compilation error on HPC

Having trouble compiling the Yambo source? Using an unusual architecture? Problems with the "configure" script? Problems in GPU architectures? This is the place to look.

Moderators: Davide Sangalli, andrea.ferretti, myrta gruning, andrea marini, Daniele Varsano, Conor Hogan, Nicola Spallanzani

Forum rules
If you have trouble compiling Yambo, please make sure to list:
(1) the compiler (vendor and release: e.g. intel 10.1)
(2) the architecture (e.g. 64-bit IBM SP5)
(3) if the problems occur compiling in serial/in parallel
(4) the version of Yambo (revision number/major release version)
(5) the relevant compiler error message
vitoversace
Posts: 58
Joined: Thu Mar 13, 2014 3:43 pm

compilation error on HPC

Post by vitoversace » Wed Jun 18, 2014 8:35 pm

Dear all,

Following problem :

when using mpiifort for compilation of yambo on HPC i get this error :
> configure: error: Found precompiler problems in processing the F90/C
> source.

I tried different compilers and preprocessors. As soon as I set FC=ifort or FC=mpiifort this problem arises, since yambo is using "-tpp7" as a default flag when it detects
an Intel Compiler, I guess the whole configuration script is not updated to Intel Compiler 2013 yet (tpp7 no longer exists as flag).

Any suggestion ?
(compilation in parallel, all intel 13.X, 3.4.1 yambo version, 64 Bit arch.)


Bests
Vito
Vito Versace
PhD student
University of York
vitoversace@yahoo.com

User avatar
Daniele Varsano
Posts: 3816
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: compilation error on HPC

Post by Daniele Varsano » Wed Jun 18, 2014 9:53 pm

Dear Vito,

you can try to overwrite the default flag.
Before your ./configure command you can put for instance:

Code: Select all

export MKL_LIB=$MKLROOT/lib/intel64
export CC=icc
export CXX=icpc
export F77=ifort
export F90=ifort
export PFC=mpiifort
export CFLAGS=-O3
export FCFLAGS=" -O2 -assume buffered_io,byterecl"
export FFLAGS=" -O2 -assume buffered_io,byterecl"
export FFLAGS_NOOPT=" -O0 -assume buffered_io,byterecl"
export UFFLAGS='-assume bscc -O0'
This script works to me.
Hope it helps,

Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

vitoversace
Posts: 58
Joined: Thu Mar 13, 2014 3:43 pm

Re: compilation error on HPC

Post by vitoversace » Thu Jun 19, 2014 8:31 am

Dear Daniele,

The script did not help.
I used this one :

./configure --with-iotk='/home/vito/espresso-5.0.2/iotk' --with-p2y=5.0 PFC=mpiifort CC=gcc FC=ifort F77=ifort
and got this error :
configure: testing preprocessor gcc -E -P
checking if precompiler works on C source... yes
checking if precompiler works on F90 source... no
configure: error: Found precompiler problems in processing the F90/C source.
It seems yambo does not like FC=ifort. When i discard FC=ifort and instead use gfortran ( that i do not want to use) , it configures but at compilation i get several other errors like :

>>>[Making modules]<<<
make[1]: Entering directory `/home/vito/megware_yambo/stable/src/modules'
mod_pars.F mod_units.F mod_stderr.F mod_parallel.F mod_wrapper.F mod_fields.F mod_collision.F mod_drivers.F mod_FFT.F mod_timing.F mod_logo mod_com.F mod_memory.F mod_R_lattice.F mod_matrix_operate.F mod_D_lattice.F mod_frequency.F mod_vec_operate.F mod_electrons.F mod_wave_func.F mod_interfaces.F mod_interpolate.F mod_X.F mod_functions.F mod_zeros.F mod_pseudo.F mod_BS.F mod_QP.F mod_TDDFT.F mod_ACFDT.F mod_wf_distribute.F mod_IO.F mod_fragments.F mod_xc_functionals.F(27): error #7013: This module file was not generated by any release of this compiler. [XC_F90_LIB_M]
use xc_f90_lib_m
-----^
mod_xc_functionals.F(34): error #6683: A kind type parameter must be a compile-time constant. [SP]
real(SP), public, allocatable:: magn(:,:)
------^
mod_xc_functionals.F(35): error #6683: A kind type parameter must be a compile-time constant. [SP]
real(SP), public, allocatable:: V_xc(:,:)
------^
mod_xc_functionals.F(36): error #6683: A kind type parameter must be a compile-time constant. [SP]
real(SP), public, allocatable:: E_xc(:)
------^
mod_xc_functionals.F(37): error #6683: A kind type parameter must be a compile-time constant. [SP]
real(SP), public, allocatable:: F_xc(:)
------^
mod_xc_functionals.F(45): error #8237: The character length in a component declaration shall either be a colon, be an initialization expression, or be a specification expression. [NAME]
character(lchlen) :: name
------------------------^
mod_xc_functionals.F(46): error #6457: This derived type name has not been declared. [XC_F90_POINTER_T]
type(xc_f90_pointer_t) :: conf
--------^
mod_xc_functionals.F(47): error #6457: This derived type name has not been declared. [XC_F90_POINTER_T]
type(xc_f90_pointer_t) :: info
--------^
mod_xc_functionals.F(80): error #7013: This module file was not generated by any release of this compiler. [XC_F90_TYPES_M]
use xc_f90_types_m
---------^
mod_xc_functionals.F(87): error #6457: This derived type name has not been declared. [XC_F90_POINTER_T]
type(xc_f90_pointer_t) :: p,info
----------^
mod_xc_functionals.F(26): error #6580: Name in only-list does not exist. [SP]
use pars, ONLY:SP,lchlen
---------------------^
mod_xc_functionals.F(45): error #6404: This name does not have a type, and must have an explicit type. [LCHLEN]
character(lchlen) :: name

