433f8a7694
receive queues are shared among all PMLs, they are declared in the base PML, and the selected PML is in charge of initializing and releasing them. The CM PML is slightly different compared with OB1 or DR. Internally it use 2 different types of requests: light and heavy. However, now with this patch both types of requests are stored in the same queue, and cast appropriately on the allocation macro. This means we might use less memory than we allocate, but in exchange we got full support for most of the parallel debuggers. Another thing with this patch, is that now for all PML (CM included) the basic PML requests start with the same fields, and they are declared in the same order in the request structure. Moreover, the fields have been moved in such a way that only one volatile/atomic will exist per line of cache (hopefully). This commit was SVN r15346.
160 строки
3.8 KiB
C
160 строки
3.8 KiB
C
/*
|
|
* Copyright (c) 2006-2007 The Trustees of Indiana University and Indiana
|
|
* University Research and Technology
|
|
* Corporation. All rights reserved.
|
|
* Copyright (c) 2004-2007 The University of Tennessee and The University
|
|
* of Tennessee Research Foundation. All rights
|
|
* reserved.
|
|
* Copyright (c) 2004-2006 The Regents of the University of California.
|
|
* All rights reserved.
|
|
* $COPYRIGHT$
|
|
*
|
|
* Additional copyrights may follow
|
|
*
|
|
* $HEADER$
|
|
*/
|
|
|
|
#include "ompi_config.h"
|
|
|
|
#include "ompi/communicator/communicator.h"
|
|
#include "opal/class/opal_list.h"
|
|
#include "ompi/mca/pml/base/pml_base_request.h"
|
|
#include "ompi/mca/pml/base/pml_base_bsend.h"
|
|
#include "ompi/mca/pml/base/base.h"
|
|
|
|
#include "pml_cm.h"
|
|
#include "pml_cm_sendreq.h"
|
|
#include "pml_cm_recvreq.h"
|
|
|
|
ompi_pml_cm_t ompi_pml_cm = {
|
|
{
|
|
mca_pml_cm_add_procs,
|
|
mca_pml_cm_del_procs,
|
|
mca_pml_cm_enable,
|
|
mca_pml_cm_progress,
|
|
mca_pml_cm_add_comm,
|
|
mca_pml_cm_del_comm,
|
|
mca_pml_cm_irecv_init,
|
|
mca_pml_cm_irecv,
|
|
mca_pml_cm_recv,
|
|
mca_pml_cm_isend_init,
|
|
mca_pml_cm_isend,
|
|
mca_pml_cm_send,
|
|
mca_pml_cm_iprobe,
|
|
mca_pml_cm_probe,
|
|
mca_pml_cm_start,
|
|
mca_pml_cm_dump,
|
|
NULL,
|
|
0,
|
|
0
|
|
}
|
|
};
|
|
|
|
|
|
int
|
|
mca_pml_cm_enable(bool enable)
|
|
{
|
|
/* BWB - FIX ME - need to have this actually do something,
|
|
maybe? */
|
|
return OMPI_SUCCESS;
|
|
}
|
|
|
|
|
|
int
|
|
mca_pml_cm_add_comm(ompi_communicator_t* comm)
|
|
{
|
|
/* setup our per-communicator data */
|
|
comm->c_pml_comm = NULL;
|
|
|
|
return OMPI_SUCCESS;
|
|
}
|
|
|
|
|
|
int
|
|
mca_pml_cm_del_comm(ompi_communicator_t* comm)
|
|
{
|
|
/* clean up our per-communicator data */
|
|
comm->c_pml_comm = NULL;
|
|
|
|
return OMPI_SUCCESS;
|
|
}
|
|
|
|
|
|
int
|
|
mca_pml_cm_add_procs(struct ompi_proc_t** procs, size_t nprocs)
|
|
{
|
|
int ret;
|
|
size_t i;
|
|
struct mca_mtl_base_endpoint_t **endpoints;
|
|
|
|
#if OMPI_ENABLE_HETEROGENEOUS_SUPPORT
|
|
for (i = 0 ; i < nprocs ; ++i) {
|
|
if (procs[i]->proc_arch != ompi_proc_local()->proc_arch) {
|
|
return OMPI_ERR_NOT_SUPPORTED;
|
|
}
|
|
}
|
|
#endif
|
|
|
|
/* make sure remote procs are using the same PML as us */
|
|
if (OMPI_SUCCESS != (ret = mca_pml_base_pml_check_selected("cm",
|
|
procs,
|
|
nprocs))) {
|
|
return ret;
|
|
}
|
|
|
|
endpoints = (struct mca_mtl_base_endpoint_t**)malloc(nprocs * sizeof(struct mca_mtl_base_endpoint_t*));
|
|
if (NULL == endpoints) return OMPI_ERROR;
|
|
|
|
#if OMPI_ENABLE_DEBUG
|
|
for (i = 0 ; i < nprocs ; ++i) {
|
|
endpoints[i] = NULL;
|
|
}
|
|
#endif
|
|
|
|
ret = OMPI_MTL_CALL(add_procs(ompi_mtl, nprocs, procs, endpoints));
|
|
if (OMPI_SUCCESS != ret) {
|
|
free(endpoints);
|
|
return ret;
|
|
}
|
|
|
|
for (i = 0 ; i < nprocs ; ++i) {
|
|
procs[i]->proc_pml = (struct mca_pml_base_endpoint_t*) endpoints[i];
|
|
}
|
|
|
|
free(endpoints);
|
|
return OMPI_SUCCESS;
|
|
}
|
|
|
|
|
|
int
|
|
mca_pml_cm_del_procs(struct ompi_proc_t** procs, size_t nprocs)
|
|
{
|
|
int ret;
|
|
size_t i;
|
|
struct mca_mtl_base_endpoint_t **endpoints;
|
|
|
|
endpoints = (struct mca_mtl_base_endpoint_t**)malloc(nprocs * sizeof(struct mca_mtl_base_endpoint_t*));
|
|
if (NULL == endpoints) return OMPI_ERROR;
|
|
|
|
for (i = 0 ; i < nprocs ; ++i) {
|
|
endpoints[i] = (struct mca_mtl_base_endpoint_t*) procs[i]->proc_pml;
|
|
}
|
|
|
|
ret = OMPI_MTL_CALL(del_procs(ompi_mtl, nprocs, procs, endpoints));
|
|
if (OMPI_SUCCESS != ret) {
|
|
free(endpoints);
|
|
return ret;
|
|
}
|
|
|
|
free(endpoints);
|
|
return OMPI_SUCCESS;
|
|
}
|
|
|
|
|
|
/* print any available useful information from this communicator */
|
|
int
|
|
mca_pml_cm_dump(struct ompi_communicator_t* comm, int verbose)
|
|
{
|
|
return OMPI_ERR_NOT_IMPLEMENTED;
|
|
}
|