TDDFT Optimization Failing (gradient: "failed to read isroots")

[Andrew Salij]

Hello,

I have been having some issues getting TDDFT excited state optimization to work. I'm getting initial DFT gradients and then excited states, but the subsequent optimization then fails in the initialization stage. I've added a "target" keyword to the TDDFT block (though not "targetsym", is that also necessary?) to little avail. Is there some absence in the input? Thank you. 

Input as below: 

title "nof"

memory 72 gb

start nof

echo

charge 0  

geometry autosym units angstrom

# Cartesian Coordinates in the form xyz

N          0.66764        3.46886       -0.03471

O          0.14102        2.42024       -0.41981

F         -0.24097        4.28862        0.51952

end

ecce_print ecce.out

basis spherical

  * library "Def2-TZVP"

END

driver  

  default

end

dft

  mult 1    # spin multiplicity. 1=singlet, 3=triplet

  XC xcamb88 1.00 lyp 0.81 vwn_5 0.19 hfexch 1.00

  cam 0.33 cam_alpha 0.19 cam_beta 0.46

  direct

  mulliken  # Method to calculate charges

end

tddft

  nroots 12

  cis

  civecs

end

task tddft optimize

task tddft energy

With error:

------------------------------------------------------------------------

 tddft_grad_init:                failed to read isroots                   0

 ------------------------------------------------------------------------

 ------------------------------------------------------------------------

  current input line : 

    43: task tddft optimize

 ------------------------------------------------------------------------

 ------------------------------------------------------------------------

 An error occured in the Runtime Database

 ------------------------------------------------------------------------

 For more information see the NWChem manual at 

 http://www.nwchem-sw.org/index.php/NWChem_Documentation

 For further details see manual section: 

                                                                                

                                                                                

                                                                                

                   

0:0:tddft_grad_init:                failed to read isroots:: -1

(rank:0 hostname:qnode5020 pid:24392):ARMCI DASSERT fail. ../../ga-5-4/armci/src/common/armci.c:ARMCI_Error():208 cond:0

(with more of the same)

[Niri Govind]

Hi Andrew,

You need a grad block within the TDDFT block and you need to specify the excited state to optimize on.

For example, to optimize on root 1, the input is as follows.

tddft

  nroots 12

  cis

  civics

  grad 

    root1 

  end

end

task tddft optimize

task tddft energy

Take a look at the manual pages: https://nwchemgit.github.io/Excited-State-Calculations.html

Best,

Niri Govind

PNNL

--
You received this message because you are subscribed to the Google Groups "NWChem Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nwchem-forum...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nwchem-forum/cffcd631-0566-4c0f-bb18-945b03ac9d4cn%40googlegroups.com.

[Andrew Salij]

Niri, thank you for your reply. I should have elaborated on my troubleshooting more in the initial conversation--I had already tested specifying a gradient block, only to get the following error. That said, thi is a different error than initially, so perhaps the initial concern was fixed.

The main concern that I've been having is that even the sample codes as provided in the documentation produce similar errors (e.g., isroots not found, as shown in first message). Could this be an issue with the NWChem implementation? To specify, all this has been run on NWChem 6.6.1. 

Error: 6:6:tddft_grad_init:                one spin state at a time:: -1

TDDFT block (all else identical to input)

tddft

  cis

  nroots 12

  civecs

  grad

    root 1

  end

end

[Niri Govind]

Hi Andrew,

Can you post or send me your output file ?

Thanks.

Best,

-Niri

Niri Govind

PNNL

niri....@gmail.com

To view this discussion on the web visit https://groups.google.com/d/msgid/nwchem-forum/ec6ce2fe-7af8-44b9-80ed-c875e5d7e89dn%40googlegroups.com.

[Niri Govind]

Hi Andrew,

I see your issue. At the gradient step, we can only work with singlets or triplets at a time.

For now, please add the notriplet keyword in the tddft block as shown below. I will log an issue and come up with a solution that will address this.   Also, please use the latest version of the code.

tddft

  nroots 12

  cis

  civics

  grad 

    root1 

  end

  notriplet

end

task tddft optimize

task tddft energy

Thanks for reporting this.

Best,

-Niri

Niri Govind

PNNL

niri....@pnnl.gov

On Wed, Feb 10, 2021 at 7:12 AM Andrew Salij <andrews...@u.northwestern.edu> wrote:

To view this discussion on the web visit https://groups.google.com/d/msgid/nwchem-forum/ec6ce2fe-7af8-44b9-80ed-c875e5d7e89dn%40googlegroups.com.

[Andrew Salij]

Once again, thank you for your reply. That was indeed causing the issue and optimization converged successfully.