1
1
openmpi/ompi/mca/coll/base/coll_base_util.c

109 строки
4.2 KiB
C
Исходник Обычный вид История

/*
* Copyright (c) 2004-2005 The Trustees of Indiana University and Indiana
* University Research and Technology
* Corporation. All rights reserved.
* Copyright (c) 2004-2015 The University of Tennessee and The University
* of Tennessee Research Foundation. All rights
* reserved.
2015-06-24 06:59:57 +03:00
* 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 (c) 2014-2015 Research Organization for Information Science
* and Technology (RIST). All rights reserved.
* $COPYRIGHT$
2015-06-24 06:59:57 +03:00
*
* Additional copyrights may follow
2015-06-24 06:59:57 +03:00
*
* $HEADER$
*/
#include "ompi_config.h"
#include "mpi.h"
#include "ompi/constants.h"
- 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 08:56:31 +04:00
#include "ompi/datatype/ompi_datatype.h"
#include "ompi/communicator/communicator.h"
#include "ompi/mca/coll/base/coll_tags.h"
#include "ompi/mca/coll/base/coll_base_functions.h"
#include "ompi/mca/pml/pml.h"
#include "coll_base_util.h"
2015-06-24 06:59:57 +03:00
int ompi_coll_base_sendrecv_nonzero_actual( void* sendbuf, size_t scount,
ompi_datatype_t* sdatatype,
int dest, int stag,
2015-06-24 06:59:57 +03:00
void* recvbuf, size_t rcount,
ompi_datatype_t* rdatatype,
int source, int rtag,
struct ompi_communicator_t* comm,
ompi_status_public_t* status )
{ /* post receive first, then send, then waitall... should be fast (I hope) */
int err, line = 0, nreqs = 0;
size_t typesize;
ompi_request_t* reqs[2], **req = reqs;
ompi_status_public_t statuses[2];
/* post new irecv */
ompi_datatype_type_size(rdatatype, &typesize);
if (0 != rcount && 0 != typesize) {
2015-06-24 06:59:57 +03:00
err = MCA_PML_CALL(irecv( recvbuf, rcount, rdatatype, source, rtag,
comm, req++));
++nreqs;
if (err != MPI_SUCCESS) { line = __LINE__; goto error_handler; }
}
/* send data to children */
ompi_datatype_type_size(sdatatype, &typesize);
if (0 != scount && 0 != typesize) {
2015-06-24 06:59:57 +03:00
err = MCA_PML_CALL(isend( sendbuf, scount, sdatatype, dest, stag,
MCA_PML_BASE_SEND_STANDARD, comm, req++));
++nreqs;
if (err != MPI_SUCCESS) { line = __LINE__; goto error_handler; }
}
if (0 != nreqs) {
err = ompi_request_wait_all( nreqs, reqs, statuses );
if( MPI_ERR_IN_STATUS == err ) {
/* As we use wait_all we will get MPI_ERR_IN_STATUS which is not an error
* code that we can propagate up the stack. Instead, look for the real
* error code from the MPI_ERROR in the status.
*/
int err_index = 0;
if( MPI_SUCCESS == statuses[0].MPI_ERROR ) {
err_index = 1;
}
if (MPI_STATUS_IGNORE != status) {
*status = statuses[err_index];
}
err = statuses[err_index].MPI_ERROR;
OPAL_OUTPUT ((ompi_coll_base_framework.framework_output, "%s:%d: Error %d occurred in the %s"
" stage of ompi_coll_base_sendrecv_zero\n",
__FILE__, line, err, (0 == err_index ? "receive" : "send")));
return err;
}
if (err != MPI_SUCCESS) { line = __LINE__; goto error_handler; }
if (MPI_STATUS_IGNORE != status) {
*status = statuses[0];
}
} else {
if( MPI_STATUS_IGNORE != status )
*status = ompi_status_empty;
}
return (MPI_SUCCESS);
error_handler:
/* Error discovered during the posting of the irecv or isend,
* and no status is available.
*/
OPAL_OUTPUT ((ompi_coll_base_framework.framework_output, "%s:%d: Error %d occurred\n",
__FILE__, line, err));
if (MPI_STATUS_IGNORE != status) {
status->MPI_ERROR = err;
}
return (err);
}