/* -*- Mode: C; c-basic-offset:4 ; indent-tabs-mode:nil ; -*- */
/*
* (C) 2010 by Argonne National Laboratory.
* See COPYRIGHT in top-level directory.
*/
#include "mpiimpl.h"
/*
=== BEGIN_MPI_T_CVAR_INFO_BLOCK ===
cvars:
- name : MPIR_CVAR_IREDUCE_SCATTER_BLOCK_RECEXCH_KVAL
category : COLLECTIVE
type : int
default : 2
class : device
verbosity : MPI_T_VERBOSITY_USER_BASIC
scope : MPI_T_SCOPE_ALL_EQ
description : >-
k value for recursive exchange based ireduce_scatter_block
- name : MPIR_CVAR_IREDUCE_SCATTER_BLOCK_INTRA_ALGORITHM
category : COLLECTIVE
type : string
default : auto
class : device
verbosity : MPI_T_VERBOSITY_USER_BASIC
scope : MPI_T_SCOPE_ALL_EQ
description : |-
Variable to select ireduce_scatter_block algorithm
auto - Internal algorithm selection
noncommutative - Force noncommutative algorithm
recursive_doubling - Force recursive doubling algorithm
pairwise - Force pairwise algorithm
recursive_halving - Force recursive halving algorithm
recexch - Force generic transport recursive exchange algorithm
- name : MPIR_CVAR_IREDUCE_SCATTER_BLOCK_INTER_ALGORITHM
category : COLLECTIVE
type : string
default : auto
class : device
verbosity : MPI_T_VERBOSITY_USER_BASIC
scope : MPI_T_SCOPE_ALL_EQ
description : |-
Variable to select ireduce_scatter_block algorithm
auto - Internal algorithm selection
remote_reduce_local_scatterv - Force remote-reduce-local-scatterv algorithm
- name : MPIR_CVAR_IREDUCE_SCATTER_BLOCK_DEVICE_COLLECTIVE
category : COLLECTIVE
type : boolean
default : true
class : device
verbosity : MPI_T_VERBOSITY_USER_BASIC
scope : MPI_T_SCOPE_ALL_EQ
description : >-
If set to true, MPI_Ireduce_scatter_block will allow the device to override the
MPIR-level collective algorithms. The device still has the
option to call the MPIR-level algorithms manually.
If set to false, the device-level ireduce_scatter_block function will not be
called.
=== END_MPI_T_CVAR_INFO_BLOCK ===
*/
/* -- Begin Profiling Symbol Block for routine MPI_Ireduce_scatter_block */
#if defined(HAVE_PRAGMA_WEAK)
#pragma weak MPI_Ireduce_scatter_block = PMPI_Ireduce_scatter_block
#elif defined(HAVE_PRAGMA_HP_SEC_DEF)
#pragma _HP_SECONDARY_DEF PMPI_Ireduce_scatter_block MPI_Ireduce_scatter_block
#elif defined(HAVE_PRAGMA_CRI_DUP)
#pragma _CRI duplicate MPI_Ireduce_scatter_block as PMPI_Ireduce_scatter_block
#elif defined(HAVE_WEAK_ATTRIBUTE)
int MPI_Ireduce_scatter_block(const void *sendbuf, void *recvbuf, int recvcount,
MPI_Datatype datatype, MPI_Op op, MPI_Comm comm,
MPI_Request * request)
__attribute__ ((weak, alias("PMPI_Ireduce_scatter_block")));
#endif
/* -- End Profiling Symbol Block */
/* Define MPICH_MPI_FROM_PMPI if weak symbols are not supported to build
the MPI routines */
#ifndef MPICH_MPI_FROM_PMPI
