QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: rebeda on September 9, 2019, 08:18

Title: spin transfer torque - reg.
Post by: rebeda on September 9, 2019, 08:18
I have been trying to perform spin transfer torque calculation for my system following the step by step procedure as given in the link below:

https://docs.quantumwise.com/tutorials/spin_transfer_torque/spin_transfer_torque.html

But  every time in the middle of the Collinear Initial State calculations, it stops the execution showing the error below:

#Warning
#The provided initial state parameter has no pre-calculated data.
#Proceeding without setting an initial state

will you please help me....

Regards,
Rebeda
Title: Re: spin transfer torque - reg.
Post by: Petr Khomyakov on September 9, 2019, 09:24
Please attach the script for this calculation and the corresponding log file.
Title: Re: spin transfer torque - reg.
Post by: rebeda on September 9, 2019, 11:50
Thank you for your kind reply. I have attached the concerned python and log files.
Regards,
Rebeda
Title: Re: spin transfer torque - reg.
Post by: rebeda on September 13, 2019, 08:30
Dear Sir,

I would like to know the status of my issue...... :( :( :( :'( :'( :'(

Regards,

Rebeda
Title: Re: spin transfer torque - reg.
Post by: Petr Khomyakov on September 16, 2019, 13:55
I was unable to reproduce this warning with your script (para.py). I assume that you have used QATK-2019.03-SP1 for doing this calculation, as evident from your log file. I notice that this is just a warning, not error message, and your calculation seems to run, even so the message was given.
Title: Re: spin transfer torque - reg.
Post by: rebeda on September 19, 2019, 10:32
Thank you Petr for your reply  :).  However I have some doubt regarding K point sampling.....

1. For my device relaxation and transport calculation (with bulk electrodes), I have considered 11 *11 * 1 and 2*2*100 k-points respectively. Is it ok?

2. If I consider 2*2*100 k-points  during relaxation as well as in transport calculation, I think the device seem to be 1 dimensional. At this point, what will be the nature of the electrodes (bulk or nanowire)?
Title: Re: spin transfer torque - reg.
Post by: Petr Khomyakov on September 25, 2019, 09:28
Deciding on k-point sampling is always a matter of convergence with respect to the k-point density. One should look at different samplings to see if the results (physical quantities of study)  obtained converge within some tolerance~accuracy. In that sense, nobody can tell you for sure that certain k-grid is OK for your particular system of study. You should perform convergence tests - this is a routine procedure for any kind calculations where one should converge results with respect to computational settings.
Title: Re: Script file for running batch job (quantum atk 2019.03)- reg.
Post by: rebeda on October 17, 2019, 08:12
Dear sir,

With reference to the attached script file for running batch job which was made for ATK 2015.1 version, will you please provide the new script file for execution of quantum ATK 2019.03 version in batch?

Waiting for your kind reply,

Regards,
Rebeda
Title: Re: spin transfer torque - reg.
Post by: Petr Khomyakov on October 17, 2019, 22:54
Jobs can be submitted through the Job Manager in QuantumATK-2019.03. You may also use the Job Manager to generate a script for submitting jobs directly on a remote machine.
Title: Re: spin transfer torque - reg.
Post by: rebeda on December 27, 2019, 20:10
Dear sir,

I have been struggling to calculate the spin transfer torque of my molecular junction following the link

https://docs.quantumatk.com/tutorials/spin_transfer_torque/spin_transfer_torque.html

but, the calculations always gets aborted. I have attached the python file and the corresponding log file for your perusal.

will you please help me .

Regards,
rebeda
Title: Re: spin transfer torque - reg.
Post by: Petr Khomyakov on December 28, 2019, 18:58
The error message suggests that this is likely a memory issue, i.e., you may need more memory to run this calculation.