openmpi/ompi/mca/coll/tuned/coll_tuned_reduce.c

/*
 * Copyright (c) 2004-2005 The Trustees of Indiana University and Indiana
 *                         University Research and Technology
 *                         Corporation.  All rights reserved.
 * Copyright (c) 2004-2006 The University of Tennessee and The University
 *                         of Tennessee Research Foundation.  All rights
 *                         reserved.
 * Copyright (c) 2004-2005 High Performance Computing Center Stuttgart,
 *                         University of Stuttgart.  All rights reserved.
 * Copyright (c) 2004-2005 The Regents of the University of California.
 *                         All rights reserved.
 * $COPYRIGHT$
 *
 * Additional copyrights may follow
 *
 * $HEADER$
 */

#include "ompi_config.h"

#include "mpi.h"
#include "ompi/constants.h"
#include "ompi/datatype/datatype.h"
#include "ompi/communicator/communicator.h"
#include "ompi/mca/coll/coll.h"
#include "ompi/mca/coll/base/coll_tags.h"
#include "ompi/mca/pml/pml.h"
#include "ompi/op/op.h"
#include "coll_tuned.h"
#include "coll_tuned_topo.h"

/**
 * This is a generic implementation of the reduce protocol. It used the tree
 * provided as an argument and execute all operations using a segment of
 * count times a datatype.
 * For the last communication it will update the count in order to limit
 * th number of datatype to the original count (original_count)
 */
int ompi_coll_tuned_reduce_generic( void* sendbuf, void* recvbuf, int original_count,
                                    ompi_datatype_t* datatype, ompi_op_t* op,
                                    int root, ompi_communicator_t* comm,
                                    ompi_coll_tree_t* tree, int count_by_segment )
{
    char *inbuf[2] = {(char*)NULL, (char*)NULL};
    char *local_op_buffer, *accumbuf = NULL, *sendtmpbuf;
    ptrdiff_t extent, lower_bound;
    size_t typelng, realsegsize;
    ompi_request_t* reqs[2] = {MPI_REQUEST_NULL, MPI_REQUEST_NULL};
    int num_segments, line, ret, segindex, i, rank;
    int recvcount, prevcount, inbi, previnbi;

    /**
     * Determine number of segments and number of elements
     * sent per operation
     */
    ompi_ddt_get_extent( datatype, &lower_bound, &extent );
    ompi_ddt_type_size( datatype, &typelng );
    num_segments = (original_count + count_by_segment - 1) / count_by_segment;
    realsegsize = count_by_segment * extent;

    sendtmpbuf = (char*) sendbuf; 
    if( sendbuf == MPI_IN_PLACE ) { 
        sendtmpbuf = (char *)recvbuf; 
    }

    rank = ompi_comm_rank(comm);

    /* non-leaf nodes - wait for children to send me data & forward up (if needed) */
    if( tree->tree_nextsize > 0 ) {
        /* handle non existant recv buffer (i.e. its NULL.. like basic allreduce uses!) */
        accumbuf = (char*)recvbuf;
        if( NULL == accumbuf ) {
            accumbuf = (char*)malloc(realsegsize * num_segments);  /* TO BE OPTIMIZED */
	    if (accumbuf == NULL) { line = __LINE__; ret = -1; goto error_hndl; }
	}

        /* Allocate two buffers for incoming segments */
        inbuf[0] = (char*) malloc(realsegsize);
        if( inbuf[0] == NULL ) { line = __LINE__; ret = -1; goto error_hndl; }
        /* if there is chance to overlap communication -
           allocate second buffer */
        if( (num_segments > 1) || (tree->tree_nextsize > 1) ) {
            inbuf[1] = (char*) malloc(realsegsize);
            if( inbuf[1] == NULL ) { line = __LINE__; ret = -1; goto error_hndl;}
        } else {
            inbuf[1] = NULL;
        }

