[Wien] WIEN2k_9.2 - still missing/incompatible functions in lapack_lapw

[Juergen Spitaler]

Dear Prof. Blaha,
thank you for including the reported fixes in WIEN2k_09.2.

Unfortunately there are still problems with respect to
WIEN2k's lapack_lap libraries:

- SRC_lapw1:
  lapw1 compiles now using lapack_lapw,  but it gets stuck
  when executed. This had  happened to me before when
  I tried to include some version of dsytri.f in the lapack_lapw.
  So probably the problem is still due to dsytri.f, and another
  version of dsytri.f should be considered.

- SRC_dipan:
  When using lapack_lapw, several routines are missing:
  * In dgetrf, undefined references to  dgetf2, dlaswp, dlaswp, dgetf2
  * In dtrtri, undefined reference to dtrtri

  - SRC_pairhess:
  When using lapack_lapw, one routine is missing:
  * In function MAIN,  dpotri

Best wishes,
Juergen



Peter Blaha wrote:
> Thanks for the report.
>
> I fixed the two non-standard fortran lines and  I've also added
> cputim_aix.f and the missing lapack subroutines to lapack_lapw (thanks
> to M.Divis).
>
>
>
> Juergen Spitaler schrieb:
>> Dear Prof. Blaha!
>>
>> We have just finished to compile WIEN2k_09.2 on our IBM/AIX cluster
>> with an xlf fortran compiler and the options "-O5 -qarch=pwr5 -q64".
>>
>> We encountered the following problems/missing parts of the
>> new package:
>>
>> - the lapack routine "dsytri", which is used in the new version of the
>>   SRC_lapw1/jacdavblock.F, is missing in liblapack_lapw
>>
>> - the routine cputim_aix.f (together with the timing routines for
>>   many other architectures), which was included in the previous versions
>>   of WIEN2k, is missing now. For this reason in many  SRC-directories
>>   the compilation stopped with an error
>>   (SRC_lapw0, SRC_lapw1, SRC_lapw2 , SRC_mixer, SRC_dstart , SRC_qtl ,
>>   SRC_aim , SRC_optic , SRC_lapdwdm , SRC_lapwso  , SRC_irrep  ,
>> SRC_joint)
>>
>>
>> Other minor system specific issues+fixes:
>>
>>  - SRC_pairhess:  Error "constrain.f", line 293.30: 1513-041 (S)
>>    Arguments of the wrong type were specified for the INTRINSIC
>> procedure "max".
>>    => fixed by changing line (add "d0" to first argument) to:
>>      t=Enormal/max(0.0005d0,w2(k))
>>
>>  - SRC_2Doptimize: "setup2D.f", several error messages like
>>    "line 121.0: 1515-010 (S) String is missing a closing delimiter ..."
>>    => fixed by changing in the respective lines "\" to "\\"
>>
>>  - SRC_eosfit6: mysterious compile errors
>>    => fixed by using a maximum optimization of -O2
>>
>> Best wishes,
>> Juergen
>>
>>
>>
>> Peter Blaha wrote:
>>> Dear WIEN2k users,
>>>
>>> A new version, WIEN2k_09.2, has been released.
>>>
>>> For a listing of all changes and download see 
>>> http://www.wien2k.at/reg_user/updates
>>>
>>> Most importantly:
>>>
>>> new package   SRC_dipan
>>>
>>> modifications to SRC_pairhess (better starting hessian for min_lapw)
>>>
>>> a new, even faster    iterative diagonalization in lapw1
>>>
>>> a new switch   -in1ef   for all run*_lapw scripts, which I recommend
>>> to use
>>> instead of     -in1new N
>>>
>>> official support for PBEsol
>>>
>>> supercell can handle R lattices
>>>
>>> updated SRC_arrows
>>>
>>> Best regards
>>
>> _______________________________________________
>> Wien mailing list
>> Wien at zeus.theochem.tuwien.ac.at
>> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>