[CP2K:1188] Re: TS optimization

[Teodoro Laino]

Hi,

having a look at the regtests file may help you solve your problem:

tests/Fist/regtest-3/2d_pot.inp
tests/Fist/regtest-3/argon.inp
tests/Fist/regtest-3/water_2_TS_CG.inp
tests/Fist/regtest-3/water_2_TS.inp
tests/QS/regtest-gpw-2/hcn_ts.inp
tests/QS/regtest-gpw-2/hcn_ts_r.inp

-------- IMPORTANT --------
KEEP in mind that the values of the parameters is chosen in order to  
let the tests run fast!!
They are meaningless (or roughly meaningless) for production.. so..  
you need to know
what you are doing.. as I said it could become a never-ending  
optimization otherwise..
-----------------------------------

Teo

On 16 Jul 2008, at 00:37, yimanu wrote:

>
> Hi Teo and everyone
>
> It appears there is still some problems, I am running cp2k compiled
> from today's CVS but I keep getting the error
> below, saying that the line spearch method is not implemented for CG,
> I have used all the line search methods
> avialable in cp2k.
>
> *
>  *** 18:28:05 ERRORL2 in cg_utils:cg_linmin processor      0
> err=-1000  ***
>  *** UNIMPLEMENTED, Line Search type not yet implemented in CG for
> TS    ***
>  ***
> search.
> ***
>  *
>
>  ===== Routine Calling Stack =====
>
>            10 cg_linmin
>             9 cp_cg_main
>             8 geoopt_cg
>             7 cp_rot_opt
>             6 cp_eval_at_ts
>             5 cp_eval_at
>             4 cp_cg_main
>             3 geoopt_cg
>             2 cp_geo_opt
>             1 CP2K
>
> Part of the input:
>
> &MOTION
>   &GEO_OPT
>      TYPE TRANSITION_STATE
>      OPTIMIZER CG
>      MAX_ITER 500
>      &CG
>         &LINE_SEARCH
>                 TYPE 3PNT
>         &END LINE_SEARCH
>      &END CG
>      &TRANSITION_STATE
>         &DIMER
>                 &ROT_OPT
>                         OPTIMIZER CG
>                 &END ROT_OPT
>         &END DIMER
>     &END TRANSITION_STATE
>   &END GEO_OPT
> &END MOTION
>
>
> any comments is appreciated
>
> regards
>
>
> On Jul 15, 11:21 am, yimanu <duu... at gmail.com> wrote:
>> thanks Teo
>>
>> the problem was with my compilation
>>
>> ciao
>>
>> On Jul 14, 10:34 am, Teodoro Laino <teodor... at gmail.com> wrote:
>>
>>
>>
>>> Hi Lekpa,
>>
>>> In my opinion this message is clear enough:
>>
>>>>  *
>>>>  *** 10:14:06 ERRORL2 in bfgs_optimizer:geoopt_bfgs processor
>>>> 0      ***
>>>>  *** err=-1000 UNIMPLEMENTED, BFGS method not yet working with for
>>>> DIMER  ***
>>>>  *** or
>>>> CELL_OPT                                                          * 
>>>> **
>>>>  *
>>
>>> BFGS does not work for DIMER or CELL_OPT.
>>
>>> If you uncomment the lines commented here below it should work
>>> (possibly choosing the right line-search algorithm.. not all line- 
>>> search
>>> have been implemented for all optimization methods)
>>
>>>> &MOTION
>>>>   &GEO_OPT
>>>>      TYPE TRANSITION_STATE
>>>> #     &TRANSITION_STATE
>>>> #       &DIMER
>>>> #               &ROT_OPT
>>>> #                       OPTIMIZER CG
>>>> #               &ROT_OPT
>>>> #       &END DIMER
>>>> #     &END TRANSITION_STATE
>>>>      OPTIMIZER CG
>>>>      MAX_ITER 500
>>>>   &END GEO_OPT
>>>> &END MOTION
>>>> Is it that TS optimization cannot be done yet?
>>
>>> Dimer method has been implemented and somehow tested (Luca can
>>> comment more on this.. at least better than I can..)..
>>> But as usual the most important thing is the choice of the  
>>> parameters
>>> (for the Dimer).. if you choose them wrong it could be a never- 
>>> ending
>>> story..
>>
>>> Teo
>>
>>>> thanks
>>
>>>> Lekpa- Hide quoted text -
>>
>> - Show quoted text -
> >