libsuperlu-dist-dev 5.3.0+dfsg1-1 / usr / include / superlu-dist / superlu_defs.h

/*! \file
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.
*/
/*! @file
 * \brief Definitions which are precision-neutral
 *
 * <pre>
 * -- Distributed SuperLU routine (version 5.2) --
 * Lawrence Berkeley National Lab, Univ. of California Berkeley.
 * November 1, 2007
 *
 * Modified:
 *     Feburary 20, 2008
 *     October 11, 2014
 * </pre>
 */

#ifndef __SUPERLU_DEFS /* allow multiple inclusions */
#define __SUPERLU_DEFS

/*
 * File name:	superlu_defs.h
 * Purpose:     Definitions which are precision-neutral
 */
#ifdef _CRAY
    #include <fortran.h>
#endif

#ifdef _OPENMP
   #include <omp.h>
#endif

#include <mpi.h>
#include <stdlib.h>
#include <stdio.h>
#include <limits.h>
#include <string.h>

/* Following is for vtune */
#if 0
#include <ittnotify.h>
#define USE_VTUNE
#endif

/*************************************************************************
 * Constants
 **************************************************************************/
/*
 * You can support older version of SuperLU_DIST.
 * At compile-time, you can catch the new release as:
 *   #ifdef SUPERLU_DIST_MAIN_VERSION == 5
 *       use the new interface
 *   #else
 *       use the old interface
 *   #endif
 * Versions 4.x and earlier do not include a #define'd version numbers.
 */
#define SUPERLU_DIST_MAJOR_VERSION     5
#define SUPERLU_DIST_MINOR_VERSION     3
#define SUPERLU_DIST_PATCH_VERSION     0
#define SUPERLU_DIST_RELEASE_DATE      "January 28, 2018"

#include "superlu_dist_config.h"
/* Define my integer size int_t */
#ifdef _CRAY
  typedef short int_t;
  /*#undef int   Revert back to int of default size. */
  #define mpi_int_t   MPI_SHORT
#elif defined (_LONGINT)
  typedef long long int int_t;
  #define mpi_int_t   MPI_LONG_LONG_INT
  #define IFMT "%lld"
#else /* Default */
  typedef int int_t;
  #define mpi_int_t   MPI_INT
  #define IFMT "%8d"
#endif

#include "superlu_enum_consts.h"
#include "Cnames.h"
#include "supermatrix.h"
#include "util_dist.h"
#include "psymbfact.h"

#define ISORT     /* NOTE: qsort() has bug on Mac */

/*********************************************************************** 
 * Constants
 ***********************************************************************/
/* 
 * For each block column of L, the index[] array contains both the row 
 * subscripts and the integers describing the size of the blocks.
 * The organization of index[] looks like:
 *
 *     [ BLOCK COLUMN HEADER (size BC_HEADER)
 *           number of blocks 
 *           number of row subscripts, i.e., LDA of nzval[]
 *       BLOCK 0                                        <----
 *           BLOCK DESCRIPTOR (of size LB_DESCRIPTOR)  |
 *               block number (global)                      |
 *               number of full rows in the block           |
 *           actual row subscripts                          |
 *       BLOCK 1                                            | Repeat ...
 *           BLOCK DESCRIPTOR                               | number of blocks
 *               block number (global)                      | 
 *               number of full rows in the block           |
 *           actual row subscripts                          |
 *       .                                                  |
 *       .                                                  |
 *       .                                              <----
 *     ]
 *
 * For each block row of U, the organization of index[] looks like:
 *
 *     [ BLOCK ROW HEADER (of size BR_HEADER)
 *           number of blocks 
 *           number of entries in nzval[]
 *           number of entries in index[]
 *       BLOCK 0                                        <----
 *           BLOCK DESCRIPTOR (of size UB_DESCRIPTOR)  |
 *               block number (global)                      |
 *               number of nonzeros in the block            |
 *           actual fstnz subscripts                        |
 *       BLOCK 1                                            | Repeat ...
 *           BLOCK DESCRIPTOR                               | number of blocks
 *               block number (global)                      |
 *               number of nonzeros in the block            |
 *           actual fstnz subscripts                        |
 *       .                                                  |
 *       .                                                  |
 *       .                                              <----
 *     ]
 *
 */
#define BC_HEADER      2
#define LB_DESCRIPTOR  2
#define BR_HEADER      3
#define UB_DESCRIPTOR  2
#define NBUFFERS       5

