[QE-users] R: R: QE 7.2 compilation with new Intel compilers

Paolo Giannozzi paolo.giannozzi at uniud.it
Fri Aug 25 15:49:35 CEST 2023 

On 8/24/23 23:44, Kliavinek, Sergei wrote

> ./configure F90=ifx CC=mpiicx CFLAGS=-O3 FFLAGS=-O3 MPIF90=mpiifx  
> --with-hdf5="/opt/hdf5" --enable-wannier90

I don't think the "-enable-wannier90" option exists

> checking version of mpiifx... unknown, assuming gfortran

"configure" tries to figure out which serial compiler is wrapped the 
parallel compiler: see install/m4/x_ac_qe_mpif90.m4. It is needed 
because the typical "mpif90" compiler can wrap many different compilers. 
This mechanism is not implemented for the newest Intel compiler, though. 
It's not difficult, but it requires
1. having access to the newest Intel compiler, and
2. having an idea on how to modify "configure".
(I may fulfill requirement 2 but not 1)

Paolo


> setting F90... gfortran
> setting MPIF90... mpiifx
> 
> So it seems like the code cannot find the version of mpiifx/ifx and 
> can't use them even if I declare that I want to use MPIF90=mpiifx
> Do you have any idea why I obtain  this behavior?
> For the context, I build QE inside the Docker Image based on 
> the intel/oneapi-hpckit:2023.2.1-devel-ubuntu22.04:
> 
> # Use Intel OneAPI HPC Toolkit as base image
> FROM intel/oneapi-hpckit:2023.2.1-devel-ubuntu22.04
> 
> # Install of utilities and libraries
> RUN apt-get update && apt-get install -y \
>      build-essential \
>      wget \
>      unzip \
>      libfftw3-dev \
>      libblas-dev \
>      liblapack-dev \
>      vim
> 
> # Install of Git, OpenSSH, python, pip and make
> RUN apt-get update && apt-get install -y git openssh-client make 
>   python3-pip
> RUN apt-get update && apt-get install -y libtool-bin
> 
> 
> # Source the OneAPI setvars script to make ifort available
> RUN echo 'source /opt/intel/oneapi/setvars.sh' >> ~/.bashrc
> 
> # Install of Quantum Espresso
> WORKDIR /opt
> RUN wget https://github.com/QEF/q-e/archive/refs/tags/qe-7.2.zip && \
>      unzip qe-7.2.zip && \
>      cd q-e-qe-7.2 && \
>      ./configure F90=ifx CC=mpiicx CFLAGS=-O3 FFLAGS=-O3 MPIF90=mpiifx 
>   --with-hdf5="/opt/hdf5" --enable-wannier90 && \
>      make all && \
>      rm ../qe-7.2.zip
> 
> Also, I tried to fix it manually. What I've done:
> made new config:
> 
> ./configure F90=ifx CC=mpiicx CFLAGS=-O3 FFLAGS=-O3 MPIF90="$(which 
> mpiifort) -fc=$(which ifx)"  --with-hdf5="/opt/hdf5" --enable-wannier90
> 
> 
> And after it change everywhere manually the mpiifort to the mpiifx. And 
> it's work.
> 
> I've add the resulting make.inc file
> 
> 
> 
> Best,
> Sergei
> 
>> 24 авг. 2023 г., в 2:31 AM, Pietro Davide Delugas <pdelugas at sissa.it> 
>> написал(а):
>>
>> Dear Sergei
>>
>> The configure script determines the compiler by checking the output of 
>> the MPIF90 command, which must be set  as
>> MPIF90=mpiifx.
>> In this way, in my case, the configure script detects correctly that I 
>> want to use an intel compiler but sets it wrongly toifortinstead ofifx.
>> This is a minor issue that may need to be fixed, but if the future 
>> ifort and ifx are the same compiler, it may be unnecessary, and fixing 
>> it now can cause problems in the future.
>>
>> In your case, instead, MPIF90 was pointing to a  wrapper for 
>> thegfortrancompiler, so all the settings, including the preprocessing 
>> options passed to the compiler with the F90FLAGS variable, are those 
>> forgfortranand the bad preprocessing causes all those syntax errors in 
>> that file.
