Re: [QE-users] Error in ph.x: wrong representation

[Noé Mascello]

Thank you. I had to use 'cg' because I was running into problems during 
EPW calculations if I used Davidson.



Am 26.09.2018 um 09:18 schrieb Paolo Giannozzi:
Please see here: https://gitlab.com/QEF/q-e/issues/64 . Workaround: 
use Wyckoff positions if you know them. Thank you for reporting this 
problem


Paolo

PS: don't use 'cg' unless you have good reasons to: it is usually much 
slower than Davidson


On Tue, Sep 25, 2018 at 12:24 PM, Noé Mascello > wrote:


Hi Paolo

It should be P4/ncc (#130). I checked it with FINDSYM from the
ISOTROPY suite.

Regards,

Noé


Am 25.09.2018 um 10:38 schrieb Paolo Giannozzi:

It might be yet another case of "quasi-symmetric" system. What is
the expected symetry group of your material?

Paolo

On Mon, Sep 24, 2018 at 12:07 AM, Noé Mascello
mailto:m...@student.ethz.ch>> wrote:

Dear Quantum Espresso community,

I have a structure where the phonon calculation for a single
irreducible q-point fails (it works for all other q-points),
rendering the error:

/Error in routine set_irr_sym_new/

/wrong representation/

I am using version 6.2.1.

I've seen that this problem has come up before but there
wasn't always a working solution for it.

I have tried increasing the threshold which sends the error
in set_irr_sym.f90 as sometimes suggested.

I have also tried generating different random matrices
(created by random_matrix_new()) by changing the random
number generator seed, using uniformly distributed numbers

and using normally distributed random numbers with different
mean value and variance. Unfortunately, I always get the same
error at this q-point. I also had no success with changing all

atom positions slightly.

I'd be thankful for any other suggestions.


Cheers,

Noé Mascello

ETH Zurich




scf input


    calculation   = 'scf'
    restart_mode  = 'from_scratch'
    prefix    = 'P4ncc_00'
    pseudo_dir    = '/pps'
    outdir    = './'
/

    ibrav   = 6
    A   = 5.28231
    B   = 5.28231
    C   = 7.84336
    cosAB   = 0
    cosAC   = 0
    cosBC   = 0
    nat = 16
    ntyp    = 2
    ecutwfc = 90
    occupations = 'smearing'
    degauss = 7.35d-4

/

    conv_thr    = 1.0d-8
    diagonalization = 'cg'
/

ATOMIC_SPECIES
W 183.84 W_ONCV_PBE-1.1.upf
O 15.9994 O_ONCV_PBE-1.0.upf


ATOMIC_POSITIONS crystal
 W 0.25000 0.25000 0.21969
 W 0.74977 0.74977 0.78031
 W 0.74977 0.74977 0.280300031
 W 0.25000 0.25000 0.719699969
 O 0.25000 0.25000 0.493539985
 O 0.74977 0.74977 0.506460015
 O 0.74977 0.74977 0.006460011
 O 0.25000 0.25000 0.993540015
 O 0.460340007 0.539659993 0.25000
 O 0.539659993 0.460340007 0.74985
 O 0.039660007 0.960340007 0.25000
 O 0.960340007 0.039659976 0.74985
 O 0.960340007 0.460340007 0.25000
 O 0.039659976 0.539659993 0.74985
 O 0.539659993 0.039660007 0.25000
 O 0.460340007 0.960340007 0.74985

K_POINTS automatic
8 8 6 0 0 0

___

ph.x input (2nd q-point):



    prefix   = 'P4ncc_00'
    outdir   = './q2'
    fildyn   = 'P4ncc_00.dyn'
    fildvscf = 'dvscf'
    ldisp    = .true.
    nq1  = 4
    nq2  = 4
    nq3  = 3
    asr  = .true.
    tr2_ph   = 1.0d-14
    start_q = 2
    last_q = 2
    recover = .true.
    search_sym = .false.
/





