Relevant to this is that enhancing generalized requests also came up. Part of the discussion is this: would an enhancement to generalized requests, which would have broader benefit, be a more general solution? Background: The Forum is increasingly concerned about the creation of routines to solve specific problems rather than stepping back and trying to find a more general and powerful solution. Bill William Gropp Director, Parallel Computing Institute Deputy Director for Research Institute for Advanced Computing Applications and Technologies Thomas M. Siebel Chair in Computer Science University of Illinois Urbana-Champaign On Mar 13, 2013, at 2:56 PM, Rob Latham wrote:
On Wed, Mar 13, 2013 at 02:46:34PM -0500, Quincey Koziol wrote:
Can something be done with generalized request? Someone always brings this up. :-) We should get up to speed on the GRQ routines and know what they can do...
They cannot do much. I am, reluctantly, the worlds expert on generalized requests, having written a paper about it 5(!!!) years ago.
http://dx.doi.org/10.1007/978-3-540-75416-9_33
pdf: http://www.mcs.anl.gov/~robl/papers/latham_grequest-enhance.pdf
Standard Generalized requests have a "all progress must happen outside of MPI" progress model. Perhaps in this modern surfeit of cores era, it is no longer a big deal to spawn a thread ?
The paper I linked proposes generalized request extensions designed to play better with posix, pvfs, and NTFS async i/o routines. These routines can make progress in the background, but do need someone to call a completion function (aio_suspend, PVFS_sys_wait, or WaitForMultipleObjectEx).
==rob
-- Rob Latham Mathematics and Computer Science Division Argonne National Lab, IL USA _______________________________________________ MPI3-IO mailing list [email protected] http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-io