charged system with CP2K

[S. Ling]

Hello all,

I have a charged system with two fragments, and I want to put the
excess charge on one fragment rather than the whole system (e.g. put
the excess charge on LUMO+1 orbital of the original neutral system
rather than put it on the default LUMO orbital of the original neutral
system. Is it possible to do this in CP2K? I checked the manual of
CP2K, it seems it might be possible to set the charge for one fragment
of the system with the CONFIGURATION keyword under the route of
__ROOT__%FORCE_EVAL%BSSE. Could any one tell me how to do this for my
case (suppose I have two fragments in my system)? There is a similar
example in the tests which are included in the original source code
package of CP2K (../cp2k/tests/QS/regtest-gpw-3/OH-H2O-bsse.inp), the
inputs are as following, could any one how to understand the
"CONFIGURATION" part?

Thank you very much!

Jon

***********************************************************************************************************
&BSSE
&FRAGMENT
LIST 1 3
&END FRAGMENT
&FRAGMENT
LIST 2 4 5
&END FRAGMENT
&CONFIGURATION
GLB_CONF 1 0
SUB_CONF 1 0
CHARGE 0
MULTIPLICITY 2
&END
&CONFIGURATION
GLB_CONF 1 1
SUB_CONF 1 0
CHARGE 0
MULTIPLICITY 2
&END
&CONFIGURATION
GLB_CONF 1 1
SUB_CONF 0 1
CHARGE 0
MULTIPLICITY 0
&END
&CONFIGURATION
GLB_CONF 1 1
SUB_CONF 1 1
CHARGE 0
MULTIPLICITY 2
&END
&END BSSE

[Laino Teodoro]

BSSE stands for BASIS SET SUPERPOSITION ERROR.
Having clarified that.. can you tell me why are you looking into the
BSSE section
if what you want to do has nothing to do (at least at this stage)
with the BSSE?

The only way I know to do something similar to what you want to do is
to restraint/constraint
the charge on few atoms:
http://cp2k.berlios.de/manual/CP2K_INPUT/FORCE_EVAL/DFT/QS/
DDAPC_RESTRAINT.html
http://cp2k.berlios.de/manual/CP2K_INPUT/FORCE_EVAL/DFT/QS/
MULLIKEN_RESTRAINT.html
http://cp2k.berlios.de/manual/CP2K_INPUT/FORCE_EVAL/DFT/QS/
BECKE_CONSTRAINT.html

you may also want to have a look at the regtests using those sections
(to have an idea of the working template).
Hope this helps,
Teo

[Axel]

hello,

On Oct 17, 12:37 pm, "S. Ling" <lsl.mit...@googlemail.com> wrote:
> Hello all,
>
> I have a charged system with two fragments, and I want to put the
> excess charge on one fragment rather than the whole system (e.g. put
> the excess charge on LUMO+1 orbital of the original neutral system
> rather than put it on the default LUMO orbital of the original neutral

wouldn't that constitute an excited state?

do you know a code that _can_ do this kind of thing?

> system. Is it possible to do this in CP2K?

> I checked the manual of
> CP2K, it seems it might be possible to set the charge for one fragment
> of the system with the CONFIGURATION keyword under the route of
> __ROOT__%FORCE_EVAL%BSSE. Could any one tell me how to do this for my
> case (suppose I have two fragments in my system)? There is a similar

you cannot. this is for BSSE calculations. please pick up a text book
on quantum chemistry and update your understanding of the methods you
want to apply.

> example in the tests which are included in the original source code
> package of CP2K (../cp2k/tests/QS/regtest-gpw-3/OH-H2O-bsse.inp), the
> inputs are as following, could any one how to understand the
> "CONFIGURATION" part?

it would be obvious, if you knew what you do in a BSSE calculation.

cheers,
axel.

[S. Ling]

On 10月17日, 下午5时52分, Laino Teodoro <teodoro.la...@gmail.com> wrote:
> BSSE stands for BASIS SET SUPERPOSITION ERROR.
> Having clarified that.. can you tell me why are you looking into the
> BSSE section

Thanks. I know what BSSE stands for. The only keyword I found in the
manual of CP2K which might be useful for my purpose is
"CONFIGURATION", and this keyword is under the route of BSSE, that's
why I look into this section. My system is a big biomolecular system
and it has two molecules/fragments, so I think it makes sense if we
take the BSSE into account, but since it is such a big system, I think
it would also be excusable that if we exclude the BSSE from our
calculations.