Bests
Vito
Vito Versace
PhD student
University of York
vitoversace@yahoo.com

User avatar
Daniele Varsano
Posts: 3816
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: compilation error on HPC

Post by Daniele Varsano » Thu Jun 19, 2014 10:01 am

Dear Vito,
the idea was to overwrite the default option for the FLAGS.
Did you check in the config.log if the flags were correctly set up?

The error you have with gfortran is probably related to the configuration of the libxc, did you type make clean_all before reconfiguring and recompiling?
In any case, if you posts the config.log other than the error it helps.

Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

vitoversace
Posts: 58
Joined: Thu Mar 13, 2014 3:43 pm

Re: compilation error on HPC

Post by vitoversace » Thu Jun 19, 2014 2:01 pm

Dear Daniele,

Attached is the config.log

Ok, i managed to pass configure and compilation but now when running ./yambo i get this error :
[0] MPI startup(): dapl fabric is not available and fallback fabric is not enabled
By the way during compilation in the last two lines i get this, too
>>>[Linking yambo]<<<
make[1]: Entering directory `/home/vito/megware_yambo/stable/driver'
yambo_driver.F
ifort: command line warning #10155: ignoring option '-l'; argument required
ifort: command line warning #10155: ignoring option '-l'; argument required
Bests
Vito
You do not have the required permissions to view the files attached to this post.
Vito Versace
PhD student
University of York
vitoversace@yahoo.com

User avatar
Daniele Varsano
Posts: 3816
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: compilation error on HPC

Post by Daniele Varsano » Thu Jun 19, 2014 2:15 pm

Dear VIto,
at first glance it does not look a problem of yambo, but of intel MPI and related library, just a guess.
You can try to look at this post here:
https://software.intel.com/en-us/forums/topic/290764

just to look if you are in the same situation.
Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

vitoversace
Posts: 58
Joined: Thu Mar 13, 2014 3:43 pm

Re: compilation error on HPC

Post by vitoversace » Thu Jun 19, 2014 2:22 pm

Dear Daniele,

I red this post before, but i don't really understand it.
Ok, if it is not a yambo related issue, i have to talk with HPC people to see what is going wrong.

Just to be sure, was every thing okay with config.log at first glance ?

Bests
Vito
Vito Versace
PhD student
University of York
vitoversace@yahoo.com

User avatar
Daniele Varsano
Posts: 3816
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: compilation error on HPC

Post by Daniele Varsano » Thu Jun 19, 2014 2:33 pm

Dear Vito,
usually when the configure arrives at the end it means that found all the required issue to compile.
The post just said to look at the version of some libraries that are not supported anymore with the new version of MPI, and in that case to upgrade them:
" Probably your DAPL library is quite old. Intel MPI Library 4.0 doesn't support DAPL 1.1 any more. In this case you need to install new library.
Could you provide your /etc/dat.conf file?"

Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

vitoversace
Posts: 58
Joined: Thu Mar 13, 2014 3:43 pm

Re: compilation error on HPC

Post by vitoversace » Thu Jun 19, 2014 2:41 pm

Dear Daniele,

Unfortunately i don't have this file !!! Strange !
So, in this file one should see DAPL version ?

Bests
Vito
Vito Versace
PhD student
University of York
vitoversace@yahoo.com

User avatar
Daniele Varsano
Posts: 3816
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: compilation error on HPC

Post by Daniele Varsano » Thu Jun 19, 2014 2:46 pm

Dear Vito,
no idea, may be you can ask to your system administrator for this kind of problems.
If it is a problem of yambo, may be someone else can advise you, I did not find anything strange.

Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

Post Reply