coll/base: give a boost to ompi_coll_base_sendrecv_nonzero_actual()

Based on current implementation it is faster to use a blocking
send than the non-blocking version. Switch the exchange function
used in the barrier to use the blocking version combined with
the non-blocking version of the receive.

This is similar to 223d755

Author: ggouaillardet

ompi/mca/coll/base/coll_base_util.c

@@ -39,54 +39,33 @@ int ompi_coll_base_sendrecv_nonzero_actual( void* sendbuf, size_t scount,
                                              ompi_status_public_t* status )
 
 { /* post receive first, then send, then waitall... should be fast (I hope) */
-    int err, line = 0, nreqs = 0;
-    size_t typesize;
-    ompi_request_t* reqs[2], **req = reqs;
-    ompi_status_public_t statuses[2];
+    int err, line = 0;
+    size_t rtypesize, stypesize;
+    ompi_request_t *req;
+    ompi_status_public_t rstatus;
 
     /* post new irecv */
-    ompi_datatype_type_size(rdatatype, &typesize);
-    if (0 != rcount && 0 != typesize) {
+    ompi_datatype_type_size(rdatatype, &rtypesize);
+    if (0 != rcount && 0 != rtypesize) {
         err = MCA_PML_CALL(irecv( recvbuf, rcount, rdatatype, source, rtag,
-                                  comm, req++));
-        ++nreqs;
+                                  comm, &req));
         if (err != MPI_SUCCESS) { line = __LINE__; goto error_handler; }
     }
 
     /* send data to children */
-    ompi_datatype_type_size(sdatatype, &typesize);
-    if (0 != scount && 0 != typesize) {
-        err = MCA_PML_CALL(isend( sendbuf, scount, sdatatype, dest, stag,
-                                  MCA_PML_BASE_SEND_STANDARD, comm, req++));
-        ++nreqs;
+    ompi_datatype_type_size(sdatatype, &stypesize);
+    if (0 != scount && 0 != stypesize) {
+        err = MCA_PML_CALL(send( sendbuf, scount, sdatatype, dest, stag,
+                                  MCA_PML_BASE_SEND_STANDARD, comm));
         if (err != MPI_SUCCESS) { line = __LINE__; goto error_handler; }
     }
 
-    if (0 != nreqs) {
-        err = ompi_request_wait_all( nreqs, reqs, statuses );
-        if( MPI_ERR_IN_STATUS == err ) { line = __LINE__;
-            /* As we use wait_all we will get MPI_ERR_IN_STATUS which is not an error
-             * code that we can propagate up the stack. Instead, look for the real
-             * error code from the MPI_ERROR in the status.
-             */
-            int err_index = 0;
-            if( MPI_SUCCESS == statuses[0].MPI_ERROR
-             || MPI_ERR_PENDING == statuses[0].MPI_ERROR ) {
-                err_index = 1;
-            }
-            if (MPI_STATUS_IGNORE != status) {
-                *status = statuses[err_index];
-            }
-            err = statuses[err_index].MPI_ERROR;
-            OPAL_OUTPUT ((ompi_coll_base_framework.framework_output, "%s:%d: Error %d occurred in the %s"
-                                                  " stage of ompi_coll_base_sendrecv_zero\n",
-                          __FILE__, line, err, (0 == err_index ? "receive" : "send")));
-            return err;
-        }
+    if (0 != rcount && 0 != rtypesize) {
+        err = ompi_request_wait( &req, &rstatus);
         if (err != MPI_SUCCESS) { line = __LINE__; goto error_handler; }
 
         if (MPI_STATUS_IGNORE != status) {
-            *status = statuses[0];
+            *status = rstatus;
         }
     } else {
         if( MPI_STATUS_IGNORE != status )
@@ -96,7 +75,7 @@ int ompi_coll_base_sendrecv_nonzero_actual( void* sendbuf, size_t scount,
     return (MPI_SUCCESS);
 
  error_handler:
-    /* Error discovered during the posting of the irecv or isend,
+    /* Error discovered during the posting of the irecv or send,
      * and no status is available.
      */
     OPAL_OUTPUT ((ompi_coll_base_framework.framework_output, "%s:%d: Error %d occurred\n",