[QE-users] help in understand why ph.x crashes with electron_phonon = 'interpolated'

patrizio.graziosi at cnr.it patrizio.graziosi at cnr.it
Mon May 16 15:03:44 CEST 2022


Hi Paolo,

thank you so much! I wasn't aware of that, sorry. I have followed that  
scheme and everything run smoothly, but now I'd need some support to  
understand the output files, is there a sort of tutorial for  
beginners/non-developers, or "bovine users"?

For example in the ph.out in the e-ph coupling section I see something like

      lambda(    1)=  0.0549   gamma=   24.38 GHz
      lambda(    2)=  0.0539   gamma=   32.01 GHz
      lambda(    3)=  0.0471   gamma=   52.47 GHz

what is that lambda, the lambda_q_nu of eq. 7 here  
https://arxiv.org/abs/cond-mat/0504077 ? and is it dimensionless or  
has some units? I expect gamma and omega are both in units of 1/s,  
thus if N(EF) has units of energy and volume, it should not be  
dimensionless but I can be wrong, obviously...


and in my lambda file I see

  Broadening   0.0020 lambda       0.3002 dos(Ef) 12.8744 omega_ln [K]  
    368.3226
  Broadening   0.0040 lambda       0.2909 dos(Ef) 12.8493 omega_ln [K]  
    367.0590
  Broadening   0.0060 lambda       0.2883 dos(Ef) 12.9313 omega_ln [K]  
    366.9214
  Broadening   0.0080 lambda       0.2865 dos(Ef) 12.9730 omega_ln [K]  
    366.7417
  Broadening   0.0100 lambda       0.2860 dos(Ef) 13.0371 omega_ln [K]  
    366.4294
  Broadening   0.0120 lambda       0.2870 dos(Ef) 13.1566 omega_ln [K]  
    366.0115
  Broadening   0.0140 lambda       0.2888 dos(Ef) 13.3094 omega_ln [K]  
    365.5325
  Broadening   0.0160 lambda       0.2907 dos(Ef) 13.4645 omega_ln [K]  
    365.0470
  Broadening   0.0180 lambda       0.2923 dos(Ef) 13.6008 omega_ln [K]  
    364.5968
  Broadening   0.0200 lambda       0.2936 dos(Ef) 13.7081 omega_ln [K]  
    364.2034

where lambda is always nearly the same, whereas in this tutorial  
http://indico.ictp.it/event/8301/session/96/contribution/534/material/0/0.pdf  
I see a large variation with the broadening
Broadening  0.0050  lambda  4.6503  dos(Ef)  7.4402  omega_ln[K]  58.9663
Broadening  0.0100  lambda  2.9129  dos(Ef)  5.1936  omega_ln[K]  54.7458
Broadening  0.0150  lambda  2.3700  dos(Ef)  4.2509  omega_ln[K]  49.0727
Broadening  0.0200  lambda  2.0155  dos(Ef)  3.7926  omega_ln[K]  46.7651
Broadening  0.0250  lambda  1.7924  dos(Ef)  3.5674  omega_ln[K]  46.0507 ...

is it because I'm working with FCC cobalt, which is not  
superconducting, or because I've done something silly, or...?


As I don't want to waste your time, I'd be glad if you could indicate  
to me a tutorial to understand all the output files, if exists.

Thanks,

Patrizio





Paolo Giannozzi <paolo.giannozzi at uniud.it> ha scritto:

> Have you looked at PHonon/examples/example03 and followed the  
> procedure described there?
>
> Paolo
>
> On 09/05/2022 07:46, patrizio.graziosi at cnr.it wrote:
>> Hello everybody,
>>
>> I'm experiencing a crash in ph.x but I cannot figure out the  
>> reason, so I ask you some suggestion if possible, please.
>>
>> I'm working on FCC Co with q-e/7.0 on CINECA cluster, if I use the  
>> following input file without the line electron_phonon =  
>> 'interpolated', it runs smoothly, but with the electron_phonon spec  
>> it crashes.
>>
>> &inputph
>>   verbosity='debug'
>>   tr2_ph=1.0d-17
>>   prefix='Co'
>>   ldisp=.true.
>>   epsil=.false.
>>   lqdir=.true.
>>   outdir='./tmp'
>>   fildyn='Co.dyn'
>>   fildvscf = 'dvscf'
>>   electron_phonon = 'interpolated'
>>   nq1=10, nq2=10, nq3=10,
>>   !nk1=20, nk2=20, nk3=20
>> /
>>
>>
>>
>> In the output file (attached) it looks the code runs until the  
>> following point, which looks the first e-ph calc.
>>
>> Number of q in the star =    1
>>      List of q in the star:
>>           1   0.000000000   0.000000000   0.000000000
>>
>>      Diagonalizing the dynamical matrix
>>
>>      q = (    0.000000000   0.000000000   0.000000000 )
>>
>>  **************************************************************************
>>      freq (    1) =       0.782369 [THz] =      26.097029 [cm-1]
>>      freq (    2) =       0.782369 [THz] =      26.097029 [cm-1]
>>      freq (    3) =       0.782369 [THz] =      26.097029 [cm-1]
>>  **************************************************************************
>>
>>      Mode symmetry, O_h (m-3m)  point group:
>>
>>      freq (   1-   3) =         26.1  [cm-1]   --> T_1u G_15  G_4- I
>>      electron-phonon interaction  ...
>>
>>
>>
>>
>> and in the output file of the run, I see a bunch of lines like the following
>>
>> forrtl: severe (24): end-of-file during read, unit 40, file  
>> /g100_scratch/userexternal/pgrazios/MatrixElements/Co_interp_fcc/./tmp/Co.a2Fsave Image              PC                Routine            Line         
>> Source
>> ph.x               0000000000FECD08  Unknown               Unknown  Unknown
>> ph.x               0000000001027ECA  Unknown               Unknown  Unknown
>> ph.x               00000000004339EC  elphsum_                  842   
>> elphon.f90
>> ph.x               000000000040D3A6  do_phonon_                143   
>> do_phonon.f90
>> ph.x               000000000040754B  MAIN__                     78   
>> phonon.f90
>> ph.x               00000000004074A2  Unknown               Unknown  Unknown
>> libc-2.28.so       0000147385E7E493  __libc_start_main     Unknown  Unknown
>> ph.x               00000000004073AE  Unknown               Unknown  Unknown
>> srun: error: r501n013: task 0: Exited with exit code 24
>> srun: launch/slurm: _step_signal: Terminating StepId=3470527.0
>> slurmstepd: error: *** STEP 3470527.0 ON r501n013 CANCELLED AT  
>> 2022-05-07T07:34:27 ***
>> forrtl: error (78): process killed (SIGTERM)
>> Image              PC                Routine            Line        Source
>> ph.x               0000000000FFBB1B  Unknown               Unknown  Unknown
>> libpthread-2.28.s  000014B7002C3B20  Unknown               Unknown  Unknown
>> libmpi.so.12.0.0   000014B700B10D6B  Unknown               Unknown  Unknown
>> libmpi.so.12.0.0   000014B700B04F87  Unknown               Unknown  Unknown
>> libmpi.so.12.0.0   000014B700A806C3  Unknown               Unknown  Unknown
>> libmpi.so.12.0.0   000014B700A57C51  Unknown               Unknown  Unknown
>> libmpi.so.12.0.0   000014B700A3BACA  Unknown               Unknown  Unknown
>> libmpi.so.12.0.0   000014B700B1EDE5  Unknown               Unknown  Unknown
>> libmpi.so.12.0.0   000014B700A1AD6D  MPI_Barrier           Unknown  Unknown
>> libmpifort.so.12.  000014B701DBC7BC  pmpi_barrier          Unknown  Unknown
>> ph.x               0000000000DF3B97  bcast_integer_             60   
>> mp_base.f90
>> ph.x               0000000000DD22E8  mp_mp_mp_bcast_i1         443  mp.f90
>> ph.x               000000000042AD26  elphsum_                  845   
>> elphon.f90
>> ph.x               000000000040D3A6  do_phonon_                143   
>> do_phonon.f90
>> ph.x               000000000040754B  MAIN__                     78   
>> phonon.f90
>> ph.x               00000000004074A2  Unknown               Unknown  Unknown
>> libc-2.28.so       000014B6FF76D493  __libc_start_main     Unknown  Unknown
>> ph.x               00000000004073AE  Unknown               Unknown  Unknown
>> forrtl: error (78): process killed (SIGTERM)
>>
>>
>>
>> The previous scf calc. has been done with PBE pseudopotential  
>> without spin orbit.
>>
>> Can you give me some hints on how to fix the problem, please?
>>
>> Thanks
>> Patrizio
>>
>>
>> _______________________________________________
>> 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
> Phone +39-0432-558216, fax +39-0432-558222



-- 

Patrizio Graziosi, PhD

Research Scientist

CNR - ISMN
Institute for the Study of Nanostructured Materials



More information about the users mailing list