1
1
openmpi/ompi/mca/coll/basic/coll_basic_allreduce.c
Rainer Keller 6c5532072a - Split the datatype engine into two parts: an MPI specific part in
OMPI
   and a language agnostic part in OPAL. The convertor is completely
   moved into OPAL.  This offers several benefits as described in RFC
   http://www.open-mpi.org/community/lists/devel/2009/07/6387.php
   namely:
    - Fewer basic types (int* and float* types, boolean and wchar
    - Fixing naming scheme to ompi-nomenclature.
    - Usability outside of the ompi-layer.
 - Due to the fixed nature of simple opal types, their information is
   completely
   known at compile time and therefore constified
 - With fewer datatypes (22), the actual sizes of bit-field types may be
   reduced
   from 64 to 32 bits, allowing reorganizing the opal_datatype
   structure, eliminating holes and keeping data required in convertor
   (upon send/recv) in one cacheline...
   This has implications to the convertor-datastructure and other parts
   of the code.
 - Several performance tests have been run, the netpipe latency does not
   change with
   this patch on Linux/x86-64 on the smoky cluster.
 - Extensive tests have been done to verify correctness (no new
   regressions) using:
   1. mpi_test_suite on linux/x86-64 using clean ompi-trunk and
    ompi-ddt:
    a. running both trunk and ompi-ddt resulted in no differences
       (except for MPI_SHORT_INT and MPI_TYPE_MIX_LB_UB do now run
       correctly).
    b. with --enable-memchecker and running under valgrind (one buglet
       when run with static found in test-suite, commited)
   2. ibm testsuite on linux/x86-64 using clean ompi-trunk and ompi-ddt:
      all passed (except for the dynamic/ tests failed!! as trunk/MTT)
   3. compilation and usage of HDF5 tests on Jaguar using PGI and
      PathScale compilers.
   4. compilation and usage on Scicortex.
 - Please note, that for the heterogeneous case, (-m32 compiled
   binaries/ompi), neither
   ompi-trunk, nor ompi-ddt branch would successfully launch.

