Merge pull request #3487 from ggouaillardet/topic/v2.x/a2aw_zeros

v2.x: fix MPI_Alltoallw() with zero size messages

Author: hppritcha

ompi/mca/coll/base/coll_base_alltoallv.c

@@ -14,7 +14,7 @@
  * Copyright (c) 2013      Los Alamos National Security, LLC. All Rights
  *                         reserved.
  * Copyright (c) 2013      FUJITSU LIMITED.  All rights reserved.
- * Copyright (c) 2014-2016 Research Organization for Information Science
+ * Copyright (c) 2014-2017 Research Organization for Information Science
  *                         and Technology (RIST). All rights reserved.
  * $COPYRIGHT$
  *
@@ -45,17 +45,16 @@ mca_coll_base_alltoallv_intra_basic_inplace(const void *rbuf, const int *rcounts
     int i, j, size, rank, err=MPI_SUCCESS;
     ompi_request_t *req;
     char *allocated_buffer, *tmp_buffer;
-    size_t max_size, rdtype_size;
-    OPAL_PTRDIFF_TYPE ext, gap;
+    size_t max_size;
+    OPAL_PTRDIFF_TYPE ext, gap = 0;
 
     /* Initialize. */
 
     size = ompi_comm_size(comm);
     rank = ompi_comm_rank(comm);
-    ompi_datatype_type_size(rdtype, &rdtype_size);
 
     /* If only one process, we're done. */
-    if (1 == size || 0 == rdtype_size) {
+    if (1 == size) {
         return MPI_SUCCESS;
     }
 
@@ -67,6 +66,10 @@ mca_coll_base_alltoallv_intra_basic_inplace(const void *rbuf, const int *rcounts
     }
     /* The gap will always be the same as we are working on the same datatype */
 
+    if (OPAL_UNLIKELY(0 == max_size)) {
+        return MPI_SUCCESS;
+    }
+
     /* Allocate a temporary buffer */
     allocated_buffer = calloc (max_size, 1);
     if (NULL == allocated_buffer) {
@@ -78,7 +81,7 @@ mca_coll_base_alltoallv_intra_basic_inplace(const void *rbuf, const int *rcounts
     /* in-place alltoallv slow algorithm (but works) */
     for (i = 0 ; i < size ; ++i) {
         for (j = i+1 ; j < size ; ++j) {
-            if (i == rank && rcounts[j]) {
+            if (i == rank && 0 != rcounts[j]) {
                 /* Copy the data into the temporary buffer */
                 err = ompi_datatype_copy_content_same_ddt (rdtype, rcounts[j],
                                                            tmp_buffer, (char *) rbuf + rdisps[j] * ext);
@@ -93,7 +96,7 @@ mca_coll_base_alltoallv_intra_basic_inplace(const void *rbuf, const int *rcounts
                                           j, MCA_COLL_BASE_TAG_ALLTOALLV, MCA_PML_BASE_SEND_STANDARD,
                                           comm));
                 if (MPI_SUCCESS != err) { goto error_hndl; }
-            } else if (j == rank && rcounts[i]) {
+            } else if (j == rank && 0 != rcounts[i]) {
                 /* Copy the data into the temporary buffer */
                 err = ompi_datatype_copy_content_same_ddt (rdtype, rcounts[i],
                                                            tmp_buffer, (char *) rbuf + rdisps[i] * ext);

ompi/mca/coll/libnbc/Makefile.am

@@ -12,6 +12,8 @@
 # Copyright (c) 2010      Cisco Systems, Inc.  All rights reserved.
 # Copyright (c) 2013      Los Alamos National Security, LLC. All rights
 #                         reserved.
+# Copyright (c) 2017      Research Organization for Information Science
+#                         and Technology (RIST). All rights reserved.
 # $COPYRIGHT$
 #
 # Additional copyrights may follow
@@ -37,7 +39,6 @@ sources = \
 	nbc_ialltoallw.c \
 	nbc_ibarrier.c \
 	nbc_ibcast.c \
-	nbc_ibcast_inter.c \
 	nbc_iexscan.c \
 	nbc_igather.c \
 	nbc_igatherv.c \

ompi/mca/coll/libnbc/nbc.c

@@ -10,7 +10,7 @@
  *                         rights reserved.
  * Copyright (c) 2015      Los Alamos National Security, LLC.  All rights
  *                         reserved.
- * Copyright (c) 2015-2016 Research Organization for Information Science
+ * Copyright (c) 2015-2017 Research Organization for Information Science
  *                         and Technology (RIST). All rights reserved.
  *
  * Author(s): Torsten Hoefler <[email protected]>
@@ -709,6 +709,25 @@ int NBC_Start(NBC_Handle *handle, NBC_Schedule *schedule) {
   return OMPI_SUCCESS;
 }
 
+int NBC_Schedule_request(NBC_Schedule *schedule, ompi_communicator_t *comm, ompi_coll_libnbc_module_t *module, ompi_request_t **request, void *tmpbuf) {
+  int res;
+  NBC_Handle *handle;
+  res = NBC_Init_handle (comm, &handle, module);
+  if (OPAL_UNLIKELY(OMPI_SUCCESS != res)) {
+    return res;
+  }
+  handle->tmpbuf = tmpbuf;
+
+  res = NBC_Start (handle, schedule);
+  if (OPAL_UNLIKELY(OMPI_SUCCESS != res)) {
+    NBC_Return_handle (handle);
+    return res;
+  }
+
+  *request = (ompi_request_t *) handle;
+  return OMPI_SUCCESS;
+}
+
 #ifdef NBC_CACHE_SCHEDULE
 void NBC_SchedCache_args_delete_key_dummy(void *k) {
     /* do nothing because the key and the data element are identical :-)

ompi/mca/coll/libnbc/nbc_iallgather.c

@@ -50,7 +50,6 @@ int ompi_coll_libnbc_iallgather(const void* sendbuf, int sendcount, MPI_Datatype
 #ifdef NBC_CACHE_SCHEDULE
   NBC_Allgather_args *args, *found, search;
 #endif
-  NBC_Handle *handle;
   ompi_coll_libnbc_module_t *libnbc_module = (ompi_coll_libnbc_module_t*) module;
 
