Thanks, Jim. Would it be useful to have a version of these matrices that indicated which rules applied? Bill On Oct 18, 2011, at 11:05 AM, James Dinan wrote:

Hi All,

Updated tables from today's telecon are attached.

~Jim.

On 10/14/11 10:22 PM, Torsten Hoefler wrote:

...

On Fri, Oct 14, 2011 at 10:19:01PM -0500, Pavan Balaji wrote:

...

Aren't all combinations allowed?

Yes, but rules 2+3 specify some additional semantics for those cases (e.g., that the outcome is "stable" and has to represent the written data at least on a per-byte granularity) while all other undefined cases may put random data into the accessed locations.

Torsten

< rma_op_matrix .pdf

...

<rma_op_matrix.pptx>_______________________________________________ mpi3-rma mailing list [email protected] http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma

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