Skip to content

Make datatype copy stream-aware #12570

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Make datatype copy stream-aware #12570

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 13 additions & 5 deletions ompi/datatype/ompi_datatype.h
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
/* -*- Mode: C; c-basic-offset:4 ; indent-tabs-mode:nil -*- */
/*
* Copyright (c) 2009-2019 The University of Tennessee and The University
* Copyright (c) 2009-2024 The University of Tennessee and The University
* of Tennessee Research Foundation. All rights
* reserved.
* Copyright (c) 2009 Oak Ridge National Labs. All rights reserved.
Expand Down Expand Up @@ -275,8 +275,9 @@ ompi_datatype_set_element_count( const ompi_datatype_t* type, size_t count, size
}

static inline int32_t
ompi_datatype_copy_content_same_ddt( const ompi_datatype_t* type, size_t count,
char* pDestBuf, char* pSrcBuf )
ompi_datatype_copy_content_same_ddt_stream( const ompi_datatype_t* type, size_t count,
char* pDestBuf, char* pSrcBuf,
opal_accelerator_stream_t *stream )
{
int32_t length, rc;
ptrdiff_t extent;
Expand All @@ -285,8 +286,8 @@ ompi_datatype_copy_content_same_ddt( const ompi_datatype_t* type, size_t count,
while( 0 != count ) {
length = INT_MAX;
if( ((size_t)length) > count ) length = (int32_t)count;
rc = opal_datatype_copy_content_same_ddt( &type->super, length,
pDestBuf, pSrcBuf );
rc = opal_datatype_copy_content_same_ddt_stream( &type->super, length,
pDestBuf, pSrcBuf, stream );
if( 0 != rc ) return rc;
pDestBuf += ((ptrdiff_t)length) * extent;
pSrcBuf += ((ptrdiff_t)length) * extent;
Expand All @@ -295,6 +296,13 @@ ompi_datatype_copy_content_same_ddt( const ompi_datatype_t* type, size_t count,
return 0;
}

static inline int32_t
ompi_datatype_copy_content_same_ddt( const ompi_datatype_t* type, size_t count,
char* pDestBuf, char* pSrcBuf )
{
return ompi_datatype_copy_content_same_ddt_stream(type, count, pDestBuf, pSrcBuf, NULL);
}

OMPI_DECLSPEC const ompi_datatype_t* ompi_datatype_match_size( int size, uint16_t datakind, uint16_t datalang );

/*
Expand Down
7 changes: 6 additions & 1 deletion opal/datatype/opal_datatype.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,7 @@
* Copyright (c) 2004-2006 The Trustees of Indiana University and Indiana
* University Research and Technology
* Corporation. All rights reserved.
* Copyright (c) 2004-2017 The University of Tennessee and The University
* Copyright (c) 2004-2024 The University of Tennessee and The University
* of Tennessee Research Foundation. All rights
* reserved.
* Copyright (c) 2004-2006 High Performance Computing Center Stuttgart,
Expand Down Expand Up @@ -42,6 +42,7 @@
#include <stddef.h>

#include "opal/class/opal_object.h"
#include "opal/mca/accelerator/accelerator.h"

BEGIN_C_DECLS

Expand Down Expand Up @@ -309,6 +310,10 @@ OPAL_DECLSPEC int32_t opal_datatype_copy_content_same_ddt(const opal_datatype_t
int32_t count, char *pDestBuf,
char *pSrcBuf);

OPAL_DECLSPEC int32_t opal_datatype_copy_content_same_ddt_stream(const opal_datatype_t *datatype, int32_t count,
char *destination_base, char *source_base,
opal_accelerator_stream_t *stream);

OPAL_DECLSPEC int opal_datatype_compute_ptypes(opal_datatype_t *datatype);

