This thread is locked.Only browsing is available.
Top Page > Browsing
installation problem
Date: 2014/07/31 05:18
Name: Khalid   <kimu206@gmail.com>

Dear all,
I am trying to install openmx 3.7 on my 32 bit linux laptop with intel compiler. Here is my CC, FC, and LIB:
CC = mpicc -openmp -O3 -I/opt/intel/composer_xe_2013.3.163/mkl/include/fftw -I/opt/intel/mkl/include -I/home/khalid/Software/ParallelQ/openmpi-1.8.1/ompi/include -I/opt/intel/composer_xe_2013.3.163/mkl/include
FC = mpif90 -opnemp -O3 -I/opt/intel/mkl/include -I/opt/intel/composer_xe_2013.3.163/mkl/include/fftw -I/home/khalid/Software/ParallelQ/openmpi-1.8.1/ompi/include
LIB = -L/home/khalid/ProgramFiles/FFTW/lib -lfftw3 -L/opt/intel/composer_xe_2013.3.163/mkl/lib/ia32 -lmkl_intel -lmkl_intel_thread -lmkl_core -lpthread -lgfortran

but this creates a problem saying:
icc: error #10236: File not found: 'elpa1.o'
icc: error #10236: File not found: 'solve_evp_real.o'
icc: error #10236: File not found: 'solve_evp_complex.o'
make: *** [openmx] Error 1

According to the suggestion given in http://www.openmx-square.org/forum/patio.cgi?mode=view&no=1505 I have changed FC to MPIFC:
MPIFC = mpif90 -opnemp -O3 -I/opt/intel/mkl/include -I/opt/intel/composer_xe_2013.3.163/mkl/include/fftw -I/home/khalid/Software/ParallelQ/openmpi-1.8.1/ompi/include
but this now complains about:

f77 -c elpa1.f90
make: f77: Command not found
make: *** [elpa1.o] Error 127

How can I solve this problem ???

Thank you
Khalid Ibne Masood
e
Page: [1]

Re: installation problem ( No.1 )
Date: 2014/07/31 17:38
Name: Artem  <artem.pulkin@epfl.ch>

Hi!

Initially, your object files 'elpa1.o', etc. have not been created. This means that the actual problem was before linking stage. Switch back to FC = mpif90, do "make clean" and "make openmx" and provide your full output here.

Artem
e
Re: installation problem ( No.2 )
Date: 2014/07/31 20:21
Name: Khalid  <kimu206@gmail.com>

Dear Sir,
Thank you for your response, but, the problem persists: here is the output:(The whole output can not be pasted here, it causes error when posting, I am justing posting the warning and error message)

liberi-091216/source/eri.c(133): warning #167: argument of type "double **" is incompatible with parameter of type "const double **"
ERI_LL_Overlap_d(solver, p, dp, alp1, dalp1, alp2, dalp2, cx);
^

liberi-091216/source/eri.c(133): warning #167: argument of type "double **" is incompatible with parameter of type "const double **"
ERI_LL_Overlap_d(solver, p, dp, alp1, dalp1, alp2, dalp2, cx);
^

liberi-091216/source/eri.c(228): warning #167: argument of type "double *" is incompatible with parameter of type "double (*)[3][2]"
ERI_Integral_GL_d(solver, I4, dI4, glF, glG, dglF, dglG,
^

liberi-091216/source/eri.c(228): warning #167: argument of type "double **" is incompatible with parameter of type "const double **"
ERI_Integral_GL_d(solver, I4, dI4, glF, glG, dglF, dglG,
^