#undef MPI_Ireduce_scatter_block
#define MPI_Ireduce_scatter_block PMPI_Ireduce_scatter_block
#undef FUNCNAME
#define FUNCNAME MPIR_Ireduce_scatter_block_sched_intra_auto
#undef FCNAME
#define FCNAME MPL_QUOTE(FUNCNAME)
int MPIR_Ireduce_scatter_block_sched_intra_auto(const void *sendbuf, void *recvbuf, int recvcount,
MPI_Datatype datatype, MPI_Op op,
MPIR_Comm * comm_ptr, MPIR_Sched_t s)
{
int mpi_errno = MPI_SUCCESS;
int is_commutative;
int total_count, type_size, nbytes;
int comm_size;
is_commutative = MPIR_Op_is_commutative(op);
comm_size = comm_ptr->local_size;
total_count = recvcount * comm_size;
if (total_count == 0) {
goto fn_exit;
}
MPIR_Datatype_get_size_macro(datatype, type_size);
nbytes = total_count * type_size;
/* select an appropriate algorithm based on commutivity and message size */
if (is_commutative && (nbytes < MPIR_CVAR_REDUCE_SCATTER_COMMUTATIVE_LONG_MSG_SIZE)) {
mpi_errno =
MPIR_Ireduce_scatter_block_sched_intra_recursive_halving(sendbuf, recvbuf, recvcount,
datatype, op, comm_ptr, s);
if (mpi_errno)
MPIR_ERR_POP(mpi_errno);
} else if (is_commutative && (nbytes >= MPIR_CVAR_REDUCE_SCATTER_COMMUTATIVE_LONG_MSG_SIZE)) {
mpi_errno =
MPIR_Ireduce_scatter_block_sched_intra_pairwise(sendbuf, recvbuf, recvcount, datatype,
op, comm_ptr, s);
if (mpi_errno)
MPIR_ERR_POP(mpi_errno);
} else { /* (!is_commutative) */
if (MPL_is_pof2(comm_size, NULL)) {
/* noncommutative, pof2 size */
mpi_errno =
MPIR_Ireduce_scatter_block_sched_intra_noncommutative(sendbuf, recvbuf, recvcount,
datatype, op, comm_ptr, s);
if (mpi_errno)
MPIR_ERR_POP(mpi_errno);
} else {
/* noncommutative and non-pof2, use recursive doubling. */
mpi_errno =
MPIR_Ireduce_scatter_block_sched_intra_recursive_doubling(sendbuf, recvbuf,
recvcount, datatype, op,
comm_ptr, s);
if (mpi_errno)
MPIR_ERR_POP(mpi_errno);
}
}
fn_exit:
return mpi_errno;
fn_fail:
goto fn_exit;
}
#undef FUNCNAME
#define FUNCNAME MPIR_Ireduce_scatter_block_sched_inter_auto
#undef FCNAME
#define FCNAME MPL_QUOTE(FUNCNAME)
int MPIR_Ireduce_scatter_block_sched_inter_auto(const void *sendbuf, void *recvbuf, int recvcount,
MPI_Datatype datatype, MPI_Op op,
MPIR_Comm * comm_ptr, MPIR_Sched_t s)
{
int mpi_errno = MPI_SUCCESS;
mpi_errno =
MPIR_Ireduce_scatter_block_sched_inter_remote_reduce_local_scatterv(sendbuf, recvbuf,
recvcount, datatype, op,
comm_ptr, s);
return mpi_errno;
}
#undef FUNCNAME
#define FUNCNAME MPIR_Ireduce_scatter_block_sched_impl
#undef FCNAME
#define FCNAME MPL_QUOTE(FUNCNAME)
int MPIR_Ireduce_scatter_block_sched_impl(const void *sendbuf, void *recvbuf, int recvcount,
MPI_Datatype datatype, MPI_Op op, MPIR_Comm * comm_ptr,
MPIR_Sched_t s)
{
int mpi_errno = MPI_SUCCESS;
if (comm_ptr->comm_kind == MPIR_COMM_KIND__INTRACOMM) {
