[mpiwg-tools] Questions for the tools community
hermanns at itc.rwth-aachen.de
Thu Aug 29 10:05:31 CDT 2019
sorry about the silence. As Martin already said, your mail led to some
internal discussion and we wanted to reach out beyond the tools people
on the mailing list because it seems no-one on the list is currently
able to handle Fortran descriptors, but we failed to keep you informed.
Would you be available on *Sep 19* for the tools call? That would give
us some time to contact some people outside the working group and
they'd have time to either give feedback or organize their schedules
in a way to attend the call.
On 29.08.19 16:53, Martin Schulz via mpiwg-tools wrote:
> Hi Jeff,
> Due to the summer break and vacation time, especially in Europe, we
> haven’t had a chance to address this - Marc-Andre had suggested a
> tools call on this for after EuroMPI and then we’ll know more (we need
> to get tool implementors to the table who are not part of the WG).
> Personally, though, I fear you are right with your assumption, but I
> am happy to be proven wrong.
> Prof. Dr. Martin Schulz, Chair of Computer Architecture and Parallel
> Department of Informatics, TU-Munich, Boltzmannstraße 3, D-85748 Garching
> Member of the Board of Directors at the Leibniz Supercomputing Centre
> Email: schulzm at in.tum.de <mailto:schulzm at in.tum.de>
>> On 29. Aug 2019, at 16:46, Jeff Squyres (jsquyres) via mpiwg-tools
>> <mpiwg-tools at lists.mpi-forum.org
>> <mailto:mpiwg-tools at lists.mpi-forum.org>> wrote:
>> Should I infer from the lack of reply that no tools properly handle
>> either Fortran descriptors or the descriptor-based symbols?
>>> On Aug 25, 2019, at 11:31 AM, Jeff Squyres (jsquyres) via
>>> mpiwg-tools <mpiwg-tools at lists.mpi-forum.org
>>> <mailto:mpiwg-tools at lists.mpi-forum.org>> wrote:
>>> In Open MPI, we are finally working towards supporting Fortran
>>> descriptors for choice buffers (i.e., "TYPE(*), DIMENSION(..)").
>>> Per MPI-3.1 p613 table 17.1, this means ***changing all the
>>> back-end symbols*** that are used for PMPI-style interception.
>>> A concrete example:
>>> - in mpif.h / the mpi module, MPI_SEND will become MPI_SEND_FTS
>>> - in the mpi_f08 module, MPI_SEND will become MPI_SEND_F08TS
>>> I.e., anywhere the application calls MPI_SEND, it will be
>>> transmogrified to either MPI_SEND_FTS or MPI_SEND_F08TS. And the
>>> choice buffer passed will be a descriptor (which may therefore be a
>>> subarray, not a contiguous buffer), not a naked pointer.
>>> Here's my questions:
>>> 1. Do any tools support Fortran descriptors as choice buffers?
>>> 2. Do any tools intercept the symbols as described in MPI-3.1 table
>>> Jeff Squyres
>>> jsquyres at cisco.com <mailto:jsquyres at cisco.com>
>>> mpiwg-tools mailing list
>>> mpiwg-tools at lists.mpi-forum.org
>> Jeff Squyres
>> jsquyres at cisco.com <mailto:jsquyres at cisco.com>
>> mpiwg-tools mailing list
>> mpiwg-tools at lists.mpi-forum.org
> mpiwg-tools mailing list
> mpiwg-tools at lists.mpi-forum.org
Dr. rer. nat. Marc-André Hermanns
Group: High Performance Computing
Division: Computational Science and Engineering
RWTH Aachen University
Seffenter Weg 23
Phone: +49 241 80-24381
hermanns at itc.rwth-aachen.de
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 5336 bytes
Desc: S/MIME Cryptographic Signature
More information about the mpiwg-tools