This bug applies to OpenMPI 1.4.x and 1.5.x.

Inline assembly does not work for PGI compilers.  configure disables inline assembly for PGI C, but neglects to do the same for PGI C++.  The code that disables inline assembly for PGI C needs to be copied to the section that handles inline assembly for C++.

Here's the diff -u from OpenMPI 1.4.3 (same code, same bug):

[root@hydra openmpi-1.4.3]# diff -u configure{.original,}
--- configure.original 2010-10-05 15:48:18.000000000 -0700
+++ configure 2011-05-17 18:35:04.000000000 -0700
@@ -34690,6 +34690,11 @@
     { $as_echo "$as_me:$LINENO: checking if $CXX supports GCC inline assembly" >&5
 $as_echo_n "checking if $CXX supports GCC inline assembly... " >&6; }
 
+    if test "$ompi_cv_cxx_compiler_vendor" = "portland group" ; then
+        # PGI seems to have some issues with our inline assembly.
+        # Disable for now.
+        asm_result="no (Portland Group)"
+    else
     case $host in
         *-aix*)
             # the AIX compilers and linkers really don't do gcc
@@ -34813,6 +34818,7 @@
 rm -f core conftest.err conftest.$ac_objext conftest_ipa8_conftest.oo \
       conftest$ac_exeext conftest.$ac_ext
     fi
+    fi
 
     { $as_echo "$as_me:$LINENO: result: $asm_result" >&5
 $as_echo "$asm_result" >&6; }

Larry Baker
US Geological Survey
650-329-5608

On 5 May 2011, at 7:15 AM, Jeff Squyres wrote:

Fixed the ROMIO attribute problem properly this time -- it's in the usual place:

   http://www.open-mpi.org/software/ompi/v1.4/

--
Jeff Squyres
jsquyres@cisco.com
For corporate legal information go to:
http://www.cisco.com/web/about/doing_business/legal/cri/


_______________________________________________
devel mailing list
devel@open-mpi.org
http://www.open-mpi.org/mailman/listinfo.cgi/devel