src/smpi/colls/bcast/bcast-ompi-pipeline.cpp

   1 /* Copyright (c) 2013-2023. The SimGrid Team.
   2  * All rights reserved.                                                     */
   3
   4 /* This program is free software; you can redistribute it and/or modify it
   5  * under the terms of the license (GNU LGPL) which comes with this package. */
   6
   7 #include "../coll_tuned_topo.hpp"
   8 #include "../colls_private.hpp"
   9
  10 #define MAXTREEFANOUT 32
  11
  12 namespace simgrid::smpi {
  13 int bcast__ompi_pipeline( void* buffer,
  14                           int original_count,
  15                           MPI_Datatype datatype,
  16                           int root,
  17                           MPI_Comm comm)
  18 {
  19     int count_by_segment = original_count;
  20     size_t type_size;
  21     size_t segsize =1024  << 7;
  22     //mca_coll_tuned_module_t *tuned_module = (mca_coll_tuned_module_t*) module;
  23     //mca_coll_tuned_comm_t *data = tuned_module->tuned_data;
  24
  25 //    return ompi_coll_tuned_bcast_intra_generic( buffer, count, datatype, root, comm, module,
  26 //                                                count_by_segment, data->cached_pipeline );
  27     ompi_coll_tree_t * tree = ompi_coll_tuned_topo_build_chain( 1, comm, root );
  28     int i;
  29     int rank, size;
  30     int segindex;
  31     int num_segments; /* Number of segments */
  32     int sendcount;    /* number of elements sent in this segment */
  33     size_t realsegsize;
  34     char *tmpbuf;
  35     ptrdiff_t extent;
  36     MPI_Request recv_reqs[2] = {MPI_REQUEST_NULL, MPI_REQUEST_NULL};
  37     MPI_Request* send_reqs   = nullptr;
  38     int req_index;
  39
  40     /**
  41      * Determine number of elements sent per operation.
  42      */
  43     type_size = datatype->size();
  44
  45     size = comm->size();
  46     rank = comm->rank();
  47     if(size==1)return MPI_SUCCESS;
  48
  49
  50     const double a_p16  = 3.2118e-6; /* [1 / byte] */
  51     const double b_p16  = 8.7936;
  52     const double a_p64  = 2.3679e-6; /* [1 / byte] */
  53     const double b_p64  = 1.1787;
  54     const double a_p128 = 1.6134e-6; /* [1 / byte] */
  55     const double b_p128 = 2.1102;
  56     size_t message_size;
  57
  58     /* else we need data size for decision function */
  59     message_size = type_size * (unsigned long)original_count;   /* needed for decision */
  60
  61     if (size < (a_p128 * message_size + b_p128)) {
  62             //Pipeline with 128KB segments
  63             segsize = 1024  << 7;
  64     }else if (size < (a_p64 * message_size + b_p64)) {
  65             // Pipeline with 64KB segments
  66             segsize = 1024 << 6;
  67     }else if (size < (a_p16 * message_size + b_p16)) {
  68             //Pipeline with 16KB segments
  69             segsize = 1024 << 4;
  70     }
  71
  72     COLL_TUNED_COMPUTED_SEGCOUNT( segsize, type_size, count_by_segment );
  73
  74     XBT_DEBUG("coll:tuned:bcast_intra_pipeline rank %d ss %5zu type_size %lu count_by_segment %d",
  75                  comm->rank(), segsize, (unsigned long)type_size, count_by_segment);
  76
  77
  78
  79     extent = datatype->get_extent();
  80     num_segments = (original_count + count_by_segment - 1) / count_by_segment;
  81     realsegsize = count_by_segment * extent;
  82
  83     /* Set the buffer pointers */
  84     tmpbuf = (char *) buffer;
  85
  86     if( tree->tree_nextsize != 0 ) {
  87       send_reqs = new MPI_Request[tree->tree_nextsize];
  88     }
  89
  90     /* Root code */
  91     if( rank == root ) {
  92         /*
  93            For each segment:
  94            - send segment to all children.