/*
 * Communication tags
 */
/* Return the mpi_tag assuming 5 pairs of communications and MPI_TAG_UB >= 5 *
 * for each supernodal column "num", the five communications are:            *
 * 0,1: for sending L to "right"                                             *
 * 2,3: for sending off-diagonal blocks of U "down"                          *
 * 4  : for sending the diagonal blcok down (in pxgstrf2)                    */
#define SLU_MPI_TAG(id,num) ( (5*(num)+id) % tag_ub )

    /* For numeric factorization. */
#if 0
#define NTAGS    10000
#else
#define NTAGS    INT_MAX
#endif
#define UjROW    10
#define UkSUB    11
#define UkVAL    12
#define LkSUB    13
#define LkVAL    14
#define LkkDIAG  15
    /* For triangular solves. */
#define XK_H     2  /* The header preceding each X block. */
#define LSUM_H   2  /* The header preceding each MOD block. */
#define GSUM     20 
#define Xk       21
#define Yk       22
#define LSUM     23

/* 
 * Communication scopes
 */
#define COMM_ALL      100
#define COMM_COLUMN   101
#define COMM_ROW      102

/*
 * Matrix distribution for sparse matrix-vector multiplication
 */
#define SUPER_LINEAR     11
#define SUPER_BLOCK      12

/*
 * No of marker arrays used in the symbolic factorization, each of size n
 */
#define NO_MARKER     3



/***********************************************************************
 * Macros
 ***********************************************************************/
#define IAM(comm)    { int rank; MPI_Comm_rank ( comm, &rank ); rank};
#define MYROW(iam,grid) ( (iam) / grid->npcol )
#define MYCOL(iam,grid) ( (iam) % grid->npcol )
#define BlockNum(i)     ( supno[i] )
#define FstBlockC(bnum) ( xsup[bnum] )
#define SuperSize(bnum) ( xsup[bnum+1]-xsup[bnum] )
#define LBi(bnum,grid)  ( (bnum)/grid->nprow )/* Global to local block rowwise */
#define LBj(bnum,grid)  ( (bnum)/grid->npcol )/* Global to local block columnwise*/
#define PROW(bnum,grid) ( (bnum) % grid->nprow )
#define PCOL(bnum,grid) ( (bnum) % grid->npcol )
#define PNUM(i,j,grid)  ( (i)*grid->npcol + j ) /* Process number at coord(i,j) */
#define CEILING(a,b)    ( ((a)%(b)) ? ((a)/(b) + 1) : ((a)/(b)) )
    /* For triangular solves */
#define RHS_ITERATE(i)                    \
        for (i = 0; i < nrhs; ++i)
#define X_BLK(i)                          \
        ilsum[i] * nrhs + (i+1) * XK_H
#define LSUM_BLK(i)                       \
        ilsum[i] * nrhs + (i+1) * LSUM_H

#define SuperLU_timer_  SuperLU_timer_dist_
#define LOG2(x)   (log10((double) x) / log10(2.0))


#if ( VAMPIR>=1 ) 
#define VT_TRACEON    VT_traceon()
#define VT_TRACEOFF   VT_traceoff()
#else
#define VT_TRACEON 
#define VT_TRACEOFF
#endif

/* Support Windows */
#ifndef SUPERLU_DIST_EXPORT
#if MSVC
#ifdef SUPERLU_DIST_EXPORTS
#define SUPERLU_DIST_EXPORT __declspec(dllexport)
#else
#define SUPERLU_DIST_EXPORT __declspec(dllimport)
#endif /* SUPERLU_DIST_EXPORTS */
#else
#define SUPERLU_DIST_EXPORT
#endif /* MSVC */
#endif /* SUPERLU_DIST_EXPORT */

/***********************************************************************
 * New data types
 ***********************************************************************/