>> You may add -fpp -allow nofpp_comments to the F90FLAGS variable in 
>> make.inc
>>
>> Or, much better, you can set the environment variable MPIF90=mpiifx. 
>> Then, rerun the configure script and fix the F90 variable inside the 
>> make.inc
>> and everything should work.
>> Best regards
>> Pietro
>> ------------------------------------------------------------------------
>> *Da:*users <users-bounces at lists.quantum-espresso.org> per conto di 
>> Kliavinek, Sergei <skliavin at caltech.edu>
>> *Inviato:*mercoledì 23 agosto 2023 21:06
>> *A:*Quantum ESPRESSO users Forum <users at lists.quantum-espresso.org>
>> *Oggetto:*[QE-users] R: QE 7.2 compilation with new Intel compilers
>> Dear Pietro,
>> Thanks a lot for your suggestion!
>> Indeed, I checked make.inc, and even though I explicitly wrote the 
>> `F90=ifx` option during configuration, it was ignored, and make.inc 
>> ended up writing `F90=gfortran`.  (perhaps this configuration behavior 
>> is already worth a separate analysis in light of the fact that Intel 
>> will soon abandon old compilers).
>> The previous error was indeed resolved by this, but a new one has 
>> appeared:
>>
>>
>> gram.f90(60): error #5082: Syntax error, found IDENTIFIER 'CREATE' 
>> when expecting one of: ) ( , : <END-OF-STATEMENT> ; % [ . = / =>
>>  data copy(cp_bgrp, bec_bgrp) create(ctmp, cp_tmp,bec_tmp,csc,csc2) &
>> ------------------------------^
>> gram.f90(61): error #5082: Syntax error, found IDENTIFIER 'COPYIN' 
>> when expecting one of: :: ) ( , : * <END-OF-STATEMENT> ; . % (/ + - [ 
>> ] /) . = ' ** / ...
>>  & copyin(betae, qq_nt, ofsbeta,ityp, ibgrp_g2l, nh, tvanp )
>> ---^
>> gram.f90(61): error #5082: Syntax error, found END-OF-STATEMENT when 
>> expecting one of: :: ) , :
>>  & copyin(betae, qq_nt, ofsbeta,ityp, ibgrp_g2l, nh, tvanp )
>> -------------------------------------------------------------^
>> gram.f90(72): error #5082: Syntax error, found IDENTIFIER 'PRESENT' 
>> when expecting one of: ( : % [ . = =>
>>  kernels present(ctmp, cp_bgrp)
>> ---------^
>> gram.f90(74): error #5082: Syntax error, found IDENTIFIER 'KERNELS' 
>> when expecting one of: <END-OF-STATEMENT> ; TYPE MODULE FUNCTION 
>> SUBROUTINE PROCEDURE BLOCK DO FORALL ...
>>  end kernels
>> -----^
>> gram.f90(76): error #5082: Syntax error, found IDENTIFIER 'PRESENT' 
>> when expecting one of: ( : % [ . = =>
>>  kernels present(ctmp)
>> ---------^
>> gram.f90(78): error #5082: Syntax error, found IDENTIFIER 'KERNELS' 
>> when expecting one of: <END-OF-STATEMENT> ; TYPE MODULE FUNCTION 
>> SUBROUTINE PROCEDURE BLOCK DO FORALL ...
>>  end kernels
>> -----^
>> gram.f90(90): error #5082: Syntax error, found IDENTIFIER 'USE_DEVICE' 
>> when expecting one of: ( : % [ . = =>
>>  host_data  use_device(ctmp)
>> ------------^
>> gram.f90(92): error #5082: Syntax error, found IDENTIFIER 'HOST_DATA' 
>> when expecting one of: <END-OF-STATEMENT> ; TYPE MODULE FUNCTION 
>> SUBROUTINE PROCEDURE BLOCK DO FORALL ...
>>  end host_data
>> -----^
>> gram.f90(94): error #5082: Syntax error, found IDENTIFIER 'PRESENT' 
>> when expecting one of: ( : % [ . = =>
>>  kernels present(cp_bgrp, ctmp)
>> ---------^
>> gram.f90(96): error #5082: Syntax error, found IDENTIFIER 'KERNELS' 
>> when expecting one of: <END-OF-STATEMENT> ; TYPE MODULE FUNCTION 
>> SUBROUTINE PROCEDURE BLOCK DO FORALL ...
>>  end kernels
>> -----^
>> gram.f90(98): error #5082: Syntax error, found END-OF-STATEMENT when 
>> expecting one of: ( : % [ . = =>
>>  kernels
>> ---------^
>> gram.f90(101): error #5082: Syntax error, found IDENTIFIER 'KERNELS' 
