ELPA libraries can't be recognized in ga_config...

[Akash Popat Gutal (P20CY201)]

Respected,

I am trying to install nwchem with ELPA libraries the env settings are as follows,

export NWCHEM_TOP='/home/mani/packages/venus-nwchem6.8/nw_elpa/nwchem-6.8'
export LARGE_FILES=TRUE
export ARMCI_NETWORK=MPI-PR
export NWCHEM_TARGET=LINUX64
export NWCHEM_MODULES="qm geninterface venus"
export USE_MPI=y
export USE_MPIF=y
export USE_MPIF4=y
export MPI_LOC=/opt/intel/oneapi/mpi/2021.4.0
export MPI_LIB=$MPI_LOC/lib
export MPI_INCLUDE=$MPI_LOC/include
export LIBMPI="-I/opt/intel/oneapi/mpi/2021.4.0//include -I/opt/intel/oneapi/mpi/2021.4.0/include -L/opt/intel/oneapi/mpi/2021.4.0/lib/release -L/opt/intel/oneapi/mpi/2021.4.0/lib -Xlinker --enable-new-dtags -Xlinker -rpath -Xlinker /opt/intel/oneapi/mpi/2021.4.0/lib/release -Xlinker -rpath -Xlinker /opt/intel/oneapi/mpi/2021.4.0/lib -lmpifort -lmpi -ldl -lrt -lpthread"
export CC=icc
export FC=ifort
export CXX=icpc
export HAS_BLAS=yes
export USE_SCALAPACK=y
export SCALAPACK_SIZE=8
export BLAS_SIZE=8
export BLASOPT="-L/home/mani/intel/oneapi/mkl/2021.4.0/lib/intel64 -lmkl_intel_ilp64 -lmkl_core -lmkl_sequential -lpthread -lm"
export LAPACK_LIB=" -L/home/mani/intel/oneapi/mkl/2021.4.0/lib/intel64 -lmkl_intel_ilp64 -lmkl_core -lmkl_sequential -lpthread -lm"
export SCALAPACK=" -L/home/mani/intel/oneapi/mkl/2021.4.0/lib/intel64 -lmkl_intel_ilp64 -lmkl_scalapack_ilp64  -lmkl_core -lmkl_sequential -lmkl_blacs_intelmpi_ilp64 -lpthread -lm"
export SCALAPACK_LIB="$SCALAPACK"
export ELPA_SIZE=8
export ELPA="-L/home/mani/packages/sources/elpa/intel-hsw/lib -lelpa"
cd $NWCHEM_TOP/src
make nwchem_config
make 2>&1 | tee make.log
make stubs.o

in ga configuration it is showing

configure: Checks for BLAS,LAPACK,ScaLAPACK
configure:
configure: Attempting to locate BLAS library
checking for BLAS with user-supplied flags... yes
configure: Attempting to locate LAPACK library
checking for Fortran 77 LAPACK with user-supplied flags... yes
configure: Attempting to locate SCALAPACK library
checking for SCALAPACK with user-supplied flags... yes
checking whether SCALAPACK implements pdsyevr... yes
configure: Attempting to locate ELPA library
checking for ELPA 2016 with user-supplied flags... no
checking for ELPA 2015 with user-supplied flags... no
checking for ELPA with user-supplied flags... no
checking for ELPA in generic library... no
configure: WARNING: ELPA library not found, interfaces won't be defined
configure: WARNING: ELPA 2015 library not found, interfaces won't be defined
configure: WARNING: ELPA 2016 library not found, interfaces won't be defined

I am using elpa-2021.05.001 version

please help......

[Edoardo Aprà]

I think that only the current master branch can compile the latest ELPA versions. Version 6.8 is too old.

[Akash Popat Gutal (P20CY201)]

Respected Edoardo,

Yes I agree with you....So I tried ELPA-2017.05.003. I checked config.log in /src/tools/build directory, but it seemed like the config was unable to find the ELPA include directory. I specified include directory and recompiled it, then I got following

configure: Checks for BLAS,LAPACK,ScaLAPACK
configure:
configure: Attempting to locate BLAS library
checking for BLAS with user-supplied flags... yes
configure: Attempting to locate LAPACK library
checking for Fortran 77 LAPACK with user-supplied flags... yes
configure: Attempting to locate SCALAPACK library
checking for SCALAPACK with user-supplied flags... yes
checking whether SCALAPACK implements pdsyevr... yes
configure: Attempting to locate ELPA library

checking for ELPA 2016 with user-supplied flags... yes

checking for ELPA 2015 with user-supplied flags... no

but after some time I am getting following error,

../ga-5.6.5/global/src/scalapack.F:3717:18:

            mpierr=get_elpa_communicators (elpa_comm,
                  1
Error: Type mismatch in argument ‘mpi_comm_global’ at (1); passed INTEGER(8) to INTEGER(4)
../ga-5.6.5/global/src/scalapack.F:3729:17:

          if(.not.solve_evp_real_2stage(
                 1
Error: Type mismatch in argument ‘na’ at (1); passed INTEGER(8) to INTEGER(4)
Makefile:7863: recipe for target 'global/src/scalapack.lo' failed
make[4]: *** [global/src/scalapack.lo] Error 1
Makefile:8272: recipe for target 'all-recursive' failed
make[3]: *** [all-recursive] Error 1
Makefile:5166: recipe for target 'all' failed
make[2]: *** [all] Error 2
GNUmakefile:651: recipe for target 'build/.libs/libga.a' failed
make[1]: *** [build/.libs/libga.a] Error 1
make[1]: Leaving directory '/home/mani/packages/venus-nwchem6.8/nw_elpa/nwchem-6.8/src/tools'
GNUmakefile:105: recipe for target 'libraries' failed
make: *** [libraries] Error 1
gfortran -m64 -ffast-math  -Warray-bounds -fdefault-integer-8 -fno-tree-dominator-opts  -finline-functions -O2 -g -fno-aggressive-loop-optimizations -fno-tree-dominator-opts  -g -O   -I.  -I/home/mani/packages/venus-nwchem6.8/nw_elpa/nwchem-6.8/src/include -I/home/mani/packages/venus-nwchem6.8/nw_elpa/nwchem-6.8/src/tools/install/include -DMPICH_NO_ATTR_TYPE_TAGS -DGFORTRAN -DCHKUNDFLW -DGCC4 -DGCC46 -DEXT_INT -DLINUX -DLINUX64 -DPARALLEL_DIAG -DCOMPILATION_DATE="'`date +%a_%b_%d_%H:%M:%S_%Y`'" -DCOMPILATION_DIR="'/home/mani/packages/venus-nwchem6.8/nw_elpa/nwchem-6.8'" -DNWCHEM_BRANCH="'Development'"  -c -o stubs.o stubs.F
f951: Warning: Nonexistent include directory ‘/home/mani/packages/venus-nwchem6.8/nw_elpa/nwchem-6.8/src/tools/install/include’ [-Wmissing-include-dirs]
stubs.F:16:2:

       Integer RTDB
  1    ~~~~~~~
Fatal Error: errquit.fh: No such file or directory
compilation terminated.
<builtin>: recipe for target 'stubs.o' failed
make: *** [stubs.o] Error 1

Please help.....!

--
You received this message because you are subscribed to the Google Groups "NWChem Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nwchem-forum...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nwchem-forum/2d218406-0013-489e-b3fc-559df8a471f0n%40googlegroups.com.

[Akash Popat Gutal (P20CY201)]

Please help

To unsubscribe from this group and stop receiving emails from it, send an email to nwchem-forum+unsubscribe@googlegroups.com.

[Edoardo Aprà]

You might want to move the current master branch (instead of using a four year old release). I am not sure who can help you with the source code of the 6.8 version.

To unsubscribe from this group and stop receiving emails from it, send an email to nwchem-forum...@googlegroups.com.

[Akash Popat Gutal (P20CY201)]

 Thank you....I will try.

To unsubscribe from this group and stop receiving emails from it, send an email to nwchem-forum+unsubscribe@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/nwchem-forum/2d218406-0013-489e-b3fc-559df8a471f0n%40googlegroups.com.

--

You received this message because you are subscribed to the Google Groups "NWChem Forum" group.

To unsubscribe from this group and stop receiving emails from it, send an email to nwchem-forum+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nwchem-forum/80d8d2a3-9878-4dcc-bf8d-7ba07d3d59e7n%40googlegroups.com.

[Akash Popat Gutal (P20CY201)]

I tried to compile with master branch I am getting following error

libtool: compile:  gfortran -DHAVE_CONFIG_H -I. -I../ga-5.8.1 -I/home/mani/packages/sources/elpa/gnu-hsw/include/elpa/modules -I/home/mani/packages/sources/elpa/gnu-hsw/include -I/home/mani/packages/MPI/openmpi-3.0.6/include -I/home/mani/packages/MPI/openmpi-3.0.6/include -I/home/mani/packages/MPI/openmpi-3.0.6/lib -I/home/mani/packages/MPI/openmpi-3.0.6/include -Ima -I../ga-5.8.1/ma -I../ga-5.8.1/LinAlg/lapack+blas -Iglobal/src -I../ga-5.8.1/global/src -I../ga-5.8.1/global/testing -I../ga-5.8.1/pario/dra -I../ga-5.8.1/pario/eaf -I../ga-5.8.1/pario/elio -I../ga-5.8.1/pario/sf -Igaf2c -I../ga-5.8.1/gaf2c -I../ga-5.8.1/tcgmsg -I../ga-5.8.1/tcgmsg/tcgmsg-mpi -I../ga-5.8.1/comex/src-armci -fdefault-integer-8 -fno-aggressive-loop-optimizations -fno-aggressive-loop-optimizations -c ../ga-5.8.1/global/src/scalapack.F -o global/src/scalapack.o
../ga-5.8.1/global/src/scalapack.F:3845:11:

       e => elpa_allocate(elpa_err)
           1
Error: More actual than formal arguments in procedure call at (1)
../ga-5.8.1/global/src/scalapack.F:3982:72: Error: More actual than formal arguments in procedure call at (1)
Makefile:6851: recipe for target 'global/src/scalapack.lo' failed

make[4]: *** [global/src/scalapack.lo] Error 1

Makefile:7344: recipe for target 'all-recursive' failed

make[3]: *** [all-recursive] Error 1

Makefile:4066: recipe for target 'all' failed

make[2]: *** [all] Error 2

GNUmakefile:691: recipe for target 'build/.libs/libga.a' failed

make[1]: *** [build/.libs/libga.a] Error 1

make[1]: Leaving directory '/home/mani/PO/Distros/nw-g-mpich-mkl/nw6.8/src/tools'