/*
Expand Down
53 changes: 36 additions & 17 deletions opal/datatype/opal_datatype_copy.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,7 @@
* Copyright (c) 2004-2006 The Trustees of Indiana University and Indiana
* University Research and Technology
* Corporation. All rights reserved.
* Copyright (c) 2004-2013 The University of Tennessee and The University
* Copyright (c) 2004-2024 The University of Tennessee and The University
* of Tennessee Research Foundation. All rights
* reserved.
* Copyright (c) 2004-2006 High Performance Computing Center Stuttgart,
Expand Down Expand Up @@ -55,7 +55,8 @@
} \
} while (0)

static void *opal_datatype_accelerator_memcpy(void *dest, const void *src, size_t size)
static void *opal_datatype_accelerator_memcpy(void *dest, const void *src, size_t size,
opal_accelerator_stream_t *stream)
{
int res;
int dst_type, dst_dev, src_type, src_dev;
Expand All @@ -74,21 +75,27 @@ static void *opal_datatype_accelerator_memcpy(void *dest, const void *src, size_
return memcpy(dest, src, size);
}
else if (0 >= dst_type && 0 < src_type) {
copy_type = MCA_ACCELERATOR_TRANSFER_DTOH;
copy_type = MCA_ACCELERATOR_TRANSFER_DTOH;
}
else if (0 < dst_type && 0 >= dst_type) {
copy_type = MCA_ACCELERATOR_TRANSFER_HTOD;
else if (0 < dst_type && 0 >= src_type) {
copy_type = MCA_ACCELERATOR_TRANSFER_HTOD;
}
if (NULL != stream) {
res = opal_accelerator.mem_copy_async(dst_dev, src_dev,
dest, src, size, stream, copy_type);
} else {
res = opal_accelerator.mem_copy(dst_dev, src_dev,
dest, src, size, copy_type);
}
res = opal_accelerator.mem_copy(dst_dev, src_dev,
dest, src, size, copy_type);
if (OPAL_SUCCESS != res) {
opal_output(0, "Error in accelerator memcpy");
abort();
}
return dest;
}

static void *opal_datatype_accelerator_memmove(void *dest, const void *src, size_t size)
static void *opal_datatype_accelerator_memmove(void *dest, const void *src, size_t size,
opal_accelerator_stream_t *stream)
{
int res;
int dst_type, dst_dev, src_type, src_dev;
Expand All @@ -107,13 +114,18 @@ static void *opal_datatype_accelerator_memmove(void *dest, const void *src, size
return memmove(dest, src, size);
}
else if (0 >= dst_type && 0 < src_type) {
copy_type = MCA_ACCELERATOR_TRANSFER_DTOH;
copy_type = MCA_ACCELERATOR_TRANSFER_DTOH;
}
else if (0 < dst_type && 0 >= src_type) {
copy_type = MCA_ACCELERATOR_TRANSFER_HTOD;
}
else if (0 < dst_type && 0 >= dst_type) {
copy_type = MCA_ACCELERATOR_TRANSFER_HTOD;
if (NULL == stream) {
res = opal_accelerator.mem_move(dst_dev, src_dev,
dest, src, size, copy_type);
} else {
res = opal_accelerator.mem_move_async(dst_dev, src_dev,
dest, src, size, stream, copy_type);
}
res = opal_accelerator.mem_move(dst_dev, src_dev,
dest, src, size, copy_type);
if (OPAL_SUCCESS != res) {
opal_output(0, "Error in accelerator memmove");
abort();
Expand All @@ -137,11 +149,12 @@ static void *opal_datatype_accelerator_memmove(void *dest, const void *src, size
#define MEM_OP opal_datatype_accelerator_memmove
#include "opal_datatype_copy.h"

int32_t opal_datatype_copy_content_same_ddt(const opal_datatype_t *datatype, int32_t count,
char *destination_base, char *source_base)
int32_t opal_datatype_copy_content_same_ddt_stream(const opal_datatype_t *datatype, int32_t count,
char *destination_base, char *source_base,
opal_accelerator_stream_t *stream)
{
ptrdiff_t extent;
int32_t (*fct)(const opal_datatype_t *, int32_t, char *, char *);
int32_t (*fct)(const opal_datatype_t *, int32_t, char *, char *, opal_accelerator_stream_t*);