>> when expecting one of: <END-OF-STATEMENT> ; TYPE MODULE FUNCTION 
>> SUBROUTINE PROCEDURE BLOCK DO FORALL ...
>>  end kernels
>> -----^
>> gram.f90(105): error #5082: Syntax error, found 'DATA' when expecting 
>> one of: <END-OF-STATEMENT> ; TYPE MODULE FUNCTION SUBROUTINE PROCEDURE 
>> BLOCK DO FORALL ...
>>  end data
>> -----^
>> gram.f90(143): error #5082: Syntax error, found END-OF-STATEMENT when 
>> expecting one of: ( , % . /
>>  data present(bec, cp, tvanp,ofsbeta, nh, ityp, qq_nt)
>> --------------------------------------------------------^
>> gram.f90(151): error #5082: Syntax error, found IDENTIFIER 'PRESENT' 
>> when expecting one of: ( : % [ . = =>
>>  kernels present(cp)
>> ---------^
>> gram.f90(153): error #5082: Syntax error, found IDENTIFIER 'KERNELS' 
>> when expecting one of: <END-OF-STATEMENT> ; TYPE MODULE FUNCTION 
>> SUBROUTINE PROCEDURE BLOCK DO FORALL ...
>>  end kernels
>> -----^
>> gram.f90(155): error #5082: Syntax error, found IDENTIFIER 'PRIVATE' 
>> when expecting one of: ( : % [ . = =>
>>  parallel private(ia, is, iv, jv, rsum_v) reduction (+:rsum) 
>> vector_length(32) &
>> ----------^
>> gram.f90(155): error #5082: Syntax error, found '+' when expecting one 
>> of: <IDENTIFIER>
>>  parallel private(ia, is, iv, jv, rsum_v) reduction (+:rsum) 
>> vector_length(32) &
>> -----------------------------------------------------^
>> gram.f90(157): error #5082: Syntax error, found IDENTIFIER 'GANG' when 
>> expecting one of: ( : % [ . = =>
>>  loop gang
>> ------^
>> gram.f90(164): error #5082: Syntax error, found IDENTIFIER 'VECTOR' 
>> when expecting one of: ( : % [ . = =>
>>  loop vector collapse(2) reduction(+:rsum_v)
>> ------^
>> gram.f90(164): error #5082: Syntax error, found IDENTIFIER 'REDUCTION' 
>> when expecting one of: :: ) ( , : * <END-OF-STATEMENT> ; . % (/ + - [ 
>> ] /) . = ' ** / ...
>>  loop vector collapse(2) reduction(+:rsum_v)
>> -------------------------^
>> gram.f90(176): error #5082: Syntax error, found IDENTIFIER 'PARALLEL' 
>> when expecting one of: <END-OF-STATEMENT> ; TYPE MODULE FUNCTION 
>> SUBROUTINE PROCEDURE BLOCK DO FORALL ...
>>  end parallel
>> -----^
>> gram.f90(177): error #5082: Syntax error, found 'DATA' when expecting 
>> one of: <END-OF-STATEMENT> ; TYPE MODULE FUNCTION SUBROUTINE PROCEDURE 
>> BLOCK DO FORALL ...
>>  end data
>> -----^
>> gram.f90(216): error #5082: Syntax error, found IDENTIFIER 'PRESENT' 
>> when expecting one of: ) ( , : <END-OF-STATEMENT> ; % [ . = / =>
>>  & present(ibgrp_g2l, nh, ofsbeta, qq_nt, ityp, tvanp, betae)
>> ---^
>> gram.f90(216): error #5082: Syntax error, found END-OF-STATEMENT when 
>> expecting one of: :
>>  & present(ibgrp_g2l, nh, ofsbeta, qq_nt, ityp, tvanp, betae)
>> ---------------------------------------------------------------^
>> gram.f90(217): error #5082: Syntax error, found IDENTIFIER 'PRESENT' 
>> when expecting one of: ( : % [ . = =>
>>  kernels present(csc)
>> ---------^
>> gram.f90(219): error #5082: Syntax error, found IDENTIFIER 'KERNELS' 
>> when expecting one of: <END-OF-STATEMENT> ; TYPE MODULE FUNCTION 
>> SUBROUTINE PROCEDURE BLOCK DO FORALL ...
>>  end kernels
>> -----^
>> gram.f90(221): error #5082: Syntax error, found IDENTIFIER 'PRESENT' 
>> when expecting one of: ( : % [ . = =>
>>  kernels present(cp_tmp)
>> ---------^
>> gram.f90(227): error #5082: Syntax error, found IDENTIFIER 'KERNELS' 
>> when expecting one of: <END-OF-STATEMENT> ; TYPE MODULE FUNCTION 
>> SUBROUTINE PROCEDURE BLOCK DO FORALL ...
>>  end kernels
>> -----^