> if what you want to do has nothing to do (at least at this stage)
> with the BSSE?
>
> The only way I know to do something similar to what you want to do is
> to restraint/constraint
> the charge on few atoms:http://cp2k.berlios.de/manual/CP2K_INPUT/FORCE_EVAL/DFT/QS/
> DDAPC_RESTRAINT.htmlhttp://cp2k.berlios.de/manual/CP2K_INPUT/FORCE_EVAL/DFT/QS/
> MULLIKEN_RESTRAINT.htmlhttp://cp2k.berlios.de/manual/CP2K_INPUT/FORCE_EVAL/DFT/QS/
> BECKE_CONSTRAINT.html
>

Thanks. I will have a look at these sections.

> you may also want to have a look at the regtests using those sections
> (to have an idea of the working template).
> Hope this helps,
> Teo
>

Thank you very much for your reply. It helps me a lot.

Jon

> > &END BSSE- 隐藏被引用文字 -
>
> - 显示引用的文字 -

[S. Ling]

On 10月17日, 下午7时52分, Axel <akohl...@gmail.com> wrote:
> hello,
>
> On Oct 17, 12:37 pm, "S. Ling" <lsl.mit...@googlemail.com> wrote:
>
> > Hello all,
>
> > I have a charged system with two fragments, and I want to put the
> > excess charge on one fragment rather than the whole system (e.g. put
> > the excess charge on LUMO+1 orbital of the original neutral system
> > rather than put it on the default LUMO orbital of the original neutral
>
> wouldn't that constitute an excited state?

Yes, it might constitute an excited state. I just want to learn what
will happen if we put the excess electron on the LUMO+1 orbital rather
than the LUMO orbital. Will this anionic state be bound or unbound for
this case? If the excess electron could occupy the LUMO+1 orbital,
then what will heppen next? These are what I want to know.

> do you know a code that _can_ do this kind of thing?

Yes. GAUSSIAN can do such kind of thing for molecular systems. If you
are interested, I can tell you how to do it.

> > system. Is it possible to do this in CP2K?
> > I checked the manual of
> > CP2K, it seems it might be possible to set the charge for one fragment
> > of the system with the CONFIGURATION keyword under the route of
> > __ROOT__%FORCE_EVAL%BSSE. Could any one tell me how to do this for my
> > case (suppose I have two fragments in my system)? There is a similar
>
> you cannot. this is for BSSE calculations. please pick up a text book
> on quantum chemistry and update your understanding of the methods you
> want to apply.

Thanks. I will definitely have a look at my text book. I know what
BSSE could do, but I just didn't find any other keywords in the manual
which might be helpful for the task I want to do, then I look into
this section. I have just explained my understanding in the above
reply. Maybe I am still confused and I shouldn't put these two
different things together. Anyway, I will consider other possible
choices.

> > example in the tests which are included in the original source code
> > package of CP2K (../cp2k/tests/QS/regtest-gpw-3/OH-H2O-bsse.inp), the
> > inputs are as following, could any one how to understand the
> > "CONFIGURATION" part?
>
> it would be obvious, if you knew what you do in a BSSE calculation.

Thanks. I have never performed BSSE calculations before. I just got a
general idea of what BSSE could do. I will have a look at by myself.

> cheers,
> axel.

Thank you very much for your reply.

Jon

>
>
>
>
> > Thank you very much!
>
> > Jon
>

> > ***************************************************************************-********************************

[Roberto Scipioni]