DO_DEBUG(opal_output(0, "opal_datatype_copy_content_same_ddt( %p, %d, dst %p, src %p )\n",
(void *) datatype, count, (void *) destination_base,
Expand Down Expand Up @@ -173,5 +186,11 @@ int32_t opal_datatype_copy_content_same_ddt(const opal_datatype_t *datatype, int
fct = overlap_accelerator_copy_content_same_ddt;
}
}
return fct(datatype, count, destination_base, source_base);
return fct(datatype, count, destination_base, source_base, stream);
}

int32_t opal_datatype_copy_content_same_ddt(const opal_datatype_t *datatype, int32_t count,
char *destination_base, char *source_base)
{
return opal_datatype_copy_content_same_ddt_stream(datatype, count, destination_base, source_base, NULL);
}
43 changes: 32 additions & 11 deletions opal/datatype/opal_datatype_copy.h
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
/* -*- Mode: C; c-basic-offset:4 ; -*- */
/*
* Copyright (c) 2004-2017 The University of Tennessee and The University
* Copyright (c) 2004-2024 The University of Tennessee and The University
* of Tennessee Research Foundation. All rights
* reserved.
* Copyright (c) 2009 Oak Ridge National Labs. All rights reserved.
Expand Down Expand Up @@ -44,7 +44,7 @@
static inline void _predefined_data(const dt_elem_desc_t *ELEM, const opal_datatype_t *DATATYPE,
unsigned char *SOURCE_BASE, size_t TOTAL_COUNT, size_t COUNT,
unsigned char *SOURCE, unsigned char *DESTINATION,
size_t *SPACE)
size_t *SPACE, opal_accelerator_stream_t *stream)
{
const ddt_elem_desc_t *_elem = &((ELEM)->elem);
unsigned char *_source = (SOURCE) + _elem->disp;
Expand All @@ -69,7 +69,7 @@ static inline void _predefined_data(const dt_elem_desc_t *ELEM, const opal_datat
DO_DEBUG(opal_output(0, "copy %s( %p, %p, %" PRIsize_t " ) => space %" PRIsize_t "\n",
STRINGIFY(MEM_OP_NAME), (void *) _destination, (void *) _source,
do_now_bytes, *(SPACE) -_i * do_now_bytes););
MEM_OP(_destination, _source, do_now_bytes);
MEM_OP(_destination, _source, do_now_bytes, stream);
_destination += _elem->extent;
_source += _elem->extent;
}
Expand All @@ -79,7 +79,7 @@ static inline void _predefined_data(const dt_elem_desc_t *ELEM, const opal_datat
static inline void _contiguous_loop(const dt_elem_desc_t *ELEM, const opal_datatype_t *DATATYPE,
unsigned char *SOURCE_BASE, size_t TOTAL_COUNT, size_t COUNT,
unsigned char *SOURCE, unsigned char *DESTINATION,
size_t *SPACE)
size_t *SPACE, opal_accelerator_stream_t *stream)
{
ddt_loop_desc_t *_loop = (ddt_loop_desc_t *) (ELEM);
ddt_endloop_desc_t *_end_loop = (ddt_endloop_desc_t *) ((ELEM) + _loop->items);
Expand All @@ -91,7 +91,7 @@ static inline void _contiguous_loop(const dt_elem_desc_t *ELEM, const opal_datat
_copy_loops *= _end_loop->size;
OPAL_DATATYPE_SAFEGUARD_POINTER(_source, _copy_loops, (SOURCE_BASE), (DATATYPE),
(TOTAL_COUNT));
MEM_OP(_destination, _source, _copy_loops);
MEM_OP(_destination, _source, _copy_loops, stream);
} else {
for (size_t _i = 0; _i < _copy_loops; _i++) {
OPAL_DATATYPE_SAFEGUARD_POINTER(_source, _end_loop->size, (SOURCE_BASE), (DATATYPE),
Expand All @@ -100,7 +100,7 @@ static inline void _contiguous_loop(const dt_elem_desc_t *ELEM, const opal_datat
"copy 3. %s( %p, %p, %" PRIsize_t " ) => space %" PRIsize_t "\n",
STRINGIFY(MEM_OP_NAME), (void *) _destination, (void *) _source,
_end_loop->size, *(SPACE) -_i * _end_loop->size););
MEM_OP(_destination, _source, _end_loop->size);
MEM_OP(_destination, _source, _end_loop->size, stream);
_source += _loop->extent;
_destination += _loop->extent;
}
Expand All @@ -110,7 +110,8 @@ static inline void _contiguous_loop(const dt_elem_desc_t *ELEM, const opal_datat
}

