552c9ca5a0
WHAT: Open our low-level communication infrastructure by moving all necessary components (btl/rcache/allocator/mpool) down in OPAL All the components required for inter-process communications are currently deeply integrated in the OMPI layer. Several groups/institutions have express interest in having a more generic communication infrastructure, without all the OMPI layer dependencies. This communication layer should be made available at a different software level, available to all layers in the Open MPI software stack. As an example, our ORTE layer could replace the current OOB and instead use the BTL directly, gaining access to more reactive network interfaces than TCP. Similarly, external software libraries could take advantage of our highly optimized AM (active message) communication layer for their own purpose. UTK with support from Sandia, developped a version of Open MPI where the entire communication infrastucture has been moved down to OPAL (btl/rcache/allocator/mpool). Most of the moved components have been updated to match the new schema, with few exceptions (mainly BTLs where I have no way of compiling/testing them). Thus, the completion of this RFC is tied to being able to completing this move for all BTLs. For this we need help from the rest of the Open MPI community, especially those supporting some of the BTLs. A non-exhaustive list of BTLs that qualify here is: mx, portals4, scif, udapl, ugni, usnic. This commit was SVN r32317.
79 строки
2.4 KiB
C
79 строки
2.4 KiB
C
/* -*- Mode: C; c-basic-offset:4 ; indent-tabs-mode:nil -*- */
|
|
/*
|
|
* Copyright (c) 2013 Los Alamos National Security, LLC. All rights
|
|
* reserved.
|
|
* $COPYRIGHT$
|
|
*
|
|
* Additional copyrights may follow
|
|
*
|
|
* $HEADER$
|
|
*/
|
|
|
|
#include "opal_config.h"
|
|
#include "btl_scif_frag.h"
|
|
|
|
#define lmin(a,b) ((a) < (b) ? (a) : (b))
|
|
|
|
/**
|
|
* Initiate a put operation.
|
|
*
|
|
* @param btl (IN) BTL module
|
|
* @param endpoint (IN) BTL addressing information
|
|
* @param descriptor (IN) Description of the data to be transferred
|
|
*/
|
|
int mca_btl_scif_put (struct mca_btl_base_module_t *btl,
|
|
struct mca_btl_base_endpoint_t *endpoint,
|
|
struct mca_btl_base_descriptor_t *des) {
|
|
mca_btl_scif_segment_t *src = (mca_btl_scif_segment_t *) des->des_local;
|
|
mca_btl_scif_segment_t *dst = (mca_btl_scif_segment_t *) des->des_remote;
|
|
size_t len = lmin (src->base.seg_len, dst->base.seg_len);
|
|
int rc, mark, flags = 0;
|
|
off_t roffset, loffset;
|
|
#if defined(SCIF_TIMING)
|
|
struct timespec ts;
|
|
|
|
clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &ts);
|
|
|
|
mca_btl_scif_component.put_count++;
|
|
#endif
|
|
|
|
BTL_VERBOSE(("Using DMA Put for frag %p", (void *) des));
|
|
|
|
roffset = dst->scif_offset + (off_t)(dst->orig_ptr - dst->base.seg_addr.lval);
|
|
loffset = src->scif_offset + (off_t)(src->orig_ptr - src->base.seg_addr.lval);
|
|
|
|
if (mca_btl_scif_component.rma_use_cpu) {
|
|
flags = SCIF_RMA_USECPU;
|
|
}
|
|
|
|
if (mca_btl_scif_component.rma_sync) {
|
|
flags |= SCIF_RMA_SYNC;
|
|
}
|
|
|
|
/* start the write */
|
|
rc = scif_writeto (endpoint->scif_epd, loffset, len, roffset, flags);
|
|
if (OPAL_UNLIKELY(-1 == rc)) {
|
|
return OPAL_ERROR;
|
|
}
|
|
|
|
/* always call the callback function */
|
|
des->des_flags |= MCA_BTL_DES_SEND_ALWAYS_CALLBACK;
|
|
|
|
/* according to the scif documentation is is better to use a fence rather
|
|
* than using the SCIF_RMA_SYNC flag with scif_writeto */
|
|
if (!(flags & SCIF_RMA_SYNC)) {
|
|
scif_fence_mark (endpoint->scif_epd, SCIF_FENCE_INIT_SELF, &mark);
|
|
scif_fence_wait (endpoint->scif_epd, mark);
|
|
}
|
|
|
|
#if defined(SCIF_TIMING)
|
|
SCIF_UPDATE_TIMER(mca_btl_scif_component.put_time,
|
|
mca_btl_scif_component.put_time_max, ts);
|
|
#endif
|
|
|
|
/* since we completed the fence the RMA operation is complete */
|
|
mca_btl_scif_frag_complete ((mca_btl_scif_base_frag_t *) des, OPAL_SUCCESS);
|
|
|
|
return OPAL_SUCCESS;
|
|
}
|