Name

HPL_plindx1 Compute local swapping index arrays.

Synopsis

#include <hpl.h>

void HPL_plindx1( HPL_T_panel * PANEL, const int K, const int * IPID, int * IPA, int * LINDXA, int * LINDXAU, int * IPLEN, int * IPMAP, int * IPMAPM1, int * PERMU, int * IWORK );

Description

HPL_plindx1 computes two local arrays LINDXA and LINDXAU containing the local source and final destination position resulting from the application of row interchanges. In addition, this function computes three arrays IPLEN, IPMAP and IPMAPM1 that contain the logarithmic mapping information for the spreading phase.

Arguments

PANEL   (local input/output)          HPL_T_panel *
        On entry,  PANEL  points to the data structure containing the
        panel information.

K       (global input)                const int
        On entry, K specifies the number of entries in IPID.  K is at
        least 2*N, and at most 4*N.

IPID    (global input)                const int *
        On entry,  IPID  is an array of length K. The first K entries
        of that array contain the src and final destination resulting
        from the application of the interchanges.

IPA     (global output)               int *
        On exit,  IPA  specifies  the number of rows that the current
        process row has that either belong to U  or should be swapped
        with remote rows of A.

LINDXA  (global output)               int *
        On entry, LINDXA  is an array of dimension 2*N. On exit, this
        array contains the local indexes of the rows of A I have that
        should be copied into U.

LINDXAU (global output)               int *
        On exit, LINDXAU  is an array of dimension 2*N. On exit, this
        array contains  the local destination  information encoded as
        follows.  If LINDXAU(k) >= 0, row  LINDXA(k)  of A  is  to be
        copied in U at position LINDXAU(k).  Otherwise, row LINDXA(k)
        of A should be locally copied into A(-LINDXAU(k),:).

IPLEN   (global output)               int *
        On entry, IPLEN is an array of dimension NPROW + 1. On  exit,
        this array is such that  IPLEN[i]  is the number of rows of A
        in  the  processes  before  process  IPMAP[i]  after the sort
        with the convention that IPLEN[nprow]  is the total number of
        rows of the panel.  In other words IPLEN[i+1]-IPLEN[i] is the
        local number of rows of A that should be moved to the process
        IPMAP[i]. IPLEN is such that the number of rows of the source
        process  row can be computed as  IPLEN[1] - IPLEN[0], and the
        remaining  entries  of  this  array  are  sorted  so that the
        quantities IPLEN[i+1] - IPLEN[i] are logarithmically sorted.

IPMAP   (global output)               int *
        On entry, IPMAP is an array of dimension NPROW. On exit, this
        array contains  the logarithmic mapping of the processes.  In
        other words, IPMAP[myrow] is the corresponding sorted process
        coordinate.

IPMAPM1 (global output)               int *
        On entry, IPMAPM1  is an array of dimension NPROW.  On  exit,
        this  array  contains  the inverse of the logarithmic mapping
        contained  in  IPMAP:  IPMAPM1[ IPMAP[i] ] = i,  for all i in
        [0.. NPROCS)

PERMU   (global output)               int *
        On entry,  PERMU  is an array of dimension JB. On exit, PERMU
        contains  a sequence of permutations,  that should be applied
        in increasing order to permute in place the row panel U.

IWORK   (workspace)                   int *
        On entry, IWORK is a workarray of dimension 2*JB.

See Also

HPL_pdlaswp00N, HPL_pdlaswp00T, HPL_pdlaswp01N, HPL_pdlaswp01T