/* 
 *   Define the 2D mapping of matrix blocks to process grid.
 *
 *   Process grid:
 *     Processes are numbered (0 : P-1).
 *     P = Pr x Pc, where Pr, Pc are the number of process rows and columns.
 *     (pr,pc) is the coordinate of IAM; 0 <= pr < Pr, 0 <= pc < Pc.
 *
 *   Matrix blocks:
 *     Matrix is partitioned according to supernode partitions, both
 *     column and row-wise. 
 *     The k-th block columns (rows) contains columns (rows) (s:t), where
 *             s=xsup[k], t=xsup[k+1]-1.
 *     Block A(I,J) contains
 *             rows from (xsup[I]:xsup[I+1]-1) and
 *             columns from (xsup[J]:xsup[J+1]-1)
 *
 *  Mapping of matrix entry (i,j) to matrix block (I,J):
 *     (I,J) = ( supno[i], supno[j] )
 *
 *  Mapping of matrix block (I,J) to process grid (pr,pc):
 *     (pr,pc) = ( MOD(I,NPROW), MOD(J,NPCOL) )
 *  
 *  (xsup[nsupers],supno[n]) are replicated on all processors.
 *
 */

/*-- Communication subgroup */
typedef struct {
    MPI_Comm comm;        /* MPI communicator */
    int Np;               /* number of processes */
    int Iam;              /* my process number */
} superlu_scope_t;

/*-- Process grid definition */
typedef struct {
    MPI_Comm comm;        /* MPI communicator */
    superlu_scope_t rscp; /* process scope in rowwise, horizontal directon */
    superlu_scope_t cscp; /* process scope in columnwise, vertical direction */
    int iam;              /* my process number in this scope */
    int_t nprow;          /* number of process rows */
    int_t npcol;          /* number of process columns */
} gridinfo_t;


/*
 *-- The structures are determined by SYMBFACT and used thereafter.
 *
 * (xsup,supno) describes mapping between supernode and column:
 *	xsup[s] is the leading column of the s-th supernode.
 *      supno[i] is the supernode no to which column i belongs;
 *	e.g.   supno 0 1 2 2 3 3 3 4 4 4 4 4   (n=12)
 *	        xsup 0 1 2 4 7 12
 *	Note: dfs will be performed on supernode rep. relative to the new 
 *	      row pivoting ordering
 *
 * This is allocated during symbolic factorization SYMBFACT.
 */
typedef struct {
    int_t     *xsup;
    int_t     *supno;
} Glu_persist_t;

/*
 *-- The structures are determined by SYMBFACT and used by DDISTRIBUTE.
 * 
 * (xlsub,lsub): lsub[*] contains the compressed subscript of
 *	rectangular supernodes; xlsub[j] points to the starting
 *	location of the j-th column in lsub[*]. Note that xlsub 
 *	is indexed by column.
 *	Storage: original row subscripts
 *
 *      During the course of sparse LU factorization, we also use
 *	(xlsub,lsub) for the purpose of symmetric pruning. For each
 *	supernode {s,s+1,...,t=s+r} with first column s and last
 *	column t, the subscript set
 *		lsub[j], j=xlsub[s], .., xlsub[s+1]-1
 *	is the structure of column s (i.e. structure of this supernode).
 *	It is used for the storage of numerical values.
 *	Furthermore,
 *		lsub[j], j=xlsub[t], .., xlsub[t+1]-1
 *	is the structure of the last column t of this supernode.
 *	It is for the purpose of symmetric pruning. Therefore, the
 *	structural subscripts can be rearranged without making physical
 *	interchanges among the numerical values.
 *
 *	However, if the supernode has only one column, then we
 *	only keep one set of subscripts. For any subscript interchange
 *	performed, similar interchange must be done on the numerical
 *	values.
 *
 *	The last column structures (for pruning) will be removed
 *	after the numercial LU factorization phase.
 *
 * (xusub,usub): xusub[i] points to the starting location of column i
 *      in usub[]. For each U-segment, only the row index of first nonzero
 *      is stored in usub[].
 *
 *      Each U column consists of a number of full segments. Each full segment
 *      starts from a leading nonzero, running up to the supernode (block)
 *      boundary. (Recall that the column-wise supernode partition is also
 *      imposed on the rows.) Because the segment is full, we don't store all
 *      the row indices. Instead, only the leading nonzero index is stored.
 *      The rest can be found together with xsup/supno pair.
 *      For example, 
 *          usub[xsub[j+1]] - usub[xsub[j]] = number of segments in column j.
 *          for any i in usub[], 
 *              supno[i]         = block number in which i belongs to
 *  	        xsup[supno[i]+1] = first row of the next block
 *              The nonzeros of this segment are: 
 *                  i, i+1 ... xsup[supno[i]+1]-1 (only i is stored in usub[])
 *
 */
typedef struct {
    int_t     *lsub;     /* compressed L subscripts */
    int_t     *xlsub;
    int_t     *usub;     /* compressed U subscripts */
    int_t     *xusub;
    int_t     nzlmax;    /* current max size of lsub */
    int_t     nzumax;    /*    "    "    "      usub */
    LU_space_t MemModel; /* 0 - system malloc'd; 1 - user provided */
    int_t     *llvl;     /* keep track of level in L for level-based ILU */
    int_t     *ulvl;     /* keep track of level in U for level-based ILU */
} Glu_freeable_t;


