Centre Blaise Pascal » HPL sur GPU

.TH HPL_pdmxswp 3 "September 10, 2008" "HPL 2.0" "HPL Library Functions"

.SH NAME

HPL_pdmxswp \- swaps and broacast the pivot row.

.SH SYNOPSIS

\fB\&#include "hpl.h"\fR

\fB\&void\fR

\fB\&HPL_pdmxswp(\fR

\fB\&HPL_T_panel *\fR

\fI\&PANEL\fR,

\fB\&const int\fR

\fI\&M\fR,

\fB\&const int\fR

\fI\&II\fR,

\fB\&const int\fR

\fI\&JJ\fR,

\fB\&double *\fR

\fI\&WORK\fR

\fB\&);\fR

.SH DESCRIPTION

\fB\&HPL_pdmxswp\fR

swaps  and  broadcasts  the  absolute value max row using

bi-directional exchange.  The buffer is partially set by HPL_dlocmax.

Bi-directional  exchange  is  used  to  perform  the  swap::broadcast

operations  at once  for one column in the panel.  This  results in a

lower number of slightly larger  messages than usual.  On P processes

and assuming bi-directional links,  the running time of this function

can be approximated by

   log_2( P ) * ( lat + ( 2 * N0 + 4 ) / bdwth )

where  lat and bdwth are the latency and bandwidth of the network for

double precision real elements.  Communication  only  occurs  in  one

process  column. Mono-directional links  will cause the communication

cost to double.

.SH ARGUMENTS

.TP 8

PANEL   (local input/output)    HPL_T_panel *

On entry,  PANEL  points to the data structure containing the

panel information.

.TP 8

M       (local input)           const int

On entry,  M specifies the local number of rows of the matrix

column on which this function operates.

.TP 8

II      (local input)           const int

On entry, II  specifies the row offset where the column to be

operated on starts with respect to the panel.

.TP 8

JJ      (local input)           const int

On entry, JJ  specifies the column offset where the column to

be operated on starts with respect to the panel.

.TP 8

WORK    (local workspace)       double *

On entry, WORK  is a workarray of size at least 2 * (4+2*N0).

It  is assumed that  HPL_dlocmax  was called  prior  to  this

routine to  initialize  the first four entries of this array.

On exit, the  N0  length max row is stored in WORK[4:4+N0-1];

Note that this is also the  JJth  row  (or column) of L1. The

remaining part is used as a temporary array.

.SH SEE ALSO

.BR HPL_dlocmax \ (3),

.BR HPL_dlocswpN \ (3),

.BR HPL_dlocswpT \ (3),

.BR HPL_pdpancrN \ (3),

.BR HPL_pdpancrT \ (3),

.BR HPL_pdpanllN \ (3),

.BR HPL_pdpanllT \ (3),

.BR HPL_pdpanrlN \ (3),

.BR HPL_pdpanrlT \ (3),

.BR HPL_pdrpancrN \ (3),

.BR HPL_pdrpancrT \ (3),

.BR HPL_pdrpanllN \ (3),

.BR HPL_pdrpanllT \ (3),

.BR HPL_pdrpanrlN \ (3),

.BR HPL_pdrpanrlT \ (3),

.BR HPL_pdfact \ (3).