Dear all, I have been trying to compile cp2k (just serially) on the Linux Cluster we have Suse Enterprise 9 x86_64 Intel Xeon CPUs Intel ifort 9,10 Intel MKL 9,10 but so far to no avail !! Can you havve a look at the /arch file Linux-x86-64-intel.sopt to see if there is something wrong ? Thanks for your help Roberto Scipioni ####INTEL_INC = /opt/intel/fce/9.1.036/include/ INTEL_INC = /opt/intel/fce/10.0.025/include/ CC       = icc CPP      = cpp FC       = /opt/intel/fce/10.0.025/bin/ifort -FR FC_fixed = /opt/intel/fce/10.0.025/bin/ifort -FI LD       = /opt/intel/fce/10.0.025/bin/ifort -i-static AR       = ar -r ###DFLAGS   = -D__INTEL -D__FFTSG -D__FFTW3 DFLAGS   = -D__INTEL -D__FFTSG CPPFLAGS = -C -traditional $(DFLAGS) -I$(INTEL_INC) ####FCFLAGS  = $(DFLAGS) -I$(INTEL_INC) -O2 -xW -heap-arrays 64 -fpp -free FCFLAGS  = $(DFLAGS) -I$(INTEL_INC) -O2 -unroll -Vaxlib ###LDFLAGS  = $(FCFLAGS)  -L/opt/intel/mkl/9.0/lib/em64t/ -L/root/install/fftw-3.1.3/.libs/ LDFLAGS  = $(FCFLAGS) ###LIBS    = -lmkl_em64t -lguide -lfftw3 #LIBS =  -L/opt/intel/mkl/10.0.1.014/lib/em64t/libmkl_em64t.a\ #        -L/opt/intel/mkl/10.0.1.014/lib/em64t/libguide.a LIBS  =      -L/opt/acml4.1.0/ifort64_int64/lib/libacml.a\              -L/opt/acml4.1.0/ifort64_int64/lib/libacml_mv.a\              -L/root/install/fftw-3.1.3/.libs/libfftw3.a ###LIBS = -L/opt/intel/mkl/9.0/lib/em64t -lmkl_em64t -lguide ###           -L/root/install/fftw-3.1.3/.libs/libfftw3.a OBJECTS_ARCHITECTURE = machine_intel.o ifposix.int:         touch ifposix.int ~

__________________________________________________
Do You Yahoo!?
Poco spazio e tanto spam? Yahoo! Mail ti protegge dallo spam e ti da tanto spazio gratuito per i tuoi file e i messaggi
http://mail.yahoo.it

[Roberto Scipioni]

The error was at the final linking stage R.

[Axel]

On Oct 17, 11:36 pm, Roberto Scipioni <robertino_scipi...@yahoo.it>
wrote:

> The error was at the final linking stage

roberto,

well, _what_ was the error?

you don't go to the doctor and only say "it hurts" but
and expect to get help without also telling _where_ and
_how_ it hurts, do you?

axel.

[Roberto Scipioni]

well unfortunately the synthoms change by changing the conditons (i.e. compilers, math libraries, etc ...) The general error is that the libraries give undefined references to .... for example in what follows is the concluding part of the error ---------------- /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_fft_lib.a(fftw3_lib.o)(.text+0x205): In function `ff                          tw31dm_': : undefined reference to `zdscal_' /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_fft_lib.a(fftw3_lib.o)(.text+0x256): In function `ff                          tw31dm_': : undefined reference to `dfftw_plan_many_dft_' /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_fft_lib.a(fftw3_lib.o)(.text+0x329): In function `ff                          tw33d_': : undefined reference to `dfftw_execute_' /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_fft_lib.a(fftw3_lib.o)(.text+0x335): In function `ff                          tw33d_': : undefined reference to `dfftw_destroy_plan_' /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_fft_lib.a(fftw3_lib.o)(.text+0x372): In function `ff                          tw33d_': : undefined reference to `zdscal_' /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_fft_lib.a(fftw3_lib.o)(.text+0x3a5): In function `ff                          tw33d_': : undefined reference to `zdscal_' /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_fft_lib.a(fftw3_lib.o)(.text+0x3dc): In function `ff                          tw33d_': : undefined reference to `dfftw_plan_dft_3d_' /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_fft_lib.a(fftw3_lib.o)(.text+0x410): In function `ff                          tw33d_': : undefined reference to `dfftw_plan_dft_3d_' /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_fft_lib.a(fftw3_lib.o)(.text+0x444): In function `ff                          tw33d_': : undefined reference to `dfftw_plan_dft_3d_' /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_fft_lib.a(fftw3_lib.o)(.text+0x478): In function `ff                          tw33d_': : undefined reference to `dfftw_plan_dft_3d_' /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_fft_lib.a(fftacml_lib.o)(.text+0x1a9): In function `                          fftacml1dm_': : undefined reference to `zfft1mx_' /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_fft_lib.a(fftacml_lib.o)(.text+0x227): In function `                          fftacml1dm_': : undefined reference to `zfft1mx_' /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_fft_lib.a(fftacml_lib.o)(.text+0x4aa): In function `                          fftacml3d_': : undefined reference to `zfft3dx_' /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_fft_lib.a(fftacml_lib.o)(.text+0x510): In function `                          fftacml3d_': : undefined reference to `zfft3dx_' /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_fft_lib.a(mltfftsg_tools.o)(.text+0xf): In function                           `mltfftsg_tools_mp_scaled_': : undefined reference to `dscal_' /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_base_lib.a(fast.o)(.text+0x8f2): In function `rankup                          _': : undefined reference to `zgeru_' /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_base_lib.a(fast.o)(.text+0x909): In function `rankup                          _': : undefined reference to `zscal_' /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libcp2k_base_lib.a(fast.o)(.text+0x941): In function `rankup                          _': : undefined reference to `zgeru_' make[2]: *** [/home/robechan/cp2k_new2/cp2k/makefiles/../exe/Linux-x86-64-intel/cp2k.sopt] Error 1 make[2]: Leaving directory `/home/robechan/cp2k_new2/cp2k/obj/Linux-x86-64-intel/sopt' make[1]: *** [build] Error 2 make[1]: Leaving directory `/home/robechan/cp2k_new2/cp2k/makefiles' make: *** [sopt] Error 2 ------------------------ Roberto -

