[Mpi3-ft] Summary of today's meeting
kannan.narasimhan at hp.com
Thu Oct 23 15:22:01 CDT 2008
If the application "ensures" that there are no messages in-flight, then I agree with you. I was thinking of applications that use asyncronous MPI communication in a multi-threaded env, potentially leading to unmatched messages.....
From: mpi3-ft-bounces at lists.mpi-forum.org [mailto:mpi3-ft-bounces at lists.mpi-forum.org] On Behalf Of Greg Bronevetsky
Sent: Thursday, October 23, 2008 2:59 PM
To: MPI 3.0 Fault Tolerance and Dynamic Process Control working Group; MPI 3.0 Fault Tolerance and Dynamic Process Control working Group
Subject: Re: [Mpi3-ft] Summary of today's meeting
Application-directed C/R will still require some quiescence hooks from the MPI layer (ex: asyncronous progression by the MPI layer). There was some discussion on this.
What kind of quiscence are you thinking of? It seems to me that applications will simply need to ensure that either no messages are in-flight at the time of the checkpoint or that all such messages have been logged appropriately by the application.
1028 Building 451
Lawrence Livermore National Lab
bronevetsky1 at llnl.gov
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the mpiwg-ft