1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
|
c-------------------------------------------------------------------------------c
c c
c QUICK REFERENCE c
c c
c-------------------------------------------------------------------------------c
c For convenience we list here the most important subroutines c
c in the various modules of SPARSKIT. More detailed information can be c
c found either in the body of the paper or in the documentation of the package.c
c c
c-------------------------------------------------------------------------------c
c===============================================================================c
c FORMATS Module c
c===============================================================================c
c c
c CSRDNS : converts a row-stored sparse matrix into the dense format. c
c DNSCSR : converts a dense matrix to a sparse storage format. c
c COOCSR : converts coordinate to to csr format c
c COICSR : in-place conversion of coordinate to csr format c
c CSRCOO : converts compressed sparse row to coordinate format. c
c CSRSSR : converts compressed sparse row to symmetric sparse row format. c
c SSRCSR : converts symmetric sparse row to compressed sparse row format. c
c CSRELL : converts compressed sparse row to Ellpack format c
c ELLCSR : converts Ellpack format to compressed sparse row format. c
c CSRMSR : converts compressed sparse row format to modified sparse c
c row format. c
c MSRCSR : converts modified sparse row format to compressed sparse c
c row format. c
c CSRCSC : converts compressed sparse row format to compressed sparse c
c column format (transposition). c
c CSRDIA : converts the compressed sparse row format into the diagonal c
c format. c
c DIACSR : converts the diagonal format into the compressed sparse row c
c format. c
c BSRCSR : converts the block-row sparse format into the compressed c
c sparse row format. c
c CSRBSR : converts the compressed sparse row format into the block-row c
c sparse format. c
c CSRBND : converts the compressed sparse row format into the banded c
c format (Linpack style). c
c BNDCSR : converts the banded format (Linpack style) into the compressed c
c sparse row storage. c
c CSRSSK : converts the compressed sparse row format to the symmetric c
c skyline format c
c SSKSSR : converts symmetric skyline format to symmetric sparse row format. c
c CSRJAD : converts the csr format into the jagged diagonal format c
c JADCSR : converts the jagged-diagonal format into the csr format c
c COOELL : converts the coordinate format into the Ellpack/Itpack format. c
c CSRVBR : converts the compressed sparse row format into the c
c variable block row format. c
c VBRCSR : converts the variable block row format into the c
c compressed sparse row format. c
c CSORTED : Checks if matrix in CSR format is sorted by columns. c
c c
c-------------------------------------------------------------------------------c
c===============================================================================c
c UNARY Module c
c===============================================================================c
c c
c SUBMAT : extracts a submatrix from a sparse matrix. c
c FILTER : filters elements from a matrix according to their magnitude. c
c FILTERM: Same as above, but for the MSR format. c
c TRANSP : in-place transposition routine (see also CSRCSC in formats) c
c GETELM : returns a(i,j) for any (i,j) from a CSR-stored matrix. c
c COPMAT : copies a matrix into another matrix (both stored csr). c
c MSRCOP : copies a matrix in MSR format into a matrix in MSR format. c
c GETELM : returns a(i,j) for any (i,j) from a CSR-stored matrix. c
c GETDIA : extracts a specified diagonal from a matrix. c
c GETL : extracts lower triangular part. c
c GETU : extracts upper triangular part. c
c LEVELS : gets the level scheduling structure for lower triangular matrices. c
c AMASK : extracts C = A * M c
c RPERM : permutes the rows of a matrix (B = P A) c
c CPERM : permutes the columns of a matrix (B = A Q) c
c DPERM : permutes a matrix (B = P A Q) given two permutations P, Q c
c DPERM2 : general submatrix permutation/extraction routine. c
c DMPERM : symmetric permutation of row and column (B=PAP') in MSR fmt. c
c DVPERM : permutes a vector (in-place). c
c IVPERM : permutes an integer vector (in-place). c
c RETMX : returns the max absolute value in each row of the matrix. c
c DIAPOS : returns the positions of the diagonal elements in A. c
c EXTBDG : extracts the main diagonal blocks of a matrix. c
c GETBWD : returns the bandwidth information on a matrix. c
c BLKFND : finds the block-size of a matrix. c
c BLKCHK : checks whether a given integer is the block size of A. c
c INFDIA : obtains information on the diagonals of A. c
c AMUBDG : computes the number of nonzero elements in each row of A*B. c
c APLBDG : computes the number of nonzero elements in each row of A+B. c
c RNRMS : computes the norms of the rows of A. c
c CNRMS : computes the norms of the columns of A. c
c ROSCAL : scales the rows of a matrix by their norms. c
c COSCAL : scales the columns of a matrix by their norms. c
c ADDBLK : adds a matrix B into a block of A. c
c GET1UP : collects the first elements of each row of the upper c
c triangular portion of the matrix. c
c XTROWS : extracts given rows from a matrix in CSR format. c
c CSRKVSTR : Finds block row partitioning of matrix in CSR format c