[Axel]

On Oct 18, 8:53 pm, Roberto Scipioni <robertino_scipi...@yahoo.it>
wrote:

> well unfortunately the synthoms change by changing the conditons
> (i.e. compilers, math libraries, etc ...)
>
> The general error is that the libraries give
>
> undefined references to ....

that usually means, that you specified them in the wrong order,
that the library is not readable (the /root/fftw-xxx one is
very suspicious), that you compiled the library wrong, or
specified not all required libraries (i.e. you forgot lapack)
or that you changed the arch/xxxx file and then continued compilation
without removing all previously compiled objects and libraries.

none of that is actually very specific to cp2k,
but would cause similar problems with other packages.

cheers,
axel.

[S. Ling]

This error information is very common. Besides what Axel has
mentioned, it might also be a result of mixing different compilers,
e.g. you use GNU compiler to compile the FFTW library, and you use the
Intel compiler to compile CP2K, or you compile the FFTW library in 32-
bit mode and you compile CP2K in 64-bit mode.

Hope this helps.

Cheers, Jon

On 10月19日, 上午1时53分, Roberto Scipioni <robertino_scipi...@yahoo.it>
wrote:

> well unfortunately the synthoms change by changing the conditons
> (i.e. compilers, math libraries, etc ...)
>
> The general error is that the libraries give
>
> undefined references to ....
>
> for example in what follows is the concluding part of the error
>
> ----------------

> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_fft_lib.a(fftw3_lib.o)(.text+0x205): In function `ff tw31dm_':
> : undefined reference to `zdscal_'
> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_fft_lib.a(fftw3_lib.o)(.text+0x256): In function `ff tw31dm_':
> : undefined reference to `dfftw_plan_many_dft_'
> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_fft_lib.a(fftw3_lib.o)(.text+0x329): In function `ff tw33d_':
> : undefined reference to `dfftw_execute_'
> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_fft_lib.a(fftw3_lib.o)(.text+0x335): In function `ff tw33d_':
> : undefined reference to `dfftw_destroy_plan_'
> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_fft_lib.a(fftw3_lib.o)(.text+0x372): In function `ff tw33d_':
> : undefined reference to `zdscal_'
> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_fft_lib.a(fftw3_lib.o)(.text+0x3a5): In function `ff tw33d_':
> : undefined reference to `zdscal_'
> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_fft_lib.a(fftw3_lib.o)(.text+0x3dc): In function `ff tw33d_':
> : undefined reference to `dfftw_plan_dft_3d_'
> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_fft_lib.a(fftw3_lib.o)(.text+0x410): In function `ff tw33d_':
> : undefined reference to `dfftw_plan_dft_3d_'
> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_fft_lib.a(fftw3_lib.o)(.text+0x444): In function `ff tw33d_':
> : undefined reference to `dfftw_plan_dft_3d_'
> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_fft_lib.a(fftw3_lib.o)(.text+0x478): In function `ff tw33d_':
> : undefined reference to `dfftw_plan_dft_3d_'
> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_fft_lib.a(fftacml_lib.o)(.text+0x1a9): In function ` fftacml1dm_':
> : undefined reference to `zfft1mx_'
> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_fft_lib.a(fftacml_lib.o)(.text+0x227): In function ` fftacml1dm_':
> : undefined reference to `zfft1mx_'
> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_fft_lib.a(fftacml_lib.o)(.text+0x4aa): In function ` fftacml3d_':
> : undefined reference to `zfft3dx_'
> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_fft_lib.a(fftacml_lib.o)(.text+0x510): In function ` fftacml3d_':
> : undefined reference to `zfft3dx_'
> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_fft_lib.a(mltfftsg_tools.o)(.text+0xf): In function `mltfftsg_tools_mp_scaled_':
> : undefined reference to `dscal_'
> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_base_lib.a(fast.o)(.text+0x8f2): In function `rankup _':
> : undefined reference to `zgeru_'
> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_base_lib.a(fast.o)(.text+0x909): In function `rankup _':
> : undefined reference to `zscal_'
> /home/robechan/cp2k_new2/cp2k/makefiles/../lib/Linux-x86-64-intel/sopt/libc-p2k_base_lib.a(fast.o)(.text+0x941): In function `rankup _':
> : undefined reference to `zgeru_'
> make[2]: *** [/home/robechan/cp2k_new2/cp2k/makefiles/../exe/Linux-x86-64-intel/cp2k.sop-t] Error 1

