Jeff,
I have not yet finished finding all the problems I'm seeing in
1.5rc5. All I have done so far is look for the ones that I fixed for
OpenMPI 1.4.2. For example, compiling with PGI 10.3 on Linux fails.
(Has anyone else tried PGI 10.x?) The Intel compiler issues messages
I haven't tracked down on Mac OS X, and I haven't even tried PathScale
or Intel on Linux yet.
Larry Baker
US Geological Survey
650-329-5608
ba...@usgs.gov
On Aug 20, 2010, at 8:23 AM, Open MPI wrote:
#2538: PGI compiler issues
---------------------
+------------------------------------------------------
Reporter: jsquyres | Owner:
Type: defect | Status: new
Priority: major | Milestone: Open MPI 1.5.1
Version: trunk | Keywords:
---------------------
+------------------------------------------------------
Larry Baker posted a bunch of PGI compiler issues:
* http://www.open-mpi.org/community/lists/devel/2010/08/8297.php
* http://www.open-mpi.org/community/lists/devel/2010/08/8298.php
* http://www.open-mpi.org/community/lists/devel/2010/08/8299.php
* http://www.open-mpi.org/community/lists/devel/2010/08/8300.php
* http://www.open-mpi.org/community/lists/devel/2010/08/8301.php
Some of them are Libtool issues, but some of them are OMPI issues.
None
appear to be drastic correctness issues (nor are they regressions), so
they shouldn't hold up 1.5.0. But we should fix them nonetheless (and
submit upstream to Libtool where appropriate).
--
Ticket URL: <https://svn.open-mpi.org/trac/ompi/ticket/2538>
Open MPI <http://www.open-mpi.org/>
