On Wed, 23 Feb 2005, TAK System management wrote:
> Hello,
>
> I'm sorry to disturb you with a question that has probably
> been asked before, but i can't find an answer that solves
> my problem. I'm trying to compile lam-7.1.1 with the intel
> compilers version 8.1. Compiling goes well, no errors, but
> trying to run the executables will result in a segmentation
> fault. Is there someone reading this list who has working
> configuration options for this compiler?
>
> Right now i do:
> $> mkdir log
> $> ./configure --prefix=/apps/share/lam-7.1.1-ic32 \
> --exec-prefix=/apps/share/lam-7.1.1-ic32 \
> CFLAGS="-O -tpp6 -static" \
> CXXFLAGS="-O -tpp6 -static" \
> FFLAGS="-O -tpp6 -static" \
> CXX=icpc \
> CC=icc \
> FC=ifort \
> --without-romio 2>&1 | tee log/configure.log
> $> make 2>&1 | tee make.log
> $> make install 2>&1 | tee install.log
> $> /apps/share/lam-7.1.1-ic32/bin/mpif77
> Segmentation fault
> $> cry for mama
> -bash: cry: command not found
>
> I would be much obliged if some kind soul could give me some
> suggestions or tips. The created log files are attached as
> recommended by 'the LAM/MPI Installation Guide'.
>
> Thank you very much for your time,
>
> Joost de Greef.
> --
> Don't ask for whom the <CTRL-G> tolls.
Intel compilers work fine with LAM, so it must just be something wierd.
To start; What platform are your running on? (P4, Nocona, Prescott, etc.
(Is SSE3 supported in /proc/cpuinfo?). What version of Linux? Try using
gdb to debug mpif77 (you might have to use the "-g" option when compiling
mpif77 to get more usefull info). Just "$> gdb mpif77" and then "bt" when
it segfaults to see which object caused it.
The only weirdness I have ever seen from Intel compilers is that they may
not include the correct library for your system during the final link. If
you have P4 optimized code, it might not link the P4 library, but this is
obvious from the error about missing symbols.
PS: Incompatability between the processor compiled for and the processor
executed on results in an "illegal instruction", not a segfault.
PPS: Could this be a stack size problem? It would really help if we knew
what distro you are using, or kernel version, and what architecture you
are on, as well as your resource limits.
------------------------------------------------------------
Anthony Ciani (aciani1_at_[hidden])
Computational Condensed Matter Physics
Department of Physics, University of Illinois, Chicago
http://ciani.phy.uic.edu/~tony
------------------------------------------------------------
|