[Axel]

On Oct 18, 10:25 pm, "S. Ling" <lsl.mit...@googlemail.com> wrote:
> This error information is very common. Besides what Axel has
> mentioned, it might also be a result of mixing different compilers,
> e.g. you use GNU compiler to compile the FFTW library, and you use the

please note, that this is _not_ a problem. particularly not with
FFTW3, as this only contains C code and the underscoring conventions
of gfortran and ifort are identical. only g77 (now obsolete), f2c
(predates
g77) and "exotic" compilers (Absoft) have different rules.

on linux it is in general a good idea to compile everything with the
GNU compilers except the (few) applications that require different
compilers or benefit much from it. with well written (c) code one can
actually implement compatibility to many fortran compilers through
#pragma weak directives.

> Intel compiler to compile CP2K, or you compile the FFTW library in 32-
> bit mode and you compile CP2K in 64-bit mode.

this would show in the linker step as "skipping incompatible library
XXXX"

cheers,
axel.

[Roberto Scipioni]

Hi all, Now I have been trying to make the parallel version work I used the arch file (I used ifort 10.0 and mkl 10)

CC       = icc CPP      = cpp

FC       = /opt/ompi_ifort10/bin/mpif90 -FR FC_fixed = /opt/ompi_ifort10/bin/mpif90 -FI LD       = /opt/ompi_ifort10/bin/mpif90 -i_dynamic -openmp AR       = ar -r DFLAGS   = -D__INTEL -D__FFTSG -D__parallel -D__BLACS -D__SCALAPACK -D__FFTW3 CPPFLAGS = -traditional -C $(DFLAGS) -P FCFLAGS  = $(DFLAGS) -O3 -Vaxlib LDFLAGS  = $(FCFLAGS) -L/opt/intel/mkl/10.0.1.014/lib/em64t/ LIBS=/opt/intel/mkl/10.0.1.014/lib/em64t/libmkl_scalapack.a /opt/intel/mkl/10.0.1.014/lib/em64t/libmkl_blacs_openmpi_ilp64.a /opt/intel/mkl/10.0.1.014/lib/em64t/libmkl_em64t.a /opt/FFTW/fftw-3.1.3/.libs/libfftw3.a -lguide -lpthread OBJECTS_ARCHITECTURE = machine_intel.o The compilation was fine I produced the binary cp2k.popt Then I went to the tests directory and I use the script #!/bin/sh ########################################################### #PBS -N TEST #PBS -l walltime=100:00:00 #PBS -l nodes=1:ppn=1:node0 ############################################################ cd ${PBS_O_WORKDIR} /opt/ompi_ifort10/bin/mpirun -np 1 -npernode 1 -x OMP_NUM_THREADS=1 ~/cp2k_new2/cp2k/exe/TEST_64_2_ifort10_mkl10/cp2k.popt 3H2O-ep.inp > out exit 0 ############################################################ ~ ~ But I got the following error [shizuwo:27227] *** Process received signal *** [shizuwo:27227] Signal: Segmentation fault (11) [shizuwo:27227] Signal code: Address not mapped (1) [shizuwo:27227] Failing at address: 0x100 [shizuwo:27227] [ 0] /lib64/tls/libpthread.so.0 [0x2a957bd650] [shizuwo:27227] [ 1] /home/robechan/cp2k_new2/cp2k/exe/TEST_64_2_ifort10_mkl10/cp2k.popt(cp_blacs_env_mp_cp_blacs_env_create_+0x37f) [0xe0325d] [shizuwo:27227] [ 2] /home/robechan/cp2k_new2/cp2k/exe/TEST_64_2_ifort10_mkl10/cp2k.popt(qs_environment_types_mp_qs_env_create_+0x3bc) [0x849c2c] [shizuwo:27227] [ 3] /home/robechan/cp2k_new2/cp2k/exe/TEST_64_2_ifort10_mkl10/cp2k.popt(qs_main_mp_quickstep_create_force_env_+0x7d) [0x8e1ec9] [shizuwo:27227] [ 4] /home/robechan/cp2k_new2/cp2k/exe/TEST_64_2_ifort10_mkl10/cp2k.popt(f77_interface_mp_create_force_env_+0x1431) [0x4c2ce3] [shizuwo:27227] [ 5] /home/robechan/cp2k_new2/cp2k/exe/TEST_64_2_ifort10_mkl10/cp2k.popt(ep_methods_mp_ep_env_init_low_+0x66a) [0x13fa502] [shizuwo:27227] [ 6] /home/robechan/cp2k_new2/cp2k/exe/TEST_64_2_ifort10_mkl10/cp2k.popt(cp_ep_init_+0x8) [0xf35cbc] [shizuwo:27227] [ 7] /home/robechan/cp2k_new2/cp2k/exe/TEST_64_2_ifort10_mkl10/cp2k.popt(ep_types_mp_ep_env_create_+0x21b) [0xf0b783] [shizuwo:27227] [ 8] /home/robechan/cp2k_new2/cp2k/exe/TEST_64_2_ifort10_mkl10/cp2k.popt(force_env_methods_mp_ep_create_force_env_+0x46) [0x511210] [shizuwo:27227] [ 