From owner-nwchem-users@emsl.pnl.gov Wed Nov 21 13:57:20 2007 Received: from odyssey.emsl.pnl.gov (localhost [127.0.0.1]) by odyssey.emsl.pnl.gov (8.14.1/8.14.1) with ESMTP id lALLvIl7026797 for ; Wed, 21 Nov 2007 13:57:19 -0800 (PST) Received: (from majordom@localhost) by odyssey.emsl.pnl.gov (8.14.1/8.14.1/Submit) id lALLvIlt026796 for nwchem-users-outgoing-0915; Wed, 21 Nov 2007 13:57:18 -0800 (PST) X-Authentication-Warning: odyssey.emsl.pnl.gov: majordom set sender to owner-nwchem-users@emsl.pnl.gov using -f X-IronPort-AV: E=Sophos;i="4.21,448,1188802800"; d="scan'208";a="57274507" Message-ID: <4744A9BC.2060408@pnl.gov> Date: Wed, 21 Nov 2007 13:57:16 -0800 From: Dunyou Wang User-Agent: Thunderbird 2.0.0.9 (X11/20071031) MIME-Version: 1.0 To: John C Hackett CC: nwchem-users@emsl.pnl.gov Subject: Re: [NWCHEM] NWChem on Intel Mac References: <754A2033-8B39-4561-8B18-A622DE6AEF43@vcu.edu> <200711202307.12637.sassmannshausen@tugraz.at> <47435D82.4000002@pnl.gov> <47446633.3020101@pnl.gov> <6949DD62-9ACD-4659-9036-5046321D2A1F@vcu.edu> <47449FDF.3060903@pnl.gov> In-Reply-To: Content-Type: text/plain; charset=WINDOWS-1252 Content-Transfer-Encoding: 8bit X-OriginalArrivalTime: 21 Nov 2007 21:57:16.0302 (UTC) FILETIME=[7AD59AE0:01C82C89] Sender: owner-nwchem-users@emsl.pnl.gov Precedence: bulk Hi John, I know there should be not such a problem for Intel 9 on MACX. However, the MACX I can access doesn't have Intel compiler, I could not test it. Please go to your $NWCHEM_TOP/src/config/makefile.h, go to line 1237, uncomment LINK.f and add this flag to it. --Dunyou John C Hackett wrote: > Hi Dunyou: > > Thanks for the advice, I'll make the change and try again with > gfortran and let you know how it goes. > > With regard to the Intel compilers, there was discussion on the Intel > Web forum about the "multiple definitions of symbol" warning I > experienced that this is an issue with Tiger (and corrected in Leopard) > > http://softwarecommunity.intel.com/isn/Community/en-US/forums/post/30231808.aspx > > > They recommend adding the flags "” What do you think? > > Where would I place these flags prior to the build? Cheers, JCH > > On Nov 21, 2007, at 4:15 PM, Dunyou Wang wrote: > >> John, >> >> Please change the line to: >> >> read(atomi(1:index(atomi,':')-1),'(i5)') k >> >> will solve this problem there. >> >> --Dunyou >> >> >> >> John C Hackett wrote: >>> Hi Dunyou- >>> >>> I went back to sourceforge and verified I had the latest version of >>> gfortran, when I try to compile with the following environment >>> variables: >>> >>> setenv NWCHEM_TOP /Applications/nwchem-5.0/ >>> setenv NWCHEM_TARGET MACX >>> setenv NWCHEM_MODULES all >>> cd $NWCHEM_TOP/src >>> make nwchem_config >>> make FC=gfortran >>> >>> I get the following error: >>> >>> >>> gfortran -c -malign-double -O2 -ffast-math -Wuninitialized >>> -funroll-all-loops -mtune=native -I. >>> -I/Applications/nwchem-5.0//src/include >>> -I/Applications/nwchem-5.0//src/tools/include -DMACX -DCHKUNDFLW >>> -DGCC4 -DPARALLEL_DIAG pre_input.F >>> pre_input.F:977.72: >>> >>> read(atomi(1:index(atomi,':')-1),'(i)') k >>> >>> 1 >>> Error: Nonnegative width required in format string at (1) >>> make[2]: *** >>> [/Applications/nwchem-5.0//lib/MACX/libprepar.a(pre_input.o)] Error 