/* intracommunicator */
switch (MPIR_Ireduce_scatter_block_intra_algo_choice) {
case MPIR_IREDUCE_SCATTER_BLOCK_INTRA_ALGO_NONCOMMUTATIVE:
mpi_errno = MPIR_Ireduce_scatter_block_sched_intra_noncommutative(sendbuf, recvbuf,
recvcount,
datatype, op,
comm_ptr, s);
break;
case MPIR_IREDUCE_SCATTER_BLOCK_INTRA_ALGO_PAIRWISE:
mpi_errno = MPIR_Ireduce_scatter_block_sched_intra_pairwise(sendbuf, recvbuf,
recvcount, datatype, op,
comm_ptr, s);
break;
case MPIR_IREDUCE_SCATTER_BLOCK_INTRA_ALGO_RECURSIVE_HALVING:
mpi_errno =
MPIR_Ireduce_scatter_block_sched_intra_recursive_halving(sendbuf, recvbuf,
recvcount, datatype,
op, comm_ptr, s);
break;
case MPIR_IREDUCE_SCATTER_BLOCK_INTRA_ALGO_RECURSIVE_DOUBLING:
mpi_errno =
MPIR_Ireduce_scatter_block_sched_intra_recursive_doubling(sendbuf, recvbuf,
recvcount, datatype,
op, comm_ptr, s);
break;
case MPIR_IREDUCE_SCATTER_BLOCK_INTRA_ALGO_AUTO:
MPL_FALLTHROUGH;
default:
mpi_errno = MPIR_Ireduce_scatter_block_sched_intra_auto(sendbuf, recvbuf,
recvcount, datatype, op,
comm_ptr, s);
break;
}
} else {
/* intercommunicator */
switch (MPIR_Ireduce_scatter_block_inter_algo_choice) {
case MPIR_IREDUCE_SCATTER_BLOCK_INTER_ALGO_REMOTE_REDUCE_LOCAL_SCATTERV:
mpi_errno =
MPIR_Ireduce_scatter_block_sched_inter_remote_reduce_local_scatterv(sendbuf,
recvbuf,
recvcount,
datatype,
op,
comm_ptr,
s);
break;
case MPIR_IREDUCE_SCATTER_BLOCK_INTER_ALGO_AUTO:
MPL_FALLTHROUGH;
default:
mpi_errno = MPIR_Ireduce_scatter_block_sched_inter_auto(sendbuf, recvbuf,
recvcount, datatype, op,
comm_ptr, s);
}
}
return mpi_errno;
}
#undef FUNCNAME
#define FUNCNAME MPIR_Ireduce_scatter_block_sched
#undef FCNAME
#define FCNAME MPL_QUOTE(FUNCNAME)
int MPIR_Ireduce_scatter_block_sched(const void *sendbuf, void *recvbuf, int recvcount,
MPI_Datatype datatype, MPI_Op op, MPIR_Comm * comm_ptr,
MPIR_Sched_t s)
{
int mpi_errno = MPI_SUCCESS;
if (MPIR_CVAR_IREDUCE_SCATTER_BLOCK_DEVICE_COLLECTIVE && MPIR_CVAR_DEVICE_COLLECTIVES) {
mpi_errno = MPID_Ireduce_scatter_block_sched(sendbuf, recvbuf, recvcount, datatype, op,
comm_ptr, s);
} else {
mpi_errno = MPIR_Ireduce_scatter_block_sched_impl(sendbuf, recvbuf, recvcount, datatype, op,
comm_ptr, s);
}
return mpi_errno;
}
#undef FUNCNAME
#define FUNCNAME MPIR_Ireduce_scatter_block_impl
#undef FCNAME
#define FCNAME MPL_QUOTE(FUNCNAME)
int MPIR_Ireduce_scatter_block_impl(const void *sendbuf, void *recvbuf,
int recvcount, MPI_Datatype datatype,
MPI_Op op, MPIR_Comm * comm_ptr, MPIR_Request ** request)
{
int mpi_errno = MPI_SUCCESS;
int tag = -1;
int is_commutative = MPIR_Op_is_commutative(op);
MPIR_Sched_t s = MPIR_SCHED_NULL;
*request = NULL;
/* If the user picks one of the transport-enabled algorithms, branch there
* before going down to the MPIR_Sched-based algorithms. */