9] /home/robechan/cp2k_new2/cp2k/exe/TEST_64_2_ifort10_mkl10/cp2k.popt(f77_interface_mp_create_force_env_+0x1532) [0x4c2de4] [shizuwo:27227] [10] /home/robechan/cp2k_new2/cp2k/exe/TEST_64_2_ifort10_mkl10/cp2k.popt(cp2k_runs_mp_cp2k_run_+0x1498) [0x48dc5c] [shizuwo:27227] [11] /home/robechan/cp2k_new2/cp2k/exe/TEST_64_2_ifort10_mkl10/cp2k.popt(cp2k_runs_mp_run_input_+0x10a) [0x48c762] [shizuwo:27227] [12] /home/robechan/cp2k_new2/cp2k/exe/TEST_64_2_ifort10_mkl10/cp2k.popt(MAIN__+0xda5) [0x48b7c5] [shizuwo:27227] [13] /home/robechan/cp2k_new2/cp2k/exe/TEST_64_2_ifort10_mkl10/cp2k.popt(main+0x2a) [0x48aa12] [shizuwo:27227] [14] /lib64/tls/libc.so.6(__libc_start_main+0xea) [0x2a96d91aaa] [shizuwo:27227] [15] /home/robechan/cp2k_new2/cp2k/exe/TEST_64_2_ifort10_mkl10/cp2k.popt(for_write_seq_fmt_xmit+0x42) [0x48a94a] [shizuwo:27227] *** End of error message *** mpirun noticed that job rank 0 with PID 27227 on node shizuwo exited on signal 11 (Segmentation fault). ~ and the following output in the file     out ---------------------------------------------------------------------------------------------------------------- @shizuwo has created process number                                 0     27227   **** **** ******  **  PROGRAM STARTED AT              Wed Oct 22 15:24:26 2008  ***** ** ***  *** **   PROGRAM STARTED ON                               shizuwo  **    ****   ******    PROGRAM STARTED BY  ***** **    ** ** **   PROGRAM PROCESS ID                                 27227   **** **  *******  **  PROGRAM STARTED IN /home/robechan/cp2k_new2/cp2k/tests/E                                            P  CP2K| version string:                  CP2K version 2.0.0 (Development Version)  CP2K| is freely available from                          http://cp2k.berlios.de/  CP2K| Program compiled at                          Wed Oct 22 15:11:53 JST 2008  CP2K| Program compiled on                                               shizuwo  CP2K| Program compiled for                              TEST_64_2_ifort10_mkl10  CP2K| Last CVS entry                   xc_xwpbe.F/1.12/Tue Aug 26 23:46:48 2008  CP2K| Input file name                                               3H2O-ep.inp  GLOBAL| Force Environment number                                              1  GLOBAL| Basis set file name                                           BASIS_SET  GLOBAL| Potential file name                                           POTENTIAL  GLOBAL| MM Potential file name                                     MM_POTENTIAL  GLOBAL| Coordinate file name                                      __STD_INPUT__  GLOBAL| Method name                                                        CP2K  GLOBAL| Project name                                                        H2O  GLOBAL| Preferred FFT library                                             FFTW3  GLOBAL| Run type                                                             MD  GLOBAL| All-to-all communication in single precision                          F  GLOBAL| FFTs using library dependent lengths                                  F  GLOBAL| Global print level                                                    1  GLOBAL| Total number of message passing processes                             1  GLOBAL| Number of threads for this process                                    1  GLOBAL| This output is from process                                           0  15:24:26 NOTE    in ep_methods:ep_env_init_low ep_methods:ep_env_init_low  creating main_qs_env ------------------------------------------------------------------------------- The segmentation occurs because the programs is probably looking for something and does not find it any idea ? Thanks Roberto