/* 
 *-- The structure used to store matrix A of the linear system and
 *   several vectors describing the transformations done to matrix A.
 *
 * A      (SuperMatrix*)
 *        Matrix A in A*X=B, of dimension (A->nrow, A->ncol).
 *        The number of linear equations is A->nrow. The type of A can be:
 *        Stype = SLU_NC; Dtype = SLU_D; Mtype = SLU_GE.
 *         
 * DiagScale  (DiagScale_t)
 *        Specifies the form of equilibration that was done.
 *        = NOEQUIL: No equilibration.
 *        = ROW:  Row equilibration, i.e., A was premultiplied by diag(R).
 *        = COL:  Column equilibration, i.e., A was postmultiplied by diag(C).
 *        = BOTH: Both row and column equilibration, i.e., A was replaced 
 *                 by diag(R)*A*diag(C).
 *
 * R      double*, dimension (A->nrow)
 *        The row scale factors for A.
 *        If DiagScale = ROW or BOTH, A is multiplied on the left by diag(R).
 *        If DiagScale = NOEQUIL or COL, R is not defined.
 *
 * C      double*, dimension (A->ncol)
 *        The column scale factors for A.
 *        If DiagScale = COL or BOTH, A is multiplied on the right by diag(C).
 *        If DiagScale = NOEQUIL or ROW, C is not defined.
 *         
 * perm_r (int*) dimension (A->nrow)
 *        Row permutation vector which defines the permutation matrix Pr,
 *        perm_r[i] = j means row i of A is in position j in Pr*A.
 *
 * perm_c (int*) dimension (A->ncol)
 *	  Column permutation vector, which defines the 
 *        permutation matrix Pc; perm_c[i] = j means column i of A is 
 *        in position j in A*Pc.
 *
 */
typedef struct {
    DiagScale_t DiagScale;
    double *R;
    double *C; 
    int_t  *perm_r;
    int_t  *perm_c;
} ScalePermstruct_t;

/*-- Data structure for redistribution of B and X --*/
typedef struct {
    int  *B_to_X_SendCnt;
    int  *X_to_B_SendCnt;
    int  *ptr_to_ibuf, *ptr_to_dbuf;

    /* the following are needed in the hybrid solver PDSLin */	
    int *X_to_B_iSendCnt;
    int *X_to_B_vSendCnt;
    int    *disp_ibuf;
    int_t  *send_ibuf;
    void   *send_dbuf;

    int_t  x2b, b2x;
    int_t  *send_ibuf2;
    int_t  *recv_ibuf2;
    void   *send_dbuf2;
    void   *recv_dbuf2;
} pxgstrs_comm_t;

