SuperLU Distributed 8.2.1
Distributed memory sparse direct solver
sreadhb.c File Reference

Read a FLOAT PRECISION matrix stored in Harwell-Boeing format. More...

#include <math.h>
#include <stdio.h>
#include <stdlib.h>
#include "superlu_sdefs.h"
Include dependency graph for sreadhb.c:

Functions

static void ReadVector (FILE *, int_t, int_t *, int_t, int_t)
 
static void sReadValues (FILE *, int_t, float *, int_t, int_t)
 
static void FormFullA (int_t, int_t *, float **, int_t **, int_t **)
 
static int DumpLine (FILE *)
 
static int ParseIntFormat (char *, int_t *, int_t *)
 
static int ParseFloatFormat (char *, int_t *, int_t *)
 
void sreadhb_dist (int iam, FILE *fp, int_t *nrow, int_t *ncol, int_t *nonz, float **nzval, int_t **rowind, int_t **colptr)
 

Detailed Description

Read a FLOAT PRECISION matrix stored in Harwell-Boeing format.

Copyright (c) 2003, The Regents of the University of California, through Lawrence Berkeley National Laboratory (subject to receipt of any required approvals from U.S. Dept. of Energy)

All rights reserved.

The source code is distributed under BSD license, see the file License.txt at the top-level directory.

-- Distributed SuperLU routine (version 1.0) --
Lawrence Berkeley National Lab, Univ. of California Berkeley.
September 1, 1999

Function Documentation

◆ DumpLine()

static int DumpLine ( FILE *  fp)
static
Here is the caller graph for this function:

◆ FormFullA()

static void FormFullA ( int_t  n,
int_t nonz,
float **  nzval,
int_t **  rowind,
int_t **  colptr 
)
static
On input, nonz/nzval/rowind/colptr represents lower part of a symmetric
matrix. On exit, it represents the full matrix with lower and upper parts.
Here is the caller graph for this function:

◆ ParseFloatFormat()

static int ParseFloatFormat ( char *  buf,
int_t num,
int_t size 
)
static
Here is the caller graph for this function:

◆ ParseIntFormat()

static int ParseIntFormat ( char *  buf,
int_t num,
int_t size 
)
static
Here is the caller graph for this function:

◆ ReadVector()

static void ReadVector ( FILE *  fp,
int_t  n,
int_t where,
int_t  perline,
int_t  persize 
)
static
Here is the caller graph for this function:

◆ sreadhb_dist()

void sreadhb_dist ( int  iam,
FILE *  fp,
int_t nrow,
int_t ncol,
int_t nonz,
float **  nzval,
int_t **  rowind,
int_t **  colptr 
)
Purpose
=======

Read a FLOAT PRECISION matrix stored in Harwell-Boeing format
as described below.

Line 1 (A72,A8)
    Col. 1 - 72   Title (TITLE)
 Col. 73 - 80  Key (KEY)

Line 2 (5I14)
    Col. 1 - 14   Total number of lines excluding header (TOTCRD)
    Col. 15 - 28  Number of lines for pointers (PTRCRD)
    Col. 29 - 42  Number of lines for row (or variable) indices (INDCRD)
    Col. 43 - 56  Number of lines for numerical values (VALCRD)
 Col. 57 - 70  Number of lines for right-hand sides (RHSCRD)
                   (including starting guesses and solution vectors
           if present)
                  (zero indicates no right-hand side data is present)

Line 3 (A3, 11X, 4I14)
    Col. 1 - 3    Matrix type (see below) (MXTYPE)
    Col. 15 - 28  Number of rows (or variables) (NROW)
    Col. 29 - 42  Number of columns (or elements) (NCOL)
 Col. 43 - 56  Number of row (or variable) indices (NNZERO)
               (equal to number of entries for assembled matrices)
    Col. 57 - 70  Number of elemental matrix entries (NELTVL)
               (zero in the case of assembled matrices)
Line 4 (2A16, 2A20)
    Col. 1 - 16   Format for pointers (PTRFMT)
 Col. 17 - 32  Format for row (or variable) indices (INDFMT)
 Col. 33 - 52  Format for numerical values of coefficient matrix (VALFMT)
    Col. 53 - 72 Format for numerical values of right-hand sides (RHSFMT)

Line 5 (A3, 11X, 2I14) Only present if there are right-hand sides present
    Col. 1        Right-hand side type:
              F for full storage or M for same format as matrix
    Col. 2        G if a starting vector(s) (Guess) is supplied. (RHSTYP)
    Col. 3        X if an exact solution vector(s) is supplied.
 Col. 15 - 28  Number of right-hand sides (NRHS)
 Col. 29 - 42  Number of row indices (NRHSIX)
                  (ignored in case of unassembled matrices)

The three character type field on line 3 describes the matrix type.
The following table lists the permitted values for each of the three
characters. As an example of the type field, RSA denotes that the matrix
is real, symmetric, and assembled.

First Character:
 R Real matrix
 C Complex matrix
 P Pattern only (no numerical values supplied)

Second Character:
 S Symmetric
 U Unsymmetric
 H Hermitian
 Z Skew symmetric
 R Rectangular

Third Character:
 A Assembled
 E Elemental matrices (unassembled)
Here is the call graph for this function:

◆ sReadValues()

void sReadValues ( FILE *  fp,
int_t  n,
float *  destination,
int_t  perline,
int_t  persize 
)
static
Here is the caller graph for this function: