Compilation error in yambo-5.0.3

[Bramhachari Khamari]

Dear developers,
I have been trying to install latest version of yambo i.e. yambo-5.0.3. following error is encountered during the compilation.

Code: Select all

ifort: error #10236: File not found:  '/home/pyadav/master-yambo/yambo-5.0.3/lib/external/intel/mpiifort/lib/libscalapack.a'
make[2]: *** [yambo] Error 1
make[2]: Leaving directory `/home/pyadav/master-yambo/yambo-5.0.3/driver'
make[1]: *** [yambo] Error 2
make[1]: Leaving directory `/home/pyadav/master-yambo/yambo-5.0.3'
yambo build failed

The generated log file is also attached

Regards,

[Daniele Varsano]

Dear Bramchari,

it seems that the scalapack libs you indicated in your config command have some problems.

You can try to understand what is in conflict in your pre-installed scalapack (some undefined refs), or if you intend to use parallel linear algebra,
you can activate by --enable-par-linalg without specifying --with-scalapack-libs and --with-blacs-libs. In this way yambo will download and compile the needed libraries.

Best,
Daniele

[andrea.ferretti]

additionally do Daniele's suggestions (and after having fixed the flags passed to configure),
you may also try to recompile from scratch using a limited number of threads when issuing make, i.e.
make -j2 yambo or make -j4 yambo

Andrea

[Bramhachari Khamari]

Dear sir,
Thanks a lot to both sir for the quick reply. When I use the same config file and install using intel-15 version above mentioned issue does not occur. Only yambo exicutable created. A new problem appears in ypp build which is mentioned below

Code: Select all

compilation aborted for electrons_dos.f90 (code 1)
make[2]: *** [electrons_dos.o] Error 1
make[2]: *** Waiting for unfinished jobs....
electrons_magn_factors.F electrons_dos_elemental.F electrons_dos_setup.F make[2]: Leaving directory `/home/pyadav/master-yambo/yambo-5.0.3/ypp/electrons'
make[1]: *** [ypp] Error 2
make[1]: Leaving directory `/home/pyadav/master-yambo/yambo-5.0.3'
ypp build failed

Regards,

[Bramhachari Khamari]

Dear sir,
I followed the suggestions mentioned recompile the code from scratch. I got a new error in the ypp build which says

Code: Select all

compilation aborted for electrons_dos.f90 (code 1)
make[2]: *** [electrons_dos.o] Error 1
make[2]: *** Waiting for unfinished jobs....
electrons_dos_elemental.F make[2]: Leaving directory `/home/pyadav/master-yambo/yambo-5.0.3/ypp/electrons'
make[1]: *** [ypp] Error 2
make[1]: Leaving directory `/home/pyadav/master-yambo/yambo-5.0.3'
ypp build failed

Any suggestions to rectify this will be highly appreciated.

Regards,

[andrea.ferretti]

Dear Bramchari,

In order to help it would be useful to have some extra info such as the actual error message issued by the compiler.
If I had to guess, since the compilation with the intel compiler is daily checked, mostly likely the problem you observe
is related to the use of an old version of the compiler (or to a specific problem of the specific compiler).

best
Andrea

[Bramhachari Khamari]

Dear sir,
Thanks you for the reply. I am installing on hpc where intel-15 is there. I have installed intel-20 manually and using that for the compilation of yambo-5.0.3. Before configuring the intel path is exported as

Code: Select all

export mkldir=/home/pyadav/intel2020/compilers_and_libraries_2020/linux/mkl/lib/intel64

After configuring following options are shown

Code: Select all

[ E ] FFT       : -L/home/pyadav/Software/fftw-3.3.6-pl1//lib -lfftw3
#                   -I/home/pyadav/Software/fftw-3.3.6-pl1//include
# [ E ] BLAS      : -Wl,--start-group /home/pyadav/intel2020/compilers_and_libraries_2020/linux/mkl/lib/intel64/libmkl_intel_lp64.a /home/pyadav/intel2020/compilers_and_libraries_2020/linux/mkl/lib/intel64/libmkl_sequential.a /home/pyadav/intel2020/compilers_and_libraries_2020/linux/mkl/lib/intel64/libmkl_core.a -Wl,--end-group
# [ E ] LAPACK    : -Wl,--start-group /home/pyadav/intel2020/compilers_and_libraries_2020/linux/mkl/lib/intel64/libmkl_intel_lp64.a /home/pyadav/intel2020/compilers_and_libraries_2020/linux/mkl/lib/intel64/libmkl_sequential.a /home/pyadav/intel2020/compilers_and_libraries_2020/linux/mkl/lib/intel64/libmkl_core.a -Wl,--end-group
# [ Ic] SCALAPACK : /home/pyadav/master-yambo/yambo-5.0.3/lib/external/intel/mpiifort/lib/libscalapack.a 
# [ E ] BLACS     : /home/pyadav/intel2020/compilers_and_libraries_2020/linux/mkl/lib/intel64/libmkl_blacs_intelmpi_lp64.a
# [ - ] PETSC     :  
#                    
# [ - ] SLEPC     :  
#                    
#
# > OTHERs
#
# [ Ic] LibXC     : /home/pyadav/master-yambo/yambo-5.0.3/lib/external/intel/mpiifort/lib/libxcf90.a /home/pyadav/master-yambo/yambo-5.0.3/lib/external/intel/mpiifort/lib/libxc.a
#                   -I/home/pyadav/master-yambo/yambo-5.0.3/lib/external/intel/mpiifort/include
# [ E ] MPI       : -L/opt/software/intel_2015.u2//impi/5.0.3.048/intel64/lib -Xlinker --enable-new-dtags -Xlinker -rpath -Xlinker /opt/software/intel_2015.u2//impi/5.0.3.048/intel64/lib/release_mt -Xlinker -rpath -Xlinker /opt/software/intel_2015.u2//impi/5.0.3.048/intel64/lib -Xlinker -rpath -Xlinker /opt/intel/mpi-rt/5.0/intel64/lib/release_mt -Xlinker -rpath -Xlinker /opt/intel/mpi-rt/5.0/intel64/lib -lmpifort -lmpi -lmpigi -ldl -lrt -lpthread 
#                   -I/opt/software/intel_2015.u2//impi/5.0.3.048/intel64/include 
# [ Ic] Ydriver   : 1.0.0
#
# - COMPILERS -
#
# FC kind = intel ifort version 15.0.2
# MPI kind= Intel(R) MPI Library 5.0 Update 3 for Linux* OS
#
# [ CPP ] icc -E -ansi -D_HDF5_LIB -D_HDF5_IO -D_MPI -D_FFTW  -D_SCALAPACK  -D_DOUBLE   -D_TIMING     -D_P2Y_QEXSD_HDF5
# [ FPP ] fpp -free -P -D_HDF5_LIB -D_HDF5_IO -D_MPI -D_FFTW  -D_SCALAPACK  -D_DOUBLE   -D_TIMING    
# [ CC  ] mpiicc -O2 -D_C_US -D_FORTRAN_US
# [ FC  ] mpiifort -assume bscc -O3 -g -ip      
# [ FCUF] -assume bscc -O0 -g   
# [ F77 ] ifort -assume bscc -O3 -g -ip   
# [ F77U] -assume bscc -O0 -g   
# [Cmain] -nofor_main
#

