[mpich-discuss] MPI_TYPE_MAX limit using MPT

Michael Raymond mraymond at sgi.com
Mon Oct 18 08:49:52 CDT 2010


  It's a per-shepherd shared table.

Rob Ross wrote:
> Is this a global table in the sense of every process contributing to a single table? Does everyone have a copy, or is it partitioned between processes?
> 
> -- Rob
> 
> On Oct 15, 2010, at 6:50 AM, Michael Raymond <mraymond at sgi.com> wrote:
> 
>>  That's correct.  In MPT every element of a complex type like a struct
>> or indexed type ends up taking up a type in the global table.  There are
>> good speed reasons for this but the downside is that you sometimes need
>> to preallocate a ton of space for datatypes.  This is remedied to some
>> extent in MPT 2.03.  If you have any more questions about MPT you can
>> contact me directly.

-- 
Michael A. Raymond
Message Passing Toolkit Team
Silicon Graphics Inc
(651) 683-3434



More information about the mpich-discuss mailing list