/* 
 *-- This contains the options used to control the solution process.
 *
 * Fact   (fact_t)
 *        Specifies whether or not the factored form of the matrix
 *        A is supplied on entry, and if not, how the matrix A should
 *        be factorizaed.
 *        = DOFACT: The matrix A will be factorized from scratch, and the
 *             factors will be stored in L and U.
 *        = SamePattern: The matrix A will be factorized assuming
 *             that a factorization of a matrix with the same sparsity
 *             pattern was performed prior to this one. Therefore, this
 *             factorization will reuse column permutation vector 
 *             ScalePermstruct->perm_c and the column elimination tree
 *             LUstruct->etree.
 *        = SamePattern_SameRowPerm: The matrix A will be factorized
 *             assuming that a factorization of a matrix with the same
 *             sparsity	pattern and similar numerical values was performed
 *             prior to this one. Therefore, this factorization will reuse
 *             both row and column scaling factors R and C, both row and
 *             column permutation vectors perm_r and perm_c, and the
 *             data structure set up from the previous symbolic factorization.
 *        = FACTORED: On entry, L, U, perm_r and perm_c contain the 
 *              factored form of A. If DiagScale is not NOEQUIL, the matrix
 *              A has been equilibrated with scaling factors R and C.
 *
 * Equil  (yes_no_t)
 *        Specifies whether to equilibrate the system (scale A's row and
 *        columns to have unit norm).
 *
 * ColPerm (colperm_t)
 *        Specifies what type of column permutation to use to reduce fill.
 *        = NATURAL: use the natural ordering 
 *        = MMD_ATA: use minimum degree ordering on structure of A'*A
 *        = MMD_AT_PLUS_A: use minimum degree ordering on structure of A'+A
 *        = COLAMD: use approximate minimum degree column ordering
 *        = MY_PERMC: use the ordering specified by the user
 *         
 * Trans  (trans_t)
 *        Specifies the form of the system of equations:
 *        = NOTRANS: A * X = B        (No transpose)
 *        = TRANS:   A**T * X = B     (Transpose)
 *        = CONJ:    A**H * X = B     (Transpose)
 *
 * IterRefine (IterRefine_t)
 *        Specifies whether to perform iterative refinement.
 *        = NO: no iterative refinement
 *        = SINGLE: perform iterative refinement in single precision
 *        = DOUBLE: perform iterative refinement in double precision
 *        = EXTRA: perform iterative refinement in extra precision
 *
 * DiagPivotThresh (double, in [0.0, 1.0]) (only for serial SuperLU)
 *        Specifies the threshold used for a diagonal entry to be an
 *        acceptable pivot.
 *
 * SymmetricMode (yest_no_t) (only for serial SuperLU)
 *        Specifies whether to use symmetric mode. Symmetric mode gives 
 *        preference to diagonal pivots, and uses an (A'+A)-based column
 *        permutation algorithm.
 *
 * PivotGrowth (yes_no_t)  (only for serial SuperLU)
 *        Specifies whether to compute the reciprocal pivot growth.
 *
 * ConditionNumber (ues_no_t) (only for serial SuperLU)
 *        Specifies whether to compute the reciprocal condition number.
 *
 * RowPerm (rowperm_t) (only for SuperLU_DIST or ILU in serial SuperLU)
 *        Specifies whether to permute rows of the original matrix.
 *        = NO: not to permute the rows
 *        = LargeDiag: make the diagonal large relative to the off-diagonal
 *        = MY_PERMR: use the permutation given by the user
 *
 * ILU_DropRule (int)  (only for serial SuperLU)
 *        Specifies the dropping rule:
 *	  = DROP_BASIC:   Basic dropping rule, supernodal based ILUTP(tau).
 *	  = DROP_PROWS:   Supernodal based ILUTP(p,tau), p = gamma * nnz(A)/n.