        /* reset input buffer index and receive count */
        inbi = 0;
        recvcount = 0;
        /* for each segment */
        for( segindex = 0; segindex <= num_segments; segindex++ ) {
            prevcount = recvcount;
            /* recvcount - number of elements in current segment */
            recvcount = count_by_segment;
            if( segindex == (num_segments-1) )
                recvcount = original_count - count_by_segment * segindex;

            /* for each child */
            for( i = 0; i < tree->tree_nextsize; i++ ) {
                /**
                 * We try to overlap communication:
                 * either with next segment or with the next child
                 */
                /* post irecv for current segindex on current child */
                if( segindex < num_segments ) {
                    void* local_recvbuf = inbuf[inbi];
                    if( 0 == i ) {
                        /* for the first step (1st child per segment) we might be able to
                         * irecv directly into the accumulate buffer so that we can
                         * reduce(op) this with our sendbuf in one step as ompi_op_reduce
                         * only has two buffer pointers, this avoids an extra memory copy.
                         *
                         * BUT if we are root and are USING MPI_IN_PLACE this is wrong ek!
                         * check for root might not be needed as it should be checked higher up
                         */
                        if( !((MPI_IN_PLACE == sendbuf) && (rank == tree->tree_root)) ) {
                            local_recvbuf = accumbuf + segindex * realsegsize;
                        }
                    }
                    ret = MCA_PML_CALL(irecv(local_recvbuf, recvcount,datatype, tree->tree_next[i],
                                             MCA_COLL_BASE_TAG_REDUCE, comm, &reqs[inbi]));
                    if (ret != MPI_SUCCESS) { line = __LINE__; goto error_hndl;  }
                }
                /* wait for previous req to complete, if any */
                previnbi = (inbi+1) % 2;
                /* wait on data from last child for previous segment */
                ret = ompi_request_wait_all( 1, &reqs[previnbi], MPI_STATUSES_IGNORE );
                if (ret != MPI_SUCCESS) { line = __LINE__; goto error_hndl;  }
                local_op_buffer = inbuf[previnbi];
                if( i > 0 ) {
                    /* our first operation is to combine our own [sendbuf] data with the data
                     * we recvd from down stream (but only if we are not root and not using
                     * MPI_IN_PLACE)
                     */
                    if( 1 == i ) {
                        if( !((MPI_IN_PLACE == sendbuf) && (rank == tree->tree_root)) ) {
                            local_op_buffer = sendtmpbuf + segindex * realsegsize;
                        }
                    }
                    /* apply operation */
                    ompi_op_reduce(op, local_op_buffer, accumbuf+segindex*realsegsize, recvcount, datatype );
                } else if ( segindex > 0 ) {
                    void* accumulator = accumbuf + (segindex-1) * realsegsize;

                    if( tree->tree_nextsize <= 1 ) {
                        if( !((MPI_IN_PLACE == sendbuf) && (rank == tree->tree_root)) ) {            
                            local_op_buffer = sendtmpbuf+(segindex-1)*realsegsize;
                        }
                    }
                    ompi_op_reduce(op, local_op_buffer, accumulator, prevcount, datatype );

                    /* all reduced on available data this step (i) complete, pass to
                     * the next process unless your the root
                     */
                    if (rank != tree->tree_root) {
                        /* send combined/accumulated data to parent */
                        ret = MCA_PML_CALL( send( accumulator, prevcount, datatype,
                                                  tree->tree_prev, MCA_COLL_BASE_TAG_REDUCE,
                                                  MCA_PML_BASE_SEND_STANDARD, comm) );
                        if (ret != MPI_SUCCESS) { line = __LINE__; goto error_hndl;  }
                    }

                    /* we stop when segindex = number of segments (i.e. we do num_segment+1 steps to allow for pipelining */
                    if (segindex == num_segments) break;
                }

                /* update input buffer index */
                inbi = previnbi;
            } /* end of for each child */
        } /* end of for each segment */

        /* clean up */
        if( inbuf[0] != NULL) free(inbuf[0]);
        if( inbuf[1] != NULL) free(inbuf[1]);
        if( NULL == recvbuf ) free(accumbuf);
    }

