Centre Blaise Pascal » HPL sur GPU

/* 

 * -- High Performance Computing Linpack Benchmark (HPL)                

 *    HPL - 2.0 - September 10, 2008                          

 *    Antoine P. Petitet                                                

 *    University of Tennessee, Knoxville                                

 *    Innovative Computing Laboratory                                 

 *    (C) Copyright 2000-2008 All Rights Reserved                       

 *                                                                      

 * -- Copyright notice and Licensing terms:                             

 *                                                                      

 * Redistribution  and  use in  source and binary forms, with or without

 * modification, are  permitted provided  that the following  conditions

 * are met:                                                             

 *                                                                      

 * 1. Redistributions  of  source  code  must retain the above copyright

 * notice, this list of conditions and the following disclaimer.        

 *                                                                      

 * 2. Redistributions in binary form must reproduce  the above copyright

 * notice, this list of conditions,  and the following disclaimer in the

 * documentation and/or other materials provided with the distribution. 

 *                                                                      

 * 3. All  advertising  materials  mentioning  features  or  use of this

 * software must display the following acknowledgement:                 

 * This  product  includes  software  developed  at  the  University  of

 * Tennessee, Knoxville, Innovative Computing Laboratory.             

 *                                                                      

 * 4. The name of the  University,  the name of the  Laboratory,  or the

 * names  of  its  contributors  may  not  be used to endorse or promote

 * products  derived   from   this  software  without  specific  written

 * permission.                                                          

 *                                                                      

 * -- Disclaimer:                                                       

 *                                                                      

 * THIS  SOFTWARE  IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS

 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES,  INCLUDING,  BUT NOT

 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR

 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE UNIVERSITY

 * OR  CONTRIBUTORS  BE  LIABLE FOR ANY  DIRECT,  INDIRECT,  INCIDENTAL,

 * SPECIAL,  EXEMPLARY,  OR  CONSEQUENTIAL DAMAGES  (INCLUDING,  BUT NOT

 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,

 * DATA OR PROFITS; OR BUSINESS INTERRUPTION)  HOWEVER CAUSED AND ON ANY

 * THEORY OF LIABILITY, WHETHER IN CONTRACT,  STRICT LIABILITY,  OR TORT

 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE

 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 

 * ---------------------------------------------------------------------

 */ 

/*

 * Include files

 */

#include "hpl.h"

#ifdef STDC_HEADERS

void HPL_pdpancrN

(

   HPL_T_panel *                    PANEL,

   const int                        M,

   const int                        N,

   const int                        ICOFF,

   double *                         WORK

)

#else

void HPL_pdpancrN

( PANEL, M, N, ICOFF, WORK )

   HPL_T_panel *                    PANEL;

   const int                        M;

   const int                        N;

   const int                        ICOFF;

   double *                         WORK;

#endif

{

/* 

 * Purpose

 * =======

 *

 * HPL_pdpancrN factorizes  a panel of columns that is a sub-array of a

 * larger one-dimensional panel  A using the Crout variant of the  usual

 * one-dimensional algorithm.  The lower triangular N0-by-N0 upper block

 * of the panel is stored in no-transpose form (i.e. just like the input

 * matrix itself).

 *  

 * 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 (when N is equal to N0):

 *  

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

 *    N0^2 * ( M - N0/3 ) * gam2-3

 *  

 * where M is the local number of rows of  the panel, lat and bdwth  are

 * the latency and bandwidth of the network for  double  precision  real

 * words, and gam2-3 is  an  estimate  of the  Level 2 and Level 3  BLAS

 * rate of execution. The  recursive  algorithm  allows indeed to almost

 * achieve  Level 3 BLAS  performance  in the panel factorization.  On a

 * large  number of modern machines,  this  operation is however latency

 * bound,  meaning  that its cost can  be estimated  by only the latency

 * portion N0 * log_2(P) * lat.  Mono-directional links will double this

 * communication cost.

 *  

 * Note that  one  iteration of the the main loop is unrolled. The local

 * computation of the absolute value max of the next column is performed

 * just after its update by the current column. This allows to bring the

 * current column only  once through  cache at each  step.  The  current

 * implementation  does not perform  any blocking  for  this sequence of

 * BLAS operations, however the design allows for plugging in an optimal

 * (machine-specific) specialized  BLAS-like kernel.  This idea has been

 * suggested to us by Fred Gustavson, IBM T.J. Watson Research Center.

 *

 * Arguments

 * =========

 *

 * PANEL   (local input/output)          HPL_T_panel *

 *         On entry,  PANEL  points to the data structure containing the

 *         panel information.

 *

 * M       (local input)                 const int

 *         On entry,  M specifies the local number of rows of sub(A).

 *

 * N       (local input)                 const int

 *         On entry,  N specifies the local number of columns of sub(A).

 *

 * ICOFF   (global input)                const int

 *         On entry, ICOFF specifies the row and column offset of sub(A)

 *         in A.

 *

 * WORK    (local workspace)             double *

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

 *

 * ---------------------------------------------------------------------

 */ 

/*

 * .. Local Variables ..

 */

   double                     * A, * L1, * L1ptr;

#ifdef HPL_CALL_VSIPL

   vsip_mview_d               * Av0, * Av1, * Yv1, * Xv0, * Xv1;

#endif

   int                        Mm1, Nm1, curr, ii, iip1, jj, kk=0, lda,

                              m=M, n0;

/* ..

 * .. Executable Statements ..

