On Mon, 15 Sep 2003, Wouter Brok wrote:
> After succeeding in getting Lam 7.0 to install properly last week (all
> the examples work nicely), I tried to compile an existing project using
> mpicc and mpiCC. The linker gave me back this:
This output looks somewhat fishy...
> /bin/sh ../../libtool --mode=link mpiCC -g -O2 -DWITH_MPI -O2
> -DREAL=double -o libmcbase.la -rpath /usr/local/lib mccrosssec.lo
> mcscattering.lo mcswarm.lo mcparticle.lo mcprocess12.lo mcprocesscont.lo
> mcspecies.lo mcswarmlist.lo mcstatistics.lo mcflightcontrol.lo
> mcenvironment.lo mcswarmstat.lo random.lo mpiCC -shared -nostdlib
> /usr/lib/gcc-lib/i486-suse-linux/3.2/../../../crti.o
> /usr/lib/gcc-lib/i486-suse-linux/3.2/crtbeginS.o .libs/mccrosssec.o
> .libs/mcscattering.o .libs/mcswarm.o .libs/mcparticle.o
> .libs/mcprocess12.o .libs/mcprocesscont.o .libs/mcspecies.o
> .libs/mcswarmlist.o .libs/mcstatistics.o .libs/mcflightcontrol.o
> .libs/mcenvironment.o .libs/mcswarmstat.o .libs/random.o
> -Wl,--whole-archive /usr/local/lib/liblammpi++.a /usr/local/lib/libmpi.a
> /usr/local/lib/liblam.a -Wl,--no-whole-archive -L/usr/lib
> -L/usr/local/lib -L/usr/lib/gcc-lib/i486-suse-linux/3.2
> -L/usr/lib/gcc-lib/i486-suse-linux/3.2/../../../../i486-suse-linux/lib
> -L/usr/lib/gcc-lib/i486-suse-linux/3.2/../../.. -llammpio -lutil
> /usr/lib/libstdc++.so -lm -lc -lgcc_s
> /usr/lib/gcc-lib/i486-suse-linux/3.2/crtendS.o
> /usr/lib/gcc-lib/i486-suse-linux/3.2/../../../crtn.o -Wl,-soname
> -Wl,libmcbase.so.0 -o .libs/libmcbase.so.0.0.0
I can't tell from this output -- is this one command or more? For
example, I see at least 2 mpiCC's in there...
If it's two commands (and the second command starts where the second mpiCC
starts), I'm a little confused as to why all the LAM libraries are showing
up. I'm confused as to where libtool is able to get
/usr/local/lib/libmpi.a, /usr/local/lib/liblam.a, and -llammpio -- I don't
see anything in the initial command that libtool should be able to figure
that out from...?
>From this output, I'm *guessing* that mpiCC is being used as the final
compiler, but is also being given a lot of additional flags (e.g., the
libmpi.a and liblam.a tokens) which might be causing problems. However, I
can't tell for sure because the output is somewhat confusing...
--
{+} Jeff Squyres
{+} jsquyres_at_[hidden]
{+} http://www.lam-mpi.org/
|
