cp2k 2-4 installation and regresstio test result

[jsho]

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

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20130322/84fa52fa/attachment.htm>