Re: [CP2K:4357] Digest for cp2k@googlegroups.com - 2 Messages in 1 Topic

71 views

Skip to first unread message

何智雄

unread,

Mar 24, 2013, 9:53:12 PM3/24/13

to cp...@googlegroups.com

Dear Dr. Hutter,

Thanks for your reply. I rebuilt cp2k with -D__HAS_ISO_C_BINDING, but I got the same set of failed tests.

and I checked the libint-1.1.4 installation:

./configure -prefix=/pkg/chem/lib/gcc/libint --with-libint-max-am=5 --with-libderiv-max-am1=4 --enable-deriv --enable-r12 --enable-int

I wonder what might be wrong with the configuration?

Jyh-Shyong Ho

2013/3/23 <cp...@googlegroups.com>

Today's Topic Summary

Group: http://groups.google.com/group/cp2k/topics

cp2k 2-4 installation and regresstio test result [2 Updates]

cp2k 2-4 installation and regresstio test result

jsho <jyhsh...@gmail.com> Mar 22 01:33AM -0700

Hi,

I am new to the CP2K and learning the program. I downloaded the program
last week and tried to install it on our computer,
after several attempts with intel (version 12) and gfortran (4.6), I found
that the regression test with Linux-x86-64-gfortran.popt
failed only 22 cases (using mpiexec -np 2 )

--------------------------------- summary --------------------------------
number of FAILED tests 22
number of WRONG tests 2
number of CORRECT tests 2359
number of NEW tests 0
number of tests 2383
------------------------------

The failed cases are:

QS/regtest-hybrid/H2O-hybrid-bhandh.inp.out
QS/regtest-hybrid/H2O-hybrid-bhandhlyp.inp.out
QS/regtest-hybrid/CH3-hybrid-b3lyp-lsd-g03.inp.out
QS/regtest-hybrid/CH3-hybrid-handh-lsd.inp.out
QS/regtest-hybrid/CH3-hybrid-handhlyp-lsd.inp.out
QS/regtest-hybrid/CH3-hybrid-pbe0-lsd.inp.out
QS/regtest-hfx/CH-hfx-md.inp.out
QS/regtest-hfx/CH-hfx-md-2.inp.out
QS/regtest-hfx/4H2O-mix-disk-ram-on-the-fly.inp.out :
QS/regtest-mp2/H2O-01.inp.out
QS/regtest-admm/MD-2_OT.inp.out :
QS/regtest-hfx/H2O-hfx-emd.inp.out
QS/regtest-hfx/H2O-hfx-emd-restart.inp.out
QS/regtest-hfx/H2O-hfx-atprop.inp.out
QS/regtest-hfx-periodic/H2O-trunc-auto-md.inp.out
QS/regtest-mp2/Li.inp.out
QS/regtest-mp2/Ne2.inp.out
QS/regtest-hfx-periodic/CH3-trunc-auto-md.inp.out
QS/regtest-mp2/H2O_B2PLYP.inp.out
QS/regtest-mp2/H2O_B2GPPLYP.inp.out
QS/regtest-mp2/H2O_DSD-BLYP.inp.out
QS/regtest-mp2/H2O_B2PLYP_D3.inp.out
QS/regtest-mp2/CH3_B2PLYP_allelec.inp.out
QS/regtest-mp2/H2O_ADMM_MP2.inp.out

Here is the Linux-x86-64.gfortran.popt I used:

CC = gcc
CPP =

FC = mpif90
LD = mpif90

AR = ar -r

CPPFLAGS =
DFLAGS = -D__GFORTRAN -D__FFTSG -D__LIBINT -D__FFTW3 -D__parallel
-D__SCALAPACK -D__BLACS -D__LIBXC2 \
-D__HAS_NO_ISO_C_BINDING -D__HAS_smm_dnn
FCFLAGS = -O3 -march=native -ffast-math -funroll-loops -g -ffree-form
$(DFLAGS) \
-I/pkg/chem/lib/gcc/include \
-I/pkg/chem/lib/gcc/libint/include \
-I/pkg/chem/lib/gcc/fftw332/include
LDFLAGS = $(FCFLAGS) -L/pkg/chem/lib/gcc/libint/lib
LIBS = /pkg/chem/lib/gcc/libint/lib/libint_cpp_wrapper.o -lderiv -lint
-lstdc++ \
/pkg/chem/lib/gcc/lib/libxc.a \
/pkg/chem/lib/gcc/fftw332/lib/libfftw3.a \
/pkg/chem/lib/gcc/lib/libsmm_dnn.a \
/pkg/chem/lib/gcc/lapack/libscalapack.a\
/pkg/chem/lib/gcc/lapack/liblapack.a \
/pkg/chem/lib/gcc/lapack/librefblas.a

OBJECTS_ARCHITECTURE = machine_gfortran.o

Here scalapack 2.0 was compiled with acml 5.0.2 and, mpif90 is from
mvapich2-1.8

I wonder if there can be any improvement in my compilation to pass all the
regression tests ?
Thanks for any suggestion.