liberi-091216/source/eri.c(228): warning #167: argument of type "double **" is incompatible with parameter of type "const double **"
ERI_Integral_GL_d(solver, I4, dI4, glF, glG, dglF, dglG,
^
mpicc -openmp -O3 -I/opt/intel/composer_xe_2013.3.163/mkl/include/fftw -I/opt/intel/mkl/include -I/home/khalid/Software/ParallelQ/openmpi-1.8.1/ompi/include -I/opt/intel/composer_xe_2013.3.163/mkl/include -I./liberi-091216/source openmx.o openmx_common.o Input_std.o Inputtools.o init.o LU_inverse.o ReLU_inverse.o truncation.o readfile.o FT_PAO.o FT_NLP.o FT_ProExpn_VNA.o FT_VNA.o FT_ProductPAO.o Hamiltonian_Cluster.o Overlap_Cluster.o Hamiltonian_Band.o Overlap_Band.o Hamiltonian_Cluster_NC.o Hamiltonian_Band_NC.o Hamiltonian_Cluster_SO.o Get_OneD_HS_Col.o SetPara_DFT.o XC_Ceperly_Alder.o XC_CA_LSDA.o XC_PW92C.o XC_PBE.o XC_EX.o DFT.o Mixing_DM.o Force.o Poisson.o Poisson_ESM.o Cluster_DFT.o Cluster_DFT_Dosout.o Cluster_DFT_ON2.o Band_DFT_Col.o Band_DFT_NonCol.o Band_DFT_kpath.o Band_DFT_MO.o Band_DFT_Dosout.o Set_Density_Grid.o Set_Orbitals_Grid.o Set_Aden_Grid.o Gauss_Legendre.o zero_cfrac.o xyz2spherical.o AngularF.o RadialF.o Dr_RadialF.o PhiF.o VNAF.o Dr_VNAF.o VH_AtomF.o Dr_VH_AtomF.o RF_BesselF.o QuickSort.o Nonlocal_RadialF.o KumoF.o Dr_KumoF.o Mulliken_Charge.o Occupation_Number_LDA_U.o Eff_Hub_Pot.o EulerAngle_Spin.o Smoothing_Func.o Orbital_Moment.o Pot_NeutralAtom.o Simple_Mixing_DM.o DIIS_Mixing_DM.o GR_Pulay_DM.o Kerker_Mixing_Rhok.o DIIS_Mixing_Rhok.o Total_Energy.o Contract_Hamiltonian.o Contract_iHNL.o Cont_Matrix0.o Cont_Matrix1.o Cont_Matrix2.o Cont_Matrix3.o Cont_Matrix4.o Opt_Contraction.o Initial_CntCoes.o Initial_CntCoes2.o Set_XC_Grid.o Get_Orbitals.o Get_dOrbitals.o Get_Cnt_Orbitals.o Get_Cnt_dOrbitals.o Gaunt.o Find_CGrids.o MD_pac.o RestartFileDFT.o Output_CompTime.o Merge_LogFile.o Make_FracCoord.o Make_InputFile_with_FinalCoord.o Divide_Conquer.o Krylov.o Divide_Conquer_Dosout.o Eigen_lapack.o Eigen_lapack2.o EigenBand_lapack.o Eigen_PReHH.o BroadCast_ReMatrix.o Eigen_PHH.o BroadCast_ComplexMatrix.o lapack_dstedc1.o lapack_dstedc2.o lapack_dstedc3.o lapack_dstegr1.o lapack_dstegr2.o lapack_dstegr3.o lapack_dstevx1.o lapack_dstevx2.o lapack_dstevx3.o lapack_dstevx4.o lapack_dstevx5.o lapack_dsteqr1.o Nonlocal_Basis.o Set_OLP_Kin.o Set_Nonlocal.o Set_ProExpn_VNA.o Set_Hamiltonian.o Set_Vpot.o Voronoi_Charge.o Voronoi_Orbital_Moment.o Fuzzy_Weight.o dampingF.o deri_dampingF.o Spherical_Bessel.o iterout.o iterout_md.o Allocate_Arrays.o Free_Arrays.o Init_List_YOUSO.o outputfile1.o malloc_multidimarray.o PrintMemory.o PrintMemory_Fix.o dtime.o OutData.o OutData_Binary.o init_alloc_first.o File_CntCoes.o SCF2File.o mimic_sse.o Make_Comm_Worlds.o Set_Allocate_Atom2CPU.o Cutoff.o Generating_MP_Special_Kpt.o Maketest.o Runtest.o Memory_Leak_test.o Force_test.o Show_DFT_DATA.o Generate_Wannier.o TRAN_Allocate.o TRAN_DFT.o TRAN_DFT_Dosout.o TRAN_Apply_Bias2e.o TRAN_Deallocate_Electrode_Grid.o TRAN_Deallocate_RestartFile.o TRAN_RestartFile.o TRAN_Calc_CentGreen.o TRAN_Input_std.o TRAN_Set_CentOverlap.o TRAN_Calc_CentGreenLesser.o TRAN_Input_std_Atoms.o TRAN_Set_Electrode_Grid.o TRAN_Calc_GridBound.o TRAN_Set_IntegPath.o TRAN_Output_HKS.o TRAN_Set_MP.o TRAN_Calc_SelfEnergy.o TRAN_Output_Trans_HS.o TRAN_Calc_Hopping_G.o TRAN_Calc_SurfGreen.o TRAN_Set_SurfOverlap.o TRAN_Add_Density_Lead.o TRAN_Add_ADensity_Lead.o TRAN_Set_Value.o TRAN_Poisson.o TRAN_adjust_Ngrid.o TRAN_Print.o TRAN_Print_Grid.o Lapack_LU_inverse.o TRAN_Distribute_Node.o TRAN_Output_HKS_Write_Grid.o TRAN_Credit.o TRAN_Check_Region_Lead.o TRAN_Check_Region.o TRAN_Check_Input.o DFTDvdW_init.o neb.o neb_run.o neb_check.o TRAN_Allocate_NC.o TRAN_DFT_NC.o TRAN_Set_CentOverlap_NC.o TRAN_Set_SurfOverlap_NC.o elpa1.o solve_evp_real.o solve_evp_complex.o exx.o exx_index.o exx_vector.o exx_log.o exx_step1.o exx_step2.o exx_file_overlap.o exx_file_eri.o exx_interface_openmx.o exx_debug.o exx_xc.o exx_rhox.o ./liberi-091216/source/eri.o ./liberi-091216/source/eri_ll.o ./liberi-091216/source/eri_sf.o ./liberi-091216/source/eri_interpolate.o ./liberi-091216/source/eri_gtbl.o ./liberi-091216/source/sbt/eri_sbt.o ./liberi-091216/source/sbt/log/eri_fsbt.o ./liberi-091216/source/sbt/log/eri_logfsbt.o ./liberi-091216/source/sbt/linear/eri_linfsbt.o -L/home/khalid/ProgramFiles/FFTW/lib -lfftw3 -L/opt/intel/composer_xe_2013.3.163/mkl/lib/ia32 -lmkl_intel -lmkl_intel_thread -lmkl_core -lpthread -lgfortran -lm -o openmx
icc: error #10236: File not found: 'elpa1.o'
icc: error #10236: File not found: 'solve_evp_real.o'
icc: error #10236: File not found: 'solve_evp_complex.o'
make: *** [openmx] Error 1