    /* leaf nodes */
    else {
        /* Send segmented data to parents */
        segindex = 0;
        while( original_count > 0 ) {
            if( original_count < count_by_segment ) count_by_segment = original_count;
            ret = MCA_PML_CALL( send((char*)sendbuf + segindex * realsegsize, count_by_segment,
                                     datatype, tree->tree_prev,
                                     MCA_COLL_BASE_TAG_REDUCE, MCA_PML_BASE_SEND_STANDARD, comm) );
            if (ret != MPI_SUCCESS) { line = __LINE__; goto error_hndl;  }
            segindex++;
            original_count -= count_by_segment;
        }
    }
    return OMPI_SUCCESS;

 error_hndl:  /* error handler */
    OPAL_OUTPUT (( ompi_coll_tuned_stream, "ERROR_HNDL: node %d file %s line %d error %d\n", rank, __FILE__, line, ret ));
    if( inbuf[0] != NULL ) free(inbuf[0]);
    if( inbuf[1] != NULL ) free(inbuf[1]);
    if( (NULL == recvbuf) && (NULL != accumbuf) ) free(accumbuf);
    return ret;
}

/* Attention: this version of the reduce operations does not
   work for:
   - non-commutative operations
   - segment sizes which are not multiplies of the extent of the datatype
     meaning that at least one datatype must fit in the segment !
*/

int ompi_coll_tuned_reduce_intra_chain( void *sendbuf, void *recvbuf, int count,
                                        ompi_datatype_t* datatype, ompi_op_t* op,
                                        int root, ompi_communicator_t* comm, uint32_t segsize,
                                        int fanout)
{
    int segcount;
    size_t typelng;

    OPAL_OUTPUT((ompi_coll_tuned_stream,"coll:tuned:reduce_intra_chain rank %d fo %d ss %5d", ompi_comm_rank(comm), fanout, segsize));

    COLL_TUNED_UPDATE_CHAIN( comm, root, fanout );
    /**
     * Determine number of segments and number of elements
     * sent per operation
     */
    ompi_ddt_type_size( datatype, &typelng );
    if( segsize > typelng ) {
        segcount     = (int)(segsize / typelng);
    } else  {
        segcount     = count;
    }

    return ompi_coll_tuned_reduce_generic( sendbuf, recvbuf, count, datatype, op, root, comm,
                                           comm->c_coll_selected_data->cached_chain, segcount );
}


int ompi_coll_tuned_reduce_intra_pipeline( void *sendbuf, void *recvbuf,
                                           int count, ompi_datatype_t* datatype,
                                           ompi_op_t* op, int root,
                                           ompi_communicator_t* comm, uint32_t segsize )
{
    int segcount;
    size_t typelng;

    OPAL_OUTPUT((ompi_coll_tuned_stream,"coll:tuned:reduce_intra_pipeline rank %d ss %5d",
                 ompi_comm_rank(comm), segsize));

    COLL_TUNED_UPDATE_PIPELINE( comm, root );

    /**
     * Determine number of segments and number of elements
     * sent per operation
     */
    ompi_ddt_type_size( datatype, &typelng );
    if( segsize > typelng ) {
        segcount     = (int)(segsize / typelng);
    } else  {
        segcount     = count;
    }

    return ompi_coll_tuned_reduce_generic( sendbuf, recvbuf, count, datatype, op, root, comm,
                                           comm->c_coll_selected_data->cached_pipeline, segcount );
}

int ompi_coll_tuned_reduce_intra_binary( void *sendbuf, void *recvbuf,
                                         int count, ompi_datatype_t* datatype,
                                         ompi_op_t* op, int root,
                                         ompi_communicator_t* comm, uint32_t segsize )
{
    int segcount;
    size_t typelng;

    OPAL_OUTPUT((ompi_coll_tuned_stream,"coll:tuned:reduce_intra_binary rank %d ss %5d",
                 ompi_comm_rank(comm), segsize));

    COLL_TUNED_UPDATE_BINTREE( comm, root );