___
users mailing list
users@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



___
users mailing list

Re: [QE-users] Error in ph.x: wrong representation

[Paolo Giannozzi]

Please see here: https://gitlab.com/QEF/q-e/issues/64 . Workaround: use
Wyckoff positions if you know them. Thank you for reporting this problem

Paolo

PS: don't use 'cg' unless you have good reasons to: it is usually much
slower than Davidson

On Tue, Sep 25, 2018 at 12:24 PM, Noé Mascello  wrote:

> Hi Paolo
>
> It should be P4/ncc (#130). I checked it with FINDSYM from the ISOTROPY
> suite.
>
> Regards,
>
> Noé
>
> Am 25.09.2018 um 10:38 schrieb Paolo Giannozzi:
>
> It might be yet another case of "quasi-symmetric" system. What is the
> expected symetry group of your material?
>
> Paolo
>
> On Mon, Sep 24, 2018 at 12:07 AM, Noé Mascello 
> wrote:
>
>> Dear Quantum Espresso community,
>>
>> I have a structure where the phonon calculation for a single irreducible
>> q-point fails (it works for all other q-points), rendering the error:
>>
>> *Error in routine set_irr_sym_new*
>>
>> *wrong representation*
>>
>> I am using version 6.2.1.
>>
>> I've seen that this problem has come up before but there wasn't always a
>> working solution for it.
>>
>> I have tried increasing the threshold which sends the error in
>> set_irr_sym.f90 as sometimes suggested.
>>
>> I have also tried generating different random matrices (created by
>> random_matrix_new()) by changing the random number generator seed, using
>> uniformly distributed numbers
>>
>> and using normally distributed random numbers with different mean value
>> and variance. Unfortunately, I always get the same error at this q-point. I
>> also had no success with changing all
>>
>> atom positions slightly.
>>
>> I'd be thankful for any other suggestions.
>>
>>
>> Cheers,
>>
>> Noé Mascello
>>
>> ETH Zurich
>>
>>
>> 
>>
>> scf input
>>
>> 
>> calculation   = 'scf'
>> restart_mode  = 'from_scratch'
>> prefix= 'P4ncc_00'
>> pseudo_dir= '/pps'
>> outdir= './'
>> /
>> 
>> ibrav   = 6
>> A   = 5.28231
>> B   = 5.28231
>> C   = 7.84336
>> cosAB   = 0
>> cosAC   = 0
>> cosBC   = 0
>> nat = 16
>> ntyp= 2
>> ecutwfc = 90
>> occupations = 'smearing'
>> degauss = 7.35d-4
>>
>> /
>> 
>> conv_thr= 1.0d-8
>> diagonalization = 'cg'
>> /
>>
>> ATOMIC_SPECIES
>> W 183.84 W_ONCV_PBE-1.1.upf
>> O 15.9994 O_ONCV_PBE-1.0.upf
>>
>>
>> ATOMIC_POSITIONS crystal
>>  W 0.25000 0.25000 0.21969
>>  W 0.74977 0.74977 0.78031
>>  W 0.74977 0.74977 0.280300031
>>  W 0.25000 0.25000 0.719699969
>>  O 0.25000 0.25000 0.493539985
>>  O 0.74977 0.74977 0.506460015
>>  O 0.74977 0.74977 0.006460011
>>  O 0.25000 0.25000 0.993540015
>>  O 0.460340007 0.539659993 0.25000
>>  O 0.539659993 0.460340007 0.74985
>>  O 0.039660007 0.960340007 0.25000
>>  O 0.960340007 0.039659976 0.74985
>>  O 0.960340007 0.460340007 0.25000
>>  O 0.039659976 0.539659993 0.74985
>>  O 0.539659993 0.039660007 0.25000
>>  O 0.460340007 0.960340007 0.74985
>>
>> K_POINTS automatic
>> 8 8 6 0 0 0
>>
>> ___
>>
>> ph.x input (2nd q-point):
>>
>>
>> 
>> prefix   = 'P4ncc_00'
>> outdir   = './q2'
>> fildyn   = 'P4ncc_00.dyn'
>> fildvscf = 'dvscf'
>> ldisp= .true.
>> nq1  = 4
>> nq2  = 4
>> nq3  = 3
>> asr  = .true.
>> tr2_ph   = 1.0d-14
>> start_q = 2
>> last_q = 2
>> recover = .true.
>> search_sym = .false.
>> /
>>
>>
>>
>>
>>
>> ___
>> users mailing list
>> users@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
>
>
>
> ___
> users mailing 
> listusers@lists.quantum-espresso.orghttps://lists.quantum-espresso.org/mailman/listinfo/users
>
>
>
> ___
> users mailing list
> users@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
___
users mailing list
users@lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users

