root / man / man3 / HPL_plindx10.3 @ 1
Historique | Voir | Annoter | Télécharger (2,27 ko)
1 |
.TH HPL_plindx10 3 "September 10, 2008" "HPL 2.0" "HPL Library Functions" |
---|---|
2 |
.SH NAME |
3 |
HPL_plindx10 \- Compute the logarithmic maps for the spreading. |
4 |
.SH SYNOPSIS |
5 |
\fB\&#include "hpl.h"\fR |
6 |
|
7 |
\fB\&void\fR |
8 |
\fB\&HPL_plindx10(\fR |
9 |
\fB\&HPL_T_panel *\fR |
10 |
\fI\&PANEL\fR, |
11 |
\fB\&const int\fR |
12 |
\fI\&K\fR, |
13 |
\fB\&const int *\fR |
14 |
\fI\&IPID\fR, |
15 |
\fB\&int *\fR |
16 |
\fI\&IPLEN\fR, |
17 |
\fB\&int *\fR |
18 |
\fI\&IPMAP\fR, |
19 |
\fB\&int *\fR |
20 |
\fI\&IPMAPM1\fR |
21 |
\fB\&);\fR |
22 |
.SH DESCRIPTION |
23 |
\fB\&HPL_plindx10\fR |
24 |
computes three arrays IPLEN, IPMAP and IPMAPM1 that |
25 |
contain the logarithmic mapping information for the spreading phase. |
26 |
.SH ARGUMENTS |
27 |
.TP 8 |
28 |
PANEL (local input/output) HPL_T_panel * |
29 |
On entry, PANEL points to the data structure containing the |
30 |
panel information. |
31 |
.TP 8 |
32 |
K (global input) const int |
33 |
On entry, K specifies the number of entries in IPID. K is at |
34 |
least 2*N, and at most 4*N. |
35 |
.TP 8 |
36 |
IPID (global input) const int * |
37 |
On entry, IPID is an array of length K. The first K entries |
38 |
of that array contain the src and final destination resulting |
39 |
from the application of the interchanges. |
40 |
.TP 8 |
41 |
IPLEN (global output) int * |
42 |
On entry, IPLEN is an array of dimension NPROW + 1. On exit, |
43 |
this array is such that IPLEN[i] is the number of rows of A |
44 |
in the processes before process IMAP[i] after the sort, with |
45 |
the convention that IPLEN[nprow] is the total number of rows. |
46 |
In other words, IPLEN[i+1] - IPLEN[i] is the local number of |
47 |
rows of A that should be moved for each process. IPLEN is |
48 |
such that the number of rows of the source process row can be |
49 |
computed as IPLEN[1] - IPLEN[0], and the remaining entries of |
50 |
this array are sorted so that the quantities IPLEN[i+1] - |
51 |
IPLEN[i] are logarithmically sorted. |
52 |
.TP 8 |
53 |
IPMAP (global output) int * |
54 |
On entry, IPMAP is an array of dimension NPROW. On exit, this |
55 |
array contains the logarithmic mapping of the processes. In |
56 |
other words, IPMAP[myrow] is the corresponding sorted process |
57 |
coordinate. |
58 |
.TP 8 |
59 |
IPMAPM1 (global output) int * |
60 |
On entry, IPMAPM1 is an array of dimension NPROW. On exit, |
61 |
this array contains the inverse of the logarithmic mapping |
62 |
contained in IPMAP: IPMAPM1[ IPMAP[i] ] = i, for all i in |
63 |
[0.. NPROW) |
64 |
.SH SEE ALSO |
65 |
.BR HPL_pdlaswp00N \ (3), |
66 |
.BR HPL_pdlaswp00T \ (3), |
67 |
.BR HPL_pdlaswp01N \ (3), |
68 |
.BR HPL_pdlaswp01T \ (3). |