From lmliu at imr.ac.cn Wed Jan 7 19:17:37 2004 From: lmliu at imr.ac.cn (lmliu) Date: Wed, 07 Jan 2004 19:17:37 Subject: [CPMD-list] Help on Compliling the PC-PGI-MPI Message-ID: <200401071116.MAA09162@internet-fence.zurich.ihost.com> Dear All: I can compile the PC-PGI CPMD 3.7.2 using PGI 3.2.4. But when I try to compile the parallel PC-PGI-MPI using the default "FC=mpif77 -c -fast" and "LD= mpif77 -fast", it has following error: mpif77 -c -fast -Mr8 -pc64 -byteswapio -Msignextend -Msecond -Underscore ./cpmd.f -o ./cpmd.o Error: bad option: -fast Make: ***[cpmd.o] Error 1. I have installed the LAM-MPI and PGI, and I think the reason may be that mpif77 and pgf77 should be related. If so, How shoud I set this? If not, How can I solve this error? Thanks a lot in advance! L.M. Liu From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Wed Jan 7 12:37:30 2004 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Wed, 7 Jan 2004 12:37:30 +0100 (CET) Subject: [CPMD-list] Help on Compliling the PC-PGI-MPI In-Reply-To: <200401071116.MAA09162@internet-fence.zurich.ihost.com> Message-ID: On Wed, 7 Jan 2004, lmliu wrote: LL> Dear All: LL> LL> I can compile the PC-PGI CPMD 3.7.2 using PGI 3.2.4. LL> But when I try to compile the parallel PC-PGI-MPI LL> using the default "FC=mpif77 -c -fast" and "LD= mpif77 -fast", LL> it has following error: LL> LL> mpif77 -c -fast -Mr8 -pc64 -byteswapio -Msignextend -Msecond -Underscore ./cpmd.f -o ./cpmd.o ^^^^^^^^^ if you don't run cpmd also on a big endian machine, and want to exchange restart files you don't (want and) need this. ^^^^^^^^^^^^^^^^^^ this should read -Msecond_underscore or has to be dropped, if you have compiled your lam-mpi for pgf77/pgf90 LL> LL> Error: bad option: -fast most likely, your lam-mpi has been compiled for use with g77, which does not recognize the -fast flag. you could try working around this by using env LAMHF77=pgf90 mpif77 instead of plain mpif77, you could recompile lam-mpi for pgi or use a set of wrapper scripts that enable compatibility between the various linux compilers for lam. more detailed explanations of this and other linux related cpmd problems and a set of cpmd 'compatible' lam-mpi rpms are available at: http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/cpmd-linux.html good luck, axel kohlmeyer. LL> Make: ***[cpmd.o] Error 1. LL> LL> I have installed the LAM-MPI and PGI, and I think the reason may be LL> that mpif77 and pgf77 should be related. If so, How shoud I set this? LL> If not, How can I solve this error? LL> Thanks a lot in advance! LL> LL> LL> L.M. Liu LL> LL> LL> LL> -- ======================================================================= Dr. Axel Kohlmeyer e-mail: axel.kohlmeyer at rub.de Lehrstuhl fuer Theoretische Chemie Phone: ++49 (0)234/32-26673 Ruhr-Universitaet Bochum - NC 03/53 Fax: ++49 (0)234/32-14045 D-44780 Bochum http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/ ======================================================================= From s.meloni at caspur.it Mon Jan 12 16:40:14 2004 From: s.meloni at caspur.it (Simone Meloni) Date: Mon, 12 Jan 2004 16:40:14 +0100 (CET) Subject: [CPMD-list] 64 on pwr4 Message-ID: Dear CPMDer, the Makefile produced by Configure for the POWER4 64bit case has a little bug, the CFLAGS variable does not contain -q64. When xlf_r links the object files founds incompatibility between sysdepend.o, shmem.o (32bits mode) and the rest of the object files (64 bits). Ciao Simone -- +------------------------------------------------------------------------+ PhD Simone Meloni | Achab: Lascio una scia Bianca e inquieta, CASPUR | acque pallide, facce piu' pallide Via Dei Tizii 6b | dovunque passo. Le onde invidiose 00185 Rome (Italy) | si gonfiano ai lati per sommergere Phone: +39-06-44486-711 | la mia traccia: facciano, ma prima Fax: +39-06-4957083 | io passo. From laura.orian at unipd.it Wed Jan 7 18:23:03 2004 From: laura.orian at unipd.it (Laura Orian) Date: Wed, 07 Jan 2004 18:23:03 +0100 Subject: [CPMD-list] compilation troubles Message-ID: <3FFC4077.4040908@unipd.it> Hello I am trying to compile the CPMD-3.7.2 code in my RH 9.0 Linux machine with Intel Fortran Compliler and Intel Math libraries, but at the end I get this message: dp.o mm_detsp.o mm_init.o mm_dim.o mm_qmmm_forcedr.o mm_mdmain.o mm_coninp.o mm_cpmd_add_MM_forces_f77.o mm_cpmd_esp_charges_f77.o mm_cpmd_ext_pot_f77.o vdwin.o vdw.o sysdepend.o shmemory.o -L/opt/intel/mkl61/lib/32/ -lmkl_lapack -lmkl_p4 -lguide -lpthread -Vaxlib ifc: warning: The Intel Fortran driver is now named ifort. You can suppress this message with '-quiet' envir.o(.text+0x3f): In function `envir_': : undefined reference to `hostnm__' envir.o(.text+0x4e): In function `envir_': : undefined reference to `getcwd__' envir.o(.text+0x56): In function `envir_': : undefined reference to `getpid__' envir.o(.text+0x60): In function `envir_': : undefined reference to `getuid__' make: *** [cpmd.x] Error 1 What's wrong? Thank you Laura The Makefile is SHELL = /bin/sh # #--------------- Default Configuration for PC-IFC --------------- SRC = . DEST = . BIN = . #QMMM_FLAGS = -D__QMECHCOUPL #QMMM_LIBS = -L. -lmm FFLAGS = -c -w -r8 -w90 -w95 -O3 -pc64 -axM -ip -tpp7 LFLAGS = -L/opt/intel/mkl61/lib/32/ -lmkl_lapack -lmkl_p4 \ -lguide -lpthread -Vaxlib CFLAGS = -c -O2 -Wall CPP = /lib/cpp -P -C -traditional CPPFLAGS = -D__Linux -D__PGI -DLAPACK -DFFT_DEFAULT -DLINUX_IFC NOOPT_FLAG = CC = gcc FC = /opt/intel_fc_80/bin/ifc -i_dynamic LD = /opt/intel_fc_80/bin/ifc -i_dynamic AR = -------------- next part -------------- An HTML attachment was scrubbed... URL: http://cpmd.org/pipermail/cpmd-list/attachments/20040107/7caf4052/attachment.html From hawlitzk at uni-mainz.de Thu Jan 8 12:48:40 2004 From: hawlitzk at uni-mainz.de (Michael Hawlitzky) Date: Thu, 08 Jan 2004 12:48:40 +0100 Subject: [CPMD-list] Problems defining ending times etc. Message-ID: <3FFD4398.1080404@uni-mainz.de> Dear CPMD-users, When using CPMD on my Input file which I have quoted after this text, I encounter the problem that the program seems to ignore the definition of the frequency of writing out data, and esp. the ending time of the simulation. I thought that the options MAXSTEP and STORE define the ending time and the write out frequency of the restart files, resp. But in fact after 500 steps the simulation is still running and no other files have been created than GEOMETRY, GSHELL and GEOMETRY.xyz (which are created at the beginning of the run). Could anybody help me with this problem? Thanks in advance, Michael Hawlitzky # RESTART LATEST WAVEFUNCTIONS COORDINATES ACCUMULATORS VELOCITIES # # KLEINMAN_BYLANDER # &CPMD MOLECULAR DYNAMICS INITIALIZE WAVEFUNCTIONS ATOMS CENTER MOLECULE OFF KOHN-SHAM ENERGIES OFF QUENCH BO TIMESTEP 3 TEMPERATURE 3100 TEMPCONTROL IONS 3100 80 TEMPCONTROL ELECTRONS 0.01 1.0 MAXSTEP 200 EMASS 1100.0 DUAL 4 STORE 50 RESTFILE 2 TRAJECTORY SAMPLE 1 MOVIE 10 &END &SYSTEM ANGSTROM # Units for atomic positions and Cell size SYMMETRY # cubic periodic boundary conditions 8 CELL # a a/b a/c cosa cosb cosc 10.389 1 1 0 0 0 CUTOFF 75.0 &END &ATOMS *Ge-MT-BLYP-NLCC.psp KLEINMAN_BYLANDER LMAX=P 20 10.2309619935557 2.24013752991853 2.74117454605157 8.98411281032944 3.76746539137866 0.0249278516517743 4.69132391436502 3.92551006385695 3.40769310713321 0.98600004312607 6.22586636012791 0.122561630985260 5.71836599354328 4.50171068132673 0.181880051847781 2.76684245171828 3.55440545586684 9.55707549059228 6.79416881151116 0.91150772099614 6.39430625276517 1.79565526437242 8.90963975522430 9.39151776342931 5.71710643176909 9.39468787457578 1.23240428408590 10.0274203091042 10.3879200521586 6.97316675031386 3.48107085409666 7.13778141736257 2.68083321334490 7.65099845217365 0.077321781464234 3.45388437325599 3.00567181264869 1.52232920810289 2.12863821189479 5.11569851955175 9.04661098030339 8.42587528129445 0.49325080634279 3.00005129040692 7.71357940059035 9.91463678345913 5.13925680870307 5.22482554035052 5.26531025611805 6.18828116739993 8.02123356861802 9.91760920789132 8.09812908044320 4.31036961267772 6.93682006244632 4.03641736939281 5.99263530404511 1.71103930737935 9.82974494204533 2.44393870390330 *O_MT_GIA_BLYP KLEINMAN_BYLANDER LMAX=P 40 2.38290699175273 2.71359719243478 0.801070560142205 3.74381320745770 5.58992428590523 3.39210378361544 10.3483125563688 1.62700771830701 6.24484703160610 6.69346235410913 0.87308114561032 4.67278483190428 8.16666271591485 8.78654504251254 4.07887571335522 6.08112280511783 7.73739868170545 7.68286207590923 6.07510866946267 4.84865358314640 7.30471213055613 3.30172580956213 8.68985131045680 8.65657895545190 1.99083984328141 2.65255292852893 8.28780843260222 9.77259274007937 2.52969122079708 1.03758551663091 10.2114970907838 3.60352185423919 3.82193545883199 6.96451632418329 10.2640782491384 1.87947048684094 7.30425024578138 3.91087170466234 10.1911562696341 0.58476936478056 7.60800912910842 9.57089316143564 10.1991134169248 8.62239901383592 6.16945898511749 5.93852978852207 9.25498793078470 9.87801397391021 5.64747830178208 3.99791731508147 4.73199900114780 10.2109010649065 6.43174455197998 4.10551461533019 1.95723403635711 9.63819089838433 0.720219330207004 9.07276386654675 1.05514421628378 3.19719009399737 0.38989158931704 4.33786623158558 6.50734333025446 3.66188936275834 2.42282496038092 3.48593546798630 0.61389299779294 10.0241286443495 8.53645704144670 1.43773944467393 1.12004374035687 2.94252735212135 5.16029719850783 5.82603611635875 9.69293016745478 8.20063287375033 5.04735156457443 5.42179560341680 0.34614487107698 8.81161907952966 2.89762535173258 3.63540010679722 6.20402884189327 7.19465443941048 4.23107046504472 0.45492905634027 1.54601975479307 5.69050130468192 0.0890332408345004 7.43148380164392 9.75799325883711 5.21169755103834 0.552751843499343 9.53590082875610 3.51733372539703 8.69321991867171 2.49023340261693 6.63781140121635 1.20543815399678 5.63462771162479 3.80968726495336 1.94262498698917 2.88047865604240 8.80487375340869 3.29670567265320 4.53909720174818 3.21454021444742 9.74967405125233 7.46202140614635 2.43008793729656 6.80173065747694 1.82486946296683 4.90095498725833 9.72904444576411 8.29432261565364 10.3860256764164 7.39530476836511 4.85246592634677 8.07326710091869 1.97974335632211 ISOTOPE 72.61 15.9994 &END &DFT FUNCTIONAL BLYP &END -- Dipl.