From: Zaak Beekman <zbeekman@gmail.com>
To: fortran@gcc.gnu.org
Cc: Damian Rouson <damian@sourceryinstitute.org>,
Soren R <s.c.rasmussen@gmail.com>,
"Rasmussen, Soren" <S.Rasmussen@cranfield.ac.uk>,
Daniel Celis Garza <daniel.celisgarza@hmc.ox.ac.uk>
Subject: Regressions introduced very late before 8.1 release
Date: Fri, 04 May 2018 15:13:00 -0000 [thread overview]
Message-ID: <CAAbnBwYsUz2=deKA_0gS-Q=PE_UtpFK=x2EVsZNF7pdgT-beQg@mail.gmail.com> (raw)
I'm not sure when this happened, but at some point very close to the 8.1
release date a regression was introduced (during the GCC regression fix
only freeze period) that has significantly broken distributed memory
coarray support with OpenCoarrays. Simple coarray Fortran syntax appears to
trigger an ICE in GFortran 8.1 on macOS. Here is an example from
[sameloc.f90](
https://github.com/sourceryinstitute/OpenCoarrays/blob/master/src/tests/unit/send-get/sameloc.f90#L31)
compiled on macOS:
/Users/ibeekman/Sandbox/OpenCoarrays/src/tests/unit/send-get/sameloc.f90:31:0:
a = a[1]
internal compiler error: in gfc_dep_resolver, at fortran/dependency.c:2258
libbacktrace could not find executable to open
Please submit a full bug report,
with preprocessed source if appropriate.
See <https://github.com/Homebrew/homebrew-core/issues> for instructions.
Error: comand:
`/usr/local/bin/gfortran-8
-I/Users/ibeekman/Sandbox/OpenCoarrays/build/include/OpenCoarrays-2.0.0-33-g5a250a1_GNU-8.1.0
-fcoarray=lib -Wl,-flat_namespace -Wl,-commons,use_dylibs -I
/usr/local/Cellar/mpich/3.2.1_2/include -DPREFIX_NAME=_gfortran_caf_
-DGCC_GE_7 -DGCC_GE_8 -DHAVE_MPI -DMPI_WORKING_MODULE -O3 -DNDEBUG -O3 -o
/Users/ibeekman/Sandbox/OpenCoarrays/build/bin/OpenCoarrays-2.0.0-33-g5a250a1-tests/get_self
/Users/ibeekman/Sandbox/OpenCoarrays/src/tests/unit/send-get/sameloc.f90
/Users/ibeekman/Sandbox/OpenCoarrays/build/lib/libcaf_mpi.a
/usr/local/Cellar/mpich/3.2.1_2/lib/libmpifort.dylib
/usr/local/Cellar/mpich/3.2.1_2/lib/libmpi.dylib
/usr/local/Cellar/mpich/3.2.1_2/lib/libpmpi.dylib`
failed to compile.
I have a GCC trunk build from r259376 on 2018-04-13 at 14:32:41 (not sure
if UTC or US Eastern time) that does not exhibit this behavior. I have yet
to post to bugzilla, but presumably I should. How did this ice sneak in so
late before the release? Because of this, OpenCoarrays will likely be
removed from the Homebrew package manager unless we can create a fix on the
OpenCoarrays library side, which looks unlikely.
How soon is 8.1 going to be released?
Thanks,
Zaak
next reply other threads:[~2018-05-04 15:13 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-05-04 15:13 Zaak Beekman [this message]
2018-05-04 16:15 Dominique d'Humières
2018-05-04 16:47 ` Zaak Beekman
2018-05-04 17:02 ` Steve Kargl
2018-05-04 17:03 ` Dominique d'Humières
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CAAbnBwYsUz2=deKA_0gS-Q=PE_UtpFK=x2EVsZNF7pdgT-beQg@mail.gmail.com' \
--to=zbeekman@gmail.com \
--cc=S.Rasmussen@cranfield.ac.uk \
--cc=damian@sourceryinstitute.org \
--cc=daniel.celisgarza@hmc.ox.ac.uk \
--cc=fortran@gcc.gnu.org \
--cc=s.c.rasmussen@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).