Re: [AMBER] Amber 20 compilation with the intel 2019.1.1 compiler

From: Scott Brozell <sbrozell.iqb.rutgers.edu>
Date: Tue, 20 Oct 2020 17:37:31 -0400

Hi,

The legacy build made metatwist for me (on the git master):
./configure --miniconda -verbose intel
...
[MoFT] CXX metatwist.cpp
...
icpc -std=c++0x -o /tmp/amber/bin/metatwist metatwist.o -L/tmp/amber/lib -lboost_program_options -lboost_iostreams -lboost_regex -lboost_system -lboost_timer -lboost_chrono -lboost_filesystem -lboost_graph -mkl=sequential -lz -lbz2

2972 -rwxr-xr-x 1 scott 3041344 Oct 19 22:02 ./bin/metatwist


The errors below are all boost related.
What happened during this stage of configure ?
Configuring boost (may be time-consuming)...

You can look in file AmberTools/src/boost_config.log

scott

On Tue, Oct 20, 2020 at 01:50:49PM -0400, David A Case wrote:
> On Tue, Oct 20, 2020, Vlad Cojocaru wrote:
> >However, after passing this, I am receiving lots of undefined reference
> >errors from the metatwist source code.
> >
> >The compilation is after successful ./configure -noX11 intel (intel
> >19.1.1) (no complains from configure at all, all tests for requirements
> >passed including for bzip2 and zlib).
>
> Something is broken with the legacy build and metatwist. Simple option
> is just to comment out that part of the build (in
> $AMBERHOME/src/Makefile).