-Math. Michael Hawlitzky Institut f?r Physik, Universit?t Mainz Staudingerweg 9, Raum 01-517 55099 Mainz Tel.: (+49)6131-3925158 Fax: (+49)6131-3925441 From christopher.baker at balliol.oxford.ac.uk Sun Jan 11 00:15:22 2004 From: christopher.baker at balliol.oxford.ac.uk (Christopher Baker) Date: Sat, 10 Jan 2004 23:15:22 +0000 (GMT) Subject: [CPMD-list] CPMD compilation Message-ID: An embedded and charset-unspecified text was scrubbed... Name: not available Url: http://cpmd.org/pipermail/cpmd-list/attachments/20040110/7d3a3d72/attachment.cc From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Mon Jan 19 20:31:37 2004 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Mon, 19 Jan 2004 20:31:37 +0100 (CET) Subject: [CPMD-list] compilation troubles In-Reply-To: <3FFC4077.4040908@unipd.it> Message-ID: On Wed, 7 Jan 2004, Laura Orian wrote: LO> Hello hi laura, LO> I am trying to compile the CPMD-3.7.2 code in my RH 9.0 Linux machine LO> with Intel Fortran Compliler and Intel Math libraries, but at the end I LO> get this message: LO> LO> dp.o mm_detsp.o mm_init.o mm_dim.o mm_qmmm_forcedr.o mm_mdmain.o LO> mm_coninp.o mm_cpmd_add_MM_forces_f77.o mm_cpmd_esp_charges_f77.o LO> mm_cpmd_ext_pot_f77.o vdwin.o vdw.o sysdepend.o shmemory.o LO> -L/opt/intel/mkl61/lib/32/ -lmkl_lapack -lmkl_p4 -lguide -lpthread -Vaxlib LO> ifc: warning: The Intel Fortran driver is now named ifort. You can LO> suppress this message with '-quiet' LO> envir.o(.text+0x3f): In function `envir_': LO> : undefined reference to `hostnm__' LO> envir.o(.text+0x4e): In function `envir_': LO> : undefined reference to `getcwd__' LO> envir.o(.text+0x56): In function `envir_': LO> : undefined reference to `getpid__' LO> envir.o(.text+0x60): In function `envir_': LO> : undefined reference to `getuid__' LO> make: *** [cpmd.x] Error 1 LO> LO> What's wrong? it seems, as if your .F files have been preprocessed without the -D__Linux define. try 'make clean; make' to recompile everything. if that does not help, please send the _first_ 10-20 lines of the output produced by make. LO> LO> Thank you LO> Laura LO> LO> The Makefile is LO> SHELL = /bin/sh LO> # LO> #--------------- Default Configuration for PC-IFC --------------- LO> SRC = . LO> DEST = . LO> BIN = . LO> #QMMM_FLAGS = -D__QMECHCOUPL LO> #QMMM_LIBS = -L. -lmm LO> FFLAGS = -c -w -r8 -w90 -w95 -O3 -pc64 -axM -ip -tpp7 LO> LFLAGS = -L/opt/intel/mkl61/lib/32/ -lmkl_lapack -lmkl_p4 \ LO> -lguide -lpthread -Vaxlib LO> CFLAGS = -c -O2 -Wall LO> CPP = /lib/cpp -P -C -traditional LO> CPPFLAGS = -D__Linux -D__PGI -DLAPACK -DFFT_DEFAULT -DLINUX_IFC LO> NOOPT_FLAG = LO> CC = gcc LO> FC = /opt/intel_fc_80/bin/ifc -i_dynamic LO> LD = /opt/intel_fc_80/bin/ifc -i_dynamic ^^^^^^^^^^ the intel compiler version 8 supposedly supports redhat 9 (i could not test it yet), so that you should be able to link statically, which is highly recommended. with -i_dynamics you _have_ to install mkl and and ifort everywhere. hope this helps, axel. LO> AR = LO> LO> LO> LO> -- ======================================================================= Dr. Axel Kohlmeyer e-mail: axel.kohlmeyer at rub.de Lehrstuhl fuer Theoretische Chemie Phone: ++49 (0)234/32-26673 Ruhr-Universitaet Bochum - NC 03/53 Fax: ++49 (0)234/32-14045 D-44780 Bochum http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/ ======================================================================= From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Mon Jan 19 20:42:18 2004 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Mon, 19 Jan 2004 20:42:18 +0100 (CET) Subject: [CPMD-list] Problems defining ending times etc. In-Reply-To: <3FFD4398.1080404@uni-mainz.de> Message-ID: On Thu, 8 Jan 2004, Michael Hawlitzky wrote: MH> Dear CPMD-users, MH> MH> When using CPMD on my Input file which I have quoted after this text, I MH> encounter the problem that the program seems to ignore the definition of MH> the frequency of writing out data, and esp. the ending time of the MH> simulation. MH> MH> I thought that the options MAXSTEP and STORE define the ending time and MH> the write out frequency of the restart files, resp. But in fact after MH> 500 steps the simulation is still running and no other files have been MH> created than GEOMETRY, GSHELL and GEOMETRY.xyz (which are created at the MH> beginning of the run). MH> MH> Could anybody help me with this problem? Thanks in advance, MH> Michael Hawlitzky hello michael, most likely your job is still optimizing the wavefunction to the default convergence. since you propably want to equilibrate your system first anyway, you should be able to start with a less tightly converged wavefunction (cf. keyword CONVERGENCE). MAXSTEP applies only to the MD part of the run. if you want to limit the number of steps for the wavefunction optimization (not recommended), you need to run the wavefunction optimization separately and the restart from that wavefunction. hope this helps, axel. MH> MH> MH> # RESTART LATEST WAVEFUNCTIONS COORDINATES ACCUMULATORS VELOCITIES MH> # MH> # KLEINMAN_BYLANDER MH> # MH> &CPMD MH> MOLECULAR