AMBER Archive (2009)

Subject: Re: [AMBER] Failure to compile amber9 serial ifort gcc

From: Francesco Pietra (chiendarret_at_gmail.com)
Date: Thu Apr 02 2009 - 06:16:36 CDT

I had already checked

which ifort
/opt/intel/fce/10.1.015/bin/ifort

in my .bashrc:

. /opt/intel/fce/10.1.015/bin/ifortvars.sh (and iccvars.sh)

However, there may be something wrong (I have replaced the two disks
of raid1 with larger ones, reinstalled lenny on the new ones, and I am
taking stuff from backup of previous disks when needed): I tried to
install openmpi-1.3.1 instead of using the executables openmpi-1.2.6
of previous disks. I configured as for 1.2.6

CC=/opt/intel/cce/10.1.015/bin/icc CXX=opt/intel/cce/10.1.015/bin/icpc
F77=/opt/intel/fce/10.1.015/bin/ifort
FC=/opt/intel/fce/10.1.015/bin/ifort --with-libnuma=/usr/lib

getting:

checking whether using GNU C++ compiler .. yes

dependency style icpc ... gcc3

checking how to run C++ preprocessor ... /lib/cpp

checking for compiler vendor ... intel

cheching if C++ compiler works .. NO

it is not clear to me to which compiler NO refers: gcc or intel?

Therefore I copied the executables from /usr/local/bin from previous
disks and also the amber9 installation, which passed both serial and
parallel. Unfortunately, this compilation must be not fully pathched
(the bugfix.all file still in the directory is 766114 bytes, while
current bugfix.all is 1475863). I need a fully patched compilation of
amber9. Therefore, I wonder whether the problem of the compilation is
with C++.

