Actual source code: matimpl.h
petsc-master 2020-08-25
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 (*setinf)(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)(void);
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)(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)(void);
142: PetscErrorCode (*matmultsymbolic)(Mat,Mat,PetscReal,Mat);
143: PetscErrorCode (*matmultnumeric)(Mat,Mat,Mat);
144: PetscErrorCode (*placeholder_92)(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)(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)(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)(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);
238: #if defined(PETSC_HAVE_SCALAPACK)
239: PETSC_INTERN PetscErrorCode MatConvert_Dense_ScaLAPACK(Mat,MatType,MatReuse,Mat*);
240: #endif
242: PETSC_INTERN PetscErrorCode MatProductSymbolic_AB(Mat);
243: PETSC_INTERN PetscErrorCode MatProductNumeric_AB(Mat);
244: PETSC_INTERN PetscErrorCode MatProductSymbolic_AtB(Mat);
245: PETSC_INTERN PetscErrorCode MatProductNumeric_AtB(Mat);
246: PETSC_INTERN PetscErrorCode MatProductSymbolic_ABt(Mat);
247: PETSC_INTERN PetscErrorCode MatProductNumeric_ABt(Mat);
248: PETSC_INTERN PetscErrorCode MatProductNumeric_PtAP(Mat);
249: PETSC_INTERN PetscErrorCode MatProductNumeric_RARt(Mat);
250: PETSC_INTERN PetscErrorCode MatProductSymbolic_ABC(Mat);
251: PETSC_INTERN PetscErrorCode MatProductNumeric_ABC(Mat);
252: PETSC_INTERN PetscErrorCode MatProductCreate_Private(Mat,Mat,Mat,Mat);
254: #if defined(PETSC_USE_DEBUG)
255: # define MatCheckPreallocated(A,arg) do { \
256: 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); \
257: } while (0)
258: #else
259: # define MatCheckPreallocated(A,arg) do {} while (0)
260: #endif
262: #if defined(PETSC_USE_DEBUG)
263: # define MatCheckProduct(A,arg) do { \
264: if (PetscUnlikely(!(A)->product)) SETERRQ2(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Argument %D \"%s\" is not a matrix obtained from MatProductCreate()",(arg),#A); \
265: } while (0)
266: #else
267: # define MatCheckProduct(A,arg) do {} while (0)
268: #endif
270: /*
271: The stash is used to temporarily store inserted matrix values that
272: belong to another processor. During the assembly phase the stashed
273: values are moved to the correct processor and
274: */
276: typedef struct _MatStashSpace *PetscMatStashSpace;
278: struct _MatStashSpace {
279: PetscMatStashSpace next;
280: PetscScalar *space_head,*val;
281: PetscInt *idx,*idy;
282: PetscInt total_space_size;
283: PetscInt local_used;
284: PetscInt local_remaining;
285: };
287: PETSC_EXTERN PetscErrorCode PetscMatStashSpaceGet(PetscInt,PetscInt,PetscMatStashSpace *);
288: PETSC_EXTERN PetscErrorCode PetscMatStashSpaceContiguous(PetscInt,PetscMatStashSpace *,PetscScalar *,PetscInt *,PetscInt *);
289: PETSC_EXTERN PetscErrorCode PetscMatStashSpaceDestroy(PetscMatStashSpace*);
291: typedef struct {
292: PetscInt count;
293: } MatStashHeader;
295: typedef struct {
296: void *buffer; /* Of type blocktype, dynamically constructed */
297: PetscInt count;
298: char pending;
299: } MatStashFrame;
301: typedef struct _MatStash MatStash;
302: struct _MatStash {
303: PetscInt nmax; /* maximum stash size */
304: PetscInt umax; /* user specified max-size */
305: PetscInt oldnmax; /* the nmax value used previously */
306: PetscInt n; /* stash size */
307: PetscInt bs; /* block size of the stash */
308: PetscInt reallocs; /* preserve the no of mallocs invoked */
309: PetscMatStashSpace space_head,space; /* linked list to hold stashed global row/column numbers and matrix values */
311: PetscErrorCode (*ScatterBegin)(Mat,MatStash*,PetscInt*);
312: PetscErrorCode (*ScatterGetMesg)(MatStash*,PetscMPIInt*,PetscInt**,PetscInt**,PetscScalar**,PetscInt*);
313: PetscErrorCode (*ScatterEnd)(MatStash*);
314: PetscErrorCode (*ScatterDestroy)(MatStash*);
316: /* The following variables are used for communication */
317: MPI_Comm comm;
318: PetscMPIInt size,rank;
319: PetscMPIInt tag1,tag2;
320: MPI_Request *send_waits; /* array of send requests */
321: MPI_Request *recv_waits; /* array of receive requests */
322: MPI_Status *send_status; /* array of send status */
323: PetscInt nsends,nrecvs; /* numbers of sends and receives */
324: PetscScalar *svalues; /* sending data */
325: PetscInt *sindices;
326: PetscScalar **rvalues; /* receiving data (values) */
327: PetscInt **rindices; /* receiving data (indices) */
328: PetscInt nprocessed; /* number of messages already processed */
329: PetscMPIInt *flg_v; /* indicates what messages have arrived so far and from whom */
330: PetscBool reproduce;
331: PetscInt reproduce_count;
333: /* The following variables are used for BTS communication */
334: PetscBool first_assembly_done; /* Is the first time matrix assembly done? */
335: PetscBool use_status; /* Use MPI_Status to determine number of items in each message */
336: PetscMPIInt nsendranks;
337: PetscMPIInt nrecvranks;
338: PetscMPIInt *sendranks;
339: PetscMPIInt *recvranks;
340: MatStashHeader *sendhdr,*recvhdr;
341: MatStashFrame *sendframes; /* pointers to the main messages */
342: MatStashFrame *recvframes;
343: MatStashFrame *recvframe_active;
344: PetscInt recvframe_i; /* index of block within active frame */
345: PetscMPIInt recvframe_count; /* Count actually sent for current frame */
346: PetscInt recvcount; /* Number of receives processed so far */
347: PetscMPIInt *some_indices; /* From last call to MPI_Waitsome */
348: MPI_Status *some_statuses; /* Statuses from last call to MPI_Waitsome */
349: PetscMPIInt some_count; /* Number of requests completed in last call to MPI_Waitsome */
350: PetscMPIInt some_i; /* Index of request currently being processed */
351: MPI_Request *sendreqs;
352: MPI_Request *recvreqs;
353: PetscSegBuffer segsendblocks;
354: PetscSegBuffer segrecvframe;
355: PetscSegBuffer segrecvblocks;
356: MPI_Datatype blocktype;
357: size_t blocktype_size;
358: InsertMode *insertmode; /* Pointer to check mat->insertmode and set upon message arrival in case no local values have been set. */
359: };
361: #if !defined(PETSC_HAVE_MPIUNI)
362: PETSC_INTERN PetscErrorCode MatStashScatterDestroy_BTS(MatStash*);
363: #endif
364: PETSC_INTERN PetscErrorCode MatStashCreate_Private(MPI_Comm,PetscInt,MatStash*);
365: PETSC_INTERN PetscErrorCode MatStashDestroy_Private(MatStash*);
366: PETSC_INTERN PetscErrorCode MatStashScatterEnd_Private(MatStash*);
367: PETSC_INTERN PetscErrorCode MatStashSetInitialSize_Private(MatStash*,PetscInt);
368: PETSC_INTERN PetscErrorCode MatStashGetInfo_Private(MatStash*,PetscInt*,PetscInt*);
369: PETSC_INTERN PetscErrorCode MatStashValuesRow_Private(MatStash*,PetscInt,PetscInt,const PetscInt[],const PetscScalar[],PetscBool);
370: PETSC_INTERN PetscErrorCode MatStashValuesCol_Private(MatStash*,PetscInt,PetscInt,const PetscInt[],const PetscScalar[],PetscInt,PetscBool);
371: PETSC_INTERN PetscErrorCode MatStashValuesRowBlocked_Private(MatStash*,PetscInt,PetscInt,const PetscInt[],const PetscScalar[],PetscInt,PetscInt,PetscInt);
372: PETSC_INTERN PetscErrorCode MatStashValuesColBlocked_Private(MatStash*,PetscInt,PetscInt,const PetscInt[],const PetscScalar[],PetscInt,PetscInt,PetscInt);
373: PETSC_INTERN PetscErrorCode MatStashScatterBegin_Private(Mat,MatStash*,PetscInt*);
374: PETSC_INTERN PetscErrorCode MatStashScatterGetMesg_Private(MatStash*,PetscMPIInt*,PetscInt**,PetscInt**,PetscScalar**,PetscInt*);
375: PETSC_INTERN PetscErrorCode MatGetInfo_External(Mat,MatInfoType,MatInfo*);
377: typedef struct {
378: PetscInt dim;
379: PetscInt dims[4];
380: PetscInt starts[4];
381: PetscBool noc; /* this is a single component problem, hence user will not set MatStencil.c */
382: } MatStencilInfo;
384: /* Info about using compressed row format */
385: typedef struct {
386: PetscBool use; /* indicates compressed rows have been checked and will be used */
387: PetscInt nrows; /* number of non-zero rows */