    /**
     * Determine number of segments and number of elements
     * sent per operation
     */
    ompi_ddt_type_size( datatype, &typelng );
    if( segsize > typelng ) {
        segcount     = (int)(segsize / typelng);
    } else  {
        segcount     = count;
    }

    return ompi_coll_tuned_reduce_generic( sendbuf, recvbuf, count, datatype, op, root, comm,
                                           comm->c_coll_selected_data->cached_bintree, segcount );
}

int ompi_coll_tuned_reduce_intra_binomial( void *sendbuf, void *recvbuf,
                                           int count, ompi_datatype_t* datatype,
                                           ompi_op_t* op, int root,
                                           ompi_communicator_t* comm, uint32_t segsize )
{
    int segcount;
    size_t typelng;

    OPAL_OUTPUT((ompi_coll_tuned_stream,"coll:tuned:reduce_intra_binomial rank %d ss %5d",
                 ompi_comm_rank(comm), segsize));

    COLL_TUNED_UPDATE_BMTREE( comm, root );

    /**
     * Determine number of segments and number of elements
     * sent per operation
     */
    ompi_ddt_type_size( datatype, &typelng );
    if( segsize > typelng ) {
        segcount     = (int)(segsize / typelng);
    } else  {
        segcount     = count;
    }

    return ompi_coll_tuned_reduce_generic( sendbuf, recvbuf, count, datatype, op, root, comm,
                                           comm->c_coll_selected_data->cached_bmtree, segcount );
}

/*
 * Linear functions are copied from the BASIC coll module
 * they do not segment the message and are simple implementations
 * but for some small number of nodes and/or small data sizes they 
 * are just as fast as tuned/tree based segmenting operations 
 * and as such may be selected by the decision functions
 * These are copied into this module due to the way we select modules
 * in V1. i.e. in V2 we will handle this differently and so will not
 * have to duplicate code.
 * GEF Oct05 after asking Jeff.
 */

/* copied function (with appropriate renaming) starts here */

/*
 *  reduce_lin_intra
 *
 *  Function:   - reduction using O(N) algorithm
 *  Accepts:    - same as MPI_Reduce()
 *  Returns:    - MPI_SUCCESS or error code
 */
int
ompi_coll_tuned_reduce_intra_basic_linear(void *sbuf, void *rbuf, int count,
                                          struct ompi_datatype_t *dtype,
                                          struct ompi_op_t *op,
                                          int root, struct ompi_communicator_t *comm)
{
    int i, rank, err, size;
    ptrdiff_t true_lb, true_extent, lb, extent;
    char *free_buffer = NULL;
    char *pml_buffer = NULL;
    char *inplace_temp = NULL;
    char *inbuf;

    /* Initialize */

    rank = ompi_comm_rank(comm);
    size = ompi_comm_size(comm);

    OPAL_OUTPUT((ompi_coll_tuned_stream,"coll:tuned:reduce_intra_basic_linear rank %d", rank));

    /* If not root, send data to the root. */

    if (rank != root) {
        err = MCA_PML_CALL(send(sbuf, count, dtype, root,
                                MCA_COLL_BASE_TAG_REDUCE,
                                MCA_PML_BASE_SEND_STANDARD, comm));
        return err;
    }

    /* see discussion in ompi_coll_basic_reduce_lin_intra about extent and true extend */
    /* for reducing buffer allocation lengths.... */

    ompi_ddt_get_extent(dtype, &lb, &extent);
    ompi_ddt_get_true_extent(dtype, &true_lb, &true_extent);

    if (MPI_IN_PLACE == sbuf) {
        sbuf = rbuf;
        inplace_temp = (char*)malloc(true_extent + (count - 1) * extent);
        if (NULL == inplace_temp) {
            return OMPI_ERR_OUT_OF_RESOURCE;
        }
        rbuf = inplace_temp - lb;
    }

    if (size > 1) {
        free_buffer = (char*)malloc(true_extent + (count - 1) * extent);
        if (NULL == free_buffer) {
            return OMPI_ERR_OUT_OF_RESOURCE;
        }
        pml_buffer = free_buffer - lb;
    }

