The ABINIT Users Mailing List ( CLOSED )

[matthieu verstraete <matthieu.jean.verstraete@gmail.com>]

Hi Marty,

indeed, this has been corrected in 5.8.0, but I will double check.
Thanks for taking the time

Matthieu

On Fri, May 15, 2009 at 1:53 AM, Marty Blaber <martin.blaber@gmail.com> wrote:
> Hi Matthieu,
>
> Sorry it took me so long to get back to you. Unfortunately there is a bug
> (typo) in the code. (Still present in 573)
> It is only apparent when you turn on gkqwrite, probably left over from some
> previous testing.
>
> On line 284 of 17ddb/normsq_gkq.F90
>
>  write (*,'(a,i4,2(3E16.6,3x))') 'gkk_qpt_tmp (band1,band1)', &
> &  iqptfull, gkk_qpt_tmp(:,1,:,1,1:10)
>
> should be
>
>  write (*,'(a,i4,2(3E16.6,3x))') 'gkk_qpt_tmp (band1,band1)', &
> &  iqptfull, gkk_qpt_tmp(:,1,:,1,1:elph_ds%nsppol)
>
> (I think)
>
>
> Cheers,
> Marty
>
>
>
>
>
> gkk_qpt_tmp (band1,band1)
> Segmentation Fault
> Electron Phonon Coupling
> Transport Spectral Function
> electrical resistivity
> thermal conductivity
>
>
> 2009/3/26 matthieu verstraete <matthieu.jean.verstraete@gmail.com>:
>> Hello,
>>
>> nice to see people using this bit of the code. I'm afraid the
>> symmetries are not used enough for saving of memory etc.. (and there
>> are specific difficulties in doing so here). The answer is yes, you
>> need to keep all bands, because later the transport integrals contain
>> different Fermi speeds for different bands. All of the k and q are
>> kept in memory, and here all of the bands (2 band indices) as well.
>>
>> However, you should try (with a later version of abinit, preferably)
>> to uncomment the prtgkq variable, which swaps much of the 100Gb to
>> disk (slower but ok as it is done in a structured way). This should
>> also work with the transport part.
>>
>> keep us posted.
>>
>> Matthieu
>>
>> On Thu, Mar 26, 2009 at 5:43 AM, Marty Blaber <martin.blaber@gmail.com>
>> wrote:
>>> Hi,
>>>
>>> I've been trying to perform an electron-phonon transport calculation
>>> on a 2 atom CsCl metallic system on a 16^3 kptrlatt with 35 qpoints.
>>> To perform the transport calculation we must set the tkeepbands to 1
>>> to keep all of the gkk bands in memory.  Is this really necessary? My
>>> calculations need ~100GB of memory (As I have only 32 GB at the
>>> moment, the system pages awfully.)
>>> Is there a solution to this problem, my system isn't that big... am I
>>> missing something obvious/asking a stupid question?
>>> My anddb/abinit inputs are below.
>>>
>>> Any advice would be appreciated.
>>>
>>> Cheers,
>>> Marty
>>>
>>> p.s. Using abinip 5.6.4 for GS + Pert, tried anaddb 564 and 573
>>>
>>> ############################################################
>>> ANADDB.IN
>>> ############################################################
>>> # turn on calculation of the electron-phonon quantities
>>> elphflag 1
>>> ifltransport 1
>>> tkeepbands 1
>>> doscalprod 1
>>> #gkqwrite 1
>>>
>>> # Path in reciprocal space along which the phonon linewidths
>>> #  and band structure will be calculated
>>> nqpath 1
>>> qpath
>>>  0.0   0.0   0.0
>>>  0.0   1/2   0.0
>>>  0.0   1     0.0
>>>  0.0   1/2   0.0
>>>  1/2   1/2   0.0
>>>  0.0   1/2   1/2
>>>  1/2   1/2   1/2
>>>  1/2   0.0   0.0
>>>
>>> # Irrelevant
>>> mustar 0.136
>>>
>>> # Minimalistic qpoint grid
>>> ngqpt 8 8 8
>>> nqshft 1
>>> q1shft 0.0 0.0 0.0
>>>
>>> nshiftk 1
>>> shiftk  0.0  0.0  0.0
>>> kptrlatt 16 0 0   0 16 0   0 0 16
>>> telphint 1
>>>
>>> #  impose acoustic sum rule in a symmetric way
>>> asr 2
>>> dipdip 1
>>>
>>> #  bravais lattice necessary
>>> brav 1
>>>
>>> #
>>> ifcflag 1
>>> ifcana 1
>>> #  ifc for all atoms?
>>> natifc 0
>>> atifc 1 2 3
>>>
>>> #  print dielectric matrix with freq dependence
>>> dieflag 0
>>>
>>> #  print out eigenvectors and symmetrize dyn matrix
>>> eivec 1
>>>
>>> #Wavevector list number 1 (Reduced coordinates and normalization factor)
>>> nph1l 1
>>> qph1l
>>>      0.0000  0.0000  0.0000  1
>>>
>>> symdynmat 0
>>>
>>> ############################################################
>>> /ANADDB.IN
>>> ############################################################
>>>
>>>
>>> ############################################################
>>> ABINIT.IN
>>> ############################################################
>>> ###########################################################
>>> #   Welcome to Marty Blaber's Abinit - EP Coupling input
>>> #   file generator
>>> #   Usage: python ab-build-ep-calc.py {qpoint file}
>>> ###########################################################
>>> # File    =  qp.list
>>> # nLines =  17
>>> # headerLines =  12
>>> # nqp =  5
>>>
>>> # File built on  2009-03-23 10:43:29.835767  by Martin G Blaber's
>>> abinit EP coupling input file generator
>>>
>>> #+**********************+
>>> #| DATASETS             |
>>> #+**********************+
>>>
>>> # Number of datasets =  36
>>> #ndtset     35
>>> #jdtset     2 3 4 5 6 7 8   9 10 11 12 13 14 15   16 17 18 19 20 21 22
>>>  23 24 25 26 27 28 29   30 31 32 33 34 35 36
>>> #ndtset      19
>>> #jdtset      18 19 20 21 22   23 24 25 26 27 28 29   30 31 32 33 34 35 36
>>> ndtset      6
>>> jdtset      31 32 33 34 35 36
>>>
>>> #DATASET 1 - Ground State
>>> iscf      3
>>> kptopt1   3
>>> tolwfr1   1.0d-18
>>> prtden1   1
>>> nline1    8
>>>
>>> #ds = 2 qp(0) =0.375 0.25 0.0
>>> iscf2    -3                            tolwfr2  1.0d-14 getden2  1
>>>        kptopt2  3 nqpt2  1 qpt2   0.375 0.25 0.0
>>> rfatpol3  1 1 rfdir3  1 0 0 rfphon3  1 tolvrs3   1.0e-5 getwfk3  1
>>> getwfq3   2 kptopt3  3 nqpt3  1 qpt3   0.375 0.25 0.0
>>> rfatpol4  1 1 rfdir4  0 1 0 rfphon4  1 tolvrs4   1.0e-5 getwfk4  1
>>> getwfq4   2 kptopt4  3 nqpt4  1 qpt4   0.375 0.25 0.0
>>> rfatpol5  1 1 rfdir5  0 0 1 rfphon5  1 tolvrs5   1.0e-5 getwfk5  1
>>> getwfq5   2 kptopt5  3 nqpt5  1 qpt5   0.375 0.25 0.0
>>> rfatpol6  2 2 rfdir6  1 0 0 rfphon6  1 tolvrs6   1.0e-5 getwfk6  1
>>> getwfq6   2 kptopt6  3 nqpt6  1 qpt6   0.375 0.25 0.0
>>> rfatpol7  2 2 rfdir7  0 1 0 rfphon7  1 tolvrs7   1.0e-5 getwfk7  1