/* TODO - Eventually the intention is to replace all of the
* MPIR_Sched-based algorithms with transport-enabled algorithms, but that
* will require sufficient performance testing and replacement algorithms. */
if (comm_ptr->comm_kind == MPIR_COMM_KIND__INTRACOMM) {
/* intracommunicator */
switch (MPIR_Ireduce_scatter_block_intra_algo_choice) {
case MPIR_IREDUCE_SCATTER_BLOCK_INTRA_ALGO_GENTRAN_RECEXCH:
if (is_commutative) {
mpi_errno =
MPIR_Ireduce_scatter_block_intra_recexch(sendbuf, recvbuf, recvcount,
datatype, op, comm_ptr, request);
if (mpi_errno)
MPIR_ERR_POP(mpi_errno);
goto fn_exit;
}
break;
default:
/* go down to the MPIR_Sched-based algorithms */
break;
}
}
mpi_errno = MPIR_Sched_next_tag(comm_ptr, &tag);
if (mpi_errno)
MPIR_ERR_POP(mpi_errno);
mpi_errno = MPIR_Sched_create(&s);
if (mpi_errno)
MPIR_ERR_POP(mpi_errno);
mpi_errno =
MPIR_Ireduce_scatter_block_sched(sendbuf, recvbuf, recvcount, datatype, op, comm_ptr, s);
if (mpi_errno)
MPIR_ERR_POP(mpi_errno);
mpi_errno = MPIR_Sched_start(&s, comm_ptr, tag, request);
if (mpi_errno)
MPIR_ERR_POP(mpi_errno);
fn_exit:
return mpi_errno;
fn_fail:
goto fn_exit;
}
#undef FUNCNAME
#define FUNCNAME MPIR_Ireduce_scatter_block
#undef FCNAME
#define FCNAME MPL_QUOTE(FUNCNAME)
int MPIR_Ireduce_scatter_block(const void *sendbuf, void *recvbuf,
int recvcount, MPI_Datatype datatype,
MPI_Op op, MPIR_Comm * comm_ptr, MPIR_Request ** request)
{
int mpi_errno = MPI_SUCCESS;
if (MPIR_CVAR_IREDUCE_SCATTER_BLOCK_DEVICE_COLLECTIVE && MPIR_CVAR_DEVICE_COLLECTIVES) {
mpi_errno = MPID_Ireduce_scatter_block(sendbuf, recvbuf, recvcount, datatype, op, comm_ptr,
request);
} else {
mpi_errno = MPIR_Ireduce_scatter_block_impl(sendbuf, recvbuf, recvcount, datatype, op,
comm_ptr, request);
}
return mpi_errno;
}
#endif /* MPICH_MPI_FROM_PMPI */
#undef FUNCNAME
#define FUNCNAME MPI_Ireduce_scatter_block
#undef FCNAME
#define FCNAME MPL_QUOTE(FUNCNAME)
/*@
MPI_Ireduce_scatter_block - Combines values and scatters the results in
a nonblocking way
Input Parameters:
+ sendbuf - starting address of the send buffer (choice)
. recvcount - element count per block (non-negative integer)
. datatype - data type of elements of input buffer (handle)
. op - operation (handle)
- comm - communicator (handle)
Output Parameters:
+ recvbuf - starting address of the receive buffer (choice)
- request - communication request (handle)
.N ThreadSafe
.N Fortran
.N Errors
@*/
int MPI_Ireduce_scatter_block(const void *sendbuf, void *recvbuf,
int recvcount,
MPI_Datatype datatype, MPI_Op op, MPI_Comm comm,
MPI_Request * request)
{
int mpi_errno = MPI_SUCCESS;
MPIR_Comm *comm_ptr = NULL;
MPIR_Request *request_ptr = NULL;
MPIR_FUNC_TERSE_STATE_DECL(MPID_STATE_MPI_IREDUCE_SCATTER_BLOCK);