    /* Initialize the receive buffer. */

    if (rank == (size - 1)) {
        err = ompi_ddt_copy_content_same_ddt(dtype, count, (char*)rbuf, (char*)sbuf);
    } else {
        err = MCA_PML_CALL(recv(rbuf, count, dtype, size - 1,
                                MCA_COLL_BASE_TAG_REDUCE, comm,
                                MPI_STATUS_IGNORE));
    }
    if (MPI_SUCCESS != err) {
        if (NULL != free_buffer) {
            free(free_buffer);
        }
        return err;
    }

    /* Loop receiving and calling reduction function (C or Fortran). */

    for (i = size - 2; i >= 0; --i) {
        if (rank == i) {
            inbuf = (char*)sbuf;
        } else {
            err = MCA_PML_CALL(recv(pml_buffer, count, dtype, i,
                                    MCA_COLL_BASE_TAG_REDUCE, comm,
                                    MPI_STATUS_IGNORE));
            if (MPI_SUCCESS != err) {
                if (NULL != free_buffer) {
                    free(free_buffer);
                }
                return err;
            }

            inbuf = pml_buffer;
        }

        /* Perform the reduction */
        ompi_op_reduce(op, inbuf, rbuf, count, dtype);
    }

    if (NULL != inplace_temp) {
        err = ompi_ddt_copy_content_same_ddt(dtype, count, (char*)sbuf, inplace_temp);
        free(inplace_temp);
    }
    if (NULL != free_buffer) {
        free(free_buffer);
    }

    /* All done */
    return MPI_SUCCESS;
}

/* copied function (with appropriate renaming) ends here */


/**
 * The following are used by dynamic and forced rules
 *
 * publish details of each algorithm and if its forced/fixed/locked in
 * as you add methods/algorithms you must update this and the query/map routines
 *
 * this routine is called by the component only
 * this makes sure that the mca parameters are set to their initial values and perms
 * module does not call this they call the forced_getvalues routine instead
 */
int ompi_coll_tuned_reduce_intra_check_forced_init (coll_tuned_force_algorithm_mca_param_indices_t *mca_param_indices)
{
    int rc;
    int max_alg = 5;

    ompi_coll_tuned_forced_max_algorithms[REDUCE] = max_alg;

    rc = mca_base_param_reg_int (&mca_coll_tuned_component.super.collm_version,
                                 "reduce_algorithm_count",
                                 "Number of reduce algorithms available",
                                 false, true, max_alg, NULL);


    mca_param_indices->algorithm_param_index
        = mca_base_param_reg_int(&mca_coll_tuned_component.super.collm_version,
                                 "reduce_algorithm",
                                 "Which reduce algorithm is used. Can be locked down to choice of: 0 ignore, 1 linear, 2 chain, 3 pipeline, 4 binary, 5 binomial",
                                 false, false, 0, NULL);

    mca_param_indices->segsize_param_index
        = mca_base_param_reg_int(&mca_coll_tuned_component.super.collm_version,
                                 "reduce_algorithm_segmentsize",
                                 "Segment size in bytes used by default for reduce algorithms. Only has meaning if algorithm is forced and supports segmenting. 0 bytes means no segmentation.",
                                 false, false, 0, NULL);

    mca_param_indices->tree_fanout_param_index
        = mca_base_param_reg_int(&mca_coll_tuned_component.super.collm_version,
                                 "reduce_algorithm_tree_fanout",
                                 "Fanout for n-tree used for reduce algorithms. Only has meaning if algorithm is forced and supports n-tree topo based operation.",
                                 false, false,
                                 ompi_coll_tuned_init_tree_fanout, /* get system wide default */
                                 NULL);

    mca_param_indices->chain_fanout_param_index
        = mca_base_param_reg_int(&mca_coll_tuned_component.super.collm_version,
                                 "reduce_algorithm_chain_fanout",
                                 "Fanout for chains used for reduce algorithms. Only has meaning if algorithm is forced and supports chain topo based operation.",
                                 false, false,
                                 ompi_coll_tuned_init_chain_fanout, /* get system wide default */
                                 NULL);
    return (MPI_SUCCESS);
}