Jyh-Shyong Ho

hut...@pci.uzh.ch Mar 22 10:35AM +0100

Hi

it seems that all the tests that fail are the ones using libint.
As all modern compilers do provide the Iso_C_binding I would suggest
to use -D__HAS_ISO_C_BINDING

regards

Juerg Hutter

--------------------------------------------------------------
Juerg Hutter Phone : ++41 44 635 4491
Physical Chemistry Institute FAX : ++41 44 635 6838
University of Zurich E-mail: hut...@pci.uzh.ch
Winterthurerstrasse 190
CH-8057 Zurich, Switzerland
---------------------------------------------------------------

-----cp...@googlegroups.com wrote: -----
To: cp...@googlegroups.com
From: jsho
Sent by: cp...@googlegroups.com
Date: 03/22/2013 09:33AM
Subject: [CP2K:4356] cp2k 2-4 installation and regresstio test result

Hi,

I am new to the CP2K and learning the program. I downloaded the program last week and tried to install it on our computer,
after several attempts with intel (version 12) and gfortran (4.6), I found that the regression test with Linux-x86-64-gfortran.popt
failed only 22 cases (using mpiexec -np 2 )

--------------------------------- summary --------------------------------
number of FAILED tests 22
number of WRONG   tests 2
number of CORRECT tests 2359
number of NEW     tests 0
number of         tests 2383
------------------------------

The failed cases are:

QS/regtest-hybrid/H2O-hybrid-bhandh.inp.out
QS/regtest-hybrid/H2O-hybrid-bhandhlyp.inp.out
QS/regtest-hybrid/CH3-hybrid-b3lyp-lsd-g03.inp.out
QS/regtest-hybrid/CH3-hybrid-handh-lsd.inp.out
QS/regtest-hybrid/CH3-hybrid-handhlyp-lsd.inp.out
QS/regtest-hybrid/CH3-hybrid-pbe0-lsd.inp.out
QS/regtest-hfx/CH-hfx-md.inp.out
QS/regtest-hfx/CH-hfx-md-2.inp.out
QS/regtest-hfx/4H2O-mix-disk-ram-on-the-fly.inp.out :
QS/regtest-mp2/H2O-01.inp.out
QS/regtest-admm/MD-2_OT.inp.out :
QS/regtest-hfx/H2O-hfx-emd.inp.out
QS/regtest-hfx/H2O-hfx-emd-restart.inp.out
QS/regtest-hfx/H2O-hfx-atprop.inp.out
QS/regtest-hfx-periodic/H2O-trunc-auto-md.inp.out
QS/regtest-mp2/Li.inp.out
QS/regtest-mp2/Ne2.inp.out
QS/regtest-hfx-periodic/CH3-trunc-auto-md.inp.out
QS/regtest-mp2/H2O_B2PLYP.inp.out
QS/regtest-mp2/H2O_B2GPPLYP.inp.out
QS/regtest-mp2/H2O_DSD-BLYP.inp.out
QS/regtest-mp2/H2O_B2PLYP_D3.inp.out
QS/regtest-mp2/CH3_B2PLYP_allelec.inp.out
QS/regtest-mp2/H2O_ADMM_MP2.inp.out

Here is the Linux-x86-64.gfortran.popt I used:

CC       = gcc
CPP      =

FC       = mpif90
LD       = mpif90

AR       = ar -r

CPPFLAGS =
DFLAGS   = -D__GFORTRAN -D__FFTSG -D__LIBINT -D__FFTW3 -D__parallel -D__SCALAPACK -D__BLACS -D__LIBXC2 \
           -D__HAS_NO_ISO_C_BINDING -D__HAS_smm_dnn
FCFLAGS = -O3 -march=native -ffast-math -funroll-loops -g -ffree-form $(DFLAGS) \
           -I/pkg/chem/lib/gcc/include \
           -I/pkg/chem/lib/gcc/libint/include \
           -I/pkg/chem/lib/gcc/fftw332/include
LDFLAGS = $(FCFLAGS) -L/pkg/chem/lib/gcc/libint/lib
LIBS     = /pkg/chem/lib/gcc/libint/lib/libint_cpp_wrapper.o -lderiv -lint -lstdc++ \
           /pkg/chem/lib/gcc/lib/libxc.a \
           /pkg/chem/lib/gcc/fftw332/lib/libfftw3.a \
           /pkg/chem/lib/gcc/lib/libsmm_dnn.a \
           /pkg/chem/lib/gcc/lapack/libscalapack.a\
           /pkg/chem/lib/gcc/lapack/liblapack.a \
           /pkg/chem/lib/gcc/lapack/librefblas.a

OBJECTS_ARCHITECTURE = machine_gfortran.o

Here scalapack 2.0 was compiled with acml 5.0.2 and, mpif90 is from mvapich2-1.8

I wonder if there can be any improvement in my compilation to pass all the regression tests ?
Thanks for any suggestion.

Jyh-Shyong Ho

--
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+uns...@googlegroups.com.
To post to this group, send email to cp...@googlegroups.com.
Visit this group at http://groups.google.com/group/cp2k?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.

You received this message because you are subscribed to the Google Group cp2k.
You can post via email.
To unsubscribe from this group, send an empty message.
For more options, visit this group.

--
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+uns...@googlegroups.com.
To post to this group, send email to cp...@googlegroups.com.
Visit this group at http://groups.google.com/group/cp2k?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.

Reply all

Reply to author

Forward

0 new messages