Thank you again
e
Re: installation problem ( No.3 )
Date: 2014/07/31 20:57
Name: Artem  <artem.pulkin@epfl.ch>

It would be more clear if you post everything related to compiling of elpa1.f90 and other Fortran sources.
e
Re: installation problem ( No.4 )
Date: 2014/08/01 10:07
Name: T.V.T. Duy  <duytvt@issp.u-tokyo.ac.jp>

Hi Khalid,

The cause is a typo with FC, -opnemp instead of the correct -openmp.

Please correct it and let us know the result.

Regards,
T.V.T. Duy


e
Re: installation problem ( No.5 )
Date: 2014/08/01 15:10
Name: Khalid  <kimu206@gmail.com>

Dear Sir,
thank you very much for your response and I earnestly appologize to disturb you again; I have corrected the spelling but now it complains about:
solve_evp_real.f90:(.text+0xc6): undefined reference to `mpi_comm_rank_'
solve_evp_real.f90:(.text+0xdf): undefined reference to `mpi_comm_rank_'
solve_evp_real.f90:(.text+0x140): undefined reference to `for_check_mult_overflow'
solve_evp_real.f90:(.text+0x177): undefined reference to `for_alloc_allocatable'
solve_evp_real.f90:(.text+0x1cb): undefined reference to `for_check_mult_overflow'
solve_evp_real.f90:(.text+0x1fa): undefined reference to `for_alloc_allocatable'
solve_evp_real.f90:(.text+0x1ff): undefined reference to `mpi_wtime_'
solve_evp_real.f90:(.text+0x250): undefined reference to `mpi_wtime_'
solve_evp_real.f90:(.text+0x2a9): undefined reference to `for_write_seq_lis'
solve_evp_real.f90:(.text+0x2d8): undefined reference to `for_write_seq_lis_xmit'
solve_evp_real.f90:(.text+0x2ff): undefined reference to `mpi_wtime_'
solve_evp_real.f90:(.text+0x334): undefined reference to `mpi_wtime_'
solve_evp_real.f90:(.text+0x38d): undefined reference to `for_write_seq_lis'
solve_evp_real.f90:(.text+0x3bc): undefined reference to `for_write_seq_lis_xmit'
solve_evp_real.f90:(.text+0x3e3): undefined reference to `mpi_wtime_'
solve_evp_real.f90:(.text+0x422): undefined reference to `mpi_wtime_'
solve_evp_real.f90:(.text+0x482): undefined reference to `for_write_seq_lis'
solve_evp_real.f90:(.text+0x4b6): undefined reference to `for_write_seq_lis_xmit'
solve_evp_real.f90:(.text+0x514): undefined reference to `for_dealloc_allocatable'
solve_evp_real.f90:(.text+0x556): undefined reference to `for_dealloc_allocatable'
solve_evp_complex.o: In function `solve_evp_complex_':
solve_evp_complex.f90:(.text+0x14b): undefined reference to `mpi_comm_rank_'
solve_evp_complex.f90:(.text+0x163): undefined reference to `mpi_comm_size_'
solve_evp_complex.f90:(.text+0x183): undefined reference to `mpi_comm_rank_'
solve_evp_complex.f90:(.text+0x19b): undefined reference to `mpi_comm_size_'
solve_evp_complex.f90:(.text+0x277): undefined reference to `for_check_mult_overflow'
solve_evp_complex.f90:(.text+0x2a9): undefined reference to `for_alloc_allocatable'
solve_evp_complex.f90:(.text+0x2fd): undefined reference to `for_check_mult_overflow'
solve_evp_complex.f90:(.text+0x32f): undefined reference to `for_alloc_allocatable'
solve_evp_complex.f90:(.text+0x3b6): undefined reference to `for_check_mult_overflow'
solve_evp_complex.f90:(.text+0x3e5): undefined reference to `for_alloc_allocatable'
solve_evp_complex.f90:(.text+0x3ea): undefined reference to `mpi_wtime_'
solve_evp_complex.f90:(.text+0x43f): undefined reference to `mpi_wtime_'
solve_evp_complex.f90:(.text+0x498): undefined reference to `for_write_seq_lis'
solve_evp_complex.f90:(.text+0x4c7): undefined reference to `for_write_seq_lis_xmit'
solve_evp_complex.f90:(.text+0x4ee): undefined reference to `mpi_wtime_'
solve_evp_complex.f90:(.text+0x52a): undefined reference to `mpi_wtime_'
solve_evp_complex.f90:(.text+0x583): undefined reference to `for_write_seq_lis'
solve_evp_complex.f90:(.text+0x5ae): undefined reference to `for_write_seq_lis_xmit'
solve_evp_complex.f90:(.text+0x5d5): undefined reference to `mpi_wtime_'
solve_evp_complex.f90:(.text+0x66d): undefined reference to `mpi_wtime_'
solve_evp_complex.f90:(.text+0x6cd): undefined reference to `for_write_seq_lis'
solve_evp_complex.f90:(.text+0x701): undefined reference to `for_write_seq_lis_xmit'
solve_evp_complex.f90:(.text+0x75c): undefined reference to `for_dealloc_allocatable'
solve_evp_complex.f90:(.text+0x7a1): undefined reference to `for_dealloc_allocatable'
solve_evp_complex.f90:(.text+0x7e6): undefined reference to `for_dealloc_allocatable'
make: *** [openmx] Error 1

Then according to "http://www.openmx-square.org/forum/patio.cgi?mode=view&no=1501" I have added the following in my LIB:
LIB = -L/home/khalid/ProgramFiles/FFTW/lib -lfftw3 -L/opt/intel/composer_xe_2013.3.163/mkl/lib/ia32 -lmkl_intel -lmkl_intel_thread -lmkl_core -lpthread -lgfortran -L/opt/intel/bin -L/opt/intel/lib/ia32 -lmpi_f77 -lmpi_f90 -lifcore



but it says:
ld: cannot find -lmpi_f77
ld: cannot find -lmpi_f90
make: *** [openmx] Error 1


Then I have deleted the -lmpi_f77 and -lmpi_f90, but the previous problem of


solve_evp_real.f90:(.text+0xc6): undefined reference to `mpi_comm_rank_'
solve_evp_real.f90:(.text+0xdf): undefined reference to `mpi_comm_rank_'

arises.

