Same here. ~Jim. On 7/19/12 11:17 AM, William Gropp wrote:
This sounds good to me - deferring the discussion of the semantics to the RMA chapter makes the most sense.
Bill
William Gropp Director, Parallel Computing Institute Deputy Director for Research Institute for Advanced Computing Applications and Technologies Paul and Cynthia Saylor Professor of Computer Science University of Illinois Urbana-Champaign
On Jul 19, 2012, at 11:05 AM, Schulz, Martin wrote:
Hi all,
during the final chapter reads, the following issue came up in the section "Memory Allocation" in Chapter 8, MPI Environmental Management. Here is the context
\section{Memory Allocation} \label{sec:misc-memalloc}
In some systems, message-passing and remote-memory-access (\RMA/) operations run faster when accessing specially allocated memory (e.g., memory that is shared by the other processes in the communicating group on an SMP). \MPI/ provides a mechanism for allocating and freeing such special memory. The use of such memory for message-passing or \RMA/ is not mandatory, and this memory can be used without restrictions as any other dynamically allocated memory.
This is followed by the following sentence:
However, implementations may restrict the use of the \mpifunc{MPI\_WIN\_LOCK} and \mpifunc{MPI\_WIN\_UNLOCK} functions to windows allocated in such memory (see Section~\ref{sec:1sided-lock}.)
This covers the restrictions for 2.2, but not for the new RMA chapter. Therefore, the RMA group (or a subset thereof) suggested to change that sentence to:
However, implementations may restrict the use of some RMA functionality as defined in Section~\ref{sec:1sided-lock}.
and leave the definition of the semantics to the RMA chapter.
Please reply to the group if there are any objections or concerns.
Thanks,
Martin
________________________________________________________________________ Martin Schulz, [email protected], http://people.llnl.gov/schulzm CASC @ Lawrence Livermore National Laboratory, Livermore, USA
_______________________________________________ mpi-forum mailing list [email protected] http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum
_______________________________________________ mpi-forum mailing list [email protected] http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum