I did not go through to compile ANTECHAMBER(AMBER7) using IFORT/ICC in
SUSE either. At last I used GCC to compile it.
JACK
On Mon, 6 Dec 2004 09:56:34 +0100 (CET), Andreas Svrcek-Seiler
<svrci.tbi.univie.ac.at> wrote:
> Hi,
> > Wornings:
> > icc -c -O2 -o am1bcc.o am1bcc.c
> > rotate.c(41): warning #1572: floating-point equality and inequality comparisons
> > are unreliable
> > if (yc == 0)
> > etc...
> ...these warnings are harmless and just a reminder that one *might* get
> unexpected results when relying on (float) x == (float) y.
>
>
>
> > Errors:
> > icc -O2 atomtype.o -lm -o atomtype
> > icc -c -o bondtype.o bondtype.C
> > icc bondtype.o -lm -o bondtype
> > bondtype.o(.text+0xe982): In function `get_addrs(std::vector<Point, std::allocat
> > or<Point> >)':
> > : undefined reference to `std::__default_alloc_template<true, 0>::deallocate(voi
> > d*, unsigned long)'
> > bondtype.o(.text+0xea92): In function `get_addrs(std::vector<Point, std::allocat
> > or<Point> >)':
> > etc...
>
> ...I just tried compiling the antechamber stuff with icc:
>
> a final
>
> icc bondtype.o /opt/intel_cc_81/lib/libcxa.a /opt/intel_cc_81/lib/libcprts.a /opt/intel_cc_81/lib/libunwind.a -o bondtype
>
> does the trick.
> Since I don't guess the programs in ${AMBERHOME}/src/antechamber are time
> critical (and icc doesn't really always produce faster executables anyway)
> it might be simpler to use gcc/g++ instead.
>
> Note that you migh find the above libraries in another path (most probably
> /opt/intel_cc_80/...)
>
> good luck
> Andreas
>
>
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber.scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Mon Dec 06 2004 - 15:53:00 PST