DYNAMICS MH> INITIALIZE WAVEFUNCTIONS ATOMS MH> CENTER MOLECULE OFF MH> KOHN-SHAM ENERGIES OFF MH> QUENCH BO MH> TIMESTEP MH> 3 MH> TEMPERATURE MH> 3100 MH> TEMPCONTROL IONS MH> 3100 80 MH> TEMPCONTROL ELECTRONS MH> 0.01 1.0 MH> MAXSTEP MH> 200 MH> EMASS MH> 1100.0 MH> DUAL MH> 4 MH> STORE MH> 50 MH> RESTFILE MH> 2 MH> TRAJECTORY SAMPLE MH> 1 MH> MOVIE MH> 10 MH> &END MH> MH> &SYSTEM MH> ANGSTROM # Units for atomic positions and Cell size MH> SYMMETRY # cubic periodic boundary conditions MH> 8 MH> CELL # a a/b a/c cosa cosb cosc MH> 10.389 1 1 0 0 0 MH> CUTOFF MH> 75.0 MH> &END MH> MH> &ATOMS MH> *Ge-MT-BLYP-NLCC.psp KLEINMAN_BYLANDER MH> LMAX=P MH> 20 MH> 10.2309619935557 2.24013752991853 2.74117454605157 MH> 8.98411281032944 3.76746539137866 0.0249278516517743 MH> 4.69132391436502 3.92551006385695 3.40769310713321 MH> 0.98600004312607 6.22586636012791 0.122561630985260 MH> 5.71836599354328 4.50171068132673 0.181880051847781 MH> 2.76684245171828 3.55440545586684 9.55707549059228 MH> 6.79416881151116 0.91150772099614 6.39430625276517 MH> 1.79565526437242 8.90963975522430 9.39151776342931 MH> 5.71710643176909 9.39468787457578 1.23240428408590 MH> 10.0274203091042 10.3879200521586 6.97316675031386 MH> 3.48107085409666 7.13778141736257 2.68083321334490 MH> 7.65099845217365 0.077321781464234 3.45388437325599 MH> 3.00567181264869 1.52232920810289 2.12863821189479 MH> 5.11569851955175 9.04661098030339 8.42587528129445 MH> 0.49325080634279 3.00005129040692 7.71357940059035 MH> 9.91463678345913 5.13925680870307 5.22482554035052 MH> 5.26531025611805 6.18828116739993 8.02123356861802 MH> 9.91760920789132 8.09812908044320 4.31036961267772 MH> 6.93682006244632 4.03641736939281 5.99263530404511 MH> 1.71103930737935 9.82974494204533 2.44393870390330 MH> MH> MH> *O_MT_GIA_BLYP KLEINMAN_BYLANDER MH> LMAX=P MH> 40 MH> 2.38290699175273 2.71359719243478 0.801070560142205 MH> 3.74381320745770 5.58992428590523 3.39210378361544 MH> 10.3483125563688 1.62700771830701 6.24484703160610 MH> 6.69346235410913 0.87308114561032 4.67278483190428 MH> 8.16666271591485 8.78654504251254 4.07887571335522 MH> 6.08112280511783 7.73739868170545 7.68286207590923 MH> 6.07510866946267 4.84865358314640 7.30471213055613 MH> 3.30172580956213 8.68985131045680 8.65657895545190 MH> 1.99083984328141 2.65255292852893 8.28780843260222 MH> 9.77259274007937 2.52969122079708 1.03758551663091 MH> 10.2114970907838 3.60352185423919 3.82193545883199 MH> 6.96451632418329 10.2640782491384 1.87947048684094 MH> 7.30425024578138 3.91087170466234 10.1911562696341 MH> 0.58476936478056 7.60800912910842 9.57089316143564 MH> 10.1991134169248 8.62239901383592 6.16945898511749 MH> 5.93852978852207 9.25498793078470 9.87801397391021 MH> 5.64747830178208 3.99791731508147 4.73199900114780 MH> 10.2109010649065 6.43174455197998 4.10551461533019 MH> 1.95723403635711 9.63819089838433 0.720219330207004 MH> 9.07276386654675 1.05514421628378 3.19719009399737 MH> 0.38989158931704 4.33786623158558 6.50734333025446 MH> 3.66188936275834 2.42282496038092 3.48593546798630 MH> 0.61389299779294 10.0241286443495 8.53645704144670 MH> 1.43773944467393 1.12004374035687 2.94252735212135 MH> 5.16029719850783 5.82603611635875 9.69293016745478 MH> 8.20063287375033 5.04735156457443 5.42179560341680 MH> 0.34614487107698 8.81161907952966 2.89762535173258 MH> 3.63540010679722 6.20402884189327 7.19465443941048 MH> 4.23107046504472 0.45492905634027 1.54601975479307 MH> 5.69050130468192 0.0890332408345004 7.43148380164392 MH> 9.75799325883711 5.21169755103834 0.552751843499343 MH> 9.53590082875610 3.51733372539703 8.69321991867171 MH> 2.49023340261693 6.63781140121635 1.20543815399678 MH> 5.63462771162479 3.80968726495336 1.94262498698917 MH> 2.88047865604240 8.80487375340869 3.29670567265320 MH> 4.53909720174818 3.21454021444742 9.74967405125233 MH> 7.46202140614635 2.43008793729656 6.80173065747694 MH> 1.82486946296683 4.90095498725833 9.72904444576411 MH> 8.29432261565364 10.3860256764164 7.39530476836511 MH> 4.85246592634677 8.07326710091869 1.97974335632211 MH> MH> ISOTOPE MH> 72.61 MH> 15.9994 MH> &END MH> MH> &DFT MH> FUNCTIONAL BLYP MH> &END MH> MH> MH> -- ======================================================================= Dr. Axel Kohlmeyer e-mail: axel.kohlmeyer at rub.de Lehrstuhl fuer Theoretische Chemie Phone: ++49 (0)234/32-26673 Ruhr-Universitaet Bochum - NC 03/53 Fax: ++49 (0)234/32-14045 D-44780 Bochum http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/ ======================================================================= From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Mon Jan 19 20:49:46 2004 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Mon, 19 Jan 2004 20:49:46 +0100 (CET) Subject: [CPMD-list] CPMD compilation In-Reply-To: Message-ID: On Sat, 10 Jan 2004, Christopher Baker wrote: CB> Hi, CB> CB> I am having a problem installing CPMD 3.7.2, and I was wondering if anyone might be able to help? CB> CB> When trying to compile CPMD with IFC using the makefile configured for PC-IFC I received a lot of errors of the type "undefined reference to dcopy_" etc. From reading other posts I gathered that this was