MPID_THREAD_CS_ENTER(GLOBAL, MPIR_THREAD_GLOBAL_ALLFUNC_MUTEX);
MPIR_FUNC_TERSE_ENTER(MPID_STATE_MPI_IREDUCE_SCATTER_BLOCK);
/* Validate parameters, especially handles needing to be converted */
#ifdef HAVE_ERROR_CHECKING
{
MPID_BEGIN_ERROR_CHECKS;
{
MPIR_ERRTEST_DATATYPE(datatype, "datatype", mpi_errno);
MPIR_ERRTEST_OP(op, mpi_errno);
MPIR_ERRTEST_COMM(comm, mpi_errno);
/* TODO more checks may be appropriate */
}
MPID_END_ERROR_CHECKS;
}
#endif /* HAVE_ERROR_CHECKING */
/* Convert MPI object handles to object pointers */
MPIR_Comm_get_ptr(comm, comm_ptr);
MPIR_Assert(comm_ptr != NULL);
/* Validate parameters and objects (post conversion) */
#ifdef HAVE_ERROR_CHECKING
{
MPID_BEGIN_ERROR_CHECKS;
{
MPIR_Comm_valid_ptr(comm_ptr, mpi_errno, FALSE);
if (HANDLE_GET_KIND(datatype) != HANDLE_KIND_BUILTIN) {
MPIR_Datatype *datatype_ptr = NULL;
MPIR_Datatype_get_ptr(datatype, datatype_ptr);
MPIR_Datatype_valid_ptr(datatype_ptr, mpi_errno);
if (mpi_errno != MPI_SUCCESS)
goto fn_fail;
MPIR_Datatype_committed_ptr(datatype_ptr, mpi_errno);
if (mpi_errno != MPI_SUCCESS)
goto fn_fail;
}
if (HANDLE_GET_KIND(op) != HANDLE_KIND_BUILTIN) {
MPIR_Op *op_ptr = NULL;
MPIR_Op_get_ptr(op, op_ptr);
MPIR_Op_valid_ptr(op_ptr, mpi_errno);
} else if (HANDLE_GET_KIND(op) == HANDLE_KIND_BUILTIN) {
mpi_errno = (*MPIR_OP_HDL_TO_DTYPE_FN(op)) (datatype);
}
if (mpi_errno != MPI_SUCCESS)
goto fn_fail;
MPIR_ERRTEST_ARGNULL(request, "request", mpi_errno);
if (comm_ptr->comm_kind == MPIR_COMM_KIND__INTRACOMM && sendbuf != MPI_IN_PLACE &&
recvcount != 0) {
MPIR_ERRTEST_ALIAS_COLL(sendbuf, recvbuf, mpi_errno)
}
/* TODO more checks may be appropriate (counts, in_place, etc) */
}
MPID_END_ERROR_CHECKS;
}
#endif /* HAVE_ERROR_CHECKING */
/* ... body of routine ... */
mpi_errno = MPIR_Ireduce_scatter_block(sendbuf, recvbuf, recvcount, datatype, op, comm_ptr,
&request_ptr);
if (mpi_errno)
MPIR_ERR_POP(mpi_errno);
/* create a complete request, if needed */
if (!request_ptr)
request_ptr = MPIR_Request_create_complete(MPIR_REQUEST_KIND__COLL);
/* return the handle of the request to the user */
*request = request_ptr->handle;
/* ... end of body of routine ... */
fn_exit:
MPIR_FUNC_TERSE_EXIT(MPID_STATE_MPI_IREDUCE_SCATTER_BLOCK);
MPID_THREAD_CS_EXIT(GLOBAL, MPIR_THREAD_GLOBAL_ALLFUNC_MUTEX);
return mpi_errno;
fn_fail:
/* --BEGIN ERROR HANDLING-- */
#ifdef HAVE_ERROR_CHECKING
{
mpi_errno =
MPIR_Err_create_code(mpi_errno, MPIR_ERR_RECOVERABLE, FCNAME, __LINE__, MPI_ERR_OTHER,
"**mpi_ireduce_scatter_block",
"**mpi_ireduce_scatter_block %p %p %d %D %O %C %p", sendbuf,
recvbuf, recvcount, datatype, op, comm, request);
}
#endif
mpi_errno = MPIR_Err_return_comm(comm_ptr, FCNAME, mpi_errno);
goto fn_exit;
/* --END ERROR HANDLING-- */
}