 *	  = DROP_COLUMN:  Variant of ILUTP(p,tau), for j-th column,
 *			      p = gamma * nnz(A(:,j)).
 *	  = DROP_AREA:    Variation of ILUTP, for j-th column, use
 *			      nnz(F(:,1:j)) / nnz(A(:,1:j)) to control memory.
 *	  = DROP_DYNAMIC: Modify the threshold tau during factorizaion:
 *			  If nnz(L(:,1:j)) / nnz(A(:,1:j)) > gamma
 *				  tau_L(j) := MIN(tau_0, tau_L(j-1) * 2);
 *			  Otherwise
 *				  tau_L(j) := MAX(tau_0, tau_L(j-1) / 2);
 *			  tau_U(j) uses the similar rule.
 *			  NOTE: the thresholds used by L and U are separate.
 *	  = DROP_INTERP:  Compute the second dropping threshold by
 *	                  interpolation instead of sorting (default).
 *  		          In this case, the actual fill ratio is not
 *			  guaranteed to be smaller than gamma.
 *   	  Note: DROP_PROWS, DROP_COLUMN and DROP_AREA are mutually exclusive.
 *	  ( Default: DROP_BASIC | DROP_AREA )
 *
 * ILU_DropTol (double) (only for serial SuperLU)
 *        numerical threshold for dropping.
 *
 * ILU_FillFactor (double) (only for serial SuperLU)
 *        Gamma in the secondary dropping.
 *
 * ILU_Norm (norm_t)  (only for serial SuperLU)
 *        Specify which norm to use to measure the row size in a
 *        supernode: infinity-norm, 1-norm, or 2-norm.
 *
 * ILU_FillTol (double) (only for serial SuperLU)
 *        numerical threshold for zero pivot perturbation.
 *
 * ILU_MILU (milu_t)  (only for serial SuperLU)
 *        Specifies which version of MILU to use.
 *
 * ILU_MILU_Dim (double) 
 *        Dimension of the PDE if available.
 *
 * ReplaceTinyPivot (yes_no_t) (only for SuperLU_DIST)
 *        Specifies whether to replace the tiny diagonals by
 *        sqrt(epsilon)*||A|| during LU factorization.
 *
 * SolveInitialized (yes_no_t) (only for SuperLU_DIST)
 *        Specifies whether the initialization has been performed to the
 *        triangular solve.
 *
 * RefineInitialized (yes_no_t) (only for SuperLU_DIST)
 *        Specifies whether the initialization has been performed to the
 *        sparse matrix-vector multiplication routine needed in iterative
 *        refinement.
 *
 * num_lookaheads (int) (only for SuperLU_DIST)
 *        Specifies the number of levels in the look-ahead factorization
 *
 * lookahead_etree (yes_no_t) (only for SuperLU_DIST)
 *        Specifies whether to use the elimination tree computed from the 
 *        serial symbolic factorization to perform scheduling.
 *
 * SymPattern (yes_no_t) (only for SuperLU_DIST)
 *        Gives the scheduling algorithm a hint whether the matrix
 *        would have symmetric pattern.
 *
 */
typedef struct {
    fact_t        Fact;
    yes_no_t      Equil;
    colperm_t     ColPerm;
    trans_t       Trans;
    IterRefine_t  IterRefine;
    double        DiagPivotThresh;
    yes_no_t      SymmetricMode;
    yes_no_t      PivotGrowth;
    yes_no_t      ConditionNumber;
    rowperm_t     RowPerm;
    int 	  ILU_DropRule;
    double	  ILU_DropTol;    /* threshold for dropping */
    double	  ILU_FillFactor; /* gamma in the secondary dropping */
    norm_t	  ILU_Norm;       /* infinity-norm, 1-norm, or 2-norm */
    double	  ILU_FillTol;    /* threshold for zero pivot perturbation */
    milu_t	  ILU_MILU;
    double	  ILU_MILU_Dim;   /* Dimension of PDE (if available) */
    yes_no_t      ParSymbFact;
    yes_no_t      ReplaceTinyPivot; /* used in SuperLU_DIST */
    yes_no_t      SolveInitialized;
    yes_no_t      RefineInitialized;
    yes_no_t      PrintStat;
    int           nnzL, nnzU;      /* used to store nnzs for now       */
    int           num_lookaheads;  /* num of levels in look-ahead      */
    yes_no_t      lookahead_etree; /* use etree computed from the
				      serial symbolic factorization */
    yes_no_t      SymPattern;      /* symmetric factorization          */
} superlu_dist_options_t;

