2 This source code file is part of thread_mpi.
3 Written by Sander Pronk, Erik Lindahl, and possibly others.
5 Copyright (c) 2009, Sander Pronk, Erik Lindahl.
8 Redistribution and use in source and binary forms, with or without
9 modification, are permitted provided that the following conditions are met:
10 1) Redistributions of source code must retain the above copyright
11 notice, this list of conditions and the following disclaimer.
12 2) Redistributions in binary form must reproduce the above copyright
13 notice, this list of conditions and the following disclaimer in the
14 documentation and/or other materials provided with the distribution.
15 3) Neither the name of the copyright holders nor the
16 names of its contributors may be used to endorse or promote products
17 derived from this software without specific prior written permission.
19 THIS SOFTWARE IS PROVIDED BY US ''AS IS'' AND ANY
20 EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
21 WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
22 DISCLAIMED. IN NO EVENT SHALL WE BE LIABLE FOR ANY
23 DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
24 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
25 LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
26 ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
27 (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
28 SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
30 If you want to redistribute modifications, please consider that
31 scientific software is very special. Version control is crucial -
32 bugs must be traceable. We will be happy to consider code for
33 inclusion in the official distribution, but derived work should not
34 be called official thread_mpi. Details are found in the README & COPYING
38 #ifdef HAVE_TMPI_CONFIG_H
39 #include "tmpi_config.h"
58 #include "collective.h"
62 /* run a single binary reduce operation on src_a and src_b, producing dest.
63 dest and src_a may be identical */
64 static int tMPI_Reduce_run_op(void *dest
, void *src_a
, void *src_b
,
65 tMPI_Datatype datatype
, int count
, tMPI_Op op
,
69 static int tMPI_Reduce_run_op(void *dest
, void *src_a
, void *src_b
,
70 tMPI_Datatype datatype
, int count
, tMPI_Op op
,
73 tMPI_Op_fn fn
=datatype
->op_functions
[op
];
77 return tMPI_Error(comm
, TMPI_ERR_XFER_BUF_OVERLAP
);
79 fn(dest
, src_a
, src_b
, count
);
83 int tMPI_Reduce_fast(void* sendbuf
, void* recvbuf
, int count
,
84 tMPI_Datatype datatype
, tMPI_Op op
, int root
,
87 struct tmpi_thread
*cur
=tMPI_Get_current();
88 int myrank
=tMPI_Comm_seek_rank(comm
, cur
);
90 /* this function uses a binary tree-like reduction algorithm: */
91 int N
=tMPI_Comm_N(comm
);
92 int myrank_rtr
=(N
+myrank
-root
)%N
; /* my rank relative to root */
93 int Nred
=N
; /* number of neighbours that still communicate
94 (decreases exponentially) */
95 int nbr_dist
=1; /* distance between communicating neighbours
96 (increases exponentially) */
97 int stepping
=2; /* distance between non-communicating neighbours
98 (increases exponentially) */
105 return tMPI_Error(TMPI_COMM_WORLD
, TMPI_ERR_COMM
);
109 return tMPI_Error(comm
, TMPI_ERR_BUF
);
111 if ( (!datatype
->op_functions
) || (!datatype
->op_functions
[op
]) )
113 return tMPI_Error(comm
, TMPI_ERR_OP_FN
);
116 if (sendbuf
==TMPI_IN_PLACE
)/* i.e. sendbuf == tMPI_IN_PLACE */
120 /* we set our send and recv buffer s*/
121 tMPI_Atomic_ptr_set(&(comm
->reduce_sendbuf
[myrank
]),sendbuf
);
122 tMPI_Atomic_ptr_set(&(comm
->reduce_recvbuf
[myrank
]),recvbuf
);
126 /* calculate neighbour rank (here we use the real rank) */
127 int nbr
=(myrank_rtr
%stepping
==0) ?
128 (N
+myrank
+nbr_dist
)%N
:
129 (N
+myrank
-nbr_dist
)%N
;
132 printf("%d: iteration %d: myrank_rtr=%d, stepping=%d\n",
133 myrank
, iteration
, myrank_rtr
, stepping
);
136 /* check if I'm the reducing thread in this iteration's pair: */
137 if (myrank_rtr
%stepping
== 0)
142 /* now wait for my neighbor's data to become ready.
143 First check if I actually have a neighbor. */
144 if (myrank_rtr
+nbr_dist
<N
)
147 printf("%d: waiting to reduce with %d, iteration=%d\n",
148 myrank
, nbr
, iteration
);
152 #if defined(TMPI_PROFILE) && defined(TMPI_CYCLE_COUNT)
153 tMPI_Profile_wait_start(cur
);
155 tMPI_Event_wait( &(comm
->csync
[myrank
].events
[nbr
]) );
156 tMPI_Event_process( &(comm
->csync
[myrank
].events
[nbr
]), 1);
157 #if defined(TMPI_PROFILE) && defined(TMPI_CYCLE_COUNT)
158 tMPI_Profile_wait_stop(cur
, TMPIWAIT_Reduce
);
162 printf("%d: reducing with %d, iteration=%d\n",
163 myrank
, nbr
, iteration
);
166 /* we reduce with our neighbour*/
169 /* for the first iteration, the inputs are in the
172 b
=(void*)tMPI_Atomic_ptr_get(&(comm
->reduce_sendbuf
[nbr
]));
176 /* after the first operation, they're already in
179 b
=(void*)tMPI_Atomic_ptr_get(&(comm
->reduce_recvbuf
[nbr
]));
181 /* here we check for overlapping buffers */
184 return tMPI_Error(comm
, TMPI_ERR_XFER_BUF_OVERLAP
);
188 if ((ret
=tMPI_Reduce_run_op(recvbuf
, a
, b
, datatype
,
189 count
, op
, comm
)) != TMPI_SUCCESS
)
192 /* signal to my neighbour that I'm ready. */
193 tMPI_Event_signal( &(comm
->csync
[nbr
].events
[myrank
]) );
198 printf("%d: not waiting copying buffer\n", myrank
);
201 /* we still need to put things in the right buffer for the next
202 iteration. We need to check for overlapping buffers
203 here because MPI_IN_PLACE might cause recvbuf to be the
205 if (iteration
==0 && (recvbuf
!=sendbuf
))
206 memcpy(recvbuf
, sendbuf
, datatype
->size
*count
);
212 /* the other thread is doing the reducing; we can just
213 wait and break when ready */
214 /* Awake our neighbour */
215 tMPI_Event_signal( &(comm
->csync
[nbr
].events
[myrank
]) );
219 printf("%d: signalled %d, now waiting: iteration=%d\n",
220 nbr
, myrank
, iteration
);
224 /* And wait for an incoming event from out neighbour */
225 #if defined(TMPI_PROFILE) && defined(TMPI_CYCLE_COUNT)
226 tMPI_Profile_wait_start(cur
);
228 tMPI_Event_wait( &(comm
->csync
[myrank
].events
[nbr
]) );
229 tMPI_Event_process( &(comm
->csync
[myrank
].events
[nbr
]), 1);
230 #if defined(TMPI_PROFILE) && defined(TMPI_CYCLE_COUNT)
231 tMPI_Profile_wait_stop(cur
, TMPIWAIT_Reduce
);
233 /* now we can break because our data is reduced, and
234 our neighbour goes on reducing it further. */
239 printf("%d: iteration over, iteration=%d\n", myrank
, iteration
);
243 Nred
= Nred
/2 + Nred
%2;
252 int tMPI_Reduce(void* sendbuf
, void* recvbuf
, int count
,
253 tMPI_Datatype datatype
, tMPI_Op op
, int root
, tMPI_Comm comm
)
255 struct tmpi_thread
*cur
=tMPI_Get_current();
256 int myrank
=tMPI_Comm_seek_rank(comm
, cur
);
260 tMPI_Profile_count_start(cur
);
263 tMPI_Trace_print("tMPI_Reduce(%p, %p, %d, %p, %p, %d, %p)",
264 sendbuf
, recvbuf
, count
, datatype
, op
, root
, comm
);
269 if (sendbuf
==TMPI_IN_PLACE
) /* i.e. sendbuf == TMPI_IN_PLACE */
276 #ifdef TMPI_WARN_MALLOC
277 fprintf(stderr
, "Warning: malloc during tMPI_Reduce\n");
279 recvbuf
=(void*)tMPI_Malloc(datatype
->size
*count
);
281 ret
=tMPI_Reduce_fast(sendbuf
, recvbuf
, count
, datatype
, op
, root
, comm
);
287 tMPI_Profile_count_stop(cur
, TMPIFN_Reduce
);
292 int tMPI_Allreduce(void* sendbuf
, void* recvbuf
, int count
,
293 tMPI_Datatype datatype
, tMPI_Op op
, tMPI_Comm comm
)
295 void *rootbuf
=NULL
; /* root process' receive buffer */
296 struct tmpi_thread
*cur
=tMPI_Get_current();
297 int myrank
=tMPI_Comm_seek_rank(comm
, cur
);
301 tMPI_Profile_count_start(cur
);
304 tMPI_Trace_print("tMPI_Allreduce(%p, %p, %d, %p, %p, %p)",
305 sendbuf
, recvbuf
, count
, datatype
, op
, comm
);
311 return tMPI_Error(comm
, TMPI_ERR_BUF
);
313 if (sendbuf
==TMPI_IN_PLACE
) /* i.e. sendbuf == TMPI_IN_PLACE */
318 ret
=tMPI_Reduce_fast(sendbuf
, recvbuf
, count
, datatype
, op
, 0, comm
);
319 #if defined(TMPI_PROFILE)
320 tMPI_Profile_wait_start(cur
);
322 tMPI_Barrier_wait( &(comm
->barrier
));
323 #if defined(TMPI_PROFILE) && defined(TMPI_CYCLE_COUNT)
324 tMPI_Profile_wait_stop(cur
, TMPIWAIT_Reduce
);
326 /* distribute rootbuf */
327 rootbuf
=(void*)tMPI_Atomic_ptr_get(&(comm
->reduce_recvbuf
[0]));
329 /* and now we just copy things back. We know that the root thread
330 arrives last, so there's no point in using tMPI_Scatter with
331 copy buffers, etc. */
334 if (rootbuf
==recvbuf
)
336 return tMPI_Error(comm
, TMPI_ERR_XFER_BUF_OVERLAP
);
338 memcpy(recvbuf
, rootbuf
, datatype
->size
*count
);
341 #if defined(TMPI_PROFILE) && defined(TMPI_CYCLE_COUNT)
342 tMPI_Profile_wait_start(cur
);
344 tMPI_Barrier_wait( &(comm
->barrier
));
345 #if defined(TMPI_PROFILE)
346 tMPI_Profile_wait_stop(cur
, TMPIWAIT_Reduce
);
347 tMPI_Profile_count_stop(cur
, TMPIFN_Allreduce
);