This commit was SVN r21641.
2009-07-13 04:56:31 +00:00

220 строки
7.5 KiB
C

/*
* Copyright (c) 2004-2005 The Trustees of Indiana University and Indiana
* University Research and Technology
* Corporation. All rights reserved.
* Copyright (c) 2004-2005 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 "coll_basic.h"
#include "mpi.h"
#include "ompi/constants.h"
#include "ompi/datatype/ompi_datatype.h"
#include "ompi/op/op.h"
#include "ompi/mca/coll/coll.h"
#include "ompi/mca/coll/base/coll_tags.h"
#include "coll_basic.h"
#include "ompi/mca/pml/pml.h"
/*
* allreduce_intra
*
* Function: - allreduce using other MPI collectives
* Accepts: - same as MPI_Allreduce()
* Returns: - MPI_SUCCESS or error code
*/
int
mca_coll_basic_allreduce_intra(void *sbuf, void *rbuf, int count,
struct ompi_datatype_t *dtype,
struct ompi_op_t *op,
struct ompi_communicator_t *comm,
mca_coll_base_module_t *module)
{
int err;
/* Reduce to 0 and broadcast. */
if (MPI_IN_PLACE == sbuf) {
if (0 == ompi_comm_rank(comm)) {
err = comm->c_coll.coll_reduce(MPI_IN_PLACE, rbuf, count, dtype, op, 0, comm, comm->c_coll.coll_reduce_module);
} else {
err = comm->c_coll.coll_reduce(rbuf, NULL, count, dtype, op, 0, comm, comm->c_coll.coll_reduce_module);
}
} else {
err = comm->c_coll.coll_reduce(sbuf, rbuf, count, dtype, op, 0, comm, comm->c_coll.coll_reduce_module);
}
if (MPI_SUCCESS != err) {
return err;
}
return comm->c_coll.coll_bcast(rbuf, count, dtype, 0, comm, comm->c_coll.coll_bcast_module);
}
/*
* allreduce_inter
*
* Function: - allreduce using other MPI collectives
* Accepts: - same as MPI_Allreduce()
* Returns: - MPI_SUCCESS or error code
*/
int
mca_coll_basic_allreduce_inter(void *sbuf, void *rbuf, int count,
struct ompi_datatype_t *dtype,
struct ompi_op_t *op,
struct ompi_communicator_t *comm,
mca_coll_base_module_t *module)
{
int err, i, rank, root = 0, rsize;
ptrdiff_t lb, extent;
char *tmpbuf = NULL, *pml_buffer = NULL;
ompi_request_t *req[2];
mca_coll_basic_module_t *basic_module = (mca_coll_basic_module_t*) module;
ompi_request_t **reqs = basic_module->mccb_reqs;
rank = ompi_comm_rank(comm);
rsize = ompi_comm_remote_size(comm);
/* determine result of the remote group, you cannot
* use coll_reduce for inter-communicators, since than
* you would need to determine an order between the
* two groups (e.g. which group is providing the data
* and which one enters coll_reduce with providing
* MPI_PROC_NULL as root argument etc.) Here,
* we execute the data exchange for both groups
* simultaniously. */
/*****************************************************************/
if (rank == root) {
err = ompi_datatype_get_extent(dtype, &lb, &extent);
if (OMPI_SUCCESS != err) {
return OMPI_ERROR;
}
tmpbuf = (char *) malloc(count * extent);
if (NULL == tmpbuf) {
return OMPI_ERR_OUT_OF_RESOURCE;
}
pml_buffer = tmpbuf - lb;
/* Do a send-recv between the two root procs. to avoid deadlock */
err = MCA_PML_CALL(irecv(rbuf, count, dtype, 0,
MCA_COLL_BASE_TAG_ALLREDUCE, comm,
&(req[0])));
if (OMPI_SUCCESS != err) {
goto exit;
}
err = MCA_PML_CALL(isend(sbuf, count, dtype, 0,
MCA_COLL_BASE_TAG_ALLREDUCE,
MCA_PML_BASE_SEND_STANDARD,
comm, &(req[1])));
if (OMPI_SUCCESS != err) {
goto exit;
}
err = ompi_request_wait_all(2, req, MPI_STATUSES_IGNORE);
if (OMPI_SUCCESS != err) {
goto exit;
}
/* Loop receiving and calling reduction function (C or Fortran). */
for (i = 1; i < rsize; i++) {
err = MCA_PML_CALL(recv(pml_buffer, count, dtype, i,
MCA_COLL_BASE_TAG_ALLREDUCE, comm,
MPI_STATUS_IGNORE));
if (MPI_SUCCESS != err) {
goto exit;
}
/* Perform the reduction */
ompi_op_reduce(op, pml_buffer, rbuf, count, dtype);
}
} else {
/* If not root, send data to the root. */
err = MCA_PML_CALL(send(sbuf, count, dtype, root,
MCA_COLL_BASE_TAG_ALLREDUCE,
MCA_PML_BASE_SEND_STANDARD, comm));
if (OMPI_SUCCESS != err) {
goto exit;
}
}
/* now we have on one process the result of the remote group. To distribute
* the data to all processes in the local group, we exchange the data between
* the two root processes. They then send it to every other process in the
* remote group. */
/***************************************************************************/
if (rank == root) {
/* sendrecv between the two roots */
err = MCA_PML_CALL(irecv(pml_buffer, count, dtype, 0,
MCA_COLL_BASE_TAG_ALLREDUCE,
comm, &(req[1])));
if (OMPI_SUCCESS != err) {
goto exit;
}
err = MCA_PML_CALL(isend(rbuf, count, dtype, 0,
MCA_COLL_BASE_TAG_ALLREDUCE,
MCA_PML_BASE_SEND_STANDARD, comm,
&(req[0])));
if (OMPI_SUCCESS != err) {
goto exit;
}
err = ompi_request_wait_all(2, req, MPI_STATUSES_IGNORE);
if (OMPI_SUCCESS != err) {
goto exit;
}
/* distribute the data to other processes in remote group.
* Note that we start from 1 (not from zero), since zero
* has already the correct data AND we avoid a potential
* deadlock here.
*/
if (rsize > 1) {
for (i = 1; i < rsize; i++) {
err = MCA_PML_CALL(isend(pml_buffer, count, dtype, i,
MCA_COLL_BASE_TAG_ALLREDUCE,
MCA_PML_BASE_SEND_STANDARD, comm,
&reqs[i - 1]));
if (OMPI_SUCCESS != err) {
goto exit;
}
}
err =
ompi_request_wait_all(rsize - 1, reqs,
MPI_STATUSES_IGNORE);
if (OMPI_SUCCESS != err) {
goto exit;
}
}
} else {
err = MCA_PML_CALL(recv(rbuf, count, dtype, root,
MCA_COLL_BASE_TAG_ALLREDUCE,
comm, MPI_STATUS_IGNORE));
}
exit:
if (NULL != tmpbuf) {
free(tmpbuf);
}
return err;
}