because Atlas also needed to be compiled using IFC. I then compiled Atlas with the following chages to the makefile: well, you don't need atlas as such, you need lapack and blas, but atlas is a very fast implementation of blas and some parts of lapack. if you compile altas from scratch, you'll need to recompile the missing parts from the lapack sources. look at the atlas documentation on how to do this. alternatively you can download a matching 'one-size-fits-all' library binaries from http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/cpmd-linux.html CB> CB> F77 = ifc CB> F77FLAGS = -FI -w90 -w95 -cm -O3 -unroll CB> FLINKER = $(F77) CB> FLINKFLAGS = $(F77FLAGS) ^^^^^^ that does not help. you need to modify the LFLAGS makefile variable. e.g. this is what i am currently using for an amd athlon cpu and the intel fortran compiler v8.0: FFLAGS = -c -r8 -O3 -pc64 -axiMK -tpp6 -unroll -cm -tune pn3 -w95 LFLAGS = -L. -latlas_athlon -Vaxlib -static $(QMMM_LIBS) CFLAGS = -c -O2 -Wall CPP = /lib/cpp -P -C -traditional CPPFLAGS = -D__Linux -D__PGI -DLAPACK -DFFT_DEFAULT -DLINUX_IFC NOOPT_FLAG = CC = cc FC = ifort LD = ifort hope this helps, axel. CB> CB> However, when I again tried to compile CPMD, I still received the sam eerrors of the type "undefined reference to dcopy_" etc. If there is anyone who could tell me where I am going wrong I would be extremely grateful. CB> CB> Thanks. CB> CB> Chris Baker CB> CB> _______________________________________________ CB> CPMD-list mailing list CB> CPMD-list at cpmd.org CB> http://www.cpmd.org/mailman/listinfo/cpmd-list CB> CB> -- ======================================================================= Dr. Axel Kohlmeyer e-mail: axel.kohlmeyer at rub.de Lehrstuhl fuer Theoretische Chemie Phone: ++49 (0)234/32-26673 Ruhr-Universitaet Bochum - NC 03/53 Fax: ++49 (0)234/32-14045 D-44780 Bochum http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/ ======================================================================= From hawlitzk at uni-mainz.de Mon Jan 19 20:57:21 2004 From: hawlitzk at uni-mainz.de (Michael Hawlitzky) Date: Mon, 19 Jan 2004 20:57:21 +0100 Subject: [CPMD-list] Problems defining ending times etc. Message-ID: <400C36A1.5040400@uni-mainz.de> Dear Mr. Kohlmeyer, Thanks a lot for your hint! You are right, the wavefunctions aren't converged correctly. I was also wondering why the iteration takes so many steps (more than 600 without gemax getting significantly below 1e-04), but most probably this comes from the position data which I have taken from a classical MD run. Greetings, Michael -- Dipl.-Math. Michael Hawlitzky Institut f?r Physik, Universit?t Mainz Staudingerweg 9, Raum 01-517 55099 Mainz Tel.: (+49)6131-3925158 Fax: (+49)6131-3925441 From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Tue Jan 20 10:13:09 2004 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Tue, 20 Jan 2004 10:13:09 +0100 Subject: [CPMD-list] Problems defining ending times etc. In-Reply-To: Your message of "Mon, 19 Jan 2004 20:57:21 +0100." <400C36A1.5040400@uni-mainz.de> Message-ID: <200401200913.i0K9D9a30806@yello.theochem.ruhr-uni-bochum.de> >>> "MH" == Michael Hawlitzky writes: MH> Dear Mr. Kohlmeyer, MH> Thanks a lot for your hint! MH> You are right, the wavefunctions aren't converged correctly. MH> I was also wondering why the iteration takes so many steps (more than 600 MH> without gemax getting significantly below 1e-04), but most probably MH> this comes from the position data which I have taken from a classical MH> MD run. that is exactly my impression, too. the wavefunction seems to converge properly only when your ionic positions are close to the equilibrium. although i have found that in some cases, e.g. bulk water, the classical md optimized positions are very close to the QM optimum and the wavefunction converges very good. you should be able to improve the situation by running a few short mds of a few hundred steps and then reoptimize the wavefunction with a tighter convergence until you have a 'stable' md. alternatively, you could run a geometry optimization first (i found using LBFGS with adaptive convergence to be very efficient, there are some examples for useful parameter combinations in the mailing list archives). good luck, axel kohlmeyer. MH> Greetings, MH> Michael MH> -- MH> Dipl.-Math. Michael Hawlitzky MH> Institut f?r Physik, Universit?t Mainz MH> Staudingerweg 9, Raum 01-517 MH> 55099 Mainz MH> Tel.: (+49)6131-3925158 MH> Fax: (+49)6131-3925441 MH> _______________________________________________ MH> CPMD-list mailing list MH> CPMD-list at cpmd.org MH> http://www.cpmd.org/mailman/listinfo/cpmd-list -- ======================================================================= Axel Kohlmeyer e-mail: axel.kohlmeyer at theochem.ruhr-uni-bochum.de Lehrstuhl fuer Theoretische Chemie Phone: ++49 (0)234/32-26673 Ruhr-Universitaet Bochum - NC 03/53 Fax: ++49 (0)234/32-14045 D-44780 Bochum http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/ ======================================================================= If you make something idiot-proof, the universe creates a better idiot. From