int ompi_coll_tuned_reduce_intra_do_forced(void *sbuf, void* rbuf, int count,
                                           struct ompi_datatype_t *dtype,
                                           struct ompi_op_t *op, int root,
                                           struct ompi_communicator_t *comm)
{
    OPAL_OUTPUT((ompi_coll_tuned_stream,"coll:tuned:reduce_intra_do_forced selected algorithm %d", 
                 comm->c_coll_selected_data->user_forced[REDUCE].algorithm));

    switch (comm->c_coll_selected_data->user_forced[REDUCE].algorithm) {
    case (0):   return ompi_coll_tuned_reduce_intra_dec_fixed (sbuf, rbuf, count, dtype, op, root, comm);
    case (1):   return ompi_coll_tuned_reduce_intra_basic_linear (sbuf, rbuf, count, dtype, op, root, comm);
    case (2):   return ompi_coll_tuned_reduce_intra_chain (sbuf, rbuf, count, dtype, op, root, comm,
                                                           comm->c_coll_selected_data->user_forced[REDUCE].segsize, 
                                                           comm->c_coll_selected_data->user_forced[REDUCE].chain_fanout); 
    case (3):   return ompi_coll_tuned_reduce_intra_pipeline (sbuf, rbuf, count, dtype, op, root, comm, 
                                                              comm->c_coll_selected_data->user_forced[REDUCE].segsize); 
    case (4):   return ompi_coll_tuned_reduce_intra_binary (sbuf, rbuf, count, dtype, op, root, comm, 
                                                            comm->c_coll_selected_data->user_forced[REDUCE].segsize); 
    case (5):   return ompi_coll_tuned_reduce_intra_binomial (sbuf, rbuf, count, dtype, op, root, comm, 
                                                              comm->c_coll_selected_data->user_forced[REDUCE].segsize); 
    default:
        OPAL_OUTPUT((ompi_coll_tuned_stream,"coll:tuned:reduce_intra_do_forced attempt to select algorithm %d when only 0-%d is valid?",
                     comm->c_coll_selected_data->user_forced[REDUCE].algorithm, ompi_coll_tuned_forced_max_algorithms[REDUCE]));
        return (MPI_ERR_ARG);
    } /* switch */
}


int ompi_coll_tuned_reduce_intra_do_this(void *sbuf, void* rbuf, int count,
                                         struct ompi_datatype_t *dtype,
                                         struct ompi_op_t *op, int root,
                                         struct ompi_communicator_t *comm,
                                         int algorithm, int faninout, int segsize)
{
    OPAL_OUTPUT((ompi_coll_tuned_stream,"coll:tuned:reduce_intra_do_this selected algorithm %d topo faninout %d segsize %d",
                 algorithm, faninout, segsize));

    switch (algorithm) {
    case (0):   return ompi_coll_tuned_reduce_intra_dec_fixed (sbuf, rbuf, count, dtype, op, root, comm);
    case (1):   return ompi_coll_tuned_reduce_intra_basic_linear (sbuf, rbuf, count, dtype, op, root, comm);
    case (2):   return ompi_coll_tuned_reduce_intra_chain (sbuf, rbuf, count, dtype, op, root, comm, 
                                                           segsize, faninout);
    case (3):   return ompi_coll_tuned_reduce_intra_pipeline (sbuf, rbuf, count, dtype, op, root, comm, 
                                                              segsize);
    case (4):   return ompi_coll_tuned_reduce_intra_binary (sbuf, rbuf, count, dtype, op, root, comm, 
                                                            segsize); 
    case (5):   return ompi_coll_tuned_reduce_intra_binomial (sbuf, rbuf, count, dtype, op, root, comm, 
                                                              segsize); 
    default:
        OPAL_OUTPUT((ompi_coll_tuned_stream,"coll:tuned:reduce_intra_do_this attempt to select algorithm %d when only 0-%d is valid?",
                     algorithm, ompi_coll_tuned_forced_max_algorithms[REDUCE]));
        return (MPI_ERR_ARG);
    } /* switch */
}