Can't build ictrans, idt (GNU Linux x86_64 and Ifort/gcc)

Dear ncarg-talk members,

As part of Terrain program in MM5 model (by improving performance of it, I
was suggested) , I'm installing NCar Graphics in a AMD Athlon 64 machine
with Mandriva 2007 Linux Distribution.

Well, I've tried and tried so many ways to compile and build for the
x86_64 architecture. I have no problem with ncargex and ctrans (the test
was succesfull at least).

Reading the archive almost overall, I've composed the /config/LINUX file
as this:

#define HdfDefines -DLINUX
#define StdDefines -DSYSV -D_POSIX_SOURCE -D_XOPEN_SOURCE -DByteSwapped
#define ByteSwapped
#define Cstatic
#define Cdynamic
#define CppCommand '/usr/bin/cpp -traditional'
#define CCompiler gcc
#define FCompiler ifort
#define CtoFLibraries -Wl,-Bstatic '-L/opt/intel/fc/9.1.040/lib'
-lCEPCF90 -lF90 -Wl,-Bdynamic -lm -lifcore -L/usr/lib64 -lgcc_s -lg2c
#define CcOptions -ansi -fno-builtin -m32
#define FcOptions -mp -pc32 -convert big_endian -Vaxlib -Kpic -f77rtl
#define LdOptions -i_dynamic
#define FOptimizeFlag -O2
#define XToolLibrary -lXt -lSM -lICE
#define BuildShared NO
#define XLibrary -lX11 -lXext

#define ArchRecLibSearch -L/usr/X11R6/lib64
#define ArchRecIncSearch -I/usr/X11R6/include

FC = $(F77)

This configuration is the best for me, in sense that it's what less errors
are produced.

I know that my system is 64 bits, but when I change the FCOptions and
CcOption to -m64 and -pc64 I've got an ifort's SIGSEV error.

Also, I tried to add a #define statement, named LdOptions, for ld.

My principal problem is the relocatable linking from an input format to an
output format, from elf32-i386 to x86_64-elf and viceverse, because I call
make passing TARGET param to it, and with both options there are diferent
file objects that can be linked after the compile process
(ncarg/ncarg2d/src/liboptional/ directory for example: bitops.c,
commands.c, conraq.c, conras.c, conrec.c, fdvd2d.c... and the directory
for libictrans).

I don't mention yet that I've have to use the -fno-builtin to avoid
diverse warnings...

I thought that maybe a problem with the compiler selected (intel compiler
over an AMD machine), but with f77 I've got similar results.

The last thing I would like to comment with you is a dissapoiment between
Athena library and me... The link with idt fails, but when I search libxaw
in my system I've got the next:

[root_at_HURACAN mm5i]# find / -iname "*libxaw*"
/usr/lib/jvm/java-1.4.2-kaffe-1.4.2.00/jre/lib/x86_64/libxawt-1.1.7.so
/usr/lib/jvm/java-1.4.2-kaffe-1.4.2.00/jre/lib/x86_64/libxawt.la
/usr/lib/jvm/java-1.4.2-kaffe-1.4.2.00/jre/lib/x86_64/libxawt.so
/usr/lib/libXaw.so.6
/usr/lib/libXaw6.so.6
/usr/lib/debug/usr/lib64/libXaw6.so.6.0.1.debug
/usr/lib/debug/usr/lib64/libXaw7.so.7.0.0.debug
/usr/lib/debug/usr/lib64/libXaw8.so.8.0.0.debug
/usr/lib/libXaw6.so.6.0.1
/usr/lib/libXaw.so.7
/usr/lib/libXaw7.so.7
/usr/lib/libXaw7.so.7.0.0
/usr/lib/libXaw.so.8
/usr/lib/libXaw8.so.8
/usr/lib/libXaw8.so.8.0.0
/usr/lib64/libXaw7.so.7
/usr/lib64/libXaw.so.7
/usr/lib64/libXaw7.so.7.0.0
/usr/lib64/libXaw8.so.8.0.0
/usr/lib64/libXaw.so.8
/usr/lib64/libXaw8.so.8
/usr/lib64/libXaw3d.so.7
/usr/lib64/libXaw3d.so.7.0
/usr/lib64/libXaw6.so.6.0.1
/usr/lib64/libXaw.so.6
/usr/lib64/libXaw6.so.6
/usr/lib64/libXaw.so
/usr/lib64/libXaw3d.so
/usr/lib64/libXawM.so.1
/usr/lib64/libXawM.so.1.0.0
/usr/lib64/libXaw6.la
/usr/lib64/libXaw6.so
/usr/lib64/libXaw7.la
/usr/lib64/libXaw7.so
/usr/lib64/libXaw8.la
/usr/lib64/libXaw8.so
/usr/lib64/libXaw6.a
/usr/lib64/libXaw7.a
/usr/lib64/libXaw8.a
/usr/src/debug/libXaw-1.0.2
[root_at_HURACAN mm5i]#

And, in the enviroment I find:

[root_at_HURACAN mm5i]# echo $LD_LIBRARY_PATH
/opt/intel/fc/9.1.040/lib:/usr/lib64:/usr/lib:/opt/intel/fc/9.1.040/lib:/usr/X11R6/lib64
[root_at_HURACAN mm5i]#

There are also other warnings.... but I think that it's better to send you
the make-output... maybe the errors can be related (or not, who knows).

I'll appreciate so so much your helpfull. I've installed all the modules
of the MM5 suite with success apparently, one week ago, and one week it's
just the time I've inverted in solving the problem with no good results.

Sorry so much for the extension of the message, but I want to clarify the
best as possible my problem to you
.

Waiting for your replies. Best regards.

Miriam

_______________________________________________
ncarg-talk mailing list
ncarg-talk_at_ucar.edu
http://mailman.ucar.edu/mailman/listinfo/ncarg-talk