per.andersson at foi.se Thu Jan 29 13:38:02 2004 From: per.andersson at foi.se (Per Andersson) Date: Thu, 29 Jan 2004 14:38:02 +0200 Subject: [CPMD-list] Restart files of PIMD Message-ID: <4018FEAA.5000503@foi.se> Dear Everyone I have with some success run some MD CP calculations of 54 atoms of H but I would like to try PIMD on the same system. After reading the manual and the references therein I still have severe problems. Unfortunately the cluster where I run my calculations has no network connection so I have no input or output listings to show. If I understand this correctly you have to converge the wavefunction in a static run and restart the MD using RESTART WAVEFUNCTIONS COORDINATES VELOCITIES (or at least wavefunctions). This works fine for MD CP. When I try PIMD I add INITIALIZATION in the &PIMD-section, PATH INTEGRALS in the &CPMD-section and restart with the same options as before. I have tried to restart from both the static calculation and the MD CP calculation with the same result. I get the error message RESTART INFORMATION FOUND ON FILE ./RESTART_1 ZHRWF| RESTART FILE NOT FOUND: ./RESTART_2 As far as I understand there will be one restart file for each replica after a PIMD calculation. If they are needed before as well, how do I generate them? Yours Per Andersson From cklein at pharma.ethz.ch Wed Jan 21 07:44:42 2004 From: cklein at pharma.ethz.ch (C. Klein) Date: Wed, 21 Jan 2004 07:44:42 +0100 Subject: [CPMD-list] localized orbitals: Method? Message-ID: <400E1FDA.8A7AF41C@pharma.ethz.ch> Hello, among the many nice features of CPMD, there is the "LOCALIZE" function which allows the localization of orbitals. I am just about to try this; However, various localization methods have been described in the literature, e.g. Pipek-Mezey, Edmiston-Ruedenberg, etc.. Now the question: Which one of those methods is used in cpmd? I would like to give the proper credits (citation) to the method developers. with best regards, Chris Klein From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Fri Jan 30 14:10:43 2004 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Fri, 30 Jan 2004 14:10:43 +0100 Subject: [CPMD-list] Restart files of PIMD In-Reply-To: Your message of "Thu, 29 Jan 2004 14:38:02 +0200." <4018FEAA.5000503@foi.se> Message-ID: <200401301310.i0UDAhH31261@yello.theochem.ruhr-uni-bochum.de> >>> "PA" == Per Andersson writes: PA> Dear Everyone PA> I have with some success run some MD CP calculations of 54 atoms of H PA> but I would like to try PIMD on the same system. After reading the PA> manual and the references therein I still have severe problems. PA> Unfortunately the cluster where I run my calculations has no PA> network connection so I have no input or output listings to show. PA> If I understand this correctly you have to converge the wavefunction PA> in a static run and restart the MD using PA> RESTART WAVEFUNCTIONS COORDINATES VELOCITIES (or at least PA> wavefunctions). This works fine for MD CP. When I try PIMD I add PA> INITIALIZATION in the &PIMD-section, PATH INTEGRALS in the PA> &CPMD-section and restart with the same PA> options as before. I have tried to restart from both the PA> static calculation and the MD CP calculation with the same result. PA> I get the error message PA> RESTART INFORMATION FOUND ON FILE ./RESTART_1 PA> ZHRWF| RESTART FILE NOT FOUND: ./RESTART_2 PA> As far as I understand there will be one restart file for PA> each replica after a PIMD calculation. If they are needed PA> before as well, how do I generate them? dear per, please have a look at the path integral example in the cpmd-tests distribution (available from cpmd.org). there is an example to start an PIMD run from coordinates in the input file (you can grab those from the GEOMETRY file). hope this helps, axel kohlmeyer. PA> Yours PA> Per Andersson PA> _______________________________________________ PA> CPMD-list mailing list PA> CPMD-list at cpmd.org PA> http://www.cpmd.org/mailman/listinfo/cpmd-list -- ======================================================================= Axel Kohlmeyer e-mail: axel.kohlmeyer at theochem.ruhr-uni-bochum.de Lehrstuhl fuer Theoretische Chemie Phone: ++49 (0)234/32-26673 Ruhr-Universitaet Bochum - NC 03/53 Fax: ++49 (0)234/32-14045 D-44780 Bochum http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/ ======================================================================= If you make something idiot-proof, the universe creates a better idiot. From per.andersson at foi.se Fri Jan 30 13:51:48 2004 From: per.andersson at foi.se (Per Andersson) Date: Fri, 30 Jan 2004 14:51:48 +0200 Subject: [CPMD-list] Restart files of PIMD References: <200401301310.i0UDAhH31261@yello.theochem.ruhr-uni-bochum.de> Message-ID: <401A5364.9020005@foi.se> Axel Kohlmeyer wrote: > > PA> As far as I understand there will be one restart file for > PA> each replica after a PIMD calculation. If they are needed > PA> before as well, how do I generate them? > > dear per, > > please have a look at the path integral example in the > cpmd-tests distribution (available from cpmd.org). there is > an example to start an PIMD run from coordinates