ERROR in compilation of input_kernel.f90

Submitted by koudai on Thu, 08/18/2016 - 11:49

Forums 

Installation

Dear all,

I would like to install BGW-1.2-RC with intel MPI 5.1.3 and MKL in CentOS 7.2.

During 'make all-flavors', I met an error below:

cpp -ansi -I../Common -DINTEL -DMPI -DUSESCALAPACK -DDEBUG -DVERBOSE input_kernel.f90 > input_kernel.p.f
mpiifort -free -I ../Common -I /opt/intel/compilers_and_libraries_2016.3.210/linux/mkl/include/fftw -c -O3 input_kernel.p.f -o input_kernel.o -module./
ifort: command line warning #10159: invalid argument for option '-m'
(1): warning #5117: Bad # preprocessor line
# 1 "/usr/include/stdc-predef.h" 1 3 4
-----------------------------------^
input_kernel.f90(343): error #6404: This name does not have a type, and must have an explicit type. [KPQ]
write(6,*) 'kp%ngkmax,kpq%ngkmax',kp%ngkmax,kpq%ngkmax
------------------------------------------------^
input_kernel.f90(343): error #6460: This is not a field name that is defined in the encompassing structure. [NGKMAX]
write(6,*) 'kp%ngkmax,kpq%ngkmax',kp%ngkmax,kpq%ngkmax
----------------------------------------------------^
compilation aborted for input_kernel.p.f (code 1)
make[3]: *** [input_kernel.o] Error 1
make[3]: Leaving directory `/home/user/BGW-1.2-beta/BSE'
make[2]: *** [bse-all] Error 2
make[2]: Leaving directory `/home/user/BGW-1.2-beta'
make[1]: *** [all] Error 2
make[1]: Leaving directory `/home/user/BGW-1.2-beta'
make: *** [all-flavors] Error 2

My arch.mk is

COMPFLAG = -DINTEL
PARAFLAG = -DMPI
MATHFLAG = -DUSESCALAPACK
DEBUGFLAG = -DDEBUG -DVERBOSE

FCPP = cpp -ansi
F90free = mpiifort -free
LINK = mpiifort -mkl=cluster
FOPTS = -O3
FNOOPTS = $(FOPTS)
MOD_OPT = -module
INCFLAG = -I

C_PARAFLAG = -DPARA -DMPICH_IGNORE_CXX_SEEK
CC_COMP = mpiicpc
C_COMP = mpiicc
C_LINK = mpiicpc
C_OPTS = -O3

REMOVE = /bin/rm -f

# Math Libraries
MKLPATH=/opt/intel/compilers_and_libraries_2016.3.210/linux/mkl
FFTWLIB = -lfftw
FFTWINCLUDE = $(MKLPATH)/include/fftw
LAPACKLIB = -Wl,--start-group \
$(MKLPATH)/lib/intel64_lin/libmkl_gf_lp64.a \
$(MKLPATH)/lib/intel64_lin/libmkl_sequential.a \
$(MKLPATH)/lib/intel64_lin/libmkl_core.a \
$(MKLPATH)/lib/intel64_lin/libmkl_blacs_intelmpi_lp64.a \
-Wl,--end-group -lpthread
SCALAPACKLIB = $(MKLPATH)/lib/intel64_lin/libmkl_scalapack_lp64.a

TESTSCRIPT = sbatch stampede.scr

What shold I do to avoid this error?
Thank you in advance.

Koudai

jornada's picture

Submitted by jornada on Wed, 08/31/2016 - 18:20

Hi Koudai,

please, download the new BerkeleyGW-1.2 and use the new arch.mk for stampede. We updated the configuration scripts to deal with newer versions of cpp.

Cheers,
Felipe

Submitted by koudai on Thu, 09/01/2016 - 05:59

Dear Felipe,

Thank you for your replying.
I downloaded the new BGW 1.2 and tried to run 'make all-flavors' with the above arch.mk, but unfortunately I met the same errors.
Is my arch.mk correct?

Best,
Koudai

jornada's picture

Submitted by jornada on Thu, 09/01/2016 - 08:48

HI Koudai,

there are two things you can try in your arch.mk file:
1) Switch from "cpp -ansi" to "cpp -C"
2) Use "cpp -ansi -nostdinc -nostdinc++"

Let me know if any of them work. Can you also type "cpp --version" in your terminal and send me the output?

Cheers,
Felipe

Submitted by koudai on Fri, 09/02/2016 - 02:42

Dear Felipe,

I added '-DUSEFFTW3' in MATHFLAG, and then the compilation successfully finished.
Thank you very much for your help.

Best,
Koudai