>> /tmp/ifx0955286960Qlmjed/ifxwhM11h.i90(236): catastrophic error: Too 
>> many errors, exiting
>> compilation aborted for gram.f90 (code 1)
>> make[2]: *** [../../make.inc:16: gram.o] Error 1
>> make[2]: Leaving directory '/opt/q-e-qe-7.2/CPV/src'
>> make[1]: *** [Makefile:7: all] Error 1
>> make[1]: Leaving directory '/opt/q-e-qe-7.2/CPV'
>> make: *** [Makefile:75: cp] Error 1
>>
>>
>> As far as I understand, the compiler is signaling that a large number 
>> of syntax errors have been caused, which is strange because in that 
>> case the code would not compile with other compilers. Could you please 
>> advise what could be the issue?
>>
>> Best,
>> Sergei
>>
>>
>>> 23 авг. 2023 г., в 2:51 AM, Pietro Davide Delugas <pdelugas at sissa.it> 
>>> написал(а):
>>>
>>> Hello
>>>
>>> tried with Intel suite version 2023.2 this morning, and it works for me.
>>>
>>> There may be a slight problem in the configuration though. The 
>>> configure script recognizes that the  compiler is intel and sets in 
>>> the make.inc F90=ifort instead of F90=ifx, which may be the cause of 
>>> your problems. But  I set F90=ifx manually in the make.inc. before 
>>> compiling, and I haven't checked what happens otherwise.
>>> hope it helps
>>> Pietro
>>> ------------------------------------------------------------------------
>>> *Da:*users <users-bounces at lists.quantum-espresso.org> per conto di 
>>> Kliavinek, Sergei <skliavin at caltech.edu>
>>> *Inviato:*martedì 22 agosto 2023 21:34
>>> *A:*users at lists.quantum-espresso.org <users at lists.quantum-espresso.org>
>>> *Oggetto:*[QE-users] QE 7.2 compilation with new Intel compilers
>>> Dear QE developers,
>>>
>>> I am currently working on compiling QE using the new Intel compilers. 
>>> My compile command is as follows:
>>>
>>> wget https://github.com/QEF/q-e/archive/refs/tags/qe-7.2.zip && \
>>>     unzip qe-7.2.zip && \
>>>     cd q-e-qe-7.2 && \
>>>     ./configure F90=ifx CC=mpiicx CFLAGS=-O3 FFLAGS=-O3 
>>> MPIF90=mpiifx  --with-hdf5="/opt/hdf5" --enable-wannier90 && \
>>>     make all && \
>>>     rm ../qe-7.2.zip
>>>
>>> Unfortunately, I get an error when I run it:
>>>
>>> 1117.3 mbdlib.f90(26): error #7002: Error in opening the compiled 
>>> module file.  Check INCLUDE paths.   [MBD]
>>> 1117.3   USE mbd,              ONLY : mbd_input_t, mbd_calc_t
>>> 1117.3 ------^
>>> 1117.3 mbdlib.f90(40): error #6406: Conflicting attributes or 
>>> multiple declaration of name.   [MBD_INPUT_T]
>>> 1117.3   TYPE(mbd_input_t):: inp
>>> 1117.3 -------^
>>> 1117.3 mbdlib.f90(41): error #6406: Conflicting attributes or 
>>> multiple declaration of name.   [MBD_CALC_T]
>>> 1117.3   TYPE(mbd_calc_t):: calc
>>> 1117.3 -------^
>>> 1117.3 mbdlib.f90(26): error #6580: Name in only-list does not exist 
>>> or is not accessible.   [MBD_INPUT_T]
>>> 1117.3   USE mbd,              ONLY : mbd_input_t, mbd_calc_t
>>> 1117.3 -------------------------------^
>>> 1117.3 mbdlib.f90(26): error #6580: Name in only-list does not exist 
>>> or is not accessible.   [MBD_CALC_T]
>>> 1117.3   USE mbd,              ONLY : mbd_input_t, mbd_calc_t
>>> 1117.3 --------------------------------------------^
>>> 1117.3 mbdlib.f90(63): error #6404: This name does not have a type, 
>>> and must have an explicit type.   [INP]
>>> 1117.3   ALLOCATE(inp%atom_types(nat))
>>> 1117.3 -----------^
>>> 1117.3 mbdlib.f90(82): error #6303: The assignment operation or the 
>>> binary expression operation is invalid for the data types of the two 
>>> operands.   [TRIM]
>>> 1117.3     inp%atom_types(na) = trim(atm(ityp(na)))
>>> 1117.3 -------------------------^
