Having (as a non-expert) to compile new versions of applications (Amber10
primarily), I have taken the opportunity to upgrade Intel compilers to version
10.1 (and MKL libraries) and now I plan also to upgrade openmpi from 1.2.3 to
1.2.6. To this concern please see the answer by Tim Prins (tprins
Hello
Due to a new feature - direct mapping between machinefile and rank - I
want to try out the svn version. However, I run into compilation
probloms. Here is what I do after I've updated svn.
kb@gate:/data/kb/compile/openmpi-svn$ ./autogen.sh
kb@gate:/data/kb/compile/openmpi-svn$ mkdir build
Hello Everybody,
i got a little bit annoying situation with OMPI error messages
on a RHEL 4-something box. Every time i should see a error
message i recieve something like:
-
Sorry! You were supposed to get help about:
orterun:init-fai
On May 15, 2008, at 12:58 PM, Wen Hao Wang wrote:
After I run "./configure CC=icc CXX=icpc FC=ifort F77=ifort" and
"make all install", some libs are installed to /usr/local/lib. These
libs can not be found when compiling my program, how should I do?
Can you send all the information listed
Here are the details of my env. My machine has only one ethernet network
adapter. If anything is missed, please let me know.
[root@xblade08 wwh]# uname -a
Linux xblade08.clusters.com 2.6.18-87.el5 #1 SMP Tue Mar 25 17:28:02 EDT
2008 i686 i686 i386 GNU/Linux
[root@xblade08 wwh]# cat /etc/*release
Ah, this is a simple problem. Don't override the CC/etc. for the
lamtests suite; it should default to mpicc and friends. If you
override the CC for the lamtests configure, it'll use those compilers
directly (not mpicc and friends).
Make sense?
On May 16, 2008, at 11:44 AM, Wen Hao Wang
On May 16, 2008, at 5:06 AM, Francesco Pietra wrote:
Having (as a non-expert) to compile new versions of applications
(Amber10 primarily), I have taken the opportunity to upgrade Intel
compilers to version 10.1 (and MKL libraries) and now I plan also to
upgrade openmpi from 1.2.3 to 1.2.6.
Hmm. I just committed this feature to the trunk the other day. Could
you perchance have gotten an intermediate r number that didn't want to
compile? :-(
What r number did you get?
FWIW, my SVN HEAD compiles fine today: r18448.
On May 16, 2008, at 10:17 AM, Karsten Bolding wrote:
Hello
Thanks for the info. But even with the all the default arguments,
"./configure" of lamtests failed with similar error.
[root@xblade08 lamtests-7.1.4]# ./configure
Configuring LAM test suite version 7.1.4
checking for a BSD-compatible install... /usr/bin/install -c
checking whether build environme
Did you setup your environment (particularly LD_LIBRARY_PATH):
http://www.open-mpi.org/faq/?category=running#run-prereqs
http://www.open-mpi.org/faq/?category=running#adding-ompi-to-path
http://www.open-mpi.org/faq/?category=running#mpirun-prefix
On May 16, 2008, at 12:35 PM, Wen Ha
just did a clean check-out:
kb@gate:/data/kb/compile/openmpi-svn$ svn update
At revision 18449.
and I get the same problem:
gcc -O3 -DNDEBUG -finline-functions -fno-strict-aliasing -pthread
-fvisibility=hidden -o .libs/orted orted.o -Wl,--export-dynamic
../../../orte/.libs/libopen-rte.so -lnsl -
Truly weird. Can you send all the information here:
http://www.open-mpi.org/community/help/
On May 16, 2008, at 2:57 PM, Karsten Bolding wrote:
just did a clean check-out:
kb@gate:/data/kb/compile/openmpi-svn$ svn update
At revision 18449.
and I get the same problem:
gcc -O3 -DNDEBUG
here is comes
On Fri, May 16, 2008 at 03:07:24 -0400, Jeff Squyres wrote:
> Truly weird. Can you send all the information here:
>
> http://www.open-mpi.org/community/help/
>
>
--
--
Karsten BoldingBo
I am unable to duplicate your problem (I get a failure later in the
build with f90 stuff, but that's likely an issue of finding compatible
compiler flags between ifort and gcc...?).
I wonder if our difference is that you're using LT 1.2.26 -- not
something that I have tried in a long, long
On Fri, May 16, 2008 at 04:06:05 -0400, Jeff Squyres wrote:
> I am unable to duplicate your problem (I get a failure later in the
> build with f90 stuff, but that's likely an issue of finding compatible
> compiler flags between ifort and gcc...?).
I've tried with gfortran instead of ifort and
On May 16, 2008, at 5:17 PM, Karsten Bolding wrote:
I wonder if our difference is that you're using LT 1.2.26 -- not
something that I have tried in a long, long time (if ever -- I don't
remember the last version of the LT 1.2 series that I tried). Can
you
upgrade to LT 2.2 or higher?
its 1
16 matches
Mail list logo
