Re: Build failures of 1.2.3 on Debian hppa, mips, mipsel, s390, m68k
- To: "Carlos O'Donell" <carlos@systemhalted.org>
- Cc: "Dirk Eddelbuettel" <edd@debian.org>, devel@open-mpi.org, pkg-openmpi-maintainers@lists.alioth.debian.org, debian-hppa@lists.debian.org, debian-mips@lists.debian.org, debian-m68k@lists.debian.org
- Subject: Re: Build failures of 1.2.3 on Debian hppa, mips, mipsel, s390, m68k
- From: Dirk Eddelbuettel <edd@debian.org>
- Date: Sat, 14 Jul 2007 11:53:16 -0500
- Message-id: <18072.65404.644311.220832@basebud.nulle.part>
- In-reply-to: <119aab440707140803l77636566nfa4e4bf7ee8764ba@mail.gmail.com>
- References: <18072.56590.326202.513472@basebud.nulle.part> <119aab440707140803l77636566nfa4e4bf7ee8764ba@mail.gmail.com>
Hi Carlos,
Thanks for the quick reply.
On 14 July 2007 at 11:03, Carlos O'Donell wrote:
| On 7/14/07, Dirk Eddelbuettel <edd@debian.org> wrote:
| > i) that hppa (in the way configure sees it) is not supported:
| >
| > checking if .size is needed... yes
| > checking if .align directive takes logarithmic value... no
| > configure: error: No atomic primitives available for hppa-unknown-linux-gnu
| > make: *** [config.status] Error 1
| >
| > Now, configure and aclocal have lots of 'hppa*64' statements. Is it
| > enough to turn these into 'hppa*64*|hppa*linux*' or something similar ?
| >
| > This issue has previously been logged in the Debian Bug Tracking System,
| > see http://bugs.debian.org/431631
|
| That bug does not appear to have any relevance to the failed configure check.
| What atomic primitives does Open MPI need?
I am confused. I am not sure I understand your question. Are you aware that
configure checks for this? Eg from my x86 build logs:
checking for pre-built assembly file... yes (atomic-ia32-linux.s)
checking for atomic assembly filename... atomic-ia32-linux.s
So atomic-$foo better be there for a given architecture foo as there are
sources for some platforms:
edd@basebud:~/src/debian/SVN/tarballs/openmpi-1.2.3/opal> ls -1 asm/generated/
atomic-alpha-linux.s
atomic-amd64-linux-nongas.s
atomic-amd64-linux.s
atomic-ia32-cygwin-nongas.s
atomic-ia32-cygwin.s
atomic-ia32-linux-nongas.s
atomic-ia32-linux.s
atomic-ia32-osx.s
atomic-ia64-linux-nongas.s
atomic-ia64-linux.s
atomic-mips-irix.s
atomic-powerpc32-64-osx.s
atomic-powerpc32-aix.s
atomic-powerpc32-linux-nongas.s
atomic-powerpc32-linux.s
atomic-powerpc32-osx.s
atomic-powerpc64-aix.s
atomic-powerpc64-linux-nongas.s
atomic-powerpc64-linux.s
atomic-powerpc64-osx.s
atomic-sparc-solaris.s
atomic-sparcv9-32-solaris.s
atomic-sparcv9-64-solaris.s
Methinks we need to fill in a few blanks here, or make do with non-asm
solutions. I don't know the problem space that well (being a maintainer
rather than upstream developer) and am looking for guidance.
For what it's worth, lam (7.1.2, currently) us available on all build
architectures for Debian, but it may not push the (hardware) envelope as
hard.
Hope this helps, Dirk
--
Hell, there are no rules here - we're trying to accomplish something.
-- Thomas A. Edison
Reply to:
- Prev by Date:
Re: [OMPI devel] Build failures of 1.2.3 on Debian hppa, mips, mipsel, s390, m68k
- Next by Date:
Re: [OMPI devel] Build failures of 1.2.3 on Debian hppa, mips, mipsel, s390, m68k
- Previous by thread:
Re: Build failures of 1.2.3 on Debian hppa, mips, mipsel, s390, m68k
- Next by thread:
Re: [OMPI devel] Build failures of 1.2.3 on Debian hppa, mips, mipsel, s390, m68k
- Index(es):