>>> getwfq7   2 kptopt7  3 nqpt7  1 qpt7   0.375 0.25 0.0
>>> rfatpol8  2 2 rfdir8  0 0 1 rfphon8  1 tolvrs8   1.0e-5 getwfk8  1
>>> getwfq8   2 kptopt8  3 nqpt8  1 qpt8   0.375 0.25 0.0
>>>
>>> #ds = 9 qp(1) =0.5 0.25 0.0
>>> iscf9    -3                            tolwfr9  1.0d-14 getden9  1
>>>        kptopt9  3 nqpt9  1 qpt9   0.5 0.25 0.0
>>> rfatpol10  1 1 rfdir10  1 0 0 rfphon10  1 tolvrs10   1.0e-5 getwfk10
>>> 1 getwfq10   9 kptopt10  3 nqpt10  1 qpt10   0.5 0.25 0.0
>>> rfatpol11  1 1 rfdir11  0 1 0 rfphon11  1 tolvrs11   1.0e-5 getwfk11
>>> 1 getwfq11   9 kptopt11  3 nqpt11  1 qpt11   0.5 0.25 0.0
>>> rfatpol12  1 1 rfdir12  0 0 1 rfphon12  1 tolvrs12   1.0e-5 getwfk12
>>> 1 getwfq12   9 kptopt12  3 nqpt12  1 qpt12   0.5 0.25 0.0
>>> rfatpol13  2 2 rfdir13  1 0 0 rfphon13  1 tolvrs13   1.0e-5 getwfk13
>>> 1 getwfq13   9 kptopt13  3 nqpt13  1 qpt13   0.5 0.25 0.0
>>> rfatpol14  2 2 rfdir14  0 1 0 rfphon14  1 tolvrs14   1.0e-5 getwfk14
>>> 1 getwfq14   9 kptopt14  3 nqpt14  1 qpt14   0.5 0.25 0.0
>>> rfatpol15  2 2 rfdir15  0 0 1 rfphon15  1 tolvrs15   1.0e-5 getwfk15
>>> 1 getwfq15   9 kptopt15  3 nqpt15  1 qpt15   0.5 0.25 0.0
>>>
>>> #ds = 16 qp(2) =0.375 0.375 0.0
>>> iscf16    -3                            tolwfr16  1.0d-14 getden16  1
>>>           kptopt16  3 nqpt16  1 qpt16   0.375 0.375 0.0
>>> rfatpol17  1 1 rfdir17  1 0 0 rfphon17  1 tolvrs17   1.0e-5 getwfk17
>>> 1 getwfq17   16 kptopt17  3 nqpt17  1 qpt17   0.375 0.375 0.0
>>> rfatpol18  1 1 rfdir18  0 1 0 rfphon18  1 tolvrs18   1.0e-5 getwfk18
>>> 1 getwfq18   16 kptopt18  3 nqpt18  1 qpt18   0.375 0.375 0.0
>>> rfatpol19  1 1 rfdir19  0 0 1 rfphon19  1 tolvrs19   1.0e-5 getwfk19
>>> 1 getwfq19   16 kptopt19  3 nqpt19  1 qpt19   0.375 0.375 0.0
>>> rfatpol20  2 2 rfdir20  1 0 0 rfphon20  1 tolvrs20   1.0e-5 getwfk20
>>> 1 getwfq20   16 kptopt20  3 nqpt20  1 qpt20   0.375 0.375 0.0
>>> rfatpol21  2 2 rfdir21  0 1 0 rfphon21  1 tolvrs21   1.0e-5 getwfk21
>>> 1 getwfq21   16 kptopt21  3 nqpt21  1 qpt21   0.375 0.375 0.0
>>> rfatpol22  2 2 rfdir22  0 0 1 rfphon22  1 tolvrs22   1.0e-5 getwfk22
>>> 1 getwfq22   16 kptopt22  3 nqpt22  1 qpt22   0.375 0.375 0.0
>>>
>>> #ds = 23 qp(3) =0.5 0.375 0.0
>>> iscf23    -3                            tolwfr23  1.0d-14 getden23  1
>>>           kptopt23  3 nqpt23  1 qpt23   0.5 0.375 0.0
>>> rfatpol24  1 1 rfdir24  1 0 0 rfphon24  1 tolvrs24   1.0e-5 getwfk24
>>> 1 getwfq24   23 kptopt24  3 nqpt24  1 qpt24   0.5 0.375 0.0
>>> rfatpol25  1 1 rfdir25  0 1 0 rfphon25  1 tolvrs25   1.0e-5 getwfk25
>>> 1 getwfq25   23 kptopt25  3 nqpt25  1 qpt25   0.5 0.375 0.0
>>> rfatpol26  1 1 rfdir26  0 0 1 rfphon26  1 tolvrs26   1.0e-5 getwfk26
>>> 1 getwfq26   23 kptopt26  3 nqpt26  1 qpt26   0.5 0.375 0.0
>>> rfatpol27  2 2 rfdir27  1 0 0 rfphon27  1 tolvrs27   1.0e-5 getwfk27
>>> 1 getwfq27   23 kptopt27  3 nqpt27  1 qpt27   0.5 0.375 0.0
>>> rfatpol28  2 2 rfdir28  0 1 0 rfphon28  1 tolvrs28   1.0e-5 getwfk28