---

Scopri il Blog di Yahoo! Mail: trucchi, novità, consigli... e scrivi la tua opinione!

[Laino Teodoro]

On 22 Oct 2008, at 09:12, Roberto Scipioni wrote:

The segmentation occurs because the programs is probably looking for something and does not find it any idea ?

No unfortunately the problem looks more located in your blacs/scalapack or compiler bug..

which version is your compiler? 10.0.WHAT?

Teo

[Axel]

roberto,

there are a bunch of possible reasons for getting segmentation faults
with intel compilers. i already sent you my arch files that work well
with recent intel 10 compilers and reduce the optimization level where
the compiler has a tendency to miscompile cp2k, do larger local
allocations
on the heap and link with MKL in a way that it does no unwanted multi-
threading.

this all has been discussed in this forum repeatedly (try searching
the
google group for "intel fortran" or "ifort"...

cheers,
axel.

On Oct 22, 3:12 am, Roberto Scipioni <robertino_scipi...@yahoo.it>
wrote:

[Roberto Scipioni]

I used Intel ifort 10.0.025  Then I compiled openmpi-1.2.2 with this compiler and then the produced mpif90 was used for the compilation of cp2k Do you think the BLACS and Scalapack I used do not work with this compiler ? R. --- Mer 22/10/08, Laino Teodoro <teodor...@gmail.com> ha scritto: Da: Laino Teodoro <teodor...@gmail.com> Oggetto: [CP2K:1476] Re: Compiling cp2k serial on x86_64 Intel Xeon Intel Ifort Suse Enterprise 9 (PARALLEL) A: cp...@googlegroups.com Data: Mercoledì 22 ottobre 2008, 09:19

[Roberto Scipioni]

I eventually managed to solve the problem and fix it. The delicate point in the linkage. We should be careful to to avoid the use of the ILP version of the MKL library, such as /mkl/10.0.1.014/lib/em64t/libmkl_blacs_openmpi_ilp64.a and replace it with the *most recommended* LP version as follows /mkl/10.0.1.014/lib/em64t/libmkl_blacs_openmpi_lp64.a For more details I strongly recommend to have a look at the following www site http://www.unix.org/version2/whatsnew/lp64_wp.html Roberto --- Mer 22/10/08, Axel <akoh...@gmail.com> ha scritto: Da: Axel <akoh...@gmail.com> Oggetto: [CP2K:1477] Re: Compiling cp2k serial on x86_64 Intel Xeon Intel Ifort Suse Enterprise 9 (PARALLEL) A: "cp2k" <cp...@googlegroups.com> Data: Mercoledì 22 ottobre 2008, 18:14