typedef struct {
    float for_lu;
    float total;
    int_t expansions;
    long long int nnzL, nnzU;
} superlu_dist_mem_usage_t;

/* 
 *-- The new structures added in the hybrid CUDA + OpenMP + MPI code.
 */
typedef struct {
    int_t rukp;
    int_t iukp;
    int_t jb;
    int_t full_u_cols;

} Ublock_info_t;

typedef struct {
    int_t lptr;
    int_t ib;
    int_t FullRow;
} Remain_info_t;

typedef struct
{
    int id, key;
    void *next;
} etree_node;

struct superlu_pair
{
    int ind;
    int val;
};

/**--------**/


/***********************************************************************
 * Function prototypes
 ***********************************************************************/

#ifdef __cplusplus
extern "C" {
#endif

extern void   set_default_options_dist(superlu_dist_options_t *);
extern void   superlu_gridinit(MPI_Comm, int_t, int_t, gridinfo_t *);
extern void   superlu_gridmap(MPI_Comm, int_t, int_t, int_t [], int_t,
			      gridinfo_t *);
extern void   superlu_gridexit(gridinfo_t *);
extern void   print_options_dist(superlu_dist_options_t *);
extern void   print_sp_ienv_dist(superlu_dist_options_t *);
extern void   Destroy_CompCol_Matrix_dist(SuperMatrix *);
extern void   Destroy_SuperNode_Matrix_dist(SuperMatrix *);
extern void   Destroy_SuperMatrix_Store_dist(SuperMatrix *);
extern void   Destroy_CompCol_Permuted_dist(SuperMatrix *);
extern void   Destroy_CompRowLoc_Matrix_dist(SuperMatrix *);
extern void   Destroy_CompRow_Matrix_dist(SuperMatrix *);
extern void   sp_colorder (superlu_dist_options_t*, SuperMatrix*, int_t*, int_t*,
			   SuperMatrix*);
extern int    sp_symetree_dist(int_t *, int_t *, int_t *, int_t, int_t *);
extern int    sp_coletree_dist (int_t *, int_t *, int_t *, int_t, int_t, int_t *);
extern void   get_perm_c_dist(int_t, int_t, SuperMatrix *, int_t *);
extern void   at_plus_a_dist(const int_t, const int_t, int_t *, int_t *,
			     int_t *, int_t **, int_t **);
extern int    genmmd_dist_(int_t *, int_t *, int_t *a, 
			   int_t *, int_t *, int_t *, int_t *, 
			   int_t *, int_t *, int_t *, int_t *, int_t *);
extern void  bcast_tree(void *, int, MPI_Datatype, int, int,
			gridinfo_t *, int, int *);
extern int_t symbfact(superlu_dist_options_t *, int, SuperMatrix *, int_t *,
                      int_t *, Glu_persist_t *, Glu_freeable_t *);
extern int_t symbfact_SubInit(fact_t, void *, int_t, int_t, int_t, int_t,
			      Glu_persist_t *, Glu_freeable_t *);
extern int_t symbfact_SubXpand(int_t, int_t, int_t, MemType, int_t *,
			       Glu_freeable_t *);
extern int_t symbfact_SubFree(Glu_freeable_t *);
extern void    countnz_dist (const int_t, int_t *, 
			     long long int *, long long int *,
			     Glu_persist_t *, Glu_freeable_t *);
extern long long int fixupL_dist (const int_t, const int_t *, Glu_persist_t *,
				  Glu_freeable_t *);
extern int_t   *TreePostorder_dist (int_t, int_t *);
extern float   smach_dist(char *);
extern double  dmach_dist(char *);
extern void    *superlu_malloc_dist (size_t);
extern void    superlu_free_dist (void*);
extern int_t   *intMalloc_dist (int_t);
extern int_t   *intCalloc_dist (int_t);
extern int_t   mc64id_dist(int_t *);
extern void  arrive_at_ublock (int_t, int_t *, int_t *, int_t *,
			       int_t *, int_t *, int_t, int_t, 
			       int_t *, int_t *, int_t *, gridinfo_t *);
extern int_t estimate_bigu_size (int_t, int_t, int_t **, Glu_persist_t *,
				 gridinfo_t *, int_t *);

