root / www / HPL_plindx1.html
Historique | Voir | Annoter | Télécharger (4,48 ko)
| 1 | 1 | equemene | <HTML>
|
|---|---|---|---|
| 2 | 1 | equemene | <HEAD>
|
| 3 | 1 | equemene | <TITLE>HPL_plindx1 HPL 2.0 Library Functions September 10, 2008</TITLE> |
| 4 | 1 | equemene | </HEAD>
|
| 5 | 1 | equemene | |
| 6 | 1 | equemene | <BODY BGCOLOR="WHITE" TEXT = "#000000" LINK = "#0000ff" VLINK = "#000099" |
| 7 | 1 | equemene | ALINK = "#ffff00"> |
| 8 | 1 | equemene | |
| 9 | 1 | equemene | <H1>Name</H1> |
| 10 | 1 | equemene | <B>HPL_plindx1</B> Compute local swapping index arrays. |
| 11 | 1 | equemene | |
| 12 | 1 | equemene | <H1>Synopsis</H1> |
| 13 | 1 | equemene | <CODE>#include "hpl.h"</CODE><BR><BR> |
| 14 | 1 | equemene | <CODE>void</CODE> |
| 15 | 1 | equemene | <CODE>HPL_plindx1(</CODE> |
| 16 | 1 | equemene | <CODE>HPL_T_panel *</CODE> |
| 17 | 1 | equemene | <CODE>PANEL</CODE>, |
| 18 | 1 | equemene | <CODE>const int</CODE> |
| 19 | 1 | equemene | <CODE>K</CODE>, |
| 20 | 1 | equemene | <CODE>const int *</CODE> |
| 21 | 1 | equemene | <CODE>IPID</CODE>, |
| 22 | 1 | equemene | <CODE>int *</CODE> |
| 23 | 1 | equemene | <CODE>IPA</CODE>, |
| 24 | 1 | equemene | <CODE>int *</CODE> |
| 25 | 1 | equemene | <CODE>LINDXA</CODE>, |
| 26 | 1 | equemene | <CODE>int *</CODE> |
| 27 | 1 | equemene | <CODE>LINDXAU</CODE>, |
| 28 | 1 | equemene | <CODE>int *</CODE> |
| 29 | 1 | equemene | <CODE>IPLEN</CODE>, |
| 30 | 1 | equemene | <CODE>int *</CODE> |
| 31 | 1 | equemene | <CODE>IPMAP</CODE>, |
| 32 | 1 | equemene | <CODE>int *</CODE> |
| 33 | 1 | equemene | <CODE>IPMAPM1</CODE>, |
| 34 | 1 | equemene | <CODE>int *</CODE> |
| 35 | 1 | equemene | <CODE>PERMU</CODE>, |
| 36 | 1 | equemene | <CODE>int *</CODE> |
| 37 | 1 | equemene | <CODE>IWORK</CODE> |
| 38 | 1 | equemene | <CODE>);</CODE> |
| 39 | 1 | equemene | |
| 40 | 1 | equemene | <H1>Description</H1> |
| 41 | 1 | equemene | <B>HPL_plindx1</B> |
| 42 | 1 | equemene | computes two local arrays LINDXA and LINDXAU containing |
| 43 | 1 | equemene | the local source and final destination position resulting from the |
| 44 | 1 | equemene | application of row interchanges. In addition, this function computes |
| 45 | 1 | equemene | three arrays IPLEN, IPMAP and IPMAPM1 that contain the logarithmic |
| 46 | 1 | equemene | mapping information for the spreading phase. |
| 47 | 1 | equemene | |
| 48 | 1 | equemene | <H1>Arguments</H1> |
| 49 | 1 | equemene | <PRE>
|
| 50 | 1 | equemene | PANEL (local input/output) HPL_T_panel * |
| 51 | 1 | equemene | On entry, PANEL points to the data structure containing the |
| 52 | 1 | equemene | panel information. |
| 53 | 1 | equemene | </PRE>
|
| 54 | 1 | equemene | <PRE>
|
| 55 | 1 | equemene | K (global input) const int |
| 56 | 1 | equemene | On entry, K specifies the number of entries in IPID. K is at |
| 57 | 1 | equemene | least 2*N, and at most 4*N. |
| 58 | 1 | equemene | </PRE>
|
| 59 | 1 | equemene | <PRE>
|
| 60 | 1 | equemene | IPID (global input) const int * |
| 61 | 1 | equemene | On entry, IPID is an array of length K. The first K entries |
| 62 | 1 | equemene | of that array contain the src and final destination resulting |
| 63 | 1 | equemene | from the application of the interchanges. |
| 64 | 1 | equemene | </PRE>
|
| 65 | 1 | equemene | <PRE>
|
| 66 | 1 | equemene | IPA (global output) int * |
| 67 | 1 | equemene | On exit, IPA specifies the number of rows that the current |
| 68 | 1 | equemene | process row has that either belong to U or should be swapped |
| 69 | 1 | equemene | with remote rows of A. |
| 70 | 1 | equemene | </PRE>
|
| 71 | 1 | equemene | <PRE>