388: PetscInt *i; /* compressed row pointer */
389: PetscInt *rindex; /* compressed row index */
390: } Mat_CompressedRow;
391: PETSC_EXTERN PetscErrorCode MatCheckCompressedRow(Mat,PetscInt,Mat_CompressedRow*,PetscInt*,PetscInt,PetscReal);
393: typedef struct { /* used by MatCreateRedundantMatrix() for reusing matredundant */
394: PetscInt nzlocal,nsends,nrecvs;
395: PetscMPIInt *send_rank,*recv_rank;
396: PetscInt *sbuf_nz,*rbuf_nz,*sbuf_j,**rbuf_j;
397: PetscScalar *sbuf_a,**rbuf_a;
398: MPI_Comm subcomm; /* when user does not provide a subcomm */
399: IS isrow,iscol;
400: Mat *matseq;
401: } Mat_Redundant;
403: typedef struct { /* used by MatProduct() */
404: MatProductType type;
405: char *alg;
406: Mat A,B,C,Dwork;
407: PetscReal fill;
408: PetscBool api_user; /* used by MatProductSetFromOptions_xxx() to distinguish command line options */
410: /* Some products may display the information on the algorithm used */
411: PetscErrorCode (*view)(Mat,PetscViewer);
413: /* many products have intermediate data structures, each specific to Mat types and product type */
414: PetscBool clear; /* whether or not to clear the data structures after MatProductNumeric has been called */
415: void *data; /* where to stash those structures */
416: PetscErrorCode (*destroy)(void*); /* destroy routine */
417: } Mat_Product;
419: struct _p_Mat {
420: PETSCHEADER(struct _MatOps);
421: PetscLayout rmap,cmap;
422: void *data; /* implementation-specific data */
423: MatFactorType factortype; /* MAT_FACTOR_LU, ILU, CHOLESKY or ICC */
424: PetscBool useordering; /* factorization using ordering provide to routine (most PETSc implementations) */
425: PetscBool assembled; /* is the matrix assembled? */
426: PetscBool was_assembled; /* new values inserted into assembled mat */
427: PetscInt num_ass; /* number of times matrix has been assembled */
428: PetscObjectState nonzerostate; /* each time new nonzeros locations are introduced into the matrix this is updated */
429: PetscObjectState ass_nonzerostate; /* nonzero state at last assembly */
430: MatInfo info; /* matrix information */
431: InsertMode insertmode; /* have values been inserted in matrix or added? */
432: MatStash stash,bstash; /* used for assembling off-proc mat emements */
433: MatNullSpace nullsp; /* null space (operator is singular) */
434: MatNullSpace transnullsp; /* null space of transpose of operator */
435: MatNullSpace nearnullsp; /* near null space to be used by multigrid methods */
436: PetscInt congruentlayouts; /* are the rows and columns layouts congruent? */
437: PetscBool preallocated;
438: MatStencilInfo stencil; /* information for structured grid */
439: PetscBool symmetric,hermitian,structurally_symmetric,spd;
440: PetscBool symmetric_set,hermitian_set,structurally_symmetric_set,spd_set; /* if true, then corresponding flag is correct*/
441: PetscBool symmetric_eternal;
442: PetscBool nooffprocentries,nooffproczerorows;
443: PetscBool assembly_subset; /* set by MAT_SUBSET_OFF_PROC_ENTRIES */
444: PetscBool submat_singleis; /* for efficient PCSetUp_ASM() */
445: PetscBool structure_only;
446: PetscBool sortedfull; /* full, sorted rows are inserted */
447: #if defined(PETSC_HAVE_VIENNACL) || defined(PETSC_HAVE_CUDA)
448: PetscOffloadMask offloadmask; /* a mask which indicates where the valid matrix data is (GPU, CPU or both) */
449: PetscBool boundtocpu;
450: #endif
451: void *spptr; /* pointer for special library like SuperLU */
452: char *solvertype;
453: PetscBool checksymmetryonassembly,checknullspaceonassembly;
454: PetscReal checksymmetrytol;
455: Mat schur; /* Schur complement matrix */
456: MatFactorSchurStatus schur_status; /* status of the Schur complement matrix */
457: Mat_Redundant *redundant; /* used by MatCreateRedundantMatrix() */
458: PetscBool erroriffailure; /* Generate an error if detected (for example a zero pivot) instead of returning */
459: MatFactorError factorerrortype; /* type of error in factorization */
460: PetscReal factorerror_zeropivot_value; /* If numerical zero pivot was detected this is the computed value */
461: PetscInt factorerror_zeropivot_row; /* Row where zero pivot was detected */
462: PetscInt nblocks,*bsizes; /* support for MatSetVariableBlockSizes() */
463: char *defaultvectype;
464: Mat_Product *product;
465: };
467: PETSC_INTERN PetscErrorCode MatAXPY_Basic(Mat,PetscScalar,Mat,MatStructure);
468: PETSC_INTERN PetscErrorCode MatAXPY_BasicWithPreallocation(Mat,Mat,PetscScalar,Mat,MatStructure);
469: PETSC_INTERN PetscErrorCode MatAXPY_Basic_Preallocate(Mat,Mat,Mat*);
471: /*
472: Utility for MatFactor (Schur complement)
473: */
474: PETSC_INTERN PetscErrorCode MatFactorFactorizeSchurComplement_Private(Mat);
475: PETSC_INTERN PetscErrorCode MatFactorInvertSchurComplement_Private(Mat);
476: PETSC_INTERN PetscErrorCode MatFactorUpdateSchurStatus_Private(Mat);
477: PETSC_INTERN PetscErrorCode MatFactorSetUpInPlaceSchur_Private(Mat);
479: /*
480: Utility for MatZeroRows
481: */
482: PETSC_INTERN PetscErrorCode MatZeroRowsMapLocal_Private(Mat,PetscInt,const PetscInt*,PetscInt*,PetscInt**);
484: /*
485: Utility for MatView/MatLoad
486: */
487: PETSC_INTERN PetscErrorCode MatView_Binary_BlockSizes(Mat,PetscViewer);
488: PETSC_INTERN PetscErrorCode MatLoad_Binary_BlockSizes(Mat,PetscViewer);
491: /*
492: Object for partitioning graphs
493: */
495: typedef struct _MatPartitioningOps *MatPartitioningOps;
496: struct _MatPartitioningOps {
497: PetscErrorCode (*apply)(MatPartitioning,IS*);
498: PetscErrorCode (*applynd)(MatPartitioning,IS*);
499: PetscErrorCode (*setfromoptions)(PetscOptionItems*,MatPartitioning);
500: PetscErrorCode (*destroy)(MatPartitioning);
501: PetscErrorCode (*view)(MatPartitioning,PetscViewer);
502: PetscErrorCode (*improve)(MatPartitioning,IS*);
503: };
505: struct _p_MatPartitioning {
506: PETSCHEADER(struct _MatPartitioningOps);
507: Mat adj;
508: PetscInt *vertex_weights;
509: PetscReal *part_weights;
510: PetscInt n; /* number of partitions */
511: void *data;
512: PetscInt setupcalled;
513: PetscBool use_edge_weights; /* A flag indicates whether or not to use edge weights */
514: };
516: /* needed for parallel nested dissection by ParMetis and PTSCOTCH */
517: PETSC_INTERN PetscErrorCode MatPartitioningSizesToSep_Private(PetscInt,PetscInt[],PetscInt[],PetscInt[]);
519: /*
520: Object for coarsen graphs
521: */
522: typedef struct _MatCoarsenOps *MatCoarsenOps;
523: struct _MatCoarsenOps {
524: PetscErrorCode (*apply)(MatCoarsen);
525: PetscErrorCode (*setfromoptions)(PetscOptionItems*,MatCoarsen);
526: PetscErrorCode (*destroy)(MatCoarsen);
527: PetscErrorCode (*view)(MatCoarsen,PetscViewer);
528: };
530: struct _p_MatCoarsen {
531: PETSCHEADER(struct _MatCoarsenOps);
532: Mat graph;
533: PetscInt setupcalled;
534: void *subctx;
535: /* */
536: PetscBool strict_aggs;
537: IS perm;
538: PetscCoarsenData *agg_lists;
539: };
541: /*
542: MatFDColoring is used to compute Jacobian matrices efficiently
543: via coloring. The data structure is explained below in an example.
545: Color = 0 1 0 2 | 2 3 0
546: ---------------------------------------------------
547: 00 01 | 05
548: 10 11 | 14 15 Processor 0
549: 22 23 | 25
550: 32 33 |
551: ===================================================
552: | 44 45 46
553: 50 | 55 Processor 1
554: | 64 66
555: ---------------------------------------------------
557: ncolors = 4;
559: ncolumns = {2,1,1,0}
560: columns = {{0,2},{1},{3},{}}
561: nrows = {4,2,3,3}
562: rows = {{0,1,2,3},{0,1},{1,2,3},{0,1,2}}
563: vwscale = {dx(0),dx(1),dx(2),dx(3)} MPI Vec
564: vscale = {dx(0),dx(1),dx(2),dx(3),dx(4),dx(5)} Seq Vec
566: ncolumns = {1,0,1,1}
567: columns = {{6},{},{4},{5}}
568: nrows = {3,0,2,2}
569: rows = {{0,1,2},{},{1,2},{1,2}}
570: vwscale = {dx(4),dx(5),dx(6)} MPI Vec
571: vscale = {dx(0),dx(4),dx(5),dx(6)} Seq Vec
573: See the routine MatFDColoringApply() for how this data is used
574: to compute the Jacobian.