/* Auxiliary routines */
extern double SuperLU_timer_ ();
extern void   superlu_abort_and_exit_dist(char *);
extern int_t  sp_ienv_dist (int_t);
extern void   ifill_dist (int_t *, int_t, int_t);
extern void   super_stats_dist (int_t, int_t *);
extern void   ScalePermstructInit(const int_t, const int_t, 
				   ScalePermstruct_t *);
extern void   ScalePermstructFree(ScalePermstruct_t *);
extern void  get_diag_procs(int_t, Glu_persist_t *, gridinfo_t *, int_t *,
			    int_t **, int_t **);
extern int_t QuerySpace_dist(int_t, int_t, Glu_freeable_t *, superlu_dist_mem_usage_t *);
extern int   xerr_dist (char *, int *);
extern void  pxerr_dist (char *, gridinfo_t *, int_t);
extern void  PStatInit(SuperLUStat_t *);
extern void  PStatFree(SuperLUStat_t *);
extern void  PStatPrint(superlu_dist_options_t *, SuperLUStat_t *, gridinfo_t *);
extern void  log_memory(long long, SuperLUStat_t *);
extern void  print_memorylog(SuperLUStat_t *, char *);
extern int   superlu_dist_GetVersionNumber(int *, int *, int *);

/* Prototypes for parallel symbolic factorization */
extern float symbfact_dist
(int,  int, SuperMatrix *, int_t *, int_t *,  int_t *, int_t *,
 Pslu_freeable_t *, MPI_Comm *, MPI_Comm *,  superlu_dist_mem_usage_t *);

/* Get the column permutation using parmetis */
extern float get_perm_c_parmetis 
(SuperMatrix *, int_t *, int_t *, int, int, 
 int_t **, int_t **, gridinfo_t *, MPI_Comm *);

/* Auxiliary routines for memory expansions used during
   the parallel symbolic factorization routine */

extern int_t psymbfact_LUXpandMem
(int_t, int_t, int_t, int_t, int_t, int_t, int_t, int_t, 
 Pslu_freeable_t *, Llu_symbfact_t *,  vtcsInfo_symbfact_t *, psymbfact_stat_t *);

extern int_t psymbfact_LUXpand
(int_t, int_t, int_t, int_t, int_t *, int_t, int_t, int_t, int_t, 
 Pslu_freeable_t *, Llu_symbfact_t *,  vtcsInfo_symbfact_t *, psymbfact_stat_t *);

extern int_t psymbfact_LUXpand_RL
(int_t, int_t, int_t, int_t, int_t, int_t,
 Pslu_freeable_t *, Llu_symbfact_t *, vtcsInfo_symbfact_t *, psymbfact_stat_t *);

extern int_t psymbfact_prLUXpand
(int_t,  int_t, int, Llu_symbfact_t *, psymbfact_stat_t *);

#ifdef GPU_ACC   /* GPU related */
extern void gemm_division_cpu_gpu (int *, int *, int *, int,
				   int, int, int *, int);
extern int_t get_cublas_nb ();
extern int_t get_num_cuda_streams ();
#endif

extern int get_thread_per_process();
extern int_t get_max_buffer_size ();
extern int_t get_min (int_t *, int_t);
extern int compare_pair (const void *, const void *);
extern int_t static_partition (struct superlu_pair *, int_t, int_t *, int_t,
			       int_t *, int_t *, int);

/* Routines for debugging */
extern void  print_panel_seg_dist(int_t, int_t, int_t, int_t, int_t *, int_t *);
extern void  check_repfnz_dist(int_t, int_t, int_t, int_t *);
extern int_t CheckZeroDiagonal(int_t, int_t *, int_t *, int_t *);
extern void  PrintDouble5(char *, int_t, double *);
extern void  PrintInt10(char *, int_t, int_t *);
extern void  PrintInt32(char *, int, int *);
extern int   file_PrintInt10(FILE *, char *, int_t, int_t *);
extern int   file_PrintInt32(FILE *, char *, int, int *);
extern int   file_PrintLong10(FILE *, char *, int_t, int_t *);

#ifdef __cplusplus
  }
#endif

#endif /* __SUPERLU_DEFS */