>>> 1117.3 mbdlib.f90(84): error #6366: The shapes of the array 
>>> expressions do not conform.   [INP]
>>> 1117.3   inp%coords = tau*alat  ! HK-TODO: this one works for PW 
>>> (check if it is for CP)
>>> 1117.3 --^
>>> 1117.3 mbdlib.f90(90): error #6366: The shapes of the array 
>>> expressions do not conform.   [INP]
>>> 1117.3     inp%lattice_vectors = at*alat  ! Lattice vector in real space
>>> 1117.3 ----^
>>> 1117.3 mbdlib.f90(94): error #6366: The shapes of the array 
>>> expressions do not conform.   [INP]
>>> 1117.3       inp%k_grid = [nk1, nk2, nk3]
>>> 1117.3 ------^
>>> 1117.3 mbdlib.f90(101): error #6366: The shapes of the array 
>>> expressions do not conform.   [INP]
>>> 1117.3       inp%k_grid = [1, 1, 1] !set default k points grid
>>> 1117.3 ------^
>>> 1117.3 mbdlib.f90(122): error #6404: This name does not have a type, 
>>> and must have an explicit type.   [CALC]
>>> 1117.3   CALL calc%init(inp)
>>> 1117.3 -------^
>>> 1117.3 mbdlib.f90(184): error #6830: The argument of the ALLOCATED 
>>> intrinsic cannot be a constant or an expression.   [INP]
>>> 1117.3   IF(ALLOCATED(inp%atom_types)) DEALLOCATE(inp%atom_types)
>>> 1117.3 ---------------^
>>> 1117.3 compilation aborted for mbdlib.f90 (code 1)
>>> 1117.3 make[1]: *** [../make.inc:16: mbdlib.o] Error 1
>>> 1117.3 make[1]: Leaving directory '/opt/q-e-qe-7.2/Modules'
>>> 1117.3 make: *** [Makefile:194: mods] Error 1
>>> Could you please provide your thoughts, what could be the solution here?
>>> Also, perhaps, you have already compiled QEs with the new compilers 
>>> and know which options to use - if so, I'd appreciate that information!
>>>
>>> Best,
>>> Sergei
>>>
>>>
>>> _______________________________________________
>>> The Quantum ESPRESSO community stands by the Ukrainian
>>> people and expresses its concerns about the devastating
>>> effects that the Russian military offensive has on their
>>> country and on the free and peaceful scientific, cultural,
>>> and economic cooperation amongst peoples
>>> _______________________________________________
>>> Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
>>> users mailing list users at lists.quantum-espresso.org
>>> https://lists.quantum-espresso.org/mailman/listinfo/users
>>
>> _______________________________________________
>> The Quantum ESPRESSO community stands by the Ukrainian
>> people and expresses its concerns about the devastating
>> effects that the Russian military offensive has on their
>> country and on the free and peaceful scientific, cultural,
>> and economic cooperation amongst peoples
>> _______________________________________________
>> Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
>> users mailing list users at lists.quantum-espresso.org
>> https://lists.quantum-espresso.org/mailman/listinfo/users
> 
> 
> _______________________________________________
> The Quantum ESPRESSO community stands by the Ukrainian
> people and expresses its concerns about the devastating
> effects that the Russian military offensive has on their
> country and on the free and peaceful scientific, cultural,
> and economic cooperation amongst peoples
> _______________________________________________
> Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
> users mailing list users at lists.quantum-espresso.org
> https://lists.quantum-espresso.org/mailman/listinfo/users

-- 
Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
Univ. Udine, via delle Scienze 208, 33100 Udine Italy, +39-0432-558216

More information about the users mailing list