   NBC_IN_PLACE(sendbuf, recvbuf, inplace);
@@ -147,20 +146,12 @@ int ompi_coll_libnbc_iallgather(const void* sendbuf, int sendcount, MPI_Datatype
   }
 #endif
 
-  res = NBC_Init_handle (comm, &handle, libnbc_module);
+  res = NBC_Schedule_request(schedule, comm, libnbc_module, request, NULL);
   if (OPAL_UNLIKELY(OMPI_SUCCESS != res)) {
     OBJ_RELEASE(schedule);
     return res;
   }
 
-  res = NBC_Start (handle, schedule);
-  if (OPAL_UNLIKELY(OMPI_SUCCESS != res)) {
-    OMPI_COLL_LIBNBC_REQUEST_RETURN(handle);
-    return res;
-  }
-
-  *request = (ompi_request_t *) handle;
-
   return OMPI_SUCCESS;
 }
 
@@ -172,7 +163,6 @@ int ompi_coll_libnbc_iallgather_inter(const void* sendbuf, int sendcount, MPI_Da
   MPI_Aint rcvext;
   NBC_Schedule *schedule;
   char *rbuf;
-  NBC_Handle *handle;
   ompi_coll_libnbc_module_t *libnbc_module = (ompi_coll_libnbc_module_t*) module;
 
   res = ompi_datatype_type_extent(recvtype, &rcvext);
@@ -213,19 +203,11 @@ int ompi_coll_libnbc_iallgather_inter(const void* sendbuf, int sendcount, MPI_Da
     return res;
   }
 
-  res = NBC_Init_handle (comm, &handle, libnbc_module);
+  res = NBC_Schedule_request(schedule, comm, libnbc_module, request, NULL);
   if (OPAL_UNLIKELY(OMPI_SUCCESS != res)) {
     OBJ_RELEASE(schedule);
     return res;
   }
 
-  res = NBC_Start (handle, schedule);
-  if (OPAL_UNLIKELY(OMPI_SUCCESS != res)) {
-    OMPI_COLL_LIBNBC_REQUEST_RETURN(handle);
-    return res;
-  }
-
-  *request = (ompi_request_t *) handle;
-
   return OMPI_SUCCESS;
 }

ompi/mca/coll/libnbc/nbc_iallgatherv.c

@@ -11,7 +11,7 @@
  * Copyright (c) 2012      Oracle and/or its affiliates.  All rights reserved.
  * Copyright (c) 2013-2015 Los Alamos National Security, LLC. All rights
  *                         reserved.
- * Copyright (c) 2014-2016 Research Organization for Information Science
+ * Copyright (c) 2014-2017 Research Organization for Information Science
  *                         and Technology (RIST). All rights reserved.
  *
  */
@@ -37,7 +37,6 @@ int ompi_coll_libnbc_iallgatherv(const void* sendbuf, int sendcount, MPI_Datatyp
   MPI_Aint rcvext;
   NBC_Schedule *schedule;
   char *rbuf, *sbuf, inplace;
-  NBC_Handle *handle;
   ompi_coll_libnbc_module_t *libnbc_module = (ompi_coll_libnbc_module_t*) module;
 
   NBC_IN_PLACE(sendbuf, recvbuf, inplace);
@@ -96,20 +95,12 @@ int ompi_coll_libnbc_iallgatherv(const void* sendbuf, int sendcount, MPI_Datatyp
     return res;
   }
 
-  res = NBC_Init_handle (comm, &handle, libnbc_module);
+  res = NBC_Schedule_request (schedule, comm, libnbc_module, request, NULL);
   if (OPAL_UNLIKELY(OMPI_SUCCESS != res)) {
     OBJ_RELEASE(schedule);
     return res;
   }
 
-  res = NBC_Start (handle, schedule);
-  if (OPAL_UNLIKELY(OMPI_SUCCESS != res)) {
-    NBC_Return_handle (handle);
-    return res;
-  }
-
-  *request = (ompi_request_t *) handle;
-
   return OMPI_SUCCESS;
 }
 
@@ -120,7 +111,6 @@ int ompi_coll_libnbc_iallgatherv_inter(const void* sendbuf, int sendcount, MPI_D
   int res, rsize;
   MPI_Aint rcvext;
   NBC_Schedule *schedule;
-  NBC_Handle *handle;
   ompi_coll_libnbc_module_t *libnbc_module = (ompi_coll_libnbc_module_t*) module;
 
   rsize = ompi_comm_remote_size (comm);
@@ -165,19 +155,11 @@ int ompi_coll_libnbc_iallgatherv_inter(const void* sendbuf, int sendcount, MPI_D
     return res;
   }
 
-  res = NBC_Init_handle (comm, &handle, libnbc_module);
+  res = NBC_Schedule_request(schedule, comm, libnbc_module, request, NULL);
   if (OPAL_UNLIKELY(OMPI_SUCCESS != res)) {
     OBJ_RELEASE(schedule);
     return res;
   }
 
-  res = NBC_Start (handle, schedule);
-  if (OPAL_UNLIKELY(OMPI_SUCCESS != res)) {
-    NBC_Return_handle (handle);
-    return res;
-  }
-
-  *request = (ompi_request_t *) handle;
-
   return OMPI_SUCCESS;
 }