static inline int32_t _copy_content_same_ddt(const opal_datatype_t *datatype, int32_t count,
char *destination_base, char *source_base)
char *destination_base, char *source_base,
opal_accelerator_stream_t *stream)
{
dt_stack_t *pStack; /* pointer to the position on the stack */
int32_t stack_pos; /* index of the stack level */
Expand Down Expand Up @@ -148,13 +149,20 @@ static inline int32_t _copy_content_same_ddt(const opal_datatype_t *datatype, in
DO_DEBUG(opal_output(0, "copy c1. %s( %p, %p, %lu ) => space %lu\n",
STRINGIFY(MEM_OP_NAME), (void *) destination, (void *) source,
(unsigned long) memop_chunk, (unsigned long) total_length););
MEM_OP(destination, source, memop_chunk);
MEM_OP(destination, source, memop_chunk, stream);
destination += memop_chunk;
source += memop_chunk;
total_length -= memop_chunk;
}
return 0; /* completed */
}
opal_accelerator_stream_t *actual_stream = stream;
bool flush_stream = false;
if (NULL == stream) {
/* use the default stream to enqueue multiple ops before sync'ing */
actual_stream = MCA_ACCELERATOR_STREAM_DEFAULT;
flush_stream = true;
}
for (pos_desc = 0; (int32_t) pos_desc < count; pos_desc++) {
OPAL_DATATYPE_SAFEGUARD_POINTER(destination, datatype->size,
(unsigned char *) destination_base, datatype, count);
Expand All @@ -164,10 +172,13 @@ static inline int32_t _copy_content_same_ddt(const opal_datatype_t *datatype, in
STRINGIFY(MEM_OP_NAME), (void *) destination, (void *) source,
(unsigned long) datatype->size,
(unsigned long) (iov_len_local - (pos_desc * datatype->size))););
MEM_OP(destination, source, datatype->size);
MEM_OP(destination, source, datatype->size, actual_stream);
destination += extent;
source += extent;
}
if (flush_stream) {
opal_accelerator.sync_stream(actual_stream);
}
return 0; /* completed */
}

Expand All @@ -185,11 +196,18 @@ static inline int32_t _copy_content_same_ddt(const opal_datatype_t *datatype, in

UPDATE_INTERNAL_COUNTERS(description, 0, pElem, count_desc);

opal_accelerator_stream_t *actual_stream = stream;
bool flush_stream = false;
if (NULL == stream) {
/* use the default stream to enqueue multiple ops before sync'ing */
actual_stream = MCA_ACCELERATOR_STREAM_DEFAULT;
flush_stream = true;
}
while (1) {
while (OPAL_LIKELY(pElem->elem.common.flags & OPAL_DATATYPE_FLAG_DATA)) {
/* now here we have a basic datatype */
_predefined_data(pElem, datatype, (unsigned char *) source_base, count, count_desc,
source, destination, &iov_len_local);
source, destination, &iov_len_local, stream);
pos_desc++; /* advance to the next data */
UPDATE_INTERNAL_COUNTERS(description, pos_desc, pElem, count_desc);
}
Expand All @@ -202,6 +220,9 @@ static inline int32_t _copy_content_same_ddt(const opal_datatype_t *datatype, in
if (--(pStack->count) == 0) { /* end of loop */
if (stack_pos == 0) {
assert(iov_len_local == 0);
if (flush_stream) {
opal_accelerator.sync_stream(actual_stream);
}
return 0; /* completed */
}
stack_pos--;
Expand Down Expand Up @@ -229,7 +250,7 @@ static inline int32_t _copy_content_same_ddt(const opal_datatype_t *datatype, in
ptrdiff_t local_disp = (ptrdiff_t) source;
if (pElem->loop.common.flags & OPAL_DATATYPE_FLAG_CONTIGUOUS) {
_contiguous_loop(pElem, datatype, (unsigned char *) source_base, count, count_desc,
source, destination, &iov_len_local);
source, destination, &iov_len_local, actual_stream);
pos_desc += pElem->loop.items + 1;
goto update_loop_description;
}
Expand Down
Loading