1 >>> make[1]: *** [optimized] Error 2 >>> make: *** [libraries] Error 1 >>> >>> On Nov 21, 2007, at 12:09 PM, Dunyou Wang wrote: >>> >>>> Hi John, >>>> >>>> When you went back to version 9, have you make sure the version 9's >>>> libs >>>> is in your path, not 10? to link with? >>>> Another thing, you don't need to use -O0 for all the directories, only >>>> use it for the directory where you have that problem. >>>> >>>> Would you try this? Under your $NWCHEM_TOP/src >>>> make realclean >>>> make nwchem_config NWCHEM_MODULES=all >>>> make FC=ifort >>>> >>>> to see how it goes? >>>> >>>> If this still does not work, try to use gfortran then(download from >>>> the >>>> website I told you before). It works for me here. >>>> >>>> --Dunyou >>>> >>>> John C Hackett wrote: >>>>> Thank you for your suggestions. I checked Intel and 10.1.006 is the >>>>> latest version of the compiler for Mac OSX. I went back and tried the >>>>> latest version 9 compilers and libraries. This seems to have resolved >>>>> the problem with sym_sh_pair.F, but a new problem has arisen: >>>>> >>>>> ifort -align -mp1 -w -g -vec_report3 -O0 -g >>>>> -L/Applications/nwchem-5.0//lib/MACX >>>>> -L/Applications/nwchem-5.0//src/tools/lib/MACX -o >>>>> /Applications/nwchem-5.0//bin/MACX/nwchem nwchem.o stubs.o -lnwctask >>>>> -lccsd -lmcscf -lselci -lmp2 -lmoints -lstepper -ldriver -ldftgrad >>>>> -lnwdft -lgradients -lcphf -lesp -lddscf -lguess -lhessian -lvib >>>>> -lnwcutil -lrimp2 -lproperty -lnwints -lprepar -lnwmd -lnwpw -lpaw >>>>> -lpspw -lband -lnwpwlib -lcafe -lspace -lanalyze -lqhop -lpfft >>>>> -ldplot >>>>> -ldrdy -lvscf -lqmmm -lqmd -letrans -lpspw -ltddft -ltce -lbq -lcons >>>>> -lperfm -lneb -lnwcutil -lpario -lglobal -lma -lpeigs -lperfm -lcons >>>>> -lbq -lnwcutil -llapack -lblas -llapack -lblas -larmci -ltcgmsg >>>>> -lSystemStubs >>>>> ld: warning multiple definitions of symbol _floor >>>>> /opt/intel/fc/9.1.041/lib/libimf.a(floor_stub.o) definition of _floor >>>>> in section (__TEXT,__text) >>>>> /usr/lib/gcc/i686-apple-darwin8/4.0.1/../../..//libdl.dylib(s_floor.o) >>>>> >>>>> definition of _floor >>>>> ld: warning multiple definitions of symbol _atan >>>>> /opt/intel/fc/9.1.041/lib/libimf.a(atan_stub.o) definition of >>>>> _atan in >>>>> section (__TEXT,__text) >>>>> /usr/lib/gcc/i686-apple-darwin8/4.0.1/../../..//libdl.dylib(xmm_arcsincostan.o) >>>>> >>>>> >>>>> definition of _atan >>>>> ld: warning multiple definitions of symbol _acos >>>>> /opt/intel/fc/9.1.041/lib/libimf.a(acos_stub.o) definition of >>>>> _acos in >>>>> section (__TEXT,__text) >>>>> /usr/lib/gcc/i686-apple-darwin8/4.0.1/../../..//libdl.dylib(xmm_arcsincostan.o) >>>>> >>>>> >>>>> definition of _acos >>>>> ld: warning multiple definitions of symbol _sin >>>>> /opt/intel/fc/9.1.041/lib/libimf.a(sin_stub.o) definition of _sin in >>>>> section (__TEXT,__text) >>>>> /usr/lib/gcc/i686-apple-darwin8/4.0.1/../../..//libdl.dylib(xmm_sincostan.o) >>>>> >>>>> >>>>> definition of _sin >>>>> ld: warning multiple definitions of symbol _cos >>>>> /opt/intel/fc/9.1.041/lib/libimf.a(cos_stub.o) definition of _cos in >>>>> section (__TEXT,__text) >>>>> /usr/lib/gcc/i686-apple-darwin8/4.0.1/../../..