Make error

Zenith · Post by **Zenith** » Fri Mar 01, 2019 7:12 am

Hi everyone,
I'm installing Abinit 8.8.4 on a small cluster. I success in configuring it but not for the compilation. The last lines i had are:

Code: Select all

............................
checking for a thread-safe mkdir -p... /bin/mkdir -p
checking for gawk... gawk
checking whether make sets $(MAKE)... yes
checking whether make supports nested variables... yes
checking for a sed that does not truncate output... /bin/sed
checking for gcc... /usr/mpi/intel/openmpi-1.8.1-qlc//bin/mpicc
checking whether the C compiler works... yes
checking for C compiler default output file name... a.out
checking for suffix of executables...
checking whether we are cross compiling... configure: error: in `/users/zenith/Abinit/abinit-8.8.4/build/fallbacks/sources/bigdft-1.7.1.25':
configure: error: cannot run C compiled programs.
If you meant to cross compile, use `--host'.
See `config.log' for more details
make[2]: *** [stamps/bigdft-config-stamp] Error 1
make[2]: Leaving directory `/users/zenith/Abinit/abinit-8.8.4/build/fallbacks'
make[1]: [all-local] Error 2 (ignored)
Checking build of bigdft fallback
test -e stamps/bigdft-install-stamp
make[1]: *** [all-local] Error 1
make[1]: Leaving directory `/users/zenith/Abinit/abinit-8.8.4/build/fallbacks'
make: *** [multi] Error 2

where is the problem?

Best regards,

ebousquet · Post by **ebousquet** » Fri Mar 01, 2019 10:57 am

Dear Zenith ,
What did you set for your configure (in a configure file *.ac or directly in command line)?
Best wishes,
Eric

Zenith · Post by **Zenith** » Fri Mar 01, 2019 2:37 pm

Thank u for your response, dear ebousquet.
My .ac file is:

Code: Select all

#
prefix="/users/zenith/Abinit"

# MPI/OpenMP
with_mpi_prefix="/usr/mpi/intel/openmpi-1.8.1-qlc/"
enable_mpi="yes"
enable_mpi_io="yes"

#enable_openmp="yes"
#enable_64bit_flags="yes"

# BLAS/LAPACK 
with_linalg_flavor="mkl"
#with_linalg_libs="-L/apps/intel_2017/compilers_and_libraries_2017.0.098/linux/mkl/lib/intel6/ -lmkl_intel_lp64 -lmkl_core -lmkl_sequential -lpthread -lm -ldl"
with_fft_libs="-L/apps/intel_2017/compilers_and_libraries_2017.0.098/linux/mkl/lib/intel64 -Wl,--start-group  -lmkl_intel_lp64 -lmkl_sequential -lmkl_core -Wl,--end-group -lpthread -lm -ldl"

# FFTW3 
with_fft_flavor="fftw3"
with_fft_incs='-I/apps/intel_2017/compilers_and_libraries_2017.0.098/linux/mkl/include'
with_fft_libs="-L/apps/intel_2017/compilers_and_libraries_2017.0.098/linux/mkl/lib/intel64 -Wl,--start-group  -lmkl_intel_lp64 -lmkl_sequential -lmkl_core -Wl,--end-group -lpthread -lm -ldl"

with_dft_flavor="atompaw+libxc+wannier90+bigdft"
with_trio_flavor="netcdf"    
with_netcdf_incs="-I/users/zenith/local/intel17/include "
with_netcdf_libs="-L/users/zenith/local/intel17/lib -lnetcdf -lnetcdff -lhdf5_fortran -lhdf5_hl -lhdf5"
#
enable_gw_dpc="yes"

On posting my .ac file, i think i find my problem. it's on letting two different lines of

Code: Select all

with_dft_flavor=

.

But, please, I still have other questions here:
1- I don't understand why "LAPACK for Abinit" is downloaded from Fallbacks when i'm choosing to use MKL librairies from my machine ?!!
2- i'm confued about enabling or not of openmp and 64bit_flags. I have some E5-2670v2 Intel CPUs, in my cluster.
3- In my configuration, i used Intel compilers and libs +openmpi. In the same cluster we have as mpi: intel-mpi, openmpi, mvapich and mvapich2. Please, which one is the most suitable for my case ?

Excuse me for asking all these questions. I'm physicist and have just begin in this domain.
Best regards,

ebousquet · Post by **ebousquet** » Fri Mar 08, 2019 3:51 pm

Dear Zenith,

First of all, your compilation crashes for the bigdft, are you really going to use it? (idem for wannier90 and atompaw, don't put them if you are not going to use them).

For the mkl, in your .ac file, if I'm correct, you have commented the line where you specify the path for mkl (your have #with_linalg_libs=...) so I guess this is why it is trying to download it? Another small comment is that you define 2 times the path for fftlib (with_fft_libs=).

Otherwise you can enable the 64 bit flag and you can enable openmp, it just depends if you are going to thread or not, the CPU you mention has many threads:
https://ark.intel.com/content/www/us/en/ark/products/75275/intel-xeon-processor-e5-2670-v2-25m-cache-2-50-ghz.html

I know that open-mpi and intel-mpi are OK, you just have to compare the two on one test on your machine to see which one is the faster.

