[CP2K:6600] Errors in regtests, CP2K 2.6
bbry... at gmail.com
bbry... at gmail.com
Wed Jun 10 09:17:01 UTC 2015
Setting the values for libint_max_am and libderiv_max_am1 to what they are
in libint configuration didn't help, but removing the lines completely did.
So, this went through with no errors for serial (sopt):
CC = gcc
CPP =
FC = gfortran
LD = gfortran
AR = ar -r
DFLAGS = -D__ACML -D__FFTW3 -D__LIBINT -D__LIBXC2
CPPFLAGS = $$FFTW_INCLUDE $$ACML_INCLUDE $$LIBXC_INCLUDE $$LIBINT_INCLUDE
FCFLAGS = $(DFLAGS) -O2 --fast-math -ffree-form -ffree-line-length-none \
-ftree-vectorize -funroll-loops -mtune=native\
$$FFTW_INCLUDE $$ACML_INCLUDE $$LIBXC_INCLUDE $$LIBINT_INCLUDE
LDFLAGS = $(FCFLAGS) $$FFTW_LDFLAGS $$ACML_LDFLAGS $$LIBINT_LDFLAGS
$$LIBXC_LDFLAGS
LIBS = -lacml -lfftw3 -lxcf90 -lxc -lderiv -lint -lm -ldl
Thanks a lot!
Now time to see if removing those lines is enough to get it to work for
openmp, etc. too :)
-Birgitte
On Tuesday, June 9, 2015 at 3:06:54 PM UTC+2, bbr... at gmail.com wrote:
>
> I didn't build libint (this is not my personal computer, but a cluster).
> However, I checked in the config file for libint here, and it looks as if
> it is configured with libint_max_am=4 and libderiv_max_am1=3
>
> So I think you are right and that may very well be the reason for the
> errors! I will try to build a new version of CP2K with those values.
>
> The stack size limit is set to unlimited, so that should not be the
> problem.
>
> Thanks! I will tell you how it goes when I have tried with the new values
> for libint_max_am and libderiv_max_am1
>
> -Birgitte
>
> On Tuesday, June 9, 2015 at 2:42:48 PM UTC+2, IBethune wrote:
>>
>> Hi Birgitte,
>>
>> Looking at all those test failures, they are all failing in or around the
>> start of the HFX calculation. One thing that could cause such a failure is
>> if that you have defined:
>>
>> > -D__LIBINT_MAX_AM=7 -D__LIBDERIV_MAX_AM1=6
>>
>> Do these correspond to the values you used when configuring libint?
>>
>> --with-libint-max-am=7 --with-libderiv-max-am1=6 ??
>>
>> Another possibility is that your system has a very small stack size
>> limit. Run ‘ulimit -s’ to see what it is, and try running again after
>> setting 'ulimit -s unlimited'
>>
>> Apart from that, I would suggest running one of the tests with 'ulimit -c
>> unlimited’ to get a core file, then open it with a debugger (e.g. ‘gdb
>> /path/to/my/cp2k.sopt /path/to/my/core.file' and run ‘bt’ to identify the
>> location of the crash. If you post here we may be able to suggest a fix.
>>
>> - Iain
>>
>> --
>>
>> Iain Bethune
>> Project Manager, EPCC
>>
>> Email: ibe... at epcc.ed.ac.uk
>> Twitter: @IainBethune
>> Web: http://www2.epcc.ed.ac.uk/~ibethune
>> Tel/Fax <http://www2.epcc.ed.ac.uk/~ibethuneTel/Fax>: +44 (0)131 650
>> 5201/6555
>> Mob: +44 (0)7598317015
>> Addr: 2404 JCMB, The King's Buildings, Peter Guthrie Tait Road,
>> Edinburgh, EH9 3FD
>>
>> > On 9 Jun 2015, at 12:44, bbr... at gmail.com wrote:
>> >
>> > I am trying to compile CP2K 2.6 (sopt, ssmp, popt, psmp) on a Linux
>> Ubuntu 12.04 system.
>> >
>> > I tried compiling with Intel/MKL, but get runtime fail for all
>> regtests. I then tried GCC/MKL, and most tests again failed.