576: */
577: typedef struct {
578: PetscInt row;
579: PetscInt col;
580: PetscScalar *valaddr; /* address of value */
581: } MatEntry;
583: typedef struct {
584: PetscInt row;
585: PetscScalar *valaddr; /* address of value */
586: } MatEntry2;
588: struct _p_MatFDColoring{
589: PETSCHEADER(int);
590: PetscInt M,N,m; /* total rows, columns; local rows */
591: PetscInt rstart; /* first row owned by local processor */
592: PetscInt ncolors; /* number of colors */
593: PetscInt *ncolumns; /* number of local columns for a color */
594: PetscInt **columns; /* lists the local columns of each color (using global column numbering) */
595: IS *isa; /* these are the IS that contain the column values given in columns */
596: PetscInt *nrows; /* number of local rows for each color */
597: MatEntry *matentry; /* holds (row, column, address of value) for Jacobian matrix entry */
598: MatEntry2 *matentry2; /* holds (row, address of value) for Jacobian matrix entry */
599: PetscScalar *dy; /* store a block of F(x+dx)-F(x) when J is in BAIJ format */
600: PetscReal error_rel; /* square root of relative error in computing function */
601: PetscReal umin; /* minimum allowable u'dx value */
602: Vec w1,w2,w3; /* work vectors used in computing Jacobian */
603: PetscBool fset; /* indicates that the initial function value F(X) is set */
604: PetscErrorCode (*f)(void); /* function that defines Jacobian */
605: void *fctx; /* optional user-defined context for use by the function f */
606: Vec vscale; /* holds FD scaling, i.e. 1/dx for each perturbed column */
607: PetscInt currentcolor; /* color for which function evaluation is being done now */
608: const char *htype; /* "wp" or "ds" */
609: ISColoringType ctype; /* IS_COLORING_GLOBAL or IS_COLORING_LOCAL */
610: PetscInt brows,bcols; /* number of block rows or columns for speedup inserting the dense matrix into sparse Jacobian */
611: PetscBool setupcalled; /* true if setup has been called */
612: PetscBool viewed; /* true if the -mat_fd_coloring_view has been triggered already */
613: void (*ftn_func_pointer)(void),*ftn_func_cntx; /* serve the same purpose as *fortran_func_pointers in PETSc objects */
614: PetscObjectId matid; /* matrix this object was created with, must always be the same */
615: };
617: typedef struct _MatColoringOps *MatColoringOps;
618: struct _MatColoringOps {
619: PetscErrorCode (*destroy)(MatColoring);
620: PetscErrorCode (*setfromoptions)(PetscOptionItems*,MatColoring);
621: PetscErrorCode (*view)(MatColoring,PetscViewer);
622: PetscErrorCode (*apply)(MatColoring,ISColoring*);
623: PetscErrorCode (*weights)(MatColoring,PetscReal**,PetscInt**);
624: };
626: struct _p_MatColoring {
627: PETSCHEADER(struct _MatColoringOps);
628: Mat mat;
629: PetscInt dist; /* distance of the coloring */
630: PetscInt maxcolors; /* the maximum number of colors returned, maxcolors=1 for MIS */
631: void *data; /* inner context */
632: PetscBool valid; /* check to see if what is produced is a valid coloring */
633: MatColoringWeightType weight_type; /* type of weight computation to be performed */
634: PetscReal *user_weights; /* custom weights and permutation */
635: PetscInt *user_lperm;
636: PetscBool valid_iscoloring; /* check to see if matcoloring is produced a valid iscoloring */
637: };
639: struct _p_MatTransposeColoring{
640: PETSCHEADER(int);
641: PetscInt M,N,m; /* total rows, columns; local rows */
642: PetscInt rstart; /* first row owned by local processor */
643: PetscInt ncolors; /* number of colors */
644: PetscInt *ncolumns; /* number of local columns for a color */
645: PetscInt *nrows; /* number of local rows for each color */
646: PetscInt currentcolor; /* color for which function evaluation is being done now */
647: ISColoringType ctype; /* IS_COLORING_GLOBAL or IS_COLORING_LOCAL */
649: PetscInt *colorforrow,*colorforcol; /* pointer to rows and columns */
650: PetscInt *rows; /* lists the local rows for each color (using the local row numbering) */
651: PetscInt *den2sp; /* maps (row,color) in the dense matrix to index of sparse matrix array a->a */
652: PetscInt *columns; /* lists the local columns of each color (using global column numbering) */
653: PetscInt brows; /* number of rows for efficient implementation of MatTransColoringApplyDenToSp() */
654: PetscInt *lstart; /* array used for loop over row blocks of Csparse */
655: };
657: /*
658: Null space context for preconditioner/operators
659: */
660: struct _p_MatNullSpace {
661: PETSCHEADER(int);
662: PetscBool has_cnst;
663: PetscInt n;
664: Vec* vecs;
665: PetscScalar* alpha; /* for projections */
666: PetscErrorCode (*remove)(MatNullSpace,Vec,void*); /* for user provided removal function */
667: void* rmctx; /* context for remove() function */
668: };
670: /*
671: Checking zero pivot for LU, ILU preconditioners.
672: */
673: typedef struct {
674: PetscInt nshift,nshift_max;
675: PetscReal shift_amount,shift_lo,shift_hi,shift_top,shift_fraction;
676: PetscBool newshift;
677: PetscReal rs; /* active row sum of abs(offdiagonals) */
678: PetscScalar pv; /* pivot of the active row */
679: } FactorShiftCtx;
681: /*
682: Used by MatCreateSubMatrices_MPIXAIJ_Local()
683: */
684: #include <petscctable.h>
685: typedef struct { /* used by MatCreateSubMatrices_MPIAIJ_SingleIS_Local() and MatCreateSubMatrices_MPIAIJ_Local */
686: PetscInt id; /* index of submats, only submats[0] is responsible for deleting some arrays below */
687: PetscInt nrqs,nrqr;
688: PetscInt **rbuf1,**rbuf2,**rbuf3,**sbuf1,**sbuf2;
689: PetscInt **ptr;
690: PetscInt *tmp;
691: PetscInt *ctr;
692: PetscInt *pa; /* proc array */
693: PetscInt *req_size,*req_source1,*req_source2;
694: PetscBool allcolumns,allrows;
695: PetscBool singleis;
696: PetscInt *row2proc; /* row to proc map */
697: PetscInt nstages;
698: #if defined(PETSC_USE_CTABLE)
699: PetscTable cmap,rmap;
700: PetscInt *cmap_loc,*rmap_loc;
701: #else
702: PetscInt *cmap,*rmap;
703: #endif
705: PetscErrorCode (*destroy)(Mat);
706: } Mat_SubSppt;
708: PETSC_EXTERN PetscErrorCode MatFactorDumpMatrix(Mat);
709: PETSC_INTERN PetscErrorCode MatShift_Basic(Mat,PetscScalar);
710: PETSC_INTERN PetscErrorCode MatSetBlockSizes_Default(Mat,PetscInt,PetscInt);
712: PETSC_STATIC_INLINE PetscErrorCode MatPivotCheck_nz(Mat mat,const MatFactorInfo *info,FactorShiftCtx *sctx,PetscInt row)
713: {
714: PetscReal _rs = sctx->rs;
715: PetscReal _zero = info->zeropivot*_rs;
718: if (PetscAbsScalar(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)){
719: /* force |diag| > zeropivot*rs */
720: if (!sctx->nshift) sctx->shift_amount = info->shiftamount;
721: else sctx->shift_amount *= 2.0;
722: sctx->newshift = PETSC_TRUE;
723: (sctx->nshift)++;
724: } else {
725: sctx->newshift = PETSC_FALSE;
726: }
727: return(0);
728: }
730: PETSC_STATIC_INLINE PetscErrorCode MatPivotCheck_pd(Mat mat,const MatFactorInfo *info,FactorShiftCtx *sctx,PetscInt row)
731: {
732: PetscReal _rs = sctx->rs;
733: PetscReal _zero = info->zeropivot*_rs;
736: if (PetscRealPart(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)){
737: /* force matfactor to be diagonally dominant */
738: if (sctx->nshift == sctx->nshift_max) {
739: sctx->shift_fraction = sctx->shift_hi;
740: } else {
741: sctx->shift_lo = sctx->shift_fraction;
742: sctx->shift_fraction = (sctx->shift_hi+sctx->shift_lo)/2.;
743: }
744: sctx->shift_amount = sctx->shift_fraction * sctx->shift_top;
745: sctx->nshift++;
746: sctx->newshift = PETSC_TRUE;
747: } else {
748: sctx->newshift = PETSC_FALSE;
749: }
750: return(0);
751: }
753: PETSC_STATIC_INLINE PetscErrorCode MatPivotCheck_inblocks(Mat mat,const MatFactorInfo *info,FactorShiftCtx *sctx,PetscInt row)
754: {
755: PetscReal _zero = info->zeropivot;
758: if (PetscAbsScalar(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)){
759: sctx->pv += info->shiftamount;