Best wishes,
Eric

Zenith · Post by **Zenith** » Tue Mar 12, 2019 10:10 am

Dear Eric,
Thank you for the response.
As you said the problem is in the the compilation of BigDFT, wich is ,unfortanately, what i think i will use from the FallBacks: It concerns the wavelets.
I recompile Abinit with the corrections you suggest on .ac; i had the same error.
From the beginning of the BigDFT compilation, i had:

Code: Select all

.
Checking build of etsf_io fallback
test -e stamps/etsf_io-install-stamp
make bigdft
make[3]: Entering directory `/users/zenith/abinit/abinit-8.8.4/build/fallbacks'
cd sources && \
          gzip -cd /users/zenith/.abinit/tarballs/bigdft-1.7.1.25.tar.gz | tar xf -
No patch to apply
cd sources/bigdft-1.7.1.25 && \
          CPP="/usr/mpi/intel/openmpi-1.8.1-qlc//bin/mpicc -E" \
          CPPFLAGS=" " \
          CC="/usr/mpi/intel/openmpi-1.8.1-qlc//bin/mpicc" \
          CFLAGS="-g -O2" \
          CXX="/usr/mpi/intel/openmpi-1.8.1-qlc//bin/mpicxx" \
          CXXFLAGS="" \
          FCCPP="cpp -P -std=c99" \
          F77="/usr/mpi/intel/openmpi-1.8.1-qlc//bin/mpif90" \
          FFLAGS="-fixed -g    -O3 -g -O2 " \
          F90="/usr/mpi/intel/openmpi-1.8.1-qlc//bin/mpif90" \
          F90FLAGS="-free -g    -O3 -g -O2 " \
          FC="/usr/mpi/intel/openmpi-1.8.1-qlc//bin/mpif90" \
          FCFLAGS="-free -g    -O3 -g -O2 " \
          LDFLAGS=" " \
          LIBS=" -L/apps/intel_2017/compilers_and_libraries_2017.0.098/linux/mkl/lib/intel64 -Wl,--start-group  -lmkl_intel_lp64 -lmkl_sequential -lmkl_core -Wl,--end-group -lpthread -lm -ldl -L/users/zenith/local/intel17/lib -lnetcdf -lnetcdff -lhdf5_fortran -lhdf5_hl -lhdf5  -L/users/zenith/abinit/abinit-8.8.4/build/fallbacks/exports/lib -lxcf90 -lxc" \
          AR="ar" \
          ARFLAGS="rc" \
          RANLIB="ranlib" \
          ac_fc_srcext="F90" \
          /bin/sh ./configure \
            --prefix="/users/zenith/abinit/abinit-8.8.4/build/fallbacks/exports" \
            --bindir="/users/zenith/abinit/abinit-8.8.4/build/fallbacks/exports/bin" \
            --libdir="/users/zenith/abinit/abinit-8.8.4/build/fallbacks/exports/lib" \
            --program-suffix="-abinit" \
            --disable-binaries --disable-bindings --enable-libbigdft --without-archives --with-moduledir="/users/zenith/abinit/abinit-8.8.4/build/fallbacks/exports/include" --enable-internal-libyaml --disable-shared --disable-internal-libxc --with-libxc-incs="-I/users/zenith/abinit/abinit-8.8.4/build/fallbacks/exports/include" --with-libxc-libs="-L/users/zenith/abinit/abinit-8.8.4/build/fallbacks/exports/lib -lxcf90 -lxc" --program-suffix='-abinit'
checking for a BSD-compatible install... /usr/bin/install -c
checking whether build environment is sane... yes
checking for a thread-safe mkdir -p... /bin/mkdir -p
checking for gawk... gawk
checking whether make sets $(MAKE)... yes
checking whether make supports nested variables... yes
checking for a sed that does not truncate output... /bin/sed
checking for gcc... /usr/mpi/intel/openmpi-1.8.1-qlc//bin/mpicc
checking whether the C compiler works... yes
checking for C compiler default output file name... a.out
checking for suffix of executables...
checking whether we are cross compiling... configure: error: in `/users/zenith/abinit/abinit-8.8.4/build/fallbacks/sources/bigdft-1.7.1.25':
configure: error: cannot run C compiled programs.
If you meant to cross compile, use `--host'.
See `config.log' for more details
make[3]: *** [stamps/bigdft-config-stamp] Error 1
make[3]: Leaving directory `/users/zenith/abinit/abinit-8.8.4/build/fallbacks'
make[2]: [all-local] Error 2 (ignored)
Checking build of bigdft fallback
test -e stamps/bigdft-install-stamp
make[2]: *** [all-local] Error 1
make[2]: Leaving directory `/users/zenith/abinit/abinit-8.8.4/build/fallbacks'
make[1]: *** [multi] Error 2
make[1]: Leaving directory `/users/zenith/abinit/abinit-8.8.4/build'
make: *** [mj4] Error 2
.

What does mean "No patch to apply"? If it's looking for patch sofware, it's installed:

Code: Select all

zenith@master build]$ rpm -q patch
patch-2.6-6.el6.x86_64

Best Regards,

ABINIT Discussion Forums

Make error

Make error

Re: Make error

Re: Make error

Re: Make error

Re: Make error