<html>
<head>
<style><!--
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 10pt;
font-family:Tahoma
}
--></style>
</head>
<body class='hmmessage'>
At the MPI forum meeting this week we decided to accelerate the release
of MPI-3.0, and have a follow on MPI-3.1 when the more complete set of
features is available. Currently, nonblocking collectives and MProbe
have already been voted in. We have created a list of features that may
be ready for first reading at the next couple of meetings, which will
be used to determine the gating factors for the MPI-3.0 release. Here
is the list we generated:<br>
- RMA<br> - FT stabilization<br> - Fortran bindings<br> - MPI-T<br> - Neighborhood collectives<br> - Const<br> - MPI Count<br> - Nonblocking collective I/O<br> - Hybrid: Shared memory communicator<br> - Hybrid: Threads<br>
- Hybrid: Endpoints<br><br>If people want to add more to this list, please let me know by end of the day Friday, 5/13/2011.<br><br>Once
we have the complete list, each eligible organization should vote for
it's top three features, with respect to MPI-3.0. An organization is
eligible to vote, if someone from that organization has participated in
person at one or more meetings in 2011. The vote will take place
5/14-20/2011, and the top three items will be what gate the release of
MPI-3.0. To vote, a representative of the organization should e-mail
BOTH Jeff Squyres and myself with your vote. Other items that are voted
in before 3.0 is final, will also be included. For MPI-3.1, we will
change our process a bit, but that is still under discussion, and will
be the subject of a future e-mail discussion.<br><br>The eligible organizations include:<br> - ANL <br> - Bull<br> - Cicso<br> - Cray<br> - Fujistsu<br> - HDF<br> - HLRS<br> - HP<br> - IBM<br> - Indiana U.<br> - INRIA<br> - Intel<br> - LLNL<br> - LANL<br> - Microsoft<br> - NCSA/UIUC<br> - ORNL<br> - OSU<br> - Oracle<br> - SNL<br> - U Alabama Birmingham<br> - U Tennessee<br> - U. Tokyo<br>
<br>Rich<br> </body>
</html>