760: sctx->shift_amount = 0.0;
761: sctx->nshift++;
762: }
763: sctx->newshift = PETSC_FALSE;
764: return(0);
765: }
767: PETSC_STATIC_INLINE PetscErrorCode MatPivotCheck_none(Mat fact,Mat mat,const MatFactorInfo *info,FactorShiftCtx *sctx,PetscInt row)
768: {
769: PetscReal _zero = info->zeropivot;
773: sctx->newshift = PETSC_FALSE;
774: if (PetscAbsScalar(sctx->pv) <= _zero && !PetscIsNanScalar(sctx->pv)) {
775: if (!mat->erroriffailure) {
776: PetscInfo3(mat,"Detected zero pivot in factorization in row %D value %g tolerance %g\n",row,(double)PetscAbsScalar(sctx->pv),(double)_zero);
777: fact->factorerrortype = MAT_FACTOR_NUMERIC_ZEROPIVOT;
778: fact->factorerror_zeropivot_value = PetscAbsScalar(sctx->pv);
779: fact->factorerror_zeropivot_row = row;
780: } 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);
781: }
782: return(0);
783: }
785: PETSC_STATIC_INLINE PetscErrorCode MatPivotCheck(Mat fact,Mat mat,const MatFactorInfo *info,FactorShiftCtx *sctx,PetscInt row)
786: {
790: if (info->shifttype == (PetscReal) MAT_SHIFT_NONZERO){
791: MatPivotCheck_nz(mat,info,sctx,row);
792: } else if (info->shifttype == (PetscReal) MAT_SHIFT_POSITIVE_DEFINITE){
793: MatPivotCheck_pd(mat,info,sctx,row);
794: } else if (info->shifttype == (PetscReal) MAT_SHIFT_INBLOCKS){
795: MatPivotCheck_inblocks(mat,info,sctx,row);
796: } else {
797: MatPivotCheck_none(fact,mat,info,sctx,row);
798: }
799: return(0);
800: }
802: /*
803: Create and initialize a linked list
804: Input Parameters:
805: idx_start - starting index of the list
806: lnk_max - max value of lnk indicating the end of the list
807: nlnk - max length of the list
808: Output Parameters:
809: lnk - list initialized
810: bt - PetscBT (bitarray) with all bits set to false
811: lnk_empty - flg indicating the list is empty
812: */
813: #define PetscLLCreate(idx_start,lnk_max,nlnk,lnk,bt) \
814: (PetscMalloc1(nlnk,&lnk) || PetscBTCreate(nlnk,&(bt)) || (lnk[idx_start] = lnk_max,0))
816: #define PetscLLCreate_new(idx_start,lnk_max,nlnk,lnk,bt,lnk_empty)\
817: (PetscMalloc1(nlnk,&lnk) || PetscBTCreate(nlnk,&(bt)) || (lnk_empty = PETSC_TRUE,0) ||(lnk[idx_start] = lnk_max,0))
819: /*
820: Add an index set into a sorted linked list
821: Input Parameters:
822: nidx - number of input indices
823: indices - integer array
824: idx_start - starting index of the list
825: lnk - linked list(an integer array) that is created
826: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
827: output Parameters:
828: nlnk - number of newly added indices
829: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
830: bt - updated PetscBT (bitarray)
831: */
832: #define PetscLLAdd(nidx,indices,idx_start,nlnk,lnk,bt) 0;\
833: {\
834: PetscInt _k,_entry,_location,_lnkdata;\
835: nlnk = 0;\
836: _lnkdata = idx_start;\
837: for (_k=0; _k<nidx; _k++){\
838: _entry = indices[_k];\
839: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
840: /* search for insertion location */\
841: /* start from the beginning if _entry < previous _entry */\
842: if (_k && _entry < _lnkdata) _lnkdata = idx_start;\
843: do {\
844: _location = _lnkdata;\
845: _lnkdata = lnk[_location];\
846: } while (_entry > _lnkdata);\
847: /* insertion location is found, add entry into lnk */\
848: lnk[_location] = _entry;\
849: lnk[_entry] = _lnkdata;\
850: nlnk++;\
851: _lnkdata = _entry; /* next search starts from here if next_entry > _entry */\
852: }\
853: }\
854: }
856: /*
857: Add a permuted index set into a sorted linked list
858: Input Parameters:
859: nidx - number of input indices
860: indices - integer array
861: perm - permutation of indices
862: idx_start - starting index of the list
863: lnk - linked list(an integer array) that is created
864: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
865: output Parameters:
866: nlnk - number of newly added indices
867: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
868: bt - updated PetscBT (bitarray)
869: */
870: #define PetscLLAddPerm(nidx,indices,perm,idx_start,nlnk,lnk,bt) 0;\
871: {\
872: PetscInt _k,_entry,_location,_lnkdata;\
873: nlnk = 0;\
874: _lnkdata = idx_start;\
875: for (_k=0; _k<nidx; _k++){\
876: _entry = perm[indices[_k]];\
877: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
878: /* search for insertion location */\
879: /* start from the beginning if _entry < previous _entry */\
880: if (_k && _entry < _lnkdata) _lnkdata = idx_start;\
881: do {\
882: _location = _lnkdata;\
883: _lnkdata = lnk[_location];\
884: } while (_entry > _lnkdata);\
885: /* insertion location is found, add entry into lnk */\
886: lnk[_location] = _entry;\
887: lnk[_entry] = _lnkdata;\
888: nlnk++;\
889: _lnkdata = _entry; /* next search starts from here if next_entry > _entry */\
890: }\
891: }\
892: }
894: /*
895: Add a SORTED ascending index set into a sorted linked list - same as PetscLLAdd() bus skip 'if (_k && _entry < _lnkdata) _lnkdata = idx_start;'
896: Input Parameters:
897: nidx - number of input indices
898: indices - sorted integer array
899: idx_start - starting index of the list
900: lnk - linked list(an integer array) that is created
901: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
902: output Parameters:
903: nlnk - number of newly added indices
904: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
905: bt - updated PetscBT (bitarray)
906: */
907: #define PetscLLAddSorted(nidx,indices,idx_start,nlnk,lnk,bt) 0;\
908: {\
909: PetscInt _k,_entry,_location,_lnkdata;\
910: nlnk = 0;\
911: _lnkdata = idx_start;\
912: for (_k=0; _k<nidx; _k++){\
913: _entry = indices[_k];\
914: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
915: /* search for insertion location */\
916: do {\
917: _location = _lnkdata;\
918: _lnkdata = lnk[_location];\
919: } while (_entry > _lnkdata);\
920: /* insertion location is found, add entry into lnk */\
921: lnk[_location] = _entry;\
922: lnk[_entry] = _lnkdata;\
923: nlnk++;\
924: _lnkdata = _entry; /* next search starts from here */\
925: }\
926: }\
927: }
929: #define PetscLLAddSorted_new(nidx,indices,idx_start,lnk_empty,nlnk,lnk,bt) 0; \
930: {\
931: PetscInt _k,_entry,_location,_lnkdata;\
932: if (lnk_empty){\
933: _lnkdata = idx_start; \
934: for (_k=0; _k<nidx; _k++){ \
935: _entry = indices[_k]; \
936: PetscBTSet(bt,_entry); /* mark the new entry */ \
937: _location = _lnkdata; \
938: _lnkdata = lnk[_location]; \
939: /* insertion location is found, add entry into lnk */ \
940: lnk[_location] = _entry; \
941: lnk[_entry] = _lnkdata; \
942: _lnkdata = _entry; /* next search starts from here */ \
943: } \
944: /*\
945: lnk[indices[nidx-1]] = lnk[idx_start];\
946: lnk[idx_start] = indices[0];\
947: PetscBTSet(bt,indices[0]); \
948: for (_k=1; _k<nidx; _k++){ \
949: PetscBTSet(bt,indices[_k]); \
950: lnk[indices[_k-1]] = indices[_k]; \
951: } \
952: */\
953: nlnk = nidx;\
954: lnk_empty = PETSC_FALSE;\
955: } else {\
956: nlnk = 0; \
957: _lnkdata = idx_start; \
958: for (_k=0; _k<nidx; _k++){ \
959: _entry = indices[_k]; \
960: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */ \
961: /* search for insertion location */ \
962: do { \
963: _location = _lnkdata; \
964: _lnkdata = lnk[_location]; \
965: } while (_entry > _lnkdata); \
966: /* insertion location is found, add entry into lnk */ \
967: lnk[_location] = _entry; \
968: lnk[_entry] = _lnkdata; \
969: nlnk++; \
970: _lnkdata = _entry; /* next search starts from here */ \
971: } \
972: } \
973: } \
974: }
976: /*
977: Add a SORTED index set into a sorted linked list used for LUFactorSymbolic()
978: Same as PetscLLAddSorted() with an additional operation:
979: count the number of input indices that are no larger than 'diag'
980: Input Parameters:
981: indices - sorted integer array
982: idx_start - starting index of the list, index of pivot row
983: lnk - linked list(an integer array) that is created