in the > input file (you can grab those from the GEOMETRY file). > > hope this helps, > axel kohlmeyer. > Dear Axel Thank you for the recommendation. I will try to optimize the wavefunction from scratch using PIMD. As a temporary hack I solved the problem with a crude approach. I just made the required number of copies of the non-PIMD restart file and then did some short PIMD calculations to re-converge the wavefunction before I did the production run. Yours /Per From ivan_van_van at yahoo.com Thu Jan 29 21:06:08 2004 From: ivan_van_van at yahoo.com (Ivan Ivanov) Date: Thu, 29 Jan 2004 12:06:08 -0800 (PST) Subject: [CPMD-list] path integral In-Reply-To: <200401200913.i0K9D9a30806@yello.theochem.ruhr-uni-bochum.de> Message-ID: <20040129200608.77561.qmail@web20412.mail.yahoo.com> Dear CPMD users, 1.How do I select nuclei for PATH INTEGRAL simulations? e.g I want to treat only 1 nucleus as "quantum", others behaving classically. The manual describes the situation where all nuclei are treated as "quantum". 2. Does anyone have experience with using PATH INTEGRAL SIMULATION coupled with using EXTERNAL POTENTIAL option? More generally - Is CPMD suitable for studying quantum behaviour of nuclei under external electrostatic field? Thanks for your answers! Iv __________________________________ Do you Yahoo!? Yahoo! SiteBuilder - Free web site building tool. Try it! http://webhosting.yahoo.com/ps/sb/ From hutter at pci.unizh.ch Fri Jan 30 18:59:52 2004 From: hutter at pci.unizh.ch (Juerg Hutter) Date: Fri, 30 Jan 2004 18:59:52 +0100 (MET) Subject: [CPMD-list] localized orbitals: Method? In-Reply-To: <400E1FDA.8A7AF41C@pharma.ethz.ch> References: <400E1FDA.8A7AF41C@pharma.ethz.ch> Message-ID: Hi > Hello, > > among the many nice features of CPMD, there is the "LOCALIZE" function > which allows the localization of orbitals. I am just about to try this; > However, various localization methods have been described in the > literature, e.g. Pipek-Mezey, Edmiston-Ruedenberg, etc.. > > Now the question: Which one of those methods is used in cpmd? I would > like to give the proper credits (citation) to the method developers. The papers to look at and probably cite are @ARTICLE{wannier, author = {N. Marzari and D. Vanderbilt}, journal = PHYSRB, volume = {56}, pages = {12847}, year = {1997} } @ARTICLE{berghold2000, author = {G. Berghold and C. J. Mundy and A. H. Romero and J. Hutter and M. Parrinello}, journal = PHYSRB, volume = {61}, pages = {10040}, year = {2000} } The procedure used is equivalent (for single k-point) to Boys localization. regards Juerg Hutter > > with best regards, > > Chris Klein > _______________________________________________ > CPMD-list mailing list > CPMD-list at cpmd.org > http://www.cpmd.org/mailman/listinfo/cpmd-list > From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Sat Jan 31 11:11:53 2004 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Sat, 31 Jan 2004 11:11:53 +0100 Subject: [CPMD-list] path integral In-Reply-To: Your message of "Thu, 29 Jan 2004 12:06:08 PST." <20040129200608.77561.qmail@web20412.mail.yahoo.com> Message-ID: <200401311011.i0VABrf12215@yello.theochem.ruhr-uni-bochum.de> >>> "II" == Ivan Ivanov writes: II> Dear CPMD users, II> 1.How do I select nuclei for PATH INTEGRAL II> simulations? II> e.g I want to treat only 1 nucleus as "quantum", II> others behaving classically. hello ivan, may i ask, what would you want to achieve by that? i don't think that you would save any significant computational effort by that. the way path integral is implemented in cpmd, you basically run a 'regular' BO- or CP-MD simulation of the whole system for each replica. so it should not make a difference if only one or many of the nuclei are treated as 'quantum'. II> The manual describes the situation where all nuclei II> are II> treated as "quantum". II> 2. Does anyone have experience with using PATH II> INTEGRAL II> SIMULATION coupled with using EXTERNAL POTENTIAL II> option? i don't know whether someone has tried this yet, but due to the way path integral is implemented (see above). there is a good chance that i will work, or may only need rather small changes to the source code to make it work. II> More generally - Is CPMD suitable for studying quantum II> behaviour of nuclei under external electrostatic II> field? yes. you may want to contact Rodolphe Vuilleumier , since he has recently contributed some code for that kind of computation to the current cpmd development version. hope this helps, axel kohlmeyer. II> Thanks for your answers! II> Iv II> __________________________________ II> Do you Yahoo!? II> Yahoo! SiteBuilder - Free web site building tool. Try it! II> http://webhosting.yahoo.com/ps/sb/ II> _______________________________________________ II> CPMD-list mailing list II> CPMD-list at cpmd.org II> http://www.cpmd.org/mailman/listinfo/cpmd-list -- ======================================================================= Axel Kohlmeyer e-mail: axel.kohlmeyer at theochem.ruhr-uni-bochum.de Lehrstuhl fuer Theoretische Chemie Phone: ++49 (0)234/32-26673 Ruhr-Universitaet Bochum - NC 03/53 Fax: ++49 (0)234/32-14045 D-44780 