make it clear that starpu_mpi_task_build submits communications

doc/doxygen/chapters/410_mpi_support.doxy

@@ -509,11 +509,12 @@ to provide a dynamic policy.
 
 A function starpu_mpi_task_build() is also provided with the aim to
 only construct the task structure. All MPI nodes need to call the
-function, only the node which is to execute the task will return a
+function, which posts the required send/recv on the various nodes which have to.
+Only the node which is to execute the task will then return a
 valid task structure, others will return <c>NULL</c>. That node must submit that task.
 All nodes then need to call the function starpu_mpi_task_post_build() -- with the same
 list of arguments as starpu_mpi_task_build() -- to post all the
-necessary data communications.
+necessary data communications meant to happen after the task execution.
 
 \code{.c}
 struct starpu_task *task;

doc/doxygen/chapters/api/mpi.doxy

@@ -445,7 +445,9 @@ has been modified. The cache can be disabled (see \ref STARPU_MPI_CACHE).
 \ingroup API_MPI_Support
 Create a task corresponding to \p codelet with the following given arguments.
 The argument list must be zero-terminated. The function performs the
-first two steps of the function starpu_mpi_task_insert(). Only the MPI
+first two steps of the function starpu_mpi_task_insert(), i.e. submitting the
+MPI communications needed before the execution of the task, and the creation of
+the task on one node. Only the MPI
 node selected in the first step of the algorithm will return a valid
 task structure which can then be submitted, others will return <c>NULL</c>. The function
 starpu_mpi_task_post_build() MUST be called after that on all nodes, and after the submission of