root / src / pgesv / HPL_pdlaswp01T.c
Historique | Voir | Annoter | Télécharger (8,81 ko)
1 |
/*
|
---|---|
2 |
* -- High Performance Computing Linpack Benchmark (HPL)
|
3 |
* HPL - 2.0 - September 10, 2008
|
4 |
* Antoine P. Petitet
|
5 |
* University of Tennessee, Knoxville
|
6 |
* Innovative Computing Laboratory
|
7 |
* (C) Copyright 2000-2008 All Rights Reserved
|
8 |
*
|
9 |
* -- Copyright notice and Licensing terms:
|
10 |
*
|
11 |
* Redistribution and use in source and binary forms, with or without
|
12 |
* modification, are permitted provided that the following conditions
|
13 |
* are met:
|
14 |
*
|
15 |
* 1. Redistributions of source code must retain the above copyright
|
16 |
* notice, this list of conditions and the following disclaimer.
|
17 |
*
|
18 |
* 2. Redistributions in binary form must reproduce the above copyright
|
19 |
* notice, this list of conditions, and the following disclaimer in the
|
20 |
* documentation and/or other materials provided with the distribution.
|
21 |
*
|
22 |
* 3. All advertising materials mentioning features or use of this
|
23 |
* software must display the following acknowledgement:
|
24 |
* This product includes software developed at the University of
|
25 |
* Tennessee, Knoxville, Innovative Computing Laboratory.
|
26 |
*
|
27 |
* 4. The name of the University, the name of the Laboratory, or the
|
28 |
* names of its contributors may not be used to endorse or promote
|
29 |
* products derived from this software without specific written
|
30 |
* permission.
|
31 |
*
|
32 |
* -- Disclaimer:
|
33 |
*
|
34 |
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
|
35 |
* ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
|
36 |
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
|
37 |
* A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE UNIVERSITY
|
38 |
* OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
|
39 |
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
|
40 |
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
|
41 |
* DATA OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
|
42 |
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
|
43 |
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
|
44 |
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
45 |
* ---------------------------------------------------------------------
|
46 |
*/
|
47 |
/*
|
48 |
* Include files
|
49 |
*/
|
50 |
#include "hpl.h" |
51 |
|
52 |
#ifdef STDC_HEADERS
|
53 |
void HPL_pdlaswp01T
|
54 |
( |
55 |
HPL_T_panel * PBCST, |
56 |
int * IFLAG,
|
57 |
HPL_T_panel * PANEL, |
58 |
const int NN |
59 |
) |
60 |
#else
|
61 |
void HPL_pdlaswp01T
|
62 |
( PBCST, IFLAG, PANEL, NN ) |
63 |
HPL_T_panel * PBCST; |
64 |
int * IFLAG;
|
65 |
HPL_T_panel * PANEL; |
66 |
const int NN; |
67 |
#endif
|
68 |
{ |
69 |
/*
|
70 |
* Purpose
|
71 |
* =======
|
72 |
*
|
73 |
* HPL_pdlaswp01T applies the NB row interchanges to NN columns of the
|
74 |
* trailing submatrix and broadcast a column panel.
|
75 |
*
|
76 |
* A "Spread then roll" algorithm performs the swap :: broadcast of the
|
77 |
* row panel U at once, resulting in a minimal communication volume and
|
78 |
* a "very good" use of the connectivity if available. With P process
|
79 |
* rows and assuming bi-directional links, the running time of this
|
80 |
* function can be approximated by:
|
81 |
*
|
82 |
* (log_2(P)+(P-1)) * lat + K * NB * LocQ(N) / bdwth
|
83 |
*
|
84 |
* where NB is the number of rows of the row panel U, N is the global
|
85 |
* number of columns being updated, lat and bdwth are the latency and
|
86 |
* bandwidth of the network for double precision real words. K is
|
87 |
* a constant in (2,3] that depends on the achieved bandwidth during a
|
88 |
* simultaneous message exchange between two processes. An empirical
|
89 |
* optimistic value of K is typically 2.4.
|
90 |
*
|
91 |
* Arguments
|
92 |
* =========
|
93 |
*
|
94 |
* PBCST (local input/output) HPL_T_panel *
|
95 |
* On entry, PBCST points to the data structure containing the
|
96 |
* panel (to be broadcast) information.
|
97 |
*
|
98 |
* IFLAG (local input/output) int *
|
99 |
* On entry, IFLAG indicates whether or not the broadcast has
|
100 |
* already been completed. If not, probing will occur, and the
|
101 |
* outcome will be contained in IFLAG on exit.
|
102 |
*
|
103 |
* PANEL (local input/output) HPL_T_panel *
|
104 |
* On entry, PANEL points to the data structure containing the
|
105 |
* panel information.
|
106 |
*
|
107 |
* NN (local input) const int
|
108 |
* On entry, NN specifies the local number of columns of the
|
109 |
* trailing submatrix to be swapped and broadcast starting at
|
110 |
* the current position. NN must be at least zero.
|
111 |
*
|
112 |
* ---------------------------------------------------------------------
|
113 |
*/
|
114 |
/*
|
115 |
* .. Local Variables ..
|
116 |
*/
|
117 |
double * A, * U;
|
118 |
int * ipID, * iplen, * ipmap, * ipmapm1,
|
119 |
* iwork, * lindxA = NULL, * lindxAU,
|
120 |
* permU; |
121 |
static int equil=-1; |
122 |
int icurrow, * iflag, * ipA, * ipl, jb, k,
|
123 |
lda, myrow, n, nprow; |
124 |
#define LDU n
|
125 |
/* ..