984: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
985: diag - index of the active row in LUFactorSymbolic
986: nzbd - number of input indices with indices <= idx_start
987: im - im[idx_start] is initialized as num of nonzero entries in row=idx_start
988: output Parameters:
989: nlnk - number of newly added indices
990: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from indices
991: bt - updated PetscBT (bitarray)
992: im - im[idx_start]: unchanged if diag is not an entry
993: : num of entries with indices <= diag if diag is an entry
994: */
995: #define PetscLLAddSortedLU(indices,idx_start,nlnk,lnk,bt,diag,nzbd,im) 0;\
996: {\
997: PetscInt _k,_entry,_location,_lnkdata,_nidx;\
998: nlnk = 0;\
999: _lnkdata = idx_start;\
1000: _nidx = im[idx_start] - nzbd; /* num of entries with idx_start < index <= diag */\
1001: for (_k=0; _k<_nidx; _k++){\
1002: _entry = indices[_k];\
1003: nzbd++;\
1004: if (_entry== diag) im[idx_start] = nzbd;\
1005: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
1006: /* search for insertion location */\
1007: do {\
1008: _location = _lnkdata;\
1009: _lnkdata = lnk[_location];\
1010: } while (_entry > _lnkdata);\
1011: /* insertion location is found, add entry into lnk */\
1012: lnk[_location] = _entry;\
1013: lnk[_entry] = _lnkdata;\
1014: nlnk++;\
1015: _lnkdata = _entry; /* next search starts from here */\
1016: }\
1017: }\
1018: }
1020: /*
1021: Copy data on the list into an array, then initialize the list
1022: Input Parameters:
1023: idx_start - starting index of the list
1024: lnk_max - max value of lnk indicating the end of the list
1025: nlnk - number of data on the list to be copied
1026: lnk - linked list
1027: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1028: output Parameters:
1029: indices - array that contains the copied data
1030: lnk - linked list that is cleaned and initialize
1031: bt - PetscBT (bitarray) with all bits set to false
1032: */
1033: #define PetscLLClean(idx_start,lnk_max,nlnk,lnk,indices,bt) 0;\
1034: {\
1035: PetscInt _j,_idx=idx_start;\
1036: for (_j=0; _j<nlnk; _j++){\
1037: _idx = lnk[_idx];\
1038: indices[_j] = _idx;\
1039: PetscBTClear(bt,_idx);\
1040: }\
1041: lnk[idx_start] = lnk_max;\
1042: }
1043: /*
1044: Free memories used by the list
1045: */
1046: #define PetscLLDestroy(lnk,bt) (PetscFree(lnk) || PetscBTDestroy(&(bt)))
1048: /* Routines below are used for incomplete matrix factorization */
1049: /*
1050: Create and initialize a linked list and its levels
1051: Input Parameters:
1052: idx_start - starting index of the list
1053: lnk_max - max value of lnk indicating the end of the list
1054: nlnk - max length of the list
1055: Output Parameters:
1056: lnk - list initialized
1057: lnk_lvl - array of size nlnk for storing levels of lnk
1058: bt - PetscBT (bitarray) with all bits set to false
1059: */
1060: #define PetscIncompleteLLCreate(idx_start,lnk_max,nlnk,lnk,lnk_lvl,bt)\
1061: (PetscIntMultError(2,nlnk,NULL) || PetscMalloc1(2*nlnk,&lnk) || PetscBTCreate(nlnk,&(bt)) || (lnk[idx_start] = lnk_max,lnk_lvl = lnk + nlnk,0))
1063: /*
1064: Initialize a sorted linked list used for ILU and ICC
1065: Input Parameters:
1066: nidx - number of input idx
1067: idx - integer array used for storing column indices
1068: idx_start - starting index of the list
1069: perm - indices of an IS
1070: lnk - linked list(an integer array) that is created
1071: lnklvl - levels of lnk
1072: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1073: output Parameters:
1074: nlnk - number of newly added idx
1075: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1076: lnklvl - levels of lnk
1077: bt - updated PetscBT (bitarray)
1078: */
1079: #define PetscIncompleteLLInit(nidx,idx,idx_start,perm,nlnk,lnk,lnklvl,bt) 0;\
1080: {\
1081: PetscInt _k,_entry,_location,_lnkdata;\
1082: nlnk = 0;\
1083: _lnkdata = idx_start;\
1084: for (_k=0; _k<nidx; _k++){\
1085: _entry = perm[idx[_k]];\
1086: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
1087: /* search for insertion location */\
1088: if (_k && _entry < _lnkdata) _lnkdata = idx_start;\
1089: do {\
1090: _location = _lnkdata;\
1091: _lnkdata = lnk[_location];\
1092: } while (_entry > _lnkdata);\
1093: /* insertion location is found, add entry into lnk */\
1094: lnk[_location] = _entry;\
1095: lnk[_entry] = _lnkdata;\
1096: lnklvl[_entry] = 0;\
1097: nlnk++;\
1098: _lnkdata = _entry; /* next search starts from here if next_entry > _entry */\
1099: }\
1100: }\
1101: }
1103: /*
1104: Add a SORTED index set into a sorted linked list for ILU
1105: Input Parameters:
1106: nidx - number of input indices
1107: idx - sorted integer array used for storing column indices
1108: level - level of fill, e.g., ICC(level)
1109: idxlvl - level of idx
1110: idx_start - starting index of the list
1111: lnk - linked list(an integer array) that is created
1112: lnklvl - levels of lnk
1113: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1114: prow - the row number of idx
1115: output Parameters:
1116: nlnk - number of newly added idx
1117: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1118: lnklvl - levels of lnk
1119: bt - updated PetscBT (bitarray)
1121: Note: the level of factor(i,j) is set as lvl(i,j) = min{ lvl(i,j), lvl(i,prow)+lvl(prow,j)+1)
1122: where idx = non-zero columns of U(prow,prow+1:n-1), prow<i
1123: */
1124: #define PetscILULLAddSorted(nidx,idx,level,idxlvl,idx_start,nlnk,lnk,lnklvl,bt,lnklvl_prow) 0;\
1125: {\
1126: PetscInt _k,_entry,_location,_lnkdata,_incrlev,_lnklvl_prow=lnklvl[prow];\
1127: nlnk = 0;\
1128: _lnkdata = idx_start;\
1129: for (_k=0; _k<nidx; _k++){\
1130: _incrlev = idxlvl[_k] + _lnklvl_prow + 1;\
1131: if (_incrlev > level) continue;\
1132: _entry = idx[_k];\
1133: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
1134: /* search for insertion location */\
1135: do {\
1136: _location = _lnkdata;\
1137: _lnkdata = lnk[_location];\
1138: } while (_entry > _lnkdata);\
1139: /* insertion location is found, add entry into lnk */\
1140: lnk[_location] = _entry;\
1141: lnk[_entry] = _lnkdata;\
1142: lnklvl[_entry] = _incrlev;\
1143: nlnk++;\
1144: _lnkdata = _entry; /* next search starts from here if next_entry > _entry */\
1145: } else { /* existing entry: update lnklvl */\
1146: if (lnklvl[_entry] > _incrlev) lnklvl[_entry] = _incrlev;\
1147: }\
1148: }\
1149: }
1151: /*
1152: Add a index set into a sorted linked list
1153: Input Parameters:
1154: nidx - number of input idx
1155: idx - integer array used for storing column indices
1156: level - level of fill, e.g., ICC(level)
1157: idxlvl - level of idx
1158: idx_start - starting index of the list
1159: lnk - linked list(an integer array) that is created
1160: lnklvl - levels of lnk
1161: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1162: output Parameters:
1163: nlnk - number of newly added idx
1164: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1165: lnklvl - levels of lnk
1166: bt - updated PetscBT (bitarray)
1167: */
1168: #define PetscIncompleteLLAdd(nidx,idx,level,idxlvl,idx_start,nlnk,lnk,lnklvl,bt) 0;\
1169: {\
1170: PetscInt _k,_entry,_location,_lnkdata,_incrlev;\
1171: nlnk = 0;\
1172: _lnkdata = idx_start;\
1173: for (_k=0; _k<nidx; _k++){\
1174: _incrlev = idxlvl[_k] + 1;\
1175: if (_incrlev > level) continue;\
1176: _entry = idx[_k];\
1177: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
1178: /* search for insertion location */\
1179: if (_k && _entry < _lnkdata) _lnkdata = idx_start;\
1180: do {\
1181: _location = _lnkdata;\
1182: _lnkdata = lnk[_location];\
1183: } while (_entry > _lnkdata);\
1184: /* insertion location is found, add entry into lnk */\
1185: lnk[_location] = _entry;\
1186: lnk[_entry] = _lnkdata;\
1187: lnklvl[_entry] = _incrlev;\
1188: nlnk++;\
1189: _lnkdata = _entry; /* next search starts from here if next_entry > _entry */\
