PtAPSymbolic;
1724: PETSC_EXTERN PetscLogEvent MAT_PtAPNumeric;
1725: PETSC_EXTERN PetscLogEvent MAT_Seqstompinum;
1726: PETSC_EXTERN PetscLogEvent MAT_Seqstompisym;
1727: PETSC_EXTERN PetscLogEvent MAT_Seqstompi;
1728: PETSC_EXTERN PetscLogEvent MAT_Getlocalmat;
1729: PETSC_EXTERN PetscLogEvent MAT_RARtSymbolic;
1730: PETSC_EXTERN PetscLogEvent MAT_RARtNumeric;
1731: PETSC_EXTERN PetscLogEvent MAT_MatTransposeMultSymbolic;
1732: PETSC_EXTERN PetscLogEvent MAT_MatTransposeMultNumeric;
1733: PETSC_EXTERN PetscLogEvent MAT_TransposeMatMultSymbolic;
1734: PETSC_EXTERN PetscLogEvent MAT_TransposeMatMultNumeric;
1735: PETSC_EXTERN PetscLogEvent MAT_MatMatMultSymbolic;
1736: PETSC_EXTERN PetscLogEvent MAT_MatMatMultNumeric;
1737: PETSC_EXTERN PetscLogEvent MAT_Applypapt;
1738: PETSC_EXTERN PetscLogEvent MAT_Applypapt_symbolic;
1739: PETSC_EXTERN PetscLogEvent MAT_Applypapt_numeric;
1740: PETSC_EXTERN PetscLogEvent MAT_Getsymtranspose;
1741: PETSC_EXTERN PetscLogEvent MAT_Getsymtransreduced;
1742: PETSC_EXTERN PetscLogEvent MAT_GetSequentialNonzeroStructure;
1743: PETSC_EXTERN PetscLogEvent MATMFFD_Mult;
1744: PETSC_EXTERN PetscLogEvent MAT_GetMultiProcBlock;
1745: PETSC_EXTERN PetscLogEvent MAT_CUSPARSECopyToGPU;
1746: PETSC_EXTERN PetscLogEvent MAT_SetValuesBatch;
1747: PETSC_EXTERN PetscLogEvent MAT_ViennaCLCopyToGPU;
1748: PETSC_EXTERN PetscLogEvent MAT_DenseCopyToGPU;
1749: PETSC_EXTERN PetscLogEvent MAT_DenseCopyFromGPU;
1750: PETSC_EXTERN PetscLogEvent MAT_Merge;
1751: PETSC_EXTERN PetscLogEvent MAT_Residual;
1752: PETSC_EXTERN PetscLogEvent MAT_SetRandom;
1753: PETSC_EXTERN PetscLogEvent MAT_FactorFactS;
1754: PETSC_EXTERN PetscLogEvent MAT_FactorInvS;
1755: PETSC_EXTERN PetscLogEvent MATCOLORING_Apply;
1756: PETSC_EXTERN PetscLogEvent MATCOLORING_Comm;
1757: PETSC_EXTERN PetscLogEvent MATCOLORING_Local;
1758: PETSC_EXTERN PetscLogEvent MATCOLORING_ISCreate;
1759: PETSC_EXTERN PetscLogEvent MATCOLORING_SetUp;
1760: PETSC_EXTERN PetscLogEvent MATCOLORING_Weights;
1762: #endif