|
| 72 | 1 | equemene | LINDXA (global output) int * |
| 73 | 1 | equemene | On entry, LINDXA is an array of dimension 2*N. On exit, this |
| 74 | 1 | equemene | array contains the local indexes of the rows of A I have that |
| 75 | 1 | equemene | should be copied into U. |
| 76 | 1 | equemene | </PRE>
|
| 77 | 1 | equemene | <PRE>
|
| 78 | 1 | equemene | LINDXAU (global output) int * |
| 79 | 1 | equemene | On exit, LINDXAU is an array of dimension 2*N. On exit, this |
| 80 | 1 | equemene | array contains the local destination information encoded as |
| 81 | 1 | equemene | follows. If LINDXAU(k) >= 0, row LINDXA(k) of A is to be
|
| 82 | 1 | equemene | copied in U at position LINDXAU(k). Otherwise, row LINDXA(k) |
| 83 | 1 | equemene | of A should be locally copied into A(-LINDXAU(k),:). |
| 84 | 1 | equemene | </PRE>
|
| 85 | 1 | equemene | <PRE>
|
| 86 | 1 | equemene | IPLEN (global output) int * |
| 87 | 1 | equemene | On entry, IPLEN is an array of dimension NPROW + 1. On exit, |
| 88 | 1 | equemene | this array is such that IPLEN[i] is the number of rows of A |
| 89 | 1 | equemene | in the processes before process IPMAP[i] after the sort |
| 90 | 1 | equemene | with the convention that IPLEN[nprow] is the total number of |
| 91 | 1 | equemene | rows of the panel. In other words IPLEN[i+1]-IPLEN[i] is the |
| 92 | 1 | equemene | local number of rows of A that should be moved to the process |
| 93 | 1 | equemene | IPMAP[i]. IPLEN is such that the number of rows of the source |
| 94 | 1 | equemene | process row can be computed as IPLEN[1] - IPLEN[0], and the |
| 95 | 1 | equemene | remaining entries of this array are sorted so that the |
| 96 | 1 | equemene | quantities IPLEN[i+1] - IPLEN[i] are logarithmically sorted. |
| 97 | 1 | equemene | </PRE>
|
| 98 | 1 | equemene | <PRE>
|
| 99 | 1 | equemene | IPMAP (global output) int * |
| 100 | 1 | equemene | On entry, IPMAP is an array of dimension NPROW. On exit, this |
| 101 | 1 | equemene | array contains the logarithmic mapping of the processes. In |
| 102 | 1 | equemene | other words, IPMAP[myrow] is the corresponding sorted process |
| 103 | 1 | equemene | coordinate. |
| 104 | 1 | equemene | </PRE>
|
| 105 | 1 | equemene | <PRE>
|
| 106 | 1 | equemene | IPMAPM1 (global output) int * |
| 107 | 1 | equemene | On entry, IPMAPM1 is an array of dimension NPROW. On exit, |
| 108 | 1 | equemene | this array contains the inverse of the logarithmic mapping |
| 109 | 1 | equemene | contained in IPMAP: IPMAPM1[ IPMAP[i] ] = i, for all i in |
| 110 | 1 | equemene | [0.. NPROCS) |
| 111 | 1 | equemene | </PRE>
|
| 112 | 1 | equemene | <PRE>
|
| 113 | 1 | equemene | PERMU (global output) int * |
| 114 | 1 | equemene | On entry, PERMU is an array of dimension JB. On exit, PERMU |
| 115 | 1 | equemene | contains a sequence of permutations, that should be applied |
| 116 | 1 | equemene | in increasing order to permute in place the row panel U. |
| 117 | 1 | equemene | </PRE>
|
| 118 | 1 | equemene | <PRE>
|
| 119 | 1 | equemene | IWORK (workspace) int * |
| 120 | 1 | equemene | On entry, IWORK is a workarray of dimension 2*JB. |
| 121 | 1 | equemene | </PRE>
|
| 122 | 1 | equemene | |
| 123 | 1 | equemene | <H1>See Also</H1> |
| 124 | 1 | equemene | <A HREF="HPL_pdlaswp00N.html">HPL_pdlaswp00N</A>, |
| 125 | 1 | equemene | <A HREF="HPL_pdlaswp00T.html">HPL_pdlaswp00T</A>, |
| 126 | 1 | equemene | <A HREF="HPL_pdlaswp01N.html">HPL_pdlaswp01N</A>, |
| 127 | 1 | equemene | <A HREF="HPL_pdlaswp01T.html">HPL_pdlaswp01T</A>. |
| 128 | 1 | equemene | |
| 129 | 1 | equemene | </BODY>
|
| 130 | 1 | equemene | </HTML> |