It seems there is nothing in my computer that can correspond to -lmpi_f77 and -lmpi_f90.

Thank you
e
Re: installation problem ( No.6 )
Date: 2014/08/03 17:07
Name: Khalid  <kimu206@gmail.com>

Dear all,

Thank you very much for your support and informations. I have solved the problem and I am posting it here so that it may help others like me, who are quite new with linux. I have installed openmx with intel compiler + openmpi 1.8.1. My main problem was that I could not use the flags -lmpi_f77 and -lmpi_f90. Therefore after a lot of googling, I found the command "mpicc -showme", and it resulted:

icc -I/home/khalid/Software/ParallelQ/ompi/include -pthread -Wl,-rpath -Wl,/home/khalid/Software/ParallelQ/ompi/lib -Wl,--enable-new-dtags -L/home/khalid/Software/ParallelQ/ompi/lib -lmpi

by luck I decided to use -lmpi instead of "-lmpi_f77 and -lmpi_f90." and yes, the problem was solved. Here is my LIB:

LIB = -L/home/khalid/ProgramFiles/FFTW/lib -lfftw3 -L/home/khalid/Software/ParallelQ/ompi/lib -lmpi_usempif08 -lmpi_usempi_ignore_tkr -lmpi_mpifh -lmpi -L/opt/intel/composer_xe_2013.3.163/mkl/lib/ia32 -lmkl_intel -lmkl_intel_thread -lmkl_core -lpthread -L/opt/intel/bin -L/opt/intel/lib/ia32 -lifcore

Frankly speaking, I really don't know what these flags mean actually, but thanks to ALLAH, it worked.

Thank you all
e
Re: installation problem ( No.8 )
Date: 2014/09/05 23:49
Name: Mustapha  <muslaty2k@gmail.com>

Dear All,

While installing openmx3.7 using make install in the &source, we experience
Krylov.c: In function eGenerate_pMatrixf:
Krylov.c:1747:3: error: unknown type name e__m128df
Krylov.c: In function eGenerate_pMatrix2f:
Krylov.c:3166:3: error: unknown type name e__m128df
Krylov.c: In function eEmbedding_Matrixf:
Krylov.c:3772:3: error: unknown type name e__m128df
Krylov.c: In function eS_orthonormalize_vecf:
Krylov.c:4739:3: error: unknown type name e__m128df
Krylov.c: In function eKrylov_Col_trdf:
Krylov.c:5869:5: error: unknown type name e__m128df
Krylov.c:5870:5: error: unknown type name e__m128df
Krylov.c:5871:5: error: unknown type name e__m128df
Krylov.c:5872:5: error: unknown type name e__m128df
Krylov.c: In function eGenerate_pMatrix_trdf:
Krylov.c:6841:3: error: unknown type name e__m128df
Krylov.c: In function eGenerate_pMatrix2_trdf:
Krylov.c:8319:3: error: unknown type name e__m128df
Krylov.c: In function eEmbedding_Matrix_trdf:
Krylov.c:8938:3: error: unknown type name e__m128df
Krylov.c: In function eS_orthonormalize_vec_trdf:
Krylov.c:9988:3: error: unknown type name e__m128df
make: *** [Krylov.o] Error 1
mustapha@mustapha-HP-Pavilion-dv6-Notebook-PC:~/openmx3.7/source$

e
Re: installation problem ( No.9 )
Date: 2014/09/10 11:14
Name: T.V.T. Duy  <duytvt@issp.u-tokyo.ac.jp>

Dear Mustapha,

The errors indicate that your system does not support SSE.
Please add -Dnosse to CC and try again.

Regards,
T.V.T. Duy
e

Page: [1]