>> >
>> > I am currently compiling with:
>> >
>> > GCC 4.9.1
>> > FFTW 3.3.4
>> > Libint 1.1.4
>> > Libxc 2.2.2
>> > ACML 5.3.1
>> >
>> > This works best of the combinations I have tried until now, but I still
>> get about 5% runtime failure on the regtests for serial CP2K. The OpenMP
>> version hangs on Ar-ref-3.inp and never goes further.
>> >
>> > Linux-x86-64-gfortran.sopt:
>> >
>> > CC = gcc
>> > CPP =
>> > FC = gfortran
>> > LD = gfortran
>> > AR = ar -r
>> > DFLAGS = -D__ACML -D__FFTW3 -D__LIBINT -D__LIBXC2\
>> > -D__LIBINT_MAX_AM=7 -D__LIBDERIV_MAX_AM1=6 -D__MAX_CONTR=4
>> > CPPFLAGS = $$FFTW_INCLUDE $$ACML_INCLUDE $$LIBXC_INCLUDE
>> $$LIBINT_INCLUDE
>> > FCFLAGS = $(DFLAGS) -O2 --fast-math -ffree-form
>> -ffree-line-length-none \
>> > -ftree-vectorize -funroll-loops -mtune=native\
>> > $$FFTW_INCLUDE $$ACML_INCLUDE $$LIBXC_INCLUDE $$LIBINT_INCLUDE
>> > LDFLAGS = $(FCFLAGS) $$FFTW_LDFLAGS $$ACML_LDFLAGS $$LIBINT_LDFLAGS
>> $$LIBXC_LDFLAGS
>> > LIBS = -lacml -lfftw3 -lxcf90 -lxc -lderiv -lint -lm -ldl
>> >
>> >
>> >
>> > -------------------------------------------------------------------
>> > Linux-x86-64-gfortran.ssmp:
>> >
>> > CC = gcc
>> > CPP =
>> > FC = gfortran
>> > LD = gfortran
>> > AR = ar -r
>> > DFLAGS = -D__ACML -D__FFTW3 -D__LIBINT -D__LIBXC2\
>> > -D__LIBINT_MAX_AM=7 -D__LIBDERIV_MAX_AM1=6 -D__MAX_CONTR=4
>> > CPPFLAGS = $$FFTW_INCLUDE $$ACML_INCLUDE $$LIBXC_INCLUDE
>> $$LIBINT_INCLUDE
>> > FCFLAGS = $(DFLAGS) -O0 -g -ffree-form -ffree-line-length-none\
>> > -fopenmp -ftree-vectorize -funroll-loops\
>> > -mtune=native\
>> > $$FFTW_INCLUDE $$ACML_INCLUDE $$LIBXC_INCLUDE
>> $$LIBINT_INCLUDE
>> > LDFLAGS = $(FCFLAGS) $$FFTW_LDFLAGS $$ACML_LDFLAGS $$LIBINT_LDFLAGS
>> $$LIBXC_LDFLAGS -static-libgfortran
>> > LIBS = -lacml -lfftw3 -lfftw3_threads -lxcf90 -lxc -lderiv -lint
>> -lm -ldl
>> >
>> >
>> > I have attached the error messages and the list of status of regtest
>> for sopt.
>> >
>> > Any suggestions to what the problem might be?
>> >
>> > Thank you!
>> >
>> > -Birgitte Brydsö
>> >
>> >
>> >
>> > --
>> > You received this message because you are subscribed to the Google
>> Groups "cp2k" group.
>> > To unsubscribe from this group and stop receiving emails from it, send
>> an email to cp2k+... at googlegroups.com.
>> > To post to this group, send email to cp... at googlegroups.com.
>> > Visit this group at http://groups.google.com/group/cp2k.
>> > For more options, visit https://groups.google.com/d/optout.
>> > <output_regtest_acml_gfortran491.txt><error_summary.txt>
>>
>>
>> --
>> The University of Edinburgh is a charitable body, registered in
>> Scotland, with registration number SC005336.
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20150610/14c7923d/attachment.htm>
More information about the CP2K-user
mailing list