thanks
francesco

On Wed, Apr 1, 2009 at 11:28 PM, Alessandro Nascimento
<al.s.nascimento_at_gmail.com> wrote:
> Hi,
>
>> make[1]: ifort: Command not found
>
> make sure you have the necessary intel compiler environment variables
> set before you start compiling the source code:
> Should be something like (in my case):
>
> source /opt/intel/Compiler/11.0/081/bin/intel64/ifortvars_intel64.s
>
> Testing:
>
> $which ifort
> /usr/bin/which: no ifort in
> (/opt/gridengine/bin/lx26-amd64:/usr/kerberos/bin:/opt/gridengine/bin/lx26-amd64:/usr/java/latest/bin:/usr/local/bin:/bin:/usr/bin:/opt/eclipse:/opt/ganglia/bin:/opt/ganglia/sbin:/opt/maven/bin:/opt/openmpi/bin/:/opt/rocks/bin:/opt/rocks/sbin:/home/asn/bin)
>
> $ source /opt/intel/Compiler/11.0/081/bin/intel64/ifortvars_intel64.sh
>
> $ which ifort
> /opt/intel/Compiler/11.0/081/bin/intel64/ifort
>
> $cd $AMBERHOME/src
>
> $./configure -nopar -bintraj ifort_x86_64
>
> $ make clean; make serial
>
> Works for me with amber9.
>
> Hope that helps,
>
> --asn
>
>
>
> On Wed, Apr 1, 2009 at 6:19 PM, Francesco Pietra <chiendarret_at_gmail.com> wrote:
>> Of the various choices of yacc-like I installed bison. That issue of
>> yacc overcome, "make clean" was still unhappy, perhaps not so much.
>> Last portion:
>>
>> make[1]: *** No rule to make target `macros.make'.  Stop.
>> make[1]: Leaving directory `/usr/local/amber9/src/netcdf/src'
>> make: [clean] Error 2 (ignored)
>> cd netcdf/lib && rm -f libnetcdf.a
>> /bin/sh: line 0: cd: netcdf/lib: No such file or directory
>> make: [clean] Error 1 (ignored)
>> cd netcdf/include && rm -f *.mod
>> /bin/sh: line 0: cd: netcdf/include: No such file or directory
>> make: [clean] Error 1 (ignored)
>>
>>
>>
>>
>> but the output of makewas  as short as:
>>
>> Starting installation of Amber9 (serial) at Wed Apr  1 23:02:25 CEST 2009.
>> cd lib; make install
>> make[1]: Entering directory `/usr/local/amber9/src/lib'
>> cpp -traditional -P -DMKL  new2oldparm.f > _new2oldparm.f
>> ifort -c -w95   -mp1 -O0   -o new2oldparm.o _new2oldparm.f
>> make[1]: ifort: Command not found
>> make[1]: *** [new2oldparm.o] Error 127
>> make[1]: Leaving directory `/usr/local/amber9/src/lib'
>> make: *** [serial] Error 2
>> -----------
>> to be compared with previous attempt, when no yacc equivalent was installed:
>>
>> Starting installation of Amber9 (serial) at Wed Apr  1 12:13:20 CEST 2009.
>> cd lib; make install
>> make[1]: Entering directory `/usr/local/amber9/src/lib'
>> cpp -traditional -P -DMKL  new2oldparm.f > _new2oldparm.f
>> ifort -c -w95   -mp1 -O0   -o new2oldparm.o _new2oldparm.f
>> cpp -traditional -P -DMKL  nxtsec.f > _nxtsec.f
>> ifort -c -w95   -mp1 -O0   -o nxtsec.o _nxtsec.f
>> ifort     -o new2oldparm new2oldparm.o nxtsec.o
>> mv new2oldparm ../../exe
>> /bin/rm _*.f
>> make[1]: Leaving directory `/usr/local/amber9/src/lib'
>> cd addles; make install
>> make[1]: Entering directory `/usr/local/amber9/src/addles'
>> cpp -traditional -P -DMKL  lesmain.f > _lesmain.f
>> ..............................
>>
>> fp
>>
>>
>>
>> On Wed, Apr 1, 2009 at 1:09 PM, Alessandro Nascimento
>> <al.s.nascimento_at_gmail.com> wrote:
>>> I think make is giving you a clue!
>>>
>>>
>>>  make[2]: yacc: Command not found
>>>
>>> So, `sudo apt-get install bison`, or something similar.
>>>
>>> HTH,
>>>
>>>
>>> --asn
>>>
>>> 2009/4/1 Francesco Pietra <chiendarret_at_gmail.com>:
>>>> Hi:
>>>> On a multisocket uma-type board with dual-opterons, following failure
>>>> of one disk of raid1, I have changed both with larger disks and
>>>> installed from scratch debian linux amd64 lenny raid1, intel compilers
>>>> 10.1.015 and mkl lib.
>>>>
>>>> As I want to install a modified version of amber9, I took from
>>>> previous working disk the ifort-gcc-mkl install of amber9, replaced
>>>> modified files in sander, recompiled, ending in failure.
>>>>
>>>> In the remote hypothesis that previous installation was not fully
>>>> patched, I started from scratch, i.e.amber9 source. Patched with the
>>>> complete bugfix.all. To simplify, "chown -R francesco
>>>> /usr/local/amber9". First I tried to compile the standard
>>>> fully-patched version, ending in failure again:
>>>>
>>>> $ ./configure ifort_x86_64
>>>>
>>>> $ make clean 2>&1 | tee make.clean.out
>>>>
>>>> $ make 2>&1 | tee make.out
>>>>
>>>>
>>>>
>>>> make.clean.out reported what I guess are irrelevant errors. Last part reads:
>>>>
>>>> rules.make:142: warning: ignoring old commands for target `/man3f90'
>>>> make[1]: *** No rule to make target `macros.make'.  Stop.
>>>> make[1]: Leaving directory `/usr/local/amber9/src/netcdf/src'
>>>> make: [clean] Error 2 (ignored)
>>>> cd netcdf/lib && rm -f libnetcdf.a
>>>> /bin/sh: line 0: cd: netcdf/lib: No such file or directory
>>>> make: [clean] Error 1 (ignored)
>>>> cd netcdf/include && rm -f *.mod
>>>> /bin/sh: line 0: cd: netcdf/include: No such file or directory
>>>> make: [clean] Error 1 (ignored)
>>>>
>>>>
>>>> make.out ended in serious errors. Last part reads (attached please
>>>> find the whole out file):
>>>>
>>>> build.c: In function ‘BuildInternalsForSimpleRings’:
>>>> build.c:2185: warning: passing argument 1 of
>>>> ‘GraphUtilFindAllSmallestRingsAndRingGroups’ from incompatible pointer
>>>> type
>>>> gcc -c -I/usr/X11R6/include -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
>>>> -m64 -O2 -o elements.o elements.c
>>>> gcc -c -I/usr/X11R6/include -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
>>>> -m64 -O2 -o library.o library.c
>>>> gcc -c -I/usr/X11R6/include -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
>>>> -m64 -O2 -o chirality.o chirality.c
>>>> gcc -c -I/usr/X11R6/include -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
>>>> -m64 -O2 -o minimizer.o minimizer.c
>>>> gcc -c -I/usr/X11R6/include -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
>>>> -m64 -O2 -o model.o model.c
>>>> model.c: In function ‘zModelOrderAtoms’:
>>>> model.c:497: warning: passing argument 1 of ‘Sift’ from incompatible
>>>> pointer type
>>>> gcc -c -I/usr/X11R6/include -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
>>>> -m64 -O2 -o parmLib.o parmLib.c
>>>> gcc -c -I/usr/X11R6/include -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
>>>> -m64 -O2 -o pdbFile.o pdbFile.c
>>>> pdbFile.c: In function ‘zPdbBuildCoordinatesForContainer’:
>>>> pdbFile.c:1088: warning: passing argument 1 of ‘BuildFixInternals’
>>>> from incompatible pointer type
>>>> pdbFile.c: In function ‘zPdbAddAtom’:
>>>> pdbFile.c:1207: warning: passing argument 1 of
>>>> ‘BuildInternalsForContainer’ from incompatible pointer type
>>>> pdbFile.c: In function ‘writeTER’:
>>>> pdbFile.c:1761: warning: initialization from incompatible pointer type
>>>> gcc -c -I/usr/X11R6/include -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
>>>> -m64 -O2 -o tools.o tools.c
>>>> tools.c: In function ‘iToolDistanceSearch’:
>>>> tools.c:1643: warning: passing argument 1 of ‘SortByDouble’ from
>>>> incompatible pointer type
>>>> tools.c:1643: warning: passing argument 4 of ‘SortByDouble’ from
>>>> incompatible pointer type
>>>> tools.c: In function ‘ToolOrientPrincipleAxisAlongCoordinateAxis’:
>>>> tools.c:1714: warning: passing argument 2 of ‘MathOpDiagonalize’ from
>>>> incompatible pointer type
>>>> gcc -c -I/usr/X11R6/include -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
>>>> -m64 -O2 -o variables.o variables.c
>>>> yacc parser.y
>>>> make[2]: yacc: Command not found
>>>> make[2]: *** [parser.c] Error 127
>>>> make[2]: Leaving directory `/usr/local/amber9/src/leap/src/leap'
>>>> make[1]: *** [install] Error 2
>>>> make[1]: Leaving directory `/usr/local/amber9/src/leap'
>>>> make: *** [serial] Error 2
>>>>
>>>> As far as I understand, the errors are related to X. A frequent "ifort
>>>> command line option -tp not supported" is probably irrelevant)
>>>>
>>>> What I checked:  startx provides the X server and a minimal window
>>>> (enough for xleap in my experience). Package "xorg-dev" and the
>>>> environment for compilation, package "build-essential" are installed
>>>> correctly.
>>>>
>>>> gcc -v  4.3.2
>>>>
>>>> make -v 3.81
>>>>
>>>> Curiously, I had no problems with previous identical installation on
>>>> debian amd64 etch, which passed without any need of recompiling to
>>>> debian amd64 lenny. Unfortunately I did not save the output of make of
>>>> the original successful installation.
>>>>
>>>> thanks for help
>>>>
>>>> francesco pietra
>>>>
>>>> _______________________________________________
>>>> AMBER mailing list
>>>> AMBER_at_ambermd.org
>>>> http://lists.ambermd.org/mailman/listinfo/amber
>>>>
>>>>
>>>
>>>
>>>
>>> --
>>> [ ]s
>>>
>>> --alessandro
>>>
>>
>
>
>
> --
> [ ]s
>
> --alessandro
>

_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber