[mvapich-discuss] Intel 2018 and Fortran 2008 Support

[Subramoni, Hari]

Hi, Matt.

Thanks for the report and paying very close attention to detail :-).

We seem to have inherited this issue directly from MPICH 3.1.2. As you say, this seems to be an issue with the conftest being used to detect compatibility and the Intel 2018 compilers. Intel 2017 prints a warning, but does not cause an error allowing things to proceed (see the example below).

The warning we get is the following with the attached conftest (the one that fails). We are able to get around this by making some changes to the conftest, but I need to run it by the MPICH folks so that we can push it up the stack to them. I will get back to you soon with an update.

$ module list

Currently Loaded Modules:
  1) intel/18.0.1   2) mvapich2/2.2   3) texlive/2016   4) StdEnv

$ ifort -o conftest     conftest.f90
conftest.f90(74): warning #8100: The actual argument is an array section or assumed-shape array, corresponding dummy argument that has either the VOLATILE or ASYNCHRONOUS attribute shall be an assumed-shape array.   [IAR]
CALL TEST_ASSUMED_RANK_ASYNC(IAR(2:7))
-----------------------------^
conftest.f90(74): error #7505: If an actual argument is an array section with vector subscript and corresponding dummy argument does not have VALUE attribute, it must not have ASYNCHRONOUS / VOLATILE attribute.   [BUF]
CALL TEST_ASSUMED_RANK_ASYNC(IAR(2:7))
^
compilation aborted for conftest.f90 (code 1)

$ module load intel/17.0.2

Due to MODULEPATH changes the following have been reloaded:
  1) mvapich2/2.2

The following have been reloaded with a version change:
  1) intel/18.0.1 => intel/17.0.2

$ ifort -o conftest     conftest.f90
conftest.f90(74): warning #8100: The actual argument is an array section or assumed-shape array, corresponding dummy argument that has either the VOLATILE or ASYNCHRONOUS attribute shall be an assumed-shape array.   [IAR]
CALL TEST_ASSUMED_RANK_ASYNC(IAR(2:7))
-----------------------------^

Thx,
Hari.

-----Original Message-----
From: mvapich-discuss-bounces at cse.ohio-state.edu On Behalf Of Thompson, Matt (GSFC-610.1)[SCIENCE SYSTEMS AND APPLICATIONS INC]
Sent: Tuesday, July 24, 2018 12:36 PM
To: mvapich-discuss at cse.ohio-state.edu <mvapich-discuss at mailman.cse.ohio-state.edu>
Subject: [mvapich-discuss] Intel 2018 and Fortran 2008 Support

MVAPICH Gurus,

Upon the release of MVAPICH2 2.3, I thought I'd try and build it on a new Omni-Path cluster that I didn't have much luck with trying with
MVAPICH2 2.3rc2.

I'm building with Intel 18.0.3.222 and configuring with:

   ./configure --with-device=ch3:psm --with-pm=slurm --with-pmi=pmi2 --enable-fortran=all CC=icc CXX=icpc FC=ifort

as, I think(?) that's the right line for a SLURM Omni-Path cluster.

However, this time I was a bit more careful looking over the configure output and I saw:

   checking for Fortran 2008 support... no

This surprises me as Intel MPI 2018 and Open MPI 3.1.0 (which I built) both seemed to say yes to this and provide mpi_f08.mod.

Do you know why Intel Fortran 18.0.3 is not being considered compliant by MVAPICH2?

Matt
--
Matt Thompson, SSAI, Sr Scientific Programmer/Analyst
NASA GSFC,    Global Modeling and Assimilation Office
Code 610.1,  8800 Greenbelt Rd,  Greenbelt,  MD 20771
Phone: 301-614-6712                 Fax: 301-614-6246
http://science.gsfc.nasa.gov/sed/bio/matthew.thompson
_______________________________________________
mvapich-discuss mailing list
mvapich-discuss at cse.ohio-state.edu
http://mailman.cse.ohio-state.edu/mailman/listinfo/mvapich-discuss
-------------- next part --------------
A non-text attachment was scrubbed...
Name: winmail.dat
Type: application/ms-tnef
Size: 15193 bytes
Desc: not available
URL: <http://mailman.cse.ohio-state.edu/pipermail/mvapich-discuss/attachments/20180724/36b8b7ba/attachment-0001.bin>