Bochum http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/ ======================================================================= If you make something idiot-proof, the universe creates a better idiot. From ivan_van_van at yahoo.com Sat Jan 31 15:57:01 2004 From: ivan_van_van at yahoo.com (Ivan Ivanov) Date: Sat, 31 Jan 2004 06:57:01 -0800 (PST) Subject: [CPMD-list] path integral In-Reply-To: <200401311011.i0VABrf12215@yello.theochem.ruhr-uni-bochum.de> Message-ID: <20040131145701.21470.qmail@web20413.mail.yahoo.com> Dear Dr.Kohlmeyer, First of all - congratulations for the great resource you made available for cpmd users. now to the point: I am not trying to save computational load. The problem is that we want to study the effect of treating some nuclei classically while others as quantum particles; dissection of quantum and thermal fluctuations etc. This is part of the research. So, please, if someone has practical hints how to do this with CPMD - help us. Sincerely, Iv Axel Kohlmeyer wrote: >>> "II" == Ivan Ivanov writes: II> Dear CPMD users, II> 1.How do I select nuclei for PATH INTEGRAL II> simulations? II> e.g I want to treat only 1 nucleus as "quantum", II> others behaving classically. hello ivan, may i ask, what would you want to achieve by that? i don't think that you would save any significant computational effort by that. the way path integral is implemented in cpmd, you basically run a 'regular' BO- or CP-MD simulation of the whole system for each replica. so it should not make a difference if only one or many of the nuclei are treated as 'quantum'. II> The manual describes the situation where all nuclei II> are II> treated as "quantum". II> 2. Does anyone have experience with using PATH II> INTEGRAL II> SIMULATION coupled with using EXTERNAL POTENTIAL II> option? i don't know whether someone has tried this yet, but due to the way path integral is implemented (see above). there is a good chance that i will work, or may only need rather small changes to the source code to make it work. II> More generally - Is CPMD suitable for studying quantum II> behaviour of nuclei under external electrostatic II> field? yes. you may want to contact Rodolphe Vuilleumier , since he has recently contributed some code for that kind of computation to the current cpmd development version. hope this helps, axel kohlmeyer. II> Thanks for your answers! II> Iv II> __________________________________ II> Do you Yahoo!? II> Yahoo! SiteBuilder - Free web site building tool. Try it! II> http://webhosting.yahoo.com/ps/sb/ II> _______________________________________________ II> CPMD-list mailing list II> CPMD-list at cpmd.org II> http://www.cpmd.org/mailman/listinfo/cpmd-list -- ======================================================================= Axel Kohlmeyer e-mail: axel.kohlmeyer at theochem.ruhr-uni-bochum.de Lehrstuhl fuer Theoretische Chemie Phone: ++49 (0)234/32-26673 Ruhr-Universitaet Bochum - NC 03/53 Fax: ++49 (0)234/32-14045 D-44780 Bochum http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/ ======================================================================= If you make something idiot-proof, the universe creates a better idiot. _______________________________________________ CPMD-list mailing list CPMD-list at cpmd.org http://www.cpmd.org/mailman/listinfo/cpmd-list --------------------------------- Do you Yahoo!? Yahoo! SiteBuilder - Free web site building tool. Try it! -------------- next part -------------- An HTML attachment was scrubbed... URL: http://cpmd.org/pipermail/cpmd-list/attachments/20040131/eb42013f/attachment.html From tironem at fiu.edu Sat Jan 31 22:32:11 2004 From: tironem at fiu.edu (Massimiliano Tirone) Date: Sat, 31 Jan 2004 16:32:11 -0500 (EST) Subject: [CPMD-list] KPOINTS Message-ID: Hello, I have a very simple question, if the instruction KPOINTS is omitted from the input file, does the program consider only the Gamma point for the energy calculations? all the best to all CPMD users, Max From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Sat Jan 31 23:57:09 2004 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Sat, 31 Jan 2004 23:57:09 +0100 (CET) Subject: [CPMD-list] KPOINTS In-Reply-To: Message-ID: On Sat, 31 Jan 2004, Massimiliano Tirone wrote: MT> MT> Hello, MT> I have a very simple question, if the instruction KPOINTS is MT> omitted from the input file, does the program consider MT> only the Gamma point for the energy calculations? hello max, the answer to your question is yes. take care, axel. MT> MT> all the best to all CPMD users, MT> Max MT> MT> _______________________________________________ MT> CPMD-list mailing list MT> CPMD-list at cpmd.org MT> http://www.cpmd.org/mailman/listinfo/cpmd-list MT> MT> -- ======================================================================= Dr. Axel Kohlmeyer e-mail: axel.kohlmeyer at rub.de Lehrstuhl fuer Theoretische Chemie Phone: ++49 (0)234/32-26673 Ruhr-Universitaet Bochum - NC 03/53 Fax: ++49 (0)234/32-14045 D-44780 Bochum http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/ =======================================================================