[Wien] Wien2k installation problem

Thu Jul 16 17:17:21 CEST 2015

Since you say that the "segmentation fault (core dumped)" error occurs 
during the running of ifort, you might also want to ask about the error 
in the Intel Fortran Compiler for Linux* and Mac OS X* forum [ 
https://software.intel.com/en-us/forum ].  One of the Intel 
representatives might know if it could caused by the DPD200364057 
Fortran Internal Compiler Error when compiling with -openmp or one of 
the other bugs listed or not listed in the 2015 Composer Edition 
Compilers Fixes List [ 
https://software.intel.com/en-us/articles/intel-composer-xe-2015-compilers-fixes-list 
].

On 7/16/2015 8:50 AM, Fecher, Gerhard wrote:
> Dear Shamik,
> -  what main processor does your computer have ? (It cannot be XEON PHI as this is a co-processor on some extra card)
>
> - if you compile for the XEON Phi, Wien2k  will most probably not work on the main processor of your computer
>    that is, if you compile it native for the PHI then you have to run it on the PHI
>    check the Intel manuals and the Intel link advisor for the PHI off-load models and how to use the mic options.
>
> - the XEON PHIs are implemented since the 2013 versions (e.g. parallel_studio_xe_2013) and those have also the emt64 versions
>    of ifort (in /opt/intel/composer_xe_2013.5.192/bin/intel64), icc, mkl (in /opt/intel/composer_xe_2013.5.192/mkl/lib/intel64) etc.
>
> - the composer_xe_2015_090 has both, intel_64 and intel64_mic for ifort and icc (or did you restrict the installation somehow ?)
>   (interesting I don't have a special mic directory for the mkl of xe_2015, but may be I just didn't install it
>    because the mic directory contains libraries only for native PHI programs)
>
> Ciao
> Gerhard
>
> DEEP THOUGHT in D. Adams; Hitchhikers Guide to the Galaxy:
> "I think the problem, to be quite honest with you,
> is that you have never actually known what the question is."
>
> ====================================
> Dr. Gerhard H. Fecher
> Institut of Inorganic and Analytical Chemistry
> Johannes Gutenberg - University
> 55099 Mainz
> and
> Max Planck Institute for Chemical Physics of Solids
> 01187 Dresden
> ________________________________________
> Von: wien-bounces at zeus.theochem.tuwien.ac.at [wien-bounces at zeus.theochem.tuwien.ac.at] im Auftrag von shamik chakrabarti [shamikphy at gmail.com]
> Gesendet: Donnerstag, 16. Juli 2015 14:42
> An: A Mailing list for WIEN2k users
> Betreff: Re: [Wien] Wien2k installation problem
>
> Dear Gerhard & Gavin,
>
>                 The compiler ifort is available only in mic directory for composer_xe_2015.0.090, and that's why we have also set the path of mkl to mic as otherwise we are getting an error of not finding mkl libs by X86_64-K1om-linux-ld:
>
> However, after setting the path of mkl to mic, there were no such error appeared except,
>
> segmentation fault (core dumped).
>
> This error is again appeared due to ifort (& not by gcc)
>
> The appearance of error is independent of whether we have set $(FOPT) or not to Linker Flags.
>
> All the -pthread are converted to -lpthread in the OPTION
>
> However, the error  segmentation fault (core dumped) is still appearing after compiling & this is the only one error we are left with.
>
> Do you think that if we use some older version of compiler in which "ifort" is available in intel64...then the problem may be solved?...
>
> However, I have no idea that whether the older version of compiler is compatible with "XEON PHI" processor?
>
> Looking forward to your suggestion in this regard.
>
> with regards,
>
>
> On Wed, Jul 15, 2015 at 8:49 PM, Gavin Abo <gsabo at crimson.ua.edu<mailto:gsabo at crimson.ua.edu>> wrote:
> Dear Gerhard and Shamik,
>
> Ok, sorry, I did not think too much into it.
>
> In siteconfig,
>
> 1) selected "S   specify a system" and "I    Linux (Intel ifort compiler (12.0 and later) + mkl )"
> 2) selected "O   specify compiler options, BLAS and LAPACK", and it gave:
>
> Recommended options for system linuxifc are:
>       Compiler options:        -FR -mp1 -w -prec_div -pc80 -pad -ip -DINTEL_VML -traceback -assume buffered_io
>       Linker Flags:            $(FOPT) -L$(MKLROOT)/lib/$(MKL_TARGET_ARCH) -pthread
>       Preprocessor flags:      '-DParallel'
>       R_LIB (LAPACK+BLAS):     -lmkl_lapack95_lp64 -lmkl_intel_lp64 -lmkl_intel_thread -lmkl_core -openmp -lpthread
>
> I knew that these recommended 64 bit settings have worked fine for me on different Linux operating systems.  Until you mentioned it, I did not know that -pthread could be safely dropped from the Linker Flags.
>
> Thanks,
>
> Gavin
>
>
> why -pthread instead of -lpthread ? The threading library is libpthread and should be linked with -lpthread !
> I remembre that I made some ten years ago the remark on that: On SUSE systems (at thae time it was 9.2 and older) one can have a very comfortable setup of the linker
> such that it finds the standart libraries by itself. You can check it if you remove the -pthread switch from the linker line and find that Wien2k is still linked correctly.
> If not then you may use ldconfig to make your linker environment more comfortable.
>
>
> Below I do not understand why the linker should need the compiler options (LDFLAGS:$(FOPT)) ??
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at<mailto:Wien at zeus.theochem.tuwien.ac.at>
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> SEARCH the MAILING-LIST at:  http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
>
>
>
>
> --
> Shamik Chakrabarti
> Senior Research Fellow
> Dept. of Physics & Meteorology
> Material Processing & Solid State Ionics Lab
> IIT Kharagpur
> Kharagpur 721302
> INDIA