1190: } else { /* existing entry: update lnklvl */\
1191: if (lnklvl[_entry] > _incrlev) lnklvl[_entry] = _incrlev;\
1192: }\
1193: }\
1194: }
1196: /*
1197: Add a SORTED index set into a sorted linked list
1198: Input Parameters:
1199: nidx - number of input indices
1200: idx - sorted integer array used for storing column indices
1201: level - level of fill, e.g., ICC(level)
1202: idxlvl - level of idx
1203: idx_start - starting index of the list
1204: lnk - linked list(an integer array) that is created
1205: lnklvl - levels of lnk
1206: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1207: output Parameters:
1208: nlnk - number of newly added idx
1209: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1210: lnklvl - levels of lnk
1211: bt - updated PetscBT (bitarray)
1212: */
1213: #define PetscIncompleteLLAddSorted(nidx,idx,level,idxlvl,idx_start,nlnk,lnk,lnklvl,bt) 0;\
1214: {\
1215: PetscInt _k,_entry,_location,_lnkdata,_incrlev;\
1216: nlnk = 0;\
1217: _lnkdata = idx_start;\
1218: for (_k=0; _k<nidx; _k++){\
1219: _incrlev = idxlvl[_k] + 1;\
1220: if (_incrlev > level) continue;\
1221: _entry = idx[_k];\
1222: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
1223: /* search for insertion location */\
1224: do {\
1225: _location = _lnkdata;\
1226: _lnkdata = lnk[_location];\
1227: } while (_entry > _lnkdata);\
1228: /* insertion location is found, add entry into lnk */\
1229: lnk[_location] = _entry;\
1230: lnk[_entry] = _lnkdata;\
1231: lnklvl[_entry] = _incrlev;\
1232: nlnk++;\
1233: _lnkdata = _entry; /* next search starts from here */\
1234: } else { /* existing entry: update lnklvl */\
1235: if (lnklvl[_entry] > _incrlev) lnklvl[_entry] = _incrlev;\
1236: }\
1237: }\
1238: }
1240: /*
1241: Add a SORTED index set into a sorted linked list for ICC
1242: Input Parameters:
1243: nidx - number of input indices
1244: idx - sorted integer array used for storing column indices
1245: level - level of fill, e.g., ICC(level)
1246: idxlvl - level of idx
1247: idx_start - starting index of the list
1248: lnk - linked list(an integer array) that is created
1249: lnklvl - levels of lnk
1250: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1251: idxlvl_prow - idxlvl[prow], where prow is the row number of the idx
1252: output Parameters:
1253: nlnk - number of newly added indices
1254: lnk - the sorted(increasing order) linked list containing new and non-redundate entries from idx
1255: lnklvl - levels of lnk
1256: bt - updated PetscBT (bitarray)
1257: Note: the level of U(i,j) is set as lvl(i,j) = min{ lvl(i,j), lvl(prow,i)+lvl(prow,j)+1)
1258: where idx = non-zero columns of U(prow,prow+1:n-1), prow<i
1259: */
1260: #define PetscICCLLAddSorted(nidx,idx,level,idxlvl,idx_start,nlnk,lnk,lnklvl,bt,idxlvl_prow) 0;\
1261: {\
1262: PetscInt _k,_entry,_location,_lnkdata,_incrlev;\
1263: nlnk = 0;\
1264: _lnkdata = idx_start;\
1265: for (_k=0; _k<nidx; _k++){\
1266: _incrlev = idxlvl[_k] + idxlvl_prow + 1;\
1267: if (_incrlev > level) continue;\
1268: _entry = idx[_k];\
1269: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */\
1270: /* search for insertion location */\
1271: do {\
1272: _location = _lnkdata;\
1273: _lnkdata = lnk[_location];\
1274: } while (_entry > _lnkdata);\
1275: /* insertion location is found, add entry into lnk */\
1276: lnk[_location] = _entry;\
1277: lnk[_entry] = _lnkdata;\
1278: lnklvl[_entry] = _incrlev;\
1279: nlnk++;\
1280: _lnkdata = _entry; /* next search starts from here */\
1281: } else { /* existing entry: update lnklvl */\
1282: if (lnklvl[_entry] > _incrlev) lnklvl[_entry] = _incrlev;\
1283: }\
1284: }\
1285: }
1287: /*
1288: Copy data on the list into an array, then initialize the list
1289: Input Parameters:
1290: idx_start - starting index of the list
1291: lnk_max - max value of lnk indicating the end of the list
1292: nlnk - number of data on the list to be copied
1293: lnk - linked list
1294: lnklvl - level of lnk
1295: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1296: output Parameters:
1297: indices - array that contains the copied data
1298: lnk - linked list that is cleaned and initialize
1299: lnklvl - level of lnk that is reinitialized
1300: bt - PetscBT (bitarray) with all bits set to false
1301: */
1302: #define PetscIncompleteLLClean(idx_start,lnk_max,nlnk,lnk,lnklvl,indices,indiceslvl,bt) 0;\
1303: do {\
1304: PetscInt _j,_idx=idx_start;\
1305: for (_j=0; _j<nlnk; _j++){\
1306: _idx = lnk[_idx];\
1307: *(indices+_j) = _idx;\
1308: *(indiceslvl+_j) = lnklvl[_idx];\
1309: lnklvl[_idx] = -1;\
1310: PetscBTClear(bt,_idx);\
1311: }\
1312: lnk[idx_start] = lnk_max;\
1313: } while (0)
1314: /*
1315: Free memories used by the list
1316: */
1317: #define PetscIncompleteLLDestroy(lnk,bt) (PetscFree(lnk) || PetscBTDestroy(&(bt)))
1319: #define MatCheckSameLocalSize(A,ar1,B,ar2) do { \
1321: 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)
1323: #define MatCheckSameSize(A,ar1,B,ar2) do { \
1324: 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);\
1325: MatCheckSameLocalSize(A,ar1,B,ar2);} while (0)
1327: #define VecCheckMatCompatible(M,x,ar1,b,ar2) do { \
1328: 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); \
1329: 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)
1331: /* -------------------------------------------------------------------------------------------------------*/
1332: #include <petscbt.h>
1333: /*
1334: Create and initialize a condensed linked list -
1335: same as PetscLLCreate(), but uses a scalable array 'lnk' with size of max number of entries, not O(N).
1336: Barry suggested this approach (Dec. 6, 2011):
1337: I've thought of an alternative way of representing a linked list that is efficient but doesn't have the O(N) scaling issue
1338: (it may be faster than the O(N) even sequentially due to less crazy memory access).
1340: 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
1341: 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
1342: 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
1343: 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.
1344: 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
1345: to each other so memory access is much better than using the big array.
1347: Example:
1348: nlnk_max=5, lnk_max=36:
1349: Initial list: [0, 0 | 36, 2 | 0, 0 | 0, 0 | 0, 0 | 0, 0 | 0, 0]
1350: here, head_node has index 2 with value lnk[2]=lnk_max=36,
1351: 0-th entry is used to store the number of entries in the list,
1352: The initial lnk represents head -> tail(marked by 36) with number of entries = lnk[0]=0.
1354: Now adding a sorted set {2,4}, the list becomes
1355: [2, 0 | 36, 4 |2, 6 | 4, 2 | 0, 0 | 0, 0 | 0, 0 ]
1356: represents head -> 2 -> 4 -> tail with number of entries = lnk[0]=2.
1358: Then adding a sorted set {0,3,35}, the list
1359: [5, 0 | 36, 8 | 2, 10 | 4, 12 | 0, 4 | 3, 6 | 35, 2 ]
1360: represents head -> 0 -> 2 -> 3 -> 4 -> 35 -> tail with number of entries = lnk[0]=5.
1362: Input Parameters:
1363: nlnk_max - max length of the list
1364: lnk_max - max value of the entries
1365: Output Parameters:
1366: lnk - list created and initialized
1367: bt - PetscBT (bitarray) with all bits set to false. Note: bt has size lnk_max, not nln_max!
1368: */
1369: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedCreate(PetscInt nlnk_max,PetscInt lnk_max,PetscInt **lnk,PetscBT *bt)
1370: {
1372: PetscInt *llnk,lsize = 0;
1375: PetscIntMultError(2,nlnk_max+2,&lsize);
1376: PetscMalloc1(lsize,lnk);
1377: PetscBTCreate(lnk_max,bt);
1378: llnk = *lnk;
1379: llnk[0] = 0; /* number of entries on the list */
1380: llnk[2] = lnk_max; /* value in the head node */
1381: llnk[3] = 2; /* next for the head node */
1382: return(0);
1383: }
1385: /*
1386: Add a SORTED ascending index set into a sorted linked list. See PetscLLCondensedCreate() for detailed description.
