[CP2K-user] [CP2K:19690] Help: cp2k vibration analysis stopped in the middle, how to renew the calculation?

[Leili Rassouli]

Dear Mikhail and Eric,

Thank you for your responses. My job is completed after 10 days, but to 
enhance the computational efficiency, I am considering the use of a 
mode-selective method or the dividing the full range into several parts to 
potentially expedite the calculation process. Mikhail, I would be 
appreciative if you could provide me access to the source files, given that 
I am currently utilizing CP2K v2022.1.

I'm currently facing an issue with negative frequencies. While using the 
ground state wavefunction in the ground state geometry, frequencies are 
positive and within the expected range. However, switching to the first 
excited wavefunction yields a significant number of negative frequencies.

Enclosed is my input file for 221 supercells of Fe2O3 with 120 atoms. Your 
insights and recommendations on this matter would be greatly valued.

I've explored various parameters, including:

Adjusting the CUTOFF value of TDDFPT from 300 to 600 and 700
Refining the optimized geometry
Increasing the TDDFT convergence from 1.0e-5 to 1.0e-6 and 1.0e-7
Despite these efforts, a notable proportion of frequencies persist as 
negative values. 

I express my gratitude in advance for your time and assistance.

Best regards,
Lili









On Tuesday, January 2, 2024 at 11:21:34 AM UTC-5 Eric Patterson wrote:

> Hello Lili,
>
> I did do a little bit of testing of the mode-selective method vs the 
> normal method. The results are comparable if I choose all atoms and restart 
> until all modes are converged (or at least nearly so — I didn’t care vey 
> much about low energy modes).
>
> However, I’ve also managed to compile a working MPI version on my HPC, so 
> I am able to run on enough nodes to get through a default vibrational 
> analysis before the queue kills the job. 
>
> Best,
> Eric
>
>
> On Dec 31, 2023, at 3:24 PM, Leili Rassouli <rassoul... at gmail.com> wrote:
>
> Hi Eric,
>
> I've faced a similar issue and was considering the approach you suggested. 
> Have you had a chance to try it, and does it work well in your case? Your 
> insights would be greatly appreciated.
>
> Thank you for your assistance.
> Lili Rassouli
>
> On Friday, February 10, 2023 at 11:32:16 AM UTC-5 Eric Patterson wrote:
>
>> On a closely-related note…
>>
>> Since normal vibrational analysis is not restartable, but mode-selective 
>> vibrational analysis is… Is there any drawback to using mode-selective and 
>> choosing all atoms, and then restarting as necessary until all modes have 
>> converged? Given the queue time limit imposed by my HPC resources, this 
>> seems to me the only way to get the job done. Are there any pitfalls one 
>> should bear in mind?
>>
>> Thanks,
>> Eric
>>
>>
>> On Feb 9, 2023, at 6:03 AM, Krack Matthias <matthia... at psi.ch> wrote:
>>
>> It is not possible to restart a vibrational analysis run. Such a 
>> functionality is not implemented.
>>  
>>
>> *From: *cp... at googlegroups.com <cp... at googlegroups.com> on behalf of Zac 
>> Smith <qingxi... at gmail.com>
>> *Date: *Thursday, 9 February 2023 at 02:54
>> *To: *cp2k <cp... at googlegroups.com>
>> *Subject: *[CP2K:18425] Help: cp2k vibration analysis stopped in the 
>> middle, how to renew the calculation?
>> Hello  cp2k users,  
>> I used a virtual machine to do vibration analysis with cp2k, but the task 
>> was killed when the hard disk was full, and the output file only contains 
>> out, molden, cube, and the eig file of size 0. Can the calculation be 
>> continued in this case?
>> -- 
>> You received this message because you are subscribed to the Google Groups 
>> "cp2k" group.
>> To unsubscribe from this group and stop receiving emails from it, send an 
>> email tocp2k+uns... at googlegroups.com.
>> To view this discussion on the web visit 
>> https://groups.google.com/d/msgid/cp2k/9d914e4f-bcb1-42ae-9b02-6c14f6fec8a5n%40googlegroups.com 
>> <https://groups.google.com/d/msgid/cp2k/9d914e4f-bcb1-42ae-9b02-6c14f6fec8a5n%40googlegroups.com?utm_medium=email&utm_source=footer>
>> .
>>
>> -- 
>> You received this message because you are subscribed to the Google Groups 
>> "cp2k" group.
>> To unsubscribe from this group and stop receiving emails from it, send an 
>> email to cp2k+uns... at googlegroups.com.
>>
>> To view this discussion on the web visit 
>> https://groups.google.com/d/msgid/cp2k/ZRAP278MB082784950ADD367A217A8C43F4D99%40ZRAP278MB0827.CHEP278.PROD.OUTLOOK.COM 
>> <https://groups.google.com/d/msgid/cp2k/ZRAP278MB082784950ADD367A217A8C43F4D99%40ZRAP278MB0827.CHEP278.PROD.OUTLOOK.COM?utm_medium=email&utm_source=footer>
>> .
>>
>>
>>
> -- 
> You received this message because you are subscribed to the Google Groups 
> "cp2k" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to cp2k+uns... at googlegroups.com.
>
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/cp2k/0d3984ac-a0a6-4040-b500-6a5d67f5c556n%40googlegroups.com 
> <https://groups.google.com/d/msgid/cp2k/0d3984ac-a0a6-4040-b500-6a5d67f5c556n%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>
>
>

-- 
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+unsubscribe at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/3e2997be-cac2-48d0-aeca-78080fa420f9n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20240102/603a45ee/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Fe2O3_vib.inp
Type: chemical/x-gamess-input
Size: 6671 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20240102/603a45ee/attachment-0001.inp>