Re: [QE-users] Error in ph.x: wrong representation

[Noé Mascello]

Hi Paolo

It should be P4/ncc (#130). I checked it with FINDSYM from the ISOTROPY 
suite.


Regards,

Noé


Am 25.09.2018 um 10:38 schrieb Paolo Giannozzi:
It might be yet another case of "quasi-symmetric" system. What is the 
expected symetry group of your material?


Paolo

On Mon, Sep 24, 2018 at 12:07 AM, Noé Mascello > wrote:


Dear Quantum Espresso community,

I have a structure where the phonon calculation for a single
irreducible q-point fails (it works for all other q-points),
rendering the error:

/Error in routine set_irr_sym_new/

/wrong representation/

I am using version 6.2.1.

I've seen that this problem has come up before but there wasn't
always a working solution for it.

I have tried increasing the threshold which sends the error in
set_irr_sym.f90 as sometimes suggested.

I have also tried generating different random matrices (created by
random_matrix_new()) by changing the random number generator seed,
using uniformly distributed numbers

and using normally distributed random numbers with different mean
value and variance. Unfortunately, I always get the same error at
this q-point. I also had no success with changing all

atom positions slightly.

I'd be thankful for any other suggestions.


Cheers,

Noé Mascello

ETH Zurich




scf input


    calculation   = 'scf'
    restart_mode  = 'from_scratch'
    prefix    = 'P4ncc_00'
    pseudo_dir    = '/pps'
    outdir    = './'
/

    ibrav   = 6
    A   = 5.28231
    B   = 5.28231
    C   = 7.84336
    cosAB   = 0
    cosAC   = 0
    cosBC   = 0
    nat = 16
    ntyp    = 2
    ecutwfc = 90
    occupations = 'smearing'
    degauss = 7.35d-4

/

    conv_thr    = 1.0d-8
    diagonalization = 'cg'
/

ATOMIC_SPECIES
W 183.84 W_ONCV_PBE-1.1.upf
O 15.9994 O_ONCV_PBE-1.0.upf


ATOMIC_POSITIONS crystal
 W 0.25000 0.25000 0.21969
 W 0.74977 0.74977 0.78031
 W 0.74977 0.74977 0.280300031
 W 0.25000 0.25000 0.719699969
 O 0.25000 0.25000 0.493539985
 O 0.74977 0.74977 0.506460015
 O 0.74977 0.74977 0.006460011
 O 0.25000 0.25000 0.993540015
 O 0.460340007 0.539659993 0.25000
 O 0.539659993 0.460340007 0.74985
 O 0.039660007 0.960340007 0.25000
 O 0.960340007 0.039659976 0.74985
 O 0.960340007 0.460340007 0.25000
 O 0.039659976 0.539659993 0.74985
 O 0.539659993 0.039660007 0.25000
 O 0.460340007 0.960340007 0.74985

K_POINTS automatic
8 8 6 0 0 0

___

ph.x input (2nd q-point):



    prefix   = 'P4ncc_00'
    outdir   = './q2'
    fildyn   = 'P4ncc_00.dyn'
    fildvscf = 'dvscf'
    ldisp    = .true.
    nq1  = 4
    nq2  = 4
    nq3  = 3
    asr  = .true.
    tr2_ph   = 1.0d-14
    start_q = 2
    last_q = 2
    recover = .true.
    search_sym = .false.
/





___
users mailing list
users@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



___
users mailing list
users@lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users


___
users mailing list
users@lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users

Re: [QE-users] Error in ph.x: wrong representation

[Paolo Giannozzi]

It might be yet another case of "quasi-symmetric" system. What is the
expected symetry group of your material?