 */

#ifdef HPL_DETAILED_TIMING

   HPL_ptimer( HPL_TIMING_PFACT );

#endif

   A    = PANEL->A;   lda = PANEL->lda;

   L1   = PANEL->L1;  n0  = PANEL->jb;

   curr = (int)( PANEL->grid->myrow == PANEL->prow );

   Nm1  = N - 1; jj = ICOFF;

   if( curr != 0 ) { ii = ICOFF; iip1 = ii+1; Mm1 = m-1; }

   else            { ii = 0;     iip1 = ii;   Mm1 = m;   }

#ifdef HPL_CALL_VSIPL

/*

 * Admit the blocks

 */

   (void) vsip_blockadmit_d( PANEL->Ablock,  VSIP_TRUE );

   (void) vsip_blockadmit_d( PANEL->L1block, VSIP_TRUE );

/*

 * Create the matrix views

 */

   Av0 = vsip_mbind_d( PANEL->Ablock,  0, 1, lda,       lda, PANEL->pmat->nq );

   Xv0 = vsip_mbind_d( PANEL->L1block, 0, 1, PANEL->jb, PANEL->jb, PANEL->jb );

#endif

/*

 * Find local absolute value max in first column - initialize WORK[0:3]

 */

   HPL_dlocmax( PANEL, m, ii, jj, WORK );

   while( Nm1 > 0 )

   {

/*

 * Swap and broadcast the current row

 */

      HPL_pdmxswp(  PANEL, m, ii, jj, WORK );

      HPL_dlocswpN( PANEL,    ii, jj, WORK );

/*

 * Compute row (column) jj of L1

 */

      if( kk > 0 )

      {

         L1ptr = Mptr( L1, jj, jj+1, n0 );

#ifdef HPL_CALL_VSIPL

/*

 * Create the matrix subviews

 */

         Av1 = vsip_msubview_d( Xv0, ICOFF, jj+1,  kk, Nm1 );

         Xv1 = vsip_msubview_d( Xv0, jj,    ICOFF, 1,  kk  );

         Yv1 = vsip_msubview_d( Xv0, jj,    jj+1,  1,  Nm1 );

         vsip_gemp_d( -HPL_rone, Xv1, VSIP_MAT_NTRANS, Av1, VSIP_MAT_NTRANS,

                      HPL_rone, Yv1 );

/*

 * Destroy the matrix subviews

 */

         (void) vsip_mdestroy_d( Yv1 );

         (void) vsip_mdestroy_d( Xv1 ); 

         (void) vsip_mdestroy_d( Av1 );

#else

         HPL_dgemv( HplColumnMajor, HplTrans, kk, Nm1, -HPL_rone,

                    Mptr( L1, ICOFF, jj+1, n0 ), n0, Mptr( L1, jj,

                    ICOFF, n0 ), n0, HPL_rone, L1ptr, n0 );

#endif

         if( curr != 0 )

            HPL_dcopy( Nm1, L1ptr, n0, Mptr( A, ii, jj+1, lda ), lda );

      }

/*

 * Scale current column by its absolute value max entry  -  Update  dia-

 * diagonal and subdiagonal elements in column  A(iip1:iip1+Mm1-1, jj+1)

 * and  find local  absolute value max in  that column  (Only  one  pass

 * through cache for each current column).  This sequence of  operations

 * could benefit from a specialized blocked implementation.

 */

      if( WORK[0] != HPL_rzero )

         HPL_dscal( Mm1, HPL_rone / WORK[0], Mptr( A, iip1, jj, lda ), 1 );

#ifdef HPL_CALL_VSIPL

/*

 * Create the matrix subviews

 */

      Av1 = vsip_msubview_d( Av0, PANEL->ii+iip1, PANEL->jj+ICOFF, Mm1, kk+1 );

      Xv1 = vsip_msubview_d( Xv0, ICOFF,          jj+1,            kk+1,   1 );

      Yv1 = vsip_msubview_d( Av0, PANEL->ii+iip1, PANEL->jj+jj+1,  Mm1,    1 );

      vsip_gemp_d( -HPL_rone, Av1, VSIP_MAT_NTRANS, Xv1, VSIP_MAT_NTRANS,

                   HPL_rone, Yv1 );

/*

 * Destroy the matrix subviews

 */

      vsip_mdestroy_d( Yv1 );

      vsip_mdestroy_d( Xv1 );

      vsip_mdestroy_d( Av1 );

#else

      HPL_dgemv( HplColumnMajor, HplNoTrans, Mm1, kk+1, -HPL_rone,

                 Mptr( A, iip1, ICOFF, lda ), lda, Mptr( L1, ICOFF,

                 jj+1, n0 ), 1, HPL_rone, Mptr( A, iip1, jj+1, lda ),

                 1 );

#endif

      HPL_dlocmax( PANEL, Mm1, iip1, jj+1, WORK );

      if( curr != 0 ) { ii = iip1; iip1++; m = Mm1; Mm1--; }

      Nm1--; jj++; kk++;

   }

/*

 * Swap and broadcast last row - Scale last column by its absolute value

 * max entry

 */ 

   HPL_pdmxswp(  PANEL, m, ii, jj, WORK );

   HPL_dlocswpN( PANEL,    ii, jj, WORK );

   if( WORK[0] != HPL_rzero )

      HPL_dscal( Mm1, HPL_rone / WORK[0], Mptr( A, iip1, jj, lda ), 1 );

#ifdef HPL_CALL_VSIPL

/*

 * Release the blocks

 */

   (void) vsip_blockrelease_d( vsip_mgetblock_d( Xv0 ), VSIP_TRUE );

   (void) vsip_blockrelease_d( vsip_mgetblock_d( Av0 ), VSIP_TRUE );

/*

 * Destroy the matrix views

 */

   (void) vsip_mdestroy_d( Xv0 );

   (void) vsip_mdestroy_d( Av0 );

#endif

#ifdef HPL_DETAILED_TIMING

   HPL_ptimer( HPL_TIMING_PFACT );

#endif

/*

 * End of HPL_pdpancrN

 */

}