>>> 1 getwfq28   23 kptopt28  3 nqpt28  1 qpt28   0.5 0.375 0.0
>>> rfatpol29  2 2 rfdir29  0 0 1 rfphon29  1 tolvrs29   1.0e-5 getwfk29
>>> 1 getwfq29   23 kptopt29  3 nqpt29  1 qpt29   0.5 0.375 0.0
>>>
>>> #ds = 30 qp(4) =0.5 0.5 0.0
>>> iscf30    -3                            tolwfr30  1.0d-14 getden30  1
>>>           kptopt30  3 nqpt30  1 qpt30   0.5 0.5 0.0
>>> rfatpol31  1 1 rfdir31  1 0 0 rfphon31  1 tolvrs31   1.0e-5 getwfk31
>>> 1 getwfq31   30 kptopt31  3 nqpt31  1 qpt31   0.5 0.5 0.0
>>> rfatpol32  1 1 rfdir32  0 1 0 rfphon32  1 tolvrs32   1.0e-5 getwfk32
>>> 1 getwfq32   30 kptopt32  3 nqpt32  1 qpt32   0.5 0.5 0.0
>>> rfatpol33  1 1 rfdir33  0 0 1 rfphon33  1 tolvrs33   1.0e-5 getwfk33
>>> 1 getwfq33   30 kptopt33  3 nqpt33  1 qpt33   0.5 0.5 0.0
>>> rfatpol34  2 2 rfdir34  1 0 0 rfphon34  1 tolvrs34   1.0e-5 getwfk34
>>> 1 getwfq34   30 kptopt34  3 nqpt34  1 qpt34   0.5 0.5 0.0
>>> rfatpol35  2 2 rfdir35  0 1 0 rfphon35  1 tolvrs35   1.0e-5 getwfk35
>>> 1 getwfq35   30 kptopt35  3 nqpt35  1 qpt35   0.5 0.5 0.0
>>> rfatpol36  2 2 rfdir36  0 0 1 rfphon36  1 tolvrs36   1.0e-5 getwfk36
>>> 1 getwfq36   30 kptopt36  3 nqpt36  1 qpt36   0.5 0.5 0.0
>>>
>>> # Done
>>> #+----------------------+
>>> #| Constants            |
>>> #+----------------------+
>>>
>>> #### Grid and Smearing ####
>>> occopt  4                 # 4 = Metallic cold smear
>>> tsmear  0.001
>>> ecut    28.0
>>> nband   8
>>>
>>> #### K-POINTS ####
>>> #kptopt  3                 # 3 = No Sym
>>> nshiftk 1
>>> shiftk  0.0  0.0  0.0
>>> kptrlatt 16 0 0   0 16 0   0 0 16
>>>
>>> #### SCF ####
>>> #iscf    3
>>> nstep   1000
>>> #toldfe  1.0d-8           # Energy
>>> #tolwfr  1e-18
>>>
>>>
>>> #### CELL ####
>>> acell  3*6.4474824015
>>> rprim  1.0  0.0  0.0   # SC primitive vectors
>>>       0.0  1.0  0.0
>>>       0.0  0.0  1.0
>>>
>>> #### ATOMS ####
>>> ixc     7                 # 7 = PW92
>>> ntypat  2
>>> znucl   11  47
>>> natom   2
>>> typat   1   2
>>> xred    0.0  0.0  0.0
>>>        0.5  0.5  0.5
>>>
>>>
>>>
>>>
>>> ############################################################
>>> /ABINIT.IN
>>> ############################################################
>>>
>>
>>
>>
>> --
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> Dr. Matthieu Verstraete
>>
>> European Theoretical Spectroscopy Facility (ETSF)
>> Dpto. Fisica de Materiales,
>> U. del Pais Vasco,
>> Centro Joxe Mari Korta, Av. de Tolosa, 72,   Phone: +34-943018393
>> E-20018 Donostia-San Sebastian, Spain        Fax  : +34-943018390
>>
>> Mail : matthieu.jean.verstraete@gmail.com
>> http://www-users.york.ac.uk/~mjv500
>>
>>
>
>



-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Dr. Matthieu Verstraete

European Theoretical Spectroscopy Facility (ETSF)
Dpto. Fisica de Materiales,
U. del Pais Vasco,
Centro Joxe Mari Korta, Av. de Tolosa, 72,   Phone: +34-943018393
E-20018 Donostia-San Sebastian, Spain        Fax  : +34-943018390

Mail : matthieu.jean.verstraete@gmail.com
http://www-users.york.ac.uk/~mjv500