  95              The last segment may have less elements than other segments.
  96         */
  97         sendcount = count_by_segment;
  98         for( segindex = 0; segindex < num_segments; segindex++ ) {
  99             if( segindex == (num_segments - 1) ) {
 100                 sendcount = original_count - segindex * count_by_segment;
 101             }
 102             for( i = 0; i < tree->tree_nextsize; i++ ) {
 103                 send_reqs[i] = Request::isend(tmpbuf, sendcount, datatype,
 104                                          tree->tree_next[i],
 105                                          COLL_TAG_BCAST, comm);
 106            }
 107
 108             /* complete the sends before starting the next sends */
 109             Request::waitall( tree->tree_nextsize, send_reqs,
 110                                          MPI_STATUSES_IGNORE );
 111
 112             /* update tmp buffer */
 113             tmpbuf += realsegsize;
 114
 115         }
 116     }
 117
 118     /* Intermediate nodes code */
 119     else if( tree->tree_nextsize > 0 ) {
 120         /*
 121            Create the pipeline.
 122            1) Post the first receive
 123            2) For segments 1 .. num_segments
 124               - post new receive
 125               - wait on the previous receive to complete
 126               - send this data to children
 127            3) Wait on the last segment
 128            4) Compute number of elements in last segment.
 129            5) Send the last segment to children
 130          */
 131         req_index = 0;
 132         recv_reqs[req_index]=Request::irecv(tmpbuf, count_by_segment, datatype,
 133                            tree->tree_prev, COLL_TAG_BCAST,
 134                            comm);
 135
 136         for( segindex = 1; segindex < num_segments; segindex++ ) {
 137
 138             req_index = req_index ^ 0x1;
 139
 140             /* post new irecv */
 141             recv_reqs[req_index]= Request::irecv( tmpbuf + realsegsize, count_by_segment,
 142                                 datatype, tree->tree_prev,
 143                                 COLL_TAG_BCAST,
 144                                 comm);
 145
 146             /* wait for and forward the previous segment to children */
 147             Request::wait( &recv_reqs[req_index ^ 0x1],
 148                                      MPI_STATUSES_IGNORE );
 149
 150             for( i = 0; i < tree->tree_nextsize; i++ ) {
 151                 send_reqs[i]=Request::isend(tmpbuf, count_by_segment, datatype,
 152                                          tree->tree_next[i],
 153                                          COLL_TAG_BCAST, comm );
 154             }
 155
 156             /* complete the sends before starting the next iteration */
 157             Request::waitall( tree->tree_nextsize, send_reqs,
 158                                          MPI_STATUSES_IGNORE );
 159
 160             /* Update the receive buffer */
 161             tmpbuf += realsegsize;
 162         }
 163
 164         /* Process the last segment */
 165         Request::wait( &recv_reqs[req_index], MPI_STATUSES_IGNORE );
 166         sendcount = original_count - (num_segments - 1) * count_by_segment;
 167         for( i = 0; i < tree->tree_nextsize; i++ ) {
 168             send_reqs[i] = Request::isend(tmpbuf, sendcount, datatype,
 169                                      tree->tree_next[i],
 170                                      COLL_TAG_BCAST, comm);
 171         }
 172
 173         Request::waitall( tree->tree_nextsize, send_reqs,
 174                                      MPI_STATUSES_IGNORE );
 175     }
 176
 177     /* Leaf nodes */
 178     else {
 179         /*
 180            Receive all segments from parent in a loop:
 181            1) post irecv for the first segment
 182            2) for segments 1 .. num_segments
 183               - post irecv for the next segment
 184               - wait on the previous segment to arrive
 185            3) wait for the last segment
 186         */
 187         req_index = 0;
 188         recv_reqs[req_index] = Request::irecv(tmpbuf, count_by_segment, datatype,
 189                                  tree->tree_prev, COLL_TAG_BCAST,
 190                                  comm);
 191
 192         for( segindex = 1; segindex < num_segments; segindex++ ) {
 193             req_index = req_index ^ 0x1;
 194             tmpbuf += realsegsize;
 195             /* post receive for the next segment */
 196             recv_reqs[req_index] = Request::irecv(tmpbuf, count_by_segment, datatype,
 197                                      tree->tree_prev, COLL_TAG_BCAST,
 198                                      comm);
 199             /* wait on the previous segment */
 200             Request::wait( &recv_reqs[req_index ^ 0x1],
 201                                      MPI_STATUS_IGNORE );
 202         }
 203
 204         Request::wait( &recv_reqs[req_index], MPI_STATUS_IGNORE );
 205     }
 206
 207     delete[] send_reqs;
 208     ompi_coll_tuned_topo_destroy_tree(&tree);
 209
 210     return (MPI_SUCCESS);
 211 }
 212
 213 } // namespace simgrid::smpi