On Tue, Oct 20, 2020 at 04:46:37PM +0200, Vlad Cojocaru wrote:
> Thanks to Scott's answer I solved the previous issue that indeed was
> related to linking the intel compiler against an old gcc version.
> However, after passing this, I am receiving lots of undefined reference
> errors from the metatwist source code.
>
> The compilation is after successful ./configure -noX11 intel (intel
> 19.1.1) (no complains from configure at all, all tests for requirements
> passed including for bzip2 and zlib).
>
> Does anyone have any clue about solving these issue ?
>
> Best wishes
> Vlad
>
> ld:
> metatwist.o:(.data.rel.ro._ZTVN5boost15program_options11typed_valueISt6vectorIdSaIdEEcEE[_ZTVN5boost15program_options11typed_valueISt6vectorIdSaIdEEcEE]+0x38):
> undefined reference to
> `boost::program_options::value_semantic_codecvt_helper<char>::parse(boost::any&,
> std::vector<std::__cxx11::basic_string<char, std::char_traits<char>,
> std::allocator<char> >, std::allocator<std::__cxx11::basic_string<char,
> std::char_traits<char>, std::allocator<char> > > > const&, bool) const'
> ld:
> metatwist.o:(.data.rel.ro._ZTVN5boost15program_options11typed_valueImcEE[_ZTVN5boost15program_options11typed_valueImcEE]+0x38):
> undefined reference to
> `boost::program_options::value_semantic_codecvt_helper<char>::parse(boost::any&,
> std::vector<std::__cxx11::basic_string<char, std::char_traits<char>,
> std::allocator<char> >, std::allocator<std::__cxx11::basic_string<char,
> std::char_traits<char>, std::allocator<char> > > > const&, bool) const'
> ld:
> metatwist.o:(.data.rel.ro._ZTVN5boost15program_options20invalid_option_valueE[_ZTVN5boost15program_options20invalid_option_valueE]+0x30):
> undefined reference to
> `boost::program_options::error_with_option_name::substitute_placeholders(std::__cxx11::basic_string<char,
> std::char_traits<char>, std::allocator<char> > const&) const'
> ld:
> metatwist.o:(.data.rel.ro._ZTVN5boost16exception_detail19error_info_injectorINS_15program_options16validation_errorEEE[_ZTVN5boost16exception_detail19error_info_injectorINS_15program_options16validation_errorEEE]+0x30):
> undefined reference to
> `boost::program_options::error_with_option_name::substitute_placeholders(std::__cxx11::basic_string<char,
> std::char_traits<char>, std::allocator<char> > const&) const'
> ld:
> metatwist.o:(.data.rel.ro._ZTVN5boost16exception_detail10clone_implINS0_19error_info_injectorINS_15program_options16validation_errorEEEEE[_ZTVN5boost16exception_detail10clone_implINS0_19error_info_injectorINS_15program_options16validation_errorEEEEE]+0x38):
> undefined reference to
> `boost::program_options::error_with_option_name::substitute_placeholders(std::__cxx11::basic_string<char,
> std::char_traits<char>, std::allocator<char> > const&) const'
> ld:
> metatwist.o:(.data.rel.ro._ZTVN5boost16exception_detail19error_info_injectorINS_15program_options20invalid_option_valueEEE[_ZTVN5boost16exception_detail19error_info_injectorINS_15program_options20invalid_option_valueEEE]+0x30):
> undefined reference to
> `boost::program_options::error_with_option_name::substitute_placeholders(std::__cxx11::basic_string<char,
> std::char_traits<char>, std::allocator<char> > const&) const'
> ld:
> metatwist.o:(.data.rel.ro._ZTVN5boost16exception_detail10clone_implINS0_19error_info_injectorINS_15program_options20invalid_option_valueEEEEE[_ZTVN5boost16exception_detail10clone_implINS0_19error_info_injectorINS_15program_options20invalid_option_valueEEEEE]+0x38):
> undefined reference to
> `boost::program_options::error_with_option_name::substitute_placeholders(std::__cxx11::basic_string<char,
> std::char_traits<char>, std::allocator<char> > const&) const'
> Makefile:30: recipe for target
> '/u/vcojocar/apps/amber/20_tools-20_pnetcdf-1.12.1_impi-2019.7_intel-2019.1.1/bin/metatwist'
> failed
> make[2]: ***
> [/u/vcojocar/apps/amber/20_tools-20_pnetcdf-1.12.1_impi-2019.7_intel-2019.1.1/bin/metatwist]
> Error 1
> make[2]: Leaving directory
> '/cobra/u/vcojocar/apps/amber/20_tools-20_pnetcdf-1.12.1_impi-2019.7_intel-2019.1.1/AmberTools/src/moft'
> Makefile:36: recipe for target 'serial' failed
> make[1]: *** [serial] Error 2
> make[1]: Leaving directory
> '/cobra/u/vcojocar/apps/amber/20_tools-20_pnetcdf-1.12.1_impi-2019.7_intel-2019.1.1/AmberTools/src'
> Makefile:7: recipe for target 'install' failed
> make: *** [install] Error 2
>
>
>
> On 2020-10-20 04:10, Scott Brozell wrote:
> > Hi again,
> >
> > That bit about "on a supercomputer" didn't penetrate my noggin
> > initially. Go ahead call me Fickian (the 'F' can be pronounced
> > with a 'Th' sound as well :).
> >
> > Anyway, this now seems like a not-so-super-computer issue. Ie,
> > support there should verify their Intel and GNU compatibility.
> > Here's some context:
> >
> > Intel compilers use the GNU tools on the cluster: header files, libraries, and
> > linker. The details of Intel and GNU compatibility and interoperability can be
> > controlled in several ways. On Foo clusters the default mechanism of control is
> > based on modules. An alternative way to control compatibility and
> > interoperability is with Intel compiler options; see the 'GNU gcc
> > Interoperability' section of the Intel compiler man pages for details.
> >
> > To view the GNU compatibility, use the '-v' intel compiler option.
> >
> > scott
> >
> > On Mon, Oct 19, 2020 at 09:44:09PM -0400, Scott Brozell wrote:
> >> On Mon, Oct 19, 2020 at 07:47:43PM +0200, Vlad Cojocaru wrote:
> >>> I am trying to install Amber 20 on a supercomputer with the intel 19.1.1
> >>> compiler but the serial compilation fails at cpptraj with the error below.
> >>> Has anyone seen this before and has an idea how to fix it ?
> >>>
> >>> In file included from Traj_GmxXtc.cpp(1):
> >>> Traj_GmxXtc.h(40): error: identifier "off_t" is undefined
> >>> ?????????? std::vector<off_t> frameOffsets_; ///< Frame offsets for reading
> >>> ?????????????????????????????????? ^
> >>>
> >>> compilation aborted for Traj_GmxXtc.cpp (code 2)
> >>> ../config.h:45: recipe for target 'Traj_GmxXtc.o' failed
> >>> make[4]: *** [Traj_GmxXtc.o] Error 2
> >> I routinely build with intel; last was 19.0.5 on oct 15 on
> >> Red Hat Enterprise Linux Server release 7.7 (Maipo)
> >>
> >> I don't have 19.1.1 but i suspect your problem is GNU/OS related.
> >> off_t comes from sys/types.h via bits/types.h
> >> My system has several versions of these but they haven't changed significantly:
> >> 8 -rw-r--r-- 1 root root 6771 Apr 30 2019 /usr/include/sys/types.h
> >> 8 -rw-r--r-- 1 root root 6838 Jan 24 2014 /usr/lib/x86_64-redhat-linux6E/include/sys/types.h
> >> 8 -rw-r--r-- 1 root root 7729 Apr 30 2019 /usr/include/bits/types.h
> >> 8 -rw-r--r-- 1 root root 7601 Jan 24 2014 /usr/lib/x86_64-redhat-linux6E/include/bits/types.h
> >>
> >> My first guess is that you have intel compilers layered on top of
> >> really old gnu compilers.
> >> So check into updating your gnu with your OS's package management tool.
> >> If you send verbose and verbatim details then someone can probably quote
> >> exactly what to execute. Those details should include
> >> cat /etc/*release
> >> gcc -v
> >> icc -v
> >>
> >> For the compilers, eg, ...
> >> gcc version 8.4.0 (GCC)
> >> icc version 19.0.5.281 (gcc version 8.4.0 compatibility)
> >>
> >> scott

_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Oct 20 2020 - 15:00:02 PDT
Custom Search