1387: Input Parameters:
1388: nidx - number of input indices
1389: indices - sorted integer array
1390: lnk - condensed linked list(an integer array) that is created
1391: bt - PetscBT (bitarray), bt[idx]=true marks idx is in lnk
1392: output Parameters:
1393: lnk - the sorted(increasing order) linked list containing previous and newly added non-redundate indices
1394: bt - updated PetscBT (bitarray)
1395: */
1396: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedAddSorted(PetscInt nidx,const PetscInt indices[],PetscInt lnk[],PetscBT bt)
1397: {
1398: PetscInt _k,_entry,_location,_next,_lnkdata,_nlnk,_newnode;
1401: _nlnk = lnk[0]; /* num of entries on the input lnk */
1402: _location = 2; /* head */
1403: for (_k=0; _k<nidx; _k++){
1404: _entry = indices[_k];
1405: if (!PetscBTLookupSet(bt,_entry)){ /* new entry */
1406: /* search for insertion location */
1407: do {
1408: _next = _location + 1; /* link from previous node to next node */
1409: _location = lnk[_next]; /* idx of next node */
1410: _lnkdata = lnk[_location];/* value of next node */
1411: } while (_entry > _lnkdata);
1412: /* insertion location is found, add entry into lnk */
1413: _newnode = 2*(_nlnk+2); /* index for this new node */
1414: lnk[_next] = _newnode; /* connect previous node to the new node */
1415: lnk[_newnode] = _entry; /* set value of the new node */
1416: lnk[_newnode+1] = _location; /* connect new node to next node */
1417: _location = _newnode; /* next search starts from the new node */
1418: _nlnk++;
1419: } \
1420: }\
1421: lnk[0] = _nlnk; /* number of entries in the list */
1422: return(0);
1423: }
1425: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedClean(PetscInt lnk_max,PetscInt nidx,PetscInt *indices,PetscInt lnk[],PetscBT bt)
1426: {
1428: PetscInt _k,_next,_nlnk;
1431: _next = lnk[3]; /* head node */
1432: _nlnk = lnk[0]; /* num of entries on the list */
1433: for (_k=0; _k<_nlnk; _k++){
1434: indices[_k] = lnk[_next];
1435: _next = lnk[_next + 1];
1436: PetscBTClear(bt,indices[_k]);
1437: }
1438: lnk[0] = 0; /* num of entries on the list */
1439: lnk[2] = lnk_max; /* initialize head node */
1440: lnk[3] = 2; /* head node */
1441: return(0);
1442: }
1444: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedView(PetscInt *lnk)
1445: {
1447: PetscInt k;
1450: PetscPrintf(PETSC_COMM_SELF,"LLCondensed of size %D, (val, next)\n",lnk[0]);
1451: for (k=2; k< lnk[0]+2; k++){
1452: PetscPrintf(PETSC_COMM_SELF," %D: (%D, %D)\n",2*k,lnk[2*k],lnk[2*k+1]);
1453: }
1454: return(0);
1455: }
1457: /*
1458: Free memories used by the list
1459: */
1460: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedDestroy(PetscInt *lnk,PetscBT bt)
1461: {
1465: PetscFree(lnk);
1466: PetscBTDestroy(&bt);
1467: return(0);
1468: }
1470: /* -------------------------------------------------------------------------------------------------------*/
1471: /*
1472: Same as PetscLLCondensedCreate(), but does not use non-scalable O(lnk_max) bitarray
1473: Input Parameters:
1474: nlnk_max - max length of the list
1475: Output Parameters:
1476: lnk - list created and initialized
1477: */
1478: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedCreate_Scalable(PetscInt nlnk_max,PetscInt **lnk)
1479: {
1481: PetscInt *llnk,lsize = 0;
1484: PetscIntMultError(2,nlnk_max+2,&lsize);
1485: PetscMalloc1(lsize,lnk);
1486: llnk = *lnk;
1487: llnk[0] = 0; /* number of entries on the list */
1488: llnk[2] = PETSC_MAX_INT; /* value in the head node */
1489: llnk[3] = 2; /* next for the head node */
1490: return(0);
1491: }
1493: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedExpand_Scalable(PetscInt nlnk_max,PetscInt **lnk)
1494: {
1496: PetscInt lsize = 0;
1499: PetscIntMultError(2,nlnk_max+2,&lsize);
1500: PetscRealloc(lsize*sizeof(PetscInt),lnk);
1501: return(0);
1502: }
1504: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedAddSorted_Scalable(PetscInt nidx,const PetscInt indices[],PetscInt lnk[])
1505: {
1506: PetscInt _k,_entry,_location,_next,_lnkdata,_nlnk,_newnode;
1507: _nlnk = lnk[0]; /* num of entries on the input lnk */
1508: _location = 2; /* head */ \
1509: for (_k=0; _k<nidx; _k++){
1510: _entry = indices[_k];
1511: /* search for insertion location */
1512: do {
1513: _next = _location + 1; /* link from previous node to next node */
1514: _location = lnk[_next]; /* idx of next node */
1515: _lnkdata = lnk[_location];/* value of next node */
1516: } while (_entry > _lnkdata);
1517: if (_entry < _lnkdata) {
1518: /* insertion location is found, add entry into lnk */
1519: _newnode = 2*(_nlnk+2); /* index for this new node */
1520: lnk[_next] = _newnode; /* connect previous node to the new node */
1521: lnk[_newnode] = _entry; /* set value of the new node */
1522: lnk[_newnode+1] = _location; /* connect new node to next node */
1523: _location = _newnode; /* next search starts from the new node */
1524: _nlnk++;
1525: }
1526: }
1527: lnk[0] = _nlnk; /* number of entries in the list */
1528: return 0;
1529: }
1531: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedClean_Scalable(PetscInt nidx,PetscInt *indices,PetscInt *lnk)
1532: {
1533: PetscInt _k,_next,_nlnk;
1534: _next = lnk[3]; /* head node */
1535: _nlnk = lnk[0];
1536: for (_k=0; _k<_nlnk; _k++){
1537: indices[_k] = lnk[_next];
1538: _next = lnk[_next + 1];
1539: }
1540: lnk[0] = 0; /* num of entries on the list */
1541: lnk[3] = 2; /* head node */
1542: return 0;
1543: }
1545: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedDestroy_Scalable(PetscInt *lnk)
1546: {
1547: return PetscFree(lnk);
1548: }
1550: /* -------------------------------------------------------------------------------------------------------*/
1551: /*
1552: lnk[0] number of links
1553: lnk[1] number of entries
1554: lnk[3n] value
1555: lnk[3n+1] len
1556: lnk[3n+2] link to next value
1558: The next three are always the first link
1560: lnk[3] PETSC_MIN_INT+1
1561: lnk[4] 1
1562: lnk[5] link to first real entry
1564: The next three are always the last link
1566: lnk[6] PETSC_MAX_INT - 1
1567: lnk[7] 1
1568: lnk[8] next valid link (this is the same as lnk[0] but without the decreases)
1569: */
1571: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedCreate_fast(PetscInt nlnk_max,PetscInt **lnk)
1572: {
1574: PetscInt *llnk,lsize = 0;
1577: PetscIntMultError(3,nlnk_max+3,&lsize);
1578: PetscMalloc1(lsize,lnk);
1579: llnk = *lnk;
1580: llnk[0] = 0; /* nlnk: number of entries on the list */
1581: llnk[1] = 0; /* number of integer entries represented in list */
1582: llnk[3] = PETSC_MIN_INT+1; /* value in the first node */
1583: llnk[4] = 1; /* count for the first node */
1584: llnk[5] = 6; /* next for the first node */
1585: llnk[6] = PETSC_MAX_INT-1; /* value in the last node */
1586: llnk[7] = 1; /* count for the last node */
1587: llnk[8] = 0; /* next valid node to be used */
1588: return(0);
1589: }
1591: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedAddSorted_fast(PetscInt nidx,const PetscInt indices[],PetscInt lnk[])
1592: {
1593: PetscInt k,entry,prev,next;
1594: prev = 3; /* first value */
1595: next = lnk[prev+2];
1596: for (k=0; k<nidx; k++){
1597: entry = indices[k];
1598: /* search for insertion location */
1599: while (entry >= lnk[next]) {
1600: prev = next;
1601: next = lnk[next+2];
1602: }
1603: /* entry is in range of previous list */
1604: if (entry < lnk[prev]+lnk[prev+1]) continue;
1605: lnk[1]++;
1606: /* entry is right after previous list */
1607: if (entry == lnk[prev]+lnk[prev+1]) {
1608: lnk[prev+1]++;