c CSRKVSTC : Finds block column partitioning of matrix in CSR format c
c KVSTMERGE: Merges block partitionings, for conformal row/col pattern c
c c
c-------------------------------------------------------------------------------c
c===============================================================================c
c INOUT Module c
c===============================================================================c
c c
c READMT : reads matrices in the boeing/Harewell format. c
c PRTMT : prints matrices in the boeing/Harewell format. c
c DUMP : prints rows of a matrix, in a readable format. c
c PLTMT : produces a 'pic' file for plotting a sparse matrix. c
c PSPLTM : Generates a post-script plot of the non-zero pattern of A. c
c SMMS : Write the matrx in a format used in SMMS package. c
c READSM : Reads matrices in coordinate format (as in SMMS package). c
c READSK : Reads matrices in CSR format (simplified H/B formate). c
c SKIT : Writes matrices to a file, format same as above. c
c PRTUNF : Writes matrices (in CSR format) unformatted. c
c READUNF : Reads unformatted data of matrices (in CSR format). c
c c
c-------------------------------------------------------------------------------c
c===============================================================================c
c INFO Module c
c===============================================================================c
c c
c DINFO1 : obtains a number of statistics on a sparse matrix. c
c VBRINFO: Print info on matrix in variable block row format c
c c
c-------------------------------------------------------------------------------c
c===============================================================================c
c MATGEN Module c
c===============================================================================c
c c
c GEN57PT : generates 5-point and 7-point matrices. c
c GEN57BL : generates block 5-point and 7-point matrices. c
c GENFEA : generates finite element matrices in assembled form. c
c GENFEU : generates finite element matrices in unassembled form. c
c ASSMB1 : assembles an unassembled matrix (as produced by genfeu). c
c MATRF2 : Routines for generating sparse matrices by Zlatev et al. c
c DCN : Routines for generating sparse matrices by Zlatev et al. c
c ECN : Routines for generating sparse matrices by Zlatev et al. c
c MARKGEN : subroutine to produce a Markov chain matrix for a random walk. c
c c
c-------------------------------------------------------------------------------c
c===============================================================================c
c BLASSM Module c
c===============================================================================c
c c
c AMUB : computes C = A*B c
c APLB : computes C = A+B c
c APLSB : computes C = A + s B c
c APMBT : Computes C = A +(-) B^T c
c APLSBT : Computes C = A + s * B^T c
c DIAMUA : Computes C = Diag * A c
c AMUDIA : Computes C = A* Diag c
c APLDIA : Computes C = A + Diag c
c APLSCA : Computes A:= A + s I (s = scalar) c
c c
c-------------------------------------------------------------------------------c
c===============================================================================c
c MATVEC Module c
c===============================================================================c
c c
c AMUX : A times a vector. Compressed Sparse Row (CSR) format. c
c ATMUX : A^T times a vector. CSR format. c
c AMUXE : A times a vector. Ellpack/Itpack (ELL) format. c
c AMUXD : A times a vector. Diagonal (DIA) format. c
c AMUXJ : A times a vector. Jagged Diagonal (JAD) format. c
c VBRMV : Sparse matrix-full vector product, in VBR format c
c c
c LSOL : Unit lower triangular system solution. Compressed Sparse Row c
c (CSR) format. c
c LDSOL : Lower triangular system solution. Modified Sparse Row (MSR) format. c
c LSOL : Unit lower triangular system solution. Compressed Sparse Column c
c (CSC) format. c
c LDSOLC: Lower triangular system solution. Modif. Sparse Column (MSC) format c
c LDSOLL: Lower triangular system solution with level scheduling. MSR format. c
c USOL : Unit upper triangular system solution. Compressed Sparse Row c
c (CSR) format. c
c UDSOL : Upper triangular system solution. Modified Sparse Row (MSR) format. c
c USOLC : Unit upper triangular system solution. Compressed Sparse c
c Column (CSC) format. c
c UDSOLC: Upper triangular system solution. Modif. Sparse Column (MSC) format c
c c
c-------------------------------------------------------------------------------c
c===============================================================================c
c ORDERINGS Module c
c===============================================================================c
c c
c LEVSET : The standard Cuthill-McKee ordering algorithm. c
c COLOR : A greedy algorithm for multicoloring ordering. c
c CCN : Strongly connected components. c
c c
c-------------------------------------------------------------------------------c
c===============================================================================c
c ITSOL Module c
c===============================================================================c
c c
c ILUT : ILUT(k) preconditioned GMRES mini package. c
c ITERS : Nine Krylov iterative solvers with reverse communication. c
c c
c-------------------------------------------------------------------------------c
c===============================================================================c
|