//libdl.dylib(xmm_sincostan.o) >>>>> >>>>> >>>>> definition of _cos >>>>> ld: warning multiple definitions of symbol _exp >>>>> /opt/intel/fc/9.1.041/lib/libimf.a(exp_stub.o) definition of _exp in >>>>> section (__TEXT,__text) >>>>> /usr/lib/gcc/i686-apple-darwin8/4.0.1/../../..//libdl.dylib(xmm_exp.o) >>>>> >>>>> definition of _exp >>>>> ld: warning multiple definitions of symbol _log10 >>>>> /opt/intel/fc/9.1.041/lib/libimf.a(log10_stub.o) definition of _log10 >>>>> in section (__TEXT,__text) >>>>> /usr/lib/gcc/i686-apple-darwin8/4.0.1/../../..//libdl.dylib(xmm_log.o) >>>>> >>>>> definition of _log10 >>>>> ld: warning multiple definitions of symbol _log >>>>> /opt/intel/fc/9.1.041/lib/libimf.a(log_stub.o) definition of _log in >>>>> section (__TEXT,__text) >>>>> /usr/lib/gcc/i686-apple-darwin8/4.0.1/../../..//libdl.dylib(xmm_log.o) >>>>> >>>>> definition of _log >>>>> ld: warning multiple definitions of symbol _cbrt >>>>> /opt/intel/fc/9.1.041/lib/libimf.a(cbrt_stub.o) definition of >>>>> _cbrt in >>>>> section (__TEXT,__text) >>>>> /usr/lib/gcc/i686-apple-darwin8/4.0.1/../../..//libdl.dylib(xmm_power.o) >>>>> >>>>> >>>>> definition of _cbrt >>>>> ld: warning multiple definitions of symbol _pow >>>>> /opt/intel/fc/9.1.041/lib/libimf.a(pow_stub.o) definition of _pow in >>>>> section (__TEXT,__text) >>>>> /usr/lib/gcc/i686-apple-darwin8/4.0.1/../../..//libdl.dylib(xmm_power.o) >>>>> >>>>> >>>>> definition of _pow >>>>> ld: warning multiple definitions of symbol _powf >>>>> /opt/intel/fc/9.1.041/lib/libimf.a(powf_stub.o) definition of >>>>> _powf in >>>>> section (__TEXT,__text) >>>>> /usr/lib/gcc/i686-apple-darwin8/4.0.1/../../..//libdl.dylib(xmm_power.o) >>>>> >>>>> >>>>> definition of _powf >>>>> ld: warning multiple definitions of symbol _tanh >>>>> /opt/intel/fc/9.1.041/lib/libimf.a(tanh_stub.o) definition of >>>>> _tanh in >>>>> section (__TEXT,__text) >>>>> /usr/lib/gcc/i686-apple-darwin8/4.0.1/../../..//libdl.dylib(s_tanh.o) >>>>> definition of _tanh >>>>> ld: warning multiple definitions of symbol _rint >>>>> /opt/intel/fc/9.1.041/lib/libimf.a(rint.o) definition of _rint in >>>>> section (__TEXT,__text) >>>>> /usr/lib/gcc/i686-apple-darwin8/4.0.1/../../..//libdl.dylib(s_rint.o) >>>>> definition of _rint >>>>> ld: warning multiple definitions of symbol _exp2 >>>>> /opt/intel/fc/9.1.041/lib/libimf.a(exp2_stub.o) definition of >>>>> _exp2 in >>>>> section (__TEXT,__text) >>>>> /usr/lib/system/libmathCommon.A.dylib(single module) definition of >>>>> _exp2 >>>>> ld: Undefined symbols: >>>>> _f__xargc >>>>> _f__xargv >>>>> make: *** [all] Error 1 >>>>> >>>>> On Nov 20, 2007, at 5:19 PM, Dunyou Wang wrote: >>>>> >>>>>> Hi John, >>>>>> >>>>>> In addition to what Jörg said here, for you ifort compilation, >>>>>> Would >>>>>> you do the compilation using this option, under your directory >>>>>> $NWCHEM_TOP/src/symmetry, to see how it goes? >>>>>> make FC=ifort CC=icc OPTIMIZE="-O0" >>>>>> >>>>>> The binary for MACX on the download site was built using an older >>>>>> version of gfortran and gcc, so it's outdated. You can download >>>>>> the >>>>>> gfortran(for GCC4.3) from this website: http://hpc.sourceforge.net/ >>>>>> and >>>>>> compile the binary yourself, it works so far. >>>>>> >>>>>> Best regards >>>>>> Dunyou Wang >>>>>> >>>>>> >>>>>> Jörg Saßmannshausen wrote: >>>>>>> Dear John, >>>>>>> >>>>>>> as for the compilation, that is a problem with the Compiler, not >>>>>>> with the >>>>>>> code. Either update to the lates compiler (which is, I believe, >>>>>>> 10.1.008) or >>>>>>> try an earlier version. >>>>>>> >>>>>>> ICE are not that uncommon and reflect a compiler problem. >>>>>>> >>>>>>> I hope that helps >>>>>>> >>>>>>> Jörg >>>>>>> >>>>>>> Am Dienstag 20 November 2007 22:46 schrieben Sie: >>>>>>> >>>>>>>> NWChem Users: >>>>>>>> >>>>>>>> I am trying to compile nwchem5.0 on an Intel macintosh with >>>>>>>> version >>>>>>>> MKL 10.0.12, ifort 10.1.006, and icc 10.1.006. When I build >>>>>>>> with the >>>>>>>> following environment variables set: >>>>>>>> >>>>>>>> setenv NWCHEM_TOP /Applications/nwchem-5.0/ >>>>>>>> setenv NWCHEM_TARGET MACX >>>>>>>> setenv NWCHEM_MODULES all >>>>>>>> setenv LD_LIBRARY_PATH /Library/Frameworks/Intel_MKL.framework/ >>>>>>>> Versions/10.0.012/lib/32 >>>>>>>> cd $NWCHEM_TOP/src >>>>>>>> make nwchem_config >>>>>>>> make FC=ifort CC=icc >>>>>>>> >>>>>>>> ifort -c -align -mp1 -w -g -vec_report3 -O3 -prefetch >>>>>>>> -unroll - >>>>>>>> I. -I/Applications/nwchem-5.0//src/include -I/Applications/ >>>>>>>> nwchem-5.0//src/tools/include -DMACX -DIFCLINUX -DPARALLEL_DIAG >>>>>>>> sym_sh_pair.F >>>>>>>> sym_sh_pair.F(347): (col. 7) remark: loop was not vectorized: >>>>>>>> nonstandard loop is not a vectorization candidate. >>>>>>>> sym_sh_pair.F(310): (col. 27) remark: loop was not vectorized: >>>>>>>> nonstandard loop is not a vectorization candidate. >>>>>>>> fortcom: Severe: **Internal compiler error: specification >>>>>>>> exception >>>>>>>> signal raised** Please report this error along with the >>>>>>>> circumstances >>>>>>>> in which it occurred in a Software Problem Report. Note: File and >>>>>>>> line given may not be explicit cause of this error. >>>>>>>> >>>>>>>> compilation aborted for sym_sh_pair.F (code 3) >>>>>>>> make[2]: *** [/Applications/nwchem-5.0//lib/MACX/libnwcutil.a >>>>>>>> (sym_sh_pair.o)] Error 3 >>>>>>>> make[1]: *** [optimized] Error 2 >>>>>>>> make: *** [libraries] Error 1 >>>>>>>> >>>>>>>> Furthermore, I have downloaded the precompiled binary for Intel >>>>>>>> Mac, >>>>>>>> when I try to run it I get the following error: >>>>>>>> >>>>>>>> dyld: Library not loaded: /usr/local/lib/libgfortran.1.dylib >>>>>>>> Referenced from: /Applications/bin/./nwchem >>>>>>>> Reason: image not found >>>>>>>> Trace/BPT trap >>>>>>>> >>>>>>>> I have downloaded gfortran, but this library does not accompany >>>>>>>> it. >>>>>>>> Where do I get it? >>>>>>>> >>>>>>>> Thanks, JCH >>>>>>>> >>>>>>>> -------------------------------------------------- >>>>>>>> John C Hackett, Ph.D., R.Ph. >>>>>>>> Assistant Professor >>>>>>>> Virginia Biotech I, Suite 212 >>>>>>>> Department of Medicinal Chemistry & >>>>>>>> Institute for Structural Biology and Drug Discovery >>>>>>>> School of Pharmacy >>>>>>>> Virginia Commonwealth University Medical Center >>>>>>>> 800 E. Leigh Street >>>>>>>> P.O. Box 980133 >>>>>>>> Richmond, VA 23219-1540 >>>>>>>> Telephone 804.828.5679 >>>>>>>> Internet http://www.people.vcu.edu/~jchackett >>>>>>>> -------------------------------------------------- >>>>>>>> >>>>> >>> >