1609: if (lnk[next] == entry+1) { /* combine two contiguous strings */
1610: lnk[prev+1] += lnk[next+1];
1611: lnk[prev+2] = lnk[next+2];
1612: next = lnk[next+2];
1613: lnk[0]--;
1614: }
1615: continue;
1616: }
1617: /* entry is right before next list */
1618: if (entry == lnk[next]-1) {
1619: lnk[next]--;
1620: lnk[next+1]++;
1621: prev = next;
1622: next = lnk[prev+2];
1623: continue;
1624: }
1625: /* add entry into lnk */
1626: lnk[prev+2] = 3*((lnk[8]++)+3); /* connect previous node to the new node */
1627: prev = lnk[prev+2];
1628: lnk[prev] = entry; /* set value of the new node */
1629: lnk[prev+1] = 1; /* number of values in contiguous string is one to start */
1630: lnk[prev+2] = next; /* connect new node to next node */
1631: lnk[0]++;
1632: }
1633: return 0;
1634: }
1636: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedClean_fast(PetscInt nidx,PetscInt *indices,PetscInt *lnk)
1637: {
1638: PetscInt _k,_next,_nlnk,cnt,j;
1639: _next = lnk[5]; /* first node */
1640: _nlnk = lnk[0];
1641: cnt = 0;
1642: for (_k=0; _k<_nlnk; _k++){
1643: for (j=0; j<lnk[_next+1]; j++) {
1644: indices[cnt++] = lnk[_next] + j;
1645: }
1646: _next = lnk[_next + 2];
1647: }
1648: lnk[0] = 0; /* nlnk: number of links */
1649: lnk[1] = 0; /* number of integer entries represented in list */
1650: lnk[3] = PETSC_MIN_INT+1; /* value in the first node */
1651: lnk[4] = 1; /* count for the first node */
1652: lnk[5] = 6; /* next for the first node */
1653: lnk[6] = PETSC_MAX_INT-1; /* value in the last node */
1654: lnk[7] = 1; /* count for the last node */
1655: lnk[8] = 0; /* next valid location to make link */
1656: return 0;
1657: }
1659: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedView_fast(PetscInt *lnk)
1660: {
1661: PetscInt k,next,nlnk;
1662: next = lnk[5]; /* first node */
1663: nlnk = lnk[0];
1664: for (k=0; k<nlnk; k++){
1665: #if 0 /* Debugging code */
1666: printf("%d value %d len %d next %d\n",next,lnk[next],lnk[next+1],lnk[next+2]);
1667: #endif
1668: next = lnk[next + 2];
1669: }
1670: return 0;
1671: }
1673: PETSC_STATIC_INLINE PetscErrorCode PetscLLCondensedDestroy_fast(PetscInt *lnk)
1674: {
1675: return PetscFree(lnk);
1676: }
1678: /* this is extern because it is used in MatFDColoringUseDM() which is in the DM library */
1679: PETSC_EXTERN PetscErrorCode MatFDColoringApply_AIJ(Mat,MatFDColoring,Vec,void*);
1681: PETSC_EXTERN PetscLogEvent MAT_Mult;
1682: PETSC_EXTERN PetscLogEvent MAT_MultMatrixFree;
1683: PETSC_EXTERN PetscLogEvent MAT_Mults;
1684: PETSC_EXTERN PetscLogEvent MAT_MultConstrained;
1685: PETSC_EXTERN PetscLogEvent MAT_MultAdd;
1686: PETSC_EXTERN PetscLogEvent MAT_MultTranspose;
1687: PETSC_EXTERN PetscLogEvent MAT_MultTransposeConstrained;
1688: PETSC_EXTERN PetscLogEvent MAT_MultTransposeAdd;
1689: PETSC_EXTERN PetscLogEvent MAT_Solve;
1690: PETSC_EXTERN PetscLogEvent MAT_Solves;
1691: PETSC_EXTERN PetscLogEvent MAT_SolveAdd;
1692: PETSC_EXTERN PetscLogEvent MAT_SolveTranspose;
1693: PETSC_EXTERN PetscLogEvent MAT_SolveTransposeAdd;
1694: PETSC_EXTERN PetscLogEvent MAT_SOR;
1695: PETSC_EXTERN PetscLogEvent MAT_ForwardSolve;
1696: PETSC_EXTERN PetscLogEvent MAT_BackwardSolve;
1697: PETSC_EXTERN PetscLogEvent MAT_LUFactor;
1698: PETSC_EXTERN PetscLogEvent MAT_LUFactorSymbolic;
1699: PETSC_EXTERN PetscLogEvent MAT_LUFactorNumeric;
1700: PETSC_EXTERN PetscLogEvent MAT_CholeskyFactor;
1701: PETSC_EXTERN PetscLogEvent MAT_CholeskyFactorSymbolic;
1702: PETSC_EXTERN PetscLogEvent MAT_CholeskyFactorNumeric;
1703: PETSC_EXTERN PetscLogEvent MAT_ILUFactor;
1704: PETSC_EXTERN PetscLogEvent MAT_ILUFactorSymbolic;
1705: PETSC_EXTERN PetscLogEvent MAT_ICCFactorSymbolic;
1706: PETSC_EXTERN PetscLogEvent MAT_Copy;
1707: PETSC_EXTERN PetscLogEvent MAT_Convert;
1708: PETSC_EXTERN PetscLogEvent MAT_Scale;
1709: PETSC_EXTERN PetscLogEvent MAT_AssemblyBegin;
1710: PETSC_EXTERN PetscLogEvent MAT_AssemblyEnd;
1711: PETSC_EXTERN PetscLogEvent MAT_SetValues;
1712: PETSC_EXTERN PetscLogEvent MAT_GetValues;
1713: PETSC_EXTERN PetscLogEvent MAT_GetRow;
1714: PETSC_EXTERN PetscLogEvent MAT_GetRowIJ;
1715: PETSC_EXTERN PetscLogEvent MAT_CreateSubMats;
1716: PETSC_EXTERN PetscLogEvent MAT_GetColoring;
1717: PETSC_EXTERN PetscLogEvent MAT_GetOrdering;
1718: PETSC_EXTERN PetscLogEvent MAT_RedundantMat;
1719: PETSC_EXTERN PetscLogEvent MAT_IncreaseOverlap;
1720: PETSC_EXTERN PetscLogEvent MAT_Partitioning;
1721: PETSC_EXTERN PetscLogEvent MAT_PartitioningND;
1722: PETSC_EXTERN PetscLogEvent MAT_Coarsen;
1723: PETSC_EXTERN PetscLogEvent MAT_ZeroEntries;
1724: PETSC_EXTERN PetscLogEvent MAT_Load;
1725: PETSC_EXTERN PetscLogEvent MAT_View;
1726: PETSC_EXTERN PetscLogEvent MAT_AXPY;
1727: PETSC_EXTERN PetscLogEvent MAT_FDColoringCreate;
1728: PETSC_EXTERN PetscLogEvent MAT_TransposeColoringCreate;
1729: PETSC_EXTERN PetscLogEvent MAT_FDColoringSetUp;
1730: PETSC_EXTERN PetscLogEvent MAT_FDColoringApply;
1731: PETSC_EXTERN PetscLogEvent MAT_Transpose;
1732: PETSC_EXTERN PetscLogEvent MAT_FDColoringFunction;
1733: PETSC_EXTERN PetscLogEvent MAT_CreateSubMat;
1734: PETSC_EXTERN PetscLogEvent MAT_MatSolve;
1735: PETSC_EXTERN PetscLogEvent MAT_MatTrSolve;
1736: PETSC_EXTERN PetscLogEvent MAT_MatMultSymbolic;
1737: PETSC_EXTERN PetscLogEvent MAT_MatMultNumeric;
1738: PETSC_EXTERN PetscLogEvent MAT_Getlocalmatcondensed;
1739: PETSC_EXTERN PetscLogEvent MAT_GetBrowsOfAcols;
1740: PETSC_EXTERN PetscLogEvent MAT_GetBrowsOfAocols;
1741: PETSC_EXTERN PetscLogEvent MAT_PtAPSymbolic;
1742: PETSC_EXTERN PetscLogEvent MAT_PtAPNumeric;
1743: PETSC_EXTERN PetscLogEvent MAT_Seqstompinum;
1744: PETSC_EXTERN PetscLogEvent MAT_Seqstompisym;
1745: PETSC_EXTERN PetscLogEvent MAT_Seqstompi;
1746: PETSC_EXTERN PetscLogEvent MAT_Getlocalmat;
1747: PETSC_EXTERN PetscLogEvent MAT_RARtSymbolic;
1748: PETSC_EXTERN PetscLogEvent MAT_RARtNumeric;
1749: PETSC_EXTERN PetscLogEvent MAT_MatTransposeMultSymbolic;
1750: PETSC_EXTERN PetscLogEvent MAT_MatTransposeMultNumeric;
1751: PETSC_EXTERN PetscLogEvent MAT_TransposeMatMultSymbolic;
1752: PETSC_EXTERN PetscLogEvent MAT_TransposeMatMultNumeric;
1753: PETSC_EXTERN PetscLogEvent MAT_MatMatMultSymbolic;
1754: PETSC_EXTERN PetscLogEvent MAT_MatMatMultNumeric;
1755: PETSC_EXTERN PetscLogEvent MAT_Applypapt;
1756: PETSC_EXTERN PetscLogEvent MAT_Applypapt_symbolic;
1757: PETSC_EXTERN PetscLogEvent MAT_Applypapt_numeric;
1758: PETSC_EXTERN PetscLogEvent MAT_Getsymtranspose;
1759: PETSC_EXTERN PetscLogEvent MAT_Getsymtransreduced;
1760: PETSC_EXTERN PetscLogEvent MAT_GetSequentialNonzeroStructure;
1761: PETSC_EXTERN PetscLogEvent MATMFFD_Mult;
1762: PETSC_EXTERN PetscLogEvent MAT_GetMultiProcBlock;
1763: PETSC_EXTERN PetscLogEvent MAT_CUSPARSECopyToGPU;
1764: PETSC_EXTERN PetscLogEvent MAT_CUSPARSEGenerateTranspose;
1765: PETSC_EXTERN PetscLogEvent MAT_SetValuesBatch;
1766: PETSC_EXTERN PetscLogEvent MAT_ViennaCLCopyToGPU;
1767: PETSC_EXTERN PetscLogEvent MAT_DenseCopyToGPU;
1768: PETSC_EXTERN PetscLogEvent MAT_DenseCopyFromGPU;
1769: PETSC_EXTERN PetscLogEvent MAT_Merge;
1770: PETSC_EXTERN PetscLogEvent MAT_Residual;
1771: PETSC_EXTERN PetscLogEvent MAT_SetRandom;
1772: PETSC_EXTERN PetscLogEvent MAT_FactorFactS;
1773: PETSC_EXTERN PetscLogEvent MAT_FactorInvS;
1774: PETSC_EXTERN PetscLogEvent MATCOLORING_Apply;
1775: PETSC_EXTERN PetscLogEvent MATCOLORING_Comm;
1776: PETSC_EXTERN PetscLogEvent MATCOLORING_Local;
1777: PETSC_EXTERN PetscLogEvent MATCOLORING_ISCreate;
1778: PETSC_EXTERN PetscLogEvent MATCOLORING_SetUp;
1779: PETSC_EXTERN PetscLogEvent MATCOLORING_Weights;
1781: #endif