What I noticed from the above is
In the MPI section intel15 is taking and FC kind is intel-ifort version 15. I think I am using both intel unknowingly. One version of intel should be choosen. Am I right?. Then my question is why it is showing FC kind=intel ifort version 15.0.2 even though I have exported the path as intel-20. Any suggestion is welcome.

Regards,

[Bramhachari Khamari]

Dear Developers,
Now I updated the ifort-1 to ifort version 19 and compilation works fine. However only the following executables are created in /bin.

Code: Select all

a2y  c2y  p2y  yambo  yambo_nl  yambo_ph  yambo_rt  yambo_sc  ypp  ypp_nl  ypp_ph  ypp_rt  ypp_sc

unlike yambo-master where 18 executables are present.

when I proceed to check a test calculation and run p2y command following error appears

Code: Select all

 <---> DBs path set to : .
 <---> detected QE data format : qexsd
 <---> == PWscf v.6.x generated data (QEXSD fmt) ==
 <---> Header/K-points/Energies... done
 <---> Cell data...
 <---> [WARNING]Two or more crystal structures fit the given cell done
 <---> Atomic data... done
 <---> Symmetries...[SI yes]...[I yes]...[-I yes]...[TR no]
 <---> XC functional...Perdew, Burke & Ernzerhof(X)+Perdew, Burke & Ernzerhof(C)
 <---> EXX fraction...:  0.000000
 <---> EXX screening...:  0.000000
 <---> K-points mesh... done
 <---> RL vectors... done
 <---> IGK arrays... done
 <---> Energies... done
 <--->  Electrons             :  16.00000
 <--->  Temperature           :  0.000000 [eV]
 <--->  Lattice factors       :   4.71774   4.08568  12.29080 [a.u.]
 <--->  K-points              :   50
 <--->  Bands                 :  100
 <--->  Spinor components     :  1
 <--->  Spin polarizations    :  1
 <--->  Spin orbit coupling    : no
 <--->  Symmetries   [spatial]:  12
 <--->                 [T-rev] : no
 <--->  Max WF components     :  1028
 <--->  RL vectors        (WF):  1453
 <--->  RL vectors    (CHARGE):   8101
 <--->  XC potential           : Perdew, Burke & Ernzerhof(X)+Perdew, Burke & Ernzerhof(C)
 <--->  Atomic species        :  2
 <--->  Max atoms/species     :  2
 <---> == DB1 (Gvecs and more) ...
[ERROR] Writing File ./SAVE//ns.db1; Variable  NOT DEFINED; Permission denied

I could not understand the issue which leads to the above error. Any help is welcome.

Regards,

[andrea.ferretti]

Dear Bramchari,

mixing different versions of the compilers is surely problematic and should be avoided
(the reason why you had both ifort 19/20 and ifort 15 is that while adding a new compiler to the path, you should also remove the others).

Now, I understand this issue was taken care and you were able to compile.

Regarding the number of executables, unless you really want to use advanced features (like yambo el-ph, real-time, or scf), it would be enough for you to
compile yambo, ypp and interfaces, which can be done by issuing

Code: Select all

make [-j4] core

(-j4 is ny own addition to get the compilation faster).

Last but not least, the error in writing DB1 sounds to me as connected to a possible miscomputation of netcdf.
Are you using internally compiled netcdf ?
Have you done a make distclean before recompiling ?

Perhaps other developers have more specific suggestions.

Andrea

[Bramhachari Khamari]

Dear sir,
Thank you. I have not used any extra tag during configure so it must have chosen the default setting. I need to install yambo-5.0.3 with both NetCDF and HDF support. Could you guide me which version of it(netcdf and hdf) I will install in my system and how to add this in configure (with netcdf and hdf) .


Regards,