[HPGMG Forum] use of MPI namespace

Jed Brown jed at jedbrown.org
Wed Jul 9 16:41:24 UTC 2014


Jeff Hammond <jeff.science at gmail.com> writes:

> The use of MPI_threadingModelRequested technically violates the MPI
> standard, which reserves the MPI namespace.  While it is highly
> unlikely that MPI will ever standardize the symbol
> MPI_threadingModelRequested, HPGMG is nonetheless wrong for using it.
>
> Same for MPI_Rank, MPI_Tasks, MPI_threadingModel...

Agreed about the namespace issues.  Jeff's comments are specific to
HPGMG-FV, so it's up to Sam what he'd like to do, though I presume he
would gladly accept a pull request.

> USE_MPI_THREAD_MULTIPLE seems a bit dangerous, but I can only assume
> that the C preprocessor doesn't apply #define MPI_THREAD_MULTIPLE in
> such a way as to case USE_MPI_THREAD_MULTIPLE to be transformed into
> e.g. USE_3.

That's not how CPP works.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 818 bytes
Desc: not available
URL: <https://hpgmg.org/lists/archives/hpgmg-forum/attachments/20140709/5b189e04/attachment.bin>


More information about the HPGMG-Forum mailing list