|
126 |
* .. Executable Statements ..
|
127 |
*/
|
128 |
n = PANEL->n; n = Mmin( NN, n ); jb = PANEL->jb; |
129 |
/*
|
130 |
* Quick return if there is nothing to do
|
131 |
*/
|
132 |
if( ( n <= 0 ) || ( jb <= 0 ) ) return; |
133 |
#ifdef HPL_DETAILED_TIMING
|
134 |
HPL_ptimer( HPL_TIMING_LASWP ); |
135 |
#endif
|
136 |
/*
|
137 |
* Decide whether equilibration should be performed or not
|
138 |
*/
|
139 |
if( equil == -1 ) equil = PANEL->algo->equil; |
140 |
/*
|
141 |
* Retrieve parameters from the PANEL data structure
|
142 |
*/
|
143 |
nprow = PANEL->grid->nprow; myrow = PANEL->grid->myrow; |
144 |
A = PANEL->A; U = PANEL->U; iflag = PANEL->IWORK; |
145 |
lda = PANEL->lda; icurrow = PANEL->prow; |
146 |
/*
|
147 |
* Compute ipID (if not already done for this panel). lindxA and lindxAU
|
148 |
* are of length at most 2*jb - iplen is of size nprow+1, ipmap, ipmapm1
|
149 |
* are of size nprow, permU is of length jb, and this function needs a
|
150 |
* workspace of size max( 2 * jb (plindx1), nprow+1(equil)):
|
151 |
* 1(iflag) + 1(ipl) + 1(ipA) + 9*jb + 3*nprow + 1 + MAX(2*jb,nprow+1)
|
152 |
* i.e. 4 + 9*jb + 3*nprow + max(2*jb, nprow+1);
|
153 |
*/
|
154 |
k = (int)((unsigned int)(jb) << 1); ipl = iflag + 1; ipID = ipl + 1; |
155 |
ipA = ipID + ((unsigned int)(k) << 1); lindxA = ipA + 1; |
156 |
lindxAU = lindxA + k; iplen = lindxAU + k; ipmap = iplen + nprow + 1;
|
157 |
ipmapm1 = ipmap + nprow; permU = ipmapm1 + nprow; iwork = permU + jb; |
158 |
|
159 |
if( *iflag == -1 ) /* no index arrays have been computed so far */ |
160 |
{ |
161 |
HPL_pipid( PANEL, ipl, ipID ); |
162 |
HPL_plindx1( PANEL, *ipl, ipID, ipA, lindxA, lindxAU, iplen, |
163 |
ipmap, ipmapm1, permU, iwork ); |
164 |
*iflag = 1;
|
165 |
} |
166 |
else if( *iflag == 0 ) /* HPL_pdlaswp00T called before: reuse ipID */ |
167 |
{ |
168 |
HPL_plindx1( PANEL, *ipl, ipID, ipA, lindxA, lindxAU, iplen, |
169 |
ipmap, ipmapm1, permU, iwork ); |
170 |
*iflag = 1;
|
171 |
} |
172 |
else if( ( *iflag == 1 ) && ( equil != 0 ) ) |
173 |
{ /* HPL_pdlaswp01T was call before only re-compute IPLEN, IPMAP */
|
174 |
HPL_plindx10( PANEL, *ipl, ipID, iplen, ipmap, ipmapm1 ); |
175 |
*iflag = 1;
|
176 |
} |
177 |
/*
|
178 |
* Copy into U the rows to be spread (local to icurrow)
|
179 |
*/
|
180 |
if( myrow == icurrow )
|
181 |
{ HPL_dlaswp01T( *ipA, n, A, lda, U, LDU, lindxA, lindxAU ); } |
182 |
/*
|
183 |
* Spread U - optionally probe for column panel
|
184 |
*/
|
185 |
HPL_spreadT( PBCST, IFLAG, PANEL, HplRight, n, U, LDU, 0, iplen,
|
186 |
ipmap, ipmapm1 ); |
187 |
/*
|
188 |
* Local exchange (everywhere but in process row icurrow)
|
189 |
*/
|
190 |
if( myrow != icurrow )
|
191 |
{ |
192 |
k = ipmapm1[myrow]; |
193 |
HPL_dlaswp06T( iplen[k+1]-iplen[k], n, A, lda, Mptr( U, 0, |
194 |
iplen[k], LDU ), LDU, lindxA ); |
195 |
} |
196 |
/*
|
197 |
* Equilibration
|
198 |
*/
|
199 |
if( equil != 0 ) |
200 |
HPL_equil( PBCST, IFLAG, PANEL, HplTrans, n, U, LDU, iplen, ipmap, |
201 |
ipmapm1, iwork ); |
202 |
/*
|
203 |
* Rolling phase
|
204 |
*/
|
205 |
HPL_rollT( PBCST, IFLAG, PANEL, n, U, LDU, iplen, ipmap, ipmapm1 ); |
206 |
/*
|
207 |
* Permute U in every process row
|
208 |
*/
|
209 |
HPL_dlaswp10N( n, jb, U, LDU, permU ); |
210 |
|
211 |
#ifdef HPL_DETAILED_TIMING
|
212 |
HPL_ptimer( HPL_TIMING_LASWP ); |
213 |
#endif
|
214 |
/*
|
215 |
* End of HPL_pdlaswp01T
|
216 |
*/
|
217 |
} |