Paolo

On Mon, Sep 24, 2018 at 12:07 AM, Noé Mascello  wrote:

> Dear Quantum Espresso community,
>
> I have a structure where the phonon calculation for a single irreducible
> q-point fails (it works for all other q-points), rendering the error:
>
> *Error in routine set_irr_sym_new*
>
> *wrong representation*
>
> I am using version 6.2.1.
>
> I've seen that this problem has come up before but there wasn't always a
> working solution for it.
>
> I have tried increasing the threshold which sends the error in
> set_irr_sym.f90 as sometimes suggested.
>
> I have also tried generating different random matrices (created by
> random_matrix_new()) by changing the random number generator seed, using
> uniformly distributed numbers
>
> and using normally distributed random numbers with different mean value
> and variance. Unfortunately, I always get the same error at this q-point. I
> also had no success with changing all
>
> atom positions slightly.
>
> I'd be thankful for any other suggestions.
>
>
> Cheers,
>
> Noé Mascello
>
> ETH Zurich
>
>
> 
>
> scf input
>
> 
> calculation   = 'scf'
> restart_mode  = 'from_scratch'
> prefix= 'P4ncc_00'
> pseudo_dir= '/pps'
> outdir= './'
> /
> 
> ibrav   = 6
> A   = 5.28231
> B   = 5.28231
> C   = 7.84336
> cosAB   = 0
> cosAC   = 0
> cosBC   = 0
> nat = 16
> ntyp= 2
> ecutwfc = 90
> occupations = 'smearing'
> degauss = 7.35d-4
>
> /
> 
> conv_thr= 1.0d-8
> diagonalization = 'cg'
> /
>
> ATOMIC_SPECIES
> W 183.84 W_ONCV_PBE-1.1.upf
> O 15.9994 O_ONCV_PBE-1.0.upf
>
>
> ATOMIC_POSITIONS crystal
>  W 0.25000 0.25000 0.21969
>  W 0.74977 0.74977 0.78031
>  W 0.74977 0.74977 0.280300031
>  W 0.25000 0.25000 0.719699969
>  O 0.25000 0.25000 0.493539985
>  O 0.74977 0.74977 0.506460015
>  O 0.74977 0.74977 0.006460011
>  O 0.25000 0.25000 0.993540015
>  O 0.460340007 0.539659993 0.25000
>  O 0.539659993 0.460340007 0.74985
>  O 0.039660007 0.960340007 0.25000
>  O 0.960340007 0.039659976 0.74985
>  O 0.960340007 0.460340007 0.25000
>  O 0.039659976 0.539659993 0.74985
>  O 0.539659993 0.039660007 0.25000
>  O 0.460340007 0.960340007 0.74985
>
> K_POINTS automatic
> 8 8 6 0 0 0
>
> ___
>
> ph.x input (2nd q-point):
>
>
> 
> prefix   = 'P4ncc_00'
> outdir   = './q2'
> fildyn   = 'P4ncc_00.dyn'
> fildvscf = 'dvscf'
> ldisp= .true.
> nq1  = 4
> nq2  = 4
> nq3  = 3
> asr  = .true.
> tr2_ph   = 1.0d-14
> start_q = 2
> last_q = 2
> recover = .true.
> search_sym = .false.
> /
>
>
>
>
>
> ___
> users mailing list
> users@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
___
users mailing list
users@lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users