[OMPI devel] 1.7.5rc5 build failure [with patch]
Sorry to be the bearer of bad news... I configured on an SGI Altix UV with only "--enable-debug --enable-static --prefix=..." The configure script found the sn/xpmem.h header and threfore enabled btl:vader: --- MCA component btl:vader (m4 configuration macro) checking for MCA component btl:vader compile mode... static checking --with-xpmem value... simple ok (unspecified) checking --with-xpmem-libdir value... simple ok (unspecified) checking xpmem.h usability... no checking xpmem.h presence... no checking for xpmem.h... no checking sn/xpmem.h usability... yes checking sn/xpmem.h presence... yes checking for sn/xpmem.h... yes looking for library without search path checking for xpmem_make in -lxpmem... yes checking if MCA component btl:vader can compile... yes However, not only was btl:vader failing at runtime on the Altix when I last checked, but the version of the code in 1.7.5rc5 won't build: Making all in mca/btl/vader make[2]: Entering directory `/brashear/hargrove/OMPI/openmpi-1.7.5rc5-linux-x86_64-uv/BLD/ompi/mca/btl/vader' CC btl_vader_module.lo In file included from /usr/users/6/hargrove/SCRATCH/OMPI/openmpi-1.7.5rc5-linux-x86_64-uv/openmpi-1.7.5rc5/ompi/mca/btl/vader/btl_vader.h:76, from /usr/users/6/hargrove/SCRATCH/OMPI/openmpi-1.7.5rc5-linux-x86_64-uv/openmpi-1.7.5rc5/ompi/mca/btl/vader/btl_vader_module.c:26: /usr/users/6/hargrove/SCRATCH/OMPI/openmpi-1.7.5rc5-linux-x86_64-uv/openmpi-1.7.5rc5/ompi/mca/btl/vader/btl_vader_endpoint.h:28:19: error: xpmem.h: No such file or directory I think the safest/fastest fix is the following 1-line change to prevent locating sn/xpmem.h, which is Altix-specific: --- ompi/mca/btl/vader/configure.m4 (revision 31166) +++ ompi/mca/btl/vader/configure.m4 (working copy) @@ -23,7 +23,7 @@ AC_ARG_WITH([xpmem], [AC_HELP_STRING([--with-xpmem(=DIR)], [Build with XPMEM kernel module support, searching for headers in DIR])]) -OMPI_CHECK_WITHDIR([xpmem], [$with_xpmem], [include/xpmem.h include/sn/xpmem.h]) +OMPI_CHECK_WITHDIR([xpmem], [$with_xpmem], [include/xpmem.h]) AC_ARG_WITH([xpmem-libdir], [AC_HELP_STRING([--with-xpmem-libdir=DIR], This patch is intended for 1.7 only - lets leave the trunk as-is. -Paul -- Paul H. Hargrove phhargr...@lbl.gov Future Technologies Group Computer and Data Sciences Department Tel: +1-510-495-2352 Lawrence Berkeley National Laboratory Fax: +1-510-486-6900
[OMPI devel] 1.7.5rc5 still building oshmem by default?
Maybe I misunderstood, but I though it had been decided 1.7.5 was *not* going to build oshmem by default. If I am mistaken, then feel free to ignore the rest of this post. Though I don't configure with any options which mention oshmem, I still see (known as ticket #4262) failures linking the oshmem examples on xlc/ppc32/linux: shmemfort -g ring_oshmemfh.f90 -o ring_oshmemfh ** ring_oshmem === End of Compilation 1 === 1501-510 Compilation successful for file ring_oshmemfh.f90. ring_oshmemfh.o: In function `ring_oshmem': /gpfs-biou/phh1/OMPI/openmpi-1.7.5rc5-linux-ppc32-xlc-11.1/BLD/examples/ring_oshmemfh.f90:35: undefined reference to `shmem_put8' /gpfs-biou/phh1/OMPI/openmpi-1.7.5rc5-linux-ppc32-xlc-11.1/BLD/examples/ring_oshmemfh.f90:48: undefined reference to `shmem_int8_wait_until' /gpfs-biou/phh1/OMPI/openmpi-1.7.5rc5-linux-ppc32-xlc-11.1/BLD/examples/ring_oshmemfh.f90:55: undefined reference to `shmem_put8' -Paul -- Paul H. Hargrove phhargr...@lbl.gov Future Technologies Group Computer and Data Sciences Department Tel: +1-510-495-2352 Lawrence Berkeley National Laboratory Fax: +1-510-486-6900
Re: [OMPI devel] 1.7.5rc5 still building oshmem by default?
No, it builds by default on 1.7.5 - if we see too many issues, then we'll turn it off unless requested On Mar 19, 2014, at 10:31 PM, Paul Hargrove wrote: > Maybe I misunderstood, but I though it had been decided 1.7.5 was *not* going > to build oshmem by default. If I am mistaken, then feel free to ignore the > rest of this post. > > Though I don't configure with any options which mention oshmem, I still see > (known as ticket #4262) failures linking the oshmem examples on > xlc/ppc32/linux: > > shmemfort -g ring_oshmemfh.f90 -o ring_oshmemfh > ** ring_oshmem === End of Compilation 1 === > 1501-510 Compilation successful for file ring_oshmemfh.f90. > ring_oshmemfh.o: In function `ring_oshmem': > /gpfs-biou/phh1/OMPI/openmpi-1.7.5rc5-linux-ppc32-xlc-11.1/BLD/examples/ring_oshmemfh.f90:35: > undefined reference to `shmem_put8' > /gpfs-biou/phh1/OMPI/openmpi-1.7.5rc5-linux-ppc32-xlc-11.1/BLD/examples/ring_oshmemfh.f90:48: > undefined reference to `shmem_int8_wait_until' > /gpfs-biou/phh1/OMPI/openmpi-1.7.5rc5-linux-ppc32-xlc-11.1/BLD/examples/ring_oshmemfh.f90:55: > undefined reference to `shmem_put8' > > > -Paul > > -- > Paul H. Hargrove phhargr...@lbl.gov > Future Technologies Group > Computer and Data Sciences Department Tel: +1-510-495-2352 > Lawrence Berkeley National Laboratory Fax: +1-510-486-6900 > ___ > devel mailing list > de...@open-mpi.org > Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/devel > Link to this post: > http://www.open-mpi.org/community/lists/devel/2014/03/14377.php
[OMPI devel] Delayed SVN/Trac mails
You may have noticed that some of the SVN and Trac mails were delayed yesterday, sometimes by a few hours. IU had some email difficulties yesterday; they have been fixed. The backlog is likely cleared out by now, but a few more may trickle in over the course of today. -- Jeff Squyres jsquy...@cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/
Re: [OMPI devel] 1.7.5rc5 build failure [with patch]
Talked to Nathan and he concurs with your proposed fix, so I'll add it only to v1.7 (leaving trunk alone) Thanks Paul! On Mar 19, 2014, at 9:40 PM, Paul Hargrove wrote: > Sorry to be the bearer of bad news... > > I configured on an SGI Altix UV with only "--enable-debug --enable-static > --prefix=..." > > The configure script found the sn/xpmem.h header and threfore enabled > btl:vader: > > --- MCA component btl:vader (m4 configuration macro) > checking for MCA component btl:vader compile mode... static > checking --with-xpmem value... simple ok (unspecified) > checking --with-xpmem-libdir value... simple ok (unspecified) > checking xpmem.h usability... no > checking xpmem.h presence... no > checking for xpmem.h... no > checking sn/xpmem.h usability... yes > checking sn/xpmem.h presence... yes > checking for sn/xpmem.h... yes > looking for library without search path > checking for xpmem_make in -lxpmem... yes > checking if MCA component btl:vader can compile... yes > > However, not only was btl:vader failing at runtime on the Altix when I last > checked, but the version of the code in 1.7.5rc5 won't build: > > Making all in mca/btl/vader > make[2]: Entering directory > `/brashear/hargrove/OMPI/openmpi-1.7.5rc5-linux-x86_64-uv/BLD/ompi/mca/btl/vader' > CC btl_vader_module.lo > In file included from > /usr/users/6/hargrove/SCRATCH/OMPI/openmpi-1.7.5rc5-linux-x86_64-uv/openmpi-1.7.5rc5/ompi/mca/btl/vader/btl_vader.h:76, > from > /usr/users/6/hargrove/SCRATCH/OMPI/openmpi-1.7.5rc5-linux-x86_64-uv/openmpi-1.7.5rc5/ompi/mca/btl/vader/btl_vader_module.c:26: > /usr/users/6/hargrove/SCRATCH/OMPI/openmpi-1.7.5rc5-linux-x86_64-uv/openmpi-1.7.5rc5/ompi/mca/btl/vader/btl_vader_endpoint.h:28:19: > error: xpmem.h: No such file or directory > > I think the safest/fastest fix is the following 1-line change to prevent > locating sn/xpmem.h, which is Altix-specific: > > --- ompi/mca/btl/vader/configure.m4 (revision 31166) > +++ ompi/mca/btl/vader/configure.m4 (working copy) > @@ -23,7 +23,7 @@ > AC_ARG_WITH([xpmem], > [AC_HELP_STRING([--with-xpmem(=DIR)], > [Build with XPMEM kernel module support, searching for > headers in DIR])]) > -OMPI_CHECK_WITHDIR([xpmem], [$with_xpmem], [include/xpmem.h > include/sn/xpmem.h]) > +OMPI_CHECK_WITHDIR([xpmem], [$with_xpmem], [include/xpmem.h]) > > AC_ARG_WITH([xpmem-libdir], > [AC_HELP_STRING([--with-xpmem-libdir=DIR], > > This patch is intended for 1.7 only - lets leave the trunk as-is. > > -Paul > > -- > Paul H. Hargrove phhargr...@lbl.gov > Future Technologies Group > Computer and Data Sciences Department Tel: +1-510-495-2352 > Lawrence Berkeley National Laboratory Fax: +1-510-486-6900 > ___ > devel mailing list > de...@open-mpi.org > Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/devel > Link to this post: > http://www.open-mpi.org/community/lists/devel/2014/03/14376.php