GNUmakefile:105: recipe for target 'libraries' failed
make: *** [libraries] Error 1

gfortran -m64 -ffast-math  -Warray-bounds -fdefault-integer-8 -g -fno-aggressive-loop-optimizations -g -O0 -fno-aggressive-loop-optimizations    -I.  -I/home/mani/PO/Distros/nw-g-mpich-mkl/nw6.8/src/include -I/home/mani/PO/Distros/nw-g-mpich-mkl/nw6.8/src/tools/install/include -DGFORTRAN -DCHKUNDFLW -DGCC4 -DGCC46 -DEXT_INT -DLINUX -DLINUX64 -DPARALLEL_DIAG -DTCE_CUDA -DBLAS_NOTHREADS -DMKL -DCOMPILATION_DATE="'`date +%a_%b_%d_%H:%M:%S_%Y`'" -DCOMPILATION_DIR="'/home/mani/PO/Distros/nw-g-mpich-mkl/nw6.8'" -DNWCHEM_BRANCH="'7.0.0'"  -c -o stubs.o stubs.F
f951: Warning: Nonexistent include directory ‘/home/mani/PO/Distros/nw-g-mpich-mkl/nw6.8/src/tools/install/include’ [-Wmissing-include-dirs]
stubs.F:15:0:

 #include "errquit.fh"
 ^

Fatal Error: errquit.fh: No such file or directory
compilation terminated.
<builtin>: recipe for target 'stubs.o' failed
make: *** [stubs.o] Error 1

please help....

To unsubscribe from this group and stop receiving emails from it, send an email to nwchem-forum...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/nwchem-forum/2d218406-0013-489e-b3fc-559df8a471f0n%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "NWChem Forum" group.

To unsubscribe from this group and stop receiving emails from it, send an email to nwchem-forum...@googlegroups.com.

[Edoardo Aprà]

Could you post the following files?

$NWCHEM_TOP/src/tools/build/config.log

$NWCHEM_TOP/src/tools/build/comex/config.log

[Akash Popat Gutal (P20CY201)]

Please find attachments.

--
You received this message because you are subscribed to the Google Groups "NWChem Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nwchem-forum...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/nwchem-forum/6b05d717-23cb-4452-a3bc-53d7f0d5326fn%40googlegroups.com.

[Edoardo Aprà]

What version of ELPA are you using? The autoconf program is recognizing both  ELPA 2017 and  ELPA 2016 interfaces ... Do you have multiple ELPA installation in the same directory?

[Akash Popat Gutal (P20CY201)]

yes.

On Tuesday, January 18, 2022, Edoardo Aprà <edoard...@gmail.com> wrote:

What version of ELPA are you using? The autoconf program is recognizing both  ELPA 2017 and  ELPA 2016 interfaces ... Do you have multiple ELPA installation in the same directory?

--

You received this message because you are subscribed to the Google Groups "NWChem Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nwchem-forum...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/nwchem-forum/4d3b8caa-f1bb-4ebb-8311-7d04347e157dn%40googlegroups.com.

[Edoardo Aprà]

Having multiple ELPA version in the same directory is never going to lead to a successful build.

By the way, do you have a good reason for struggling so hard to use ELPA? What's wrong with Scalapack?

[Akash Popat Gutal (P20CY201)]

Ok I will try in different directories... I will check for Scalapack also...  And the reason that I am interested in this is because our institute got a new HPC and a couple of NVIDIA-DGX machines. I am very excited to use them with full capacity with nwchem... I am trying nwchem with different compilers/MPI/Math-libraries/ that are available for getting good speed for our VENUS-NWCHEM dynamics calculations. I want to know the performance of nwchem on the new HPC with its fullest potential with perfect blas libraries-compiler-MPI combination with additional libraries like ELPA (has GPU support) so that I can increase the possibility of doing things in optimum way.......

I have already optimized the compilation of 24 combinations of NWCHEM and have tested its speed on dynamic code. I have tried different ARMCI networks also....for one test dynamics calculation I have reduced time from 256s to 110s( on the same machine)..!!!! by installing the right combination of above accessories. Even reduction in a single second is important for us because it will reduce our trajectory time by days. 

and for HPC we can run larger molecule direct dynamics also....will ELPA be useful ? What do you suggest?

To view this discussion on the web visit https://groups.google.com/d/msgid/nwchem-forum/b41c5068-e802-4fa2-97b1-5de103c97d75n%40googlegroups.com.

[Edoardo Aprà]

Unfortunately, the NWChem eigensolvers are